Professional Documents
Culture Documents
Africa’s Infrastructure
A Time for Transformation
Africa’s Infrastructure
Africa’s Infrastructure
A Time for Transformation
The Africa Development Forum series was created in 2009 to focus on issues of significant
relevance to Sub-Saharan Africa’s social and economic development. Its aim is both to record
the state of the art on a specific topic and to contribute to ongoing local, regional, and global
policy debates. It is designed specifically to provide practitioners, scholars, and students
with the most up-to-date research results while highlighting the promise, challenges, and
opportunities that exist on the continent.
The series is sponsored by the Agence Française de Développement and the World Bank.
The manuscripts chosen for publication represent the highest quality in each institution’s
research and activity output and have been selected for their relevance to the development
agenda. Working together with a shared sense of mission and interdisciplinary purpose,
the two institutions are committed to a common search for new insights and new ways of
analyzing the development realities of the Sub-Saharan Africa Region.
World Bank
Shantayanan Devarajan, Chief Economist, Africa Region
Jorge Arbache, Senior Economist
Contents
Preface xix
Acknowledgments xxi
Abbreviations xxiii
vii
viii Contents
Note 41
References 41
Index 337
xii Contents
Boxes
I.1 The AICD Geographic Information Systems Platform for Africa 35
1.1 Introducing a Country Typology 51
2.1 Does Deficit-Financed Public Investment in Infrastructure Pay for Itself? 77
3.1 Access to Basic Infrastructure and Time Use 95
3.2 Access, Affordability, and Alternatives—Urban Public Transportation 96
3.3 Are Connection Subsidies Well Targeted to the Poor? 98
4.1 Infrastructure’s Institutional Scorecard 106
4.2 Privatization in African Ports 114
4.3 Lessons from the DAWASA Lease Contract (Tanzania) 116
4.4 Lessons from Successful SOE Reforms in Botswana Power Corporation 118
4.5 Performance Agreement for the National Water and Sewerage Corporation
(Uganda) 119
4.6 Regulation by Contract in Senegal 121
5.1 Land Issues in Tanzania 136
6.1 Not So EASSy 148
6.2 One-Stop Border Posts to Facilitate Trade 159
8.1 South Africa’s Power Supply Crisis 183
8.2 Kenya’s Success with Private Participation in Power 189
8.3 Botswana’s Success with a State-Owned Power Utility 190
8.4 CREST Spreading Good Practices 196
8.5 Rural Electrification in Mali 199
9.1 Impediments to Transit Traffic Growth in the Maputo Corridor 207
9.2 A New Attempt to Reform the Transit System in the Cameroon–Central African
Republic–Chad Corridor 208
10.1 The Role of AGETIPs 215
12.1 Private Participation and Port Efficiency: The Case of Apapa Container Terminal,
Lagos, Nigeria 257
14.1 The Metolong Dam and Water Supply Program in Lesotho 277
15.1 Assumptions for Irrigation Investment Potential Study 291
15.2 An Enabling Environment for Reform: Office du Niger 295
15.3 Nigeria’s Fadama Water User Association: Expanding Irrigation 296
16.1 Legalizing Household Water Resellers in Côte d’Ivoire 302
16.2 Standposts in Kigali, Rwanda 306
16.3 Cost Recovery, Equity, and Efficiency in Water Tariffs in Africa 308
16.4 Senegal’s Successful Experience with Private Sector Participation 312
16.5 Uganda’s Successful Case of State-Owned Enterprise Reform 314
16.6 Issues Constraining Rural Water Supply in Cross River State, Nigeria 316
17.1 What Is Improved Sanitation? 324
17.2 Ethiopia’s Success with a Community-Led Program 331
17.3 Burkina Faso’s Sanitation Tax 334
Contents xiii
Figures
O.1 Access to Household Services 3
O.2 Underlying Causes of Africa’s Power Supply Crisis 6
O.3 Burden of Infrastructure Spending Needs 8
O.4 Infrastructure Public Spending as a Percentage of GDP 9
O.5 Rehabilitation Backlog 10
O.6 Hidden Costs of Utility Inefficiencies 11
O.7 Infrastructure Funding Gap by Sector and Country Type 12
O.8 Institutional Progress across Sectors 13
O.9 Access to and Affordability of Household Services 23
I.1 Country Coverage of the Africa Infrastructure Country Diagnostic 34
I.2 Representativeness of Phase I Sampled Countries 34
1.1 Changes in Growth per Capita Caused by Changes in Growth Fundamentals 45
1.2 Changes in Growth per Capita Caused by Changes in Different Kinds of
Infrastructure 45
1.3 Contribution of Infrastructure to Total Factor Productivity of Firms 46
1.4 Growth of Africa’s Infrastructure Stocks Compared with Asia 49
1.5 Africa’s Aggregate Infrastructure Spending Needs, by Country, 2006–15 59
2.1 Sources of Financing for Infrastructure Capital Investment, by Sector and
Country Type 68
2.2 Split Investment Responsibilities between Governments and Public Enterprises,
by Type of Country and Sector 69
2.3 Rehabilitation Backlog 73
2.4 Costs of Capital by Funding Source 83
2.5 Spreading Spending over Time 84
3.1 Coverage of Network Infrastructure Services, 1990–2005 88
3.2 Expansion of Access to Infrastructure Services Each Year, Mid-1990s to Mid-2000s 89
3.3 Projected Universal Access for Piped Water for Sub-Saharan African Countries,
2050 and Beyond 89
3.4 Coverage of Modern Infrastructure Services, by Budget Quintile 89
3.5 Share of Household Budgets Spent on Infrastructure Services, by Budget Quintile 91
3.6 Population with Service Connections Who Do Not Pay for Service 91
3.7 Affordability of Subsistence Consumption Priced at Cost-Recovery Levels 91
3.8 Extent to Which Electricity and Water Subsidies Reach the Poor, by Country 93
3.9 Amount of Subsidy Needed to Maintain Affordability of Water and
Electricity Service, 2005 94
3.10 Access to Alternative Water and Sanitation Services across All Income Levels 96
3.11 Increased Industrial and Commercial Tariffs, Niger and Malawi 99
3.12 Residential Customers Using Prepayment Meters, by Utility 100
4.1 Institutional Progress across Countries, by Income Group, Aid Dependence,
and Resource Richness 108
4.2 Institutional Progress across Sectors 109
xiv Contents
Tables
O.1 Africa’s Infrastructure Deficit 3
O.2 Africa’s High-Cost Infrastructure 5
O.3 Overall Infrastructure Spending Needs for Sub-Saharan Africa 7
O.4 Infrastructure Spending on Addressing Sub-Saharan Africa’s Infrastructure Needs 9
O.5 Finding Resources: The Efficiency Gap and the Funding Gap 12
O.6 Overview of Private Participation in Infrastructure 14
I.1 AICD Background Papers 33
I.2 AICD Working Papers 41
1.1 Links between Infrastructure and Growth in Africa: What the Research Says 44
1.2 Evidence on Links between Infrastructure and MDGs in Africa 47
1.3 International Perspective on Africa’s Infrastructure Deficit 48
1.4 Intraregional Perspective on Africa’s Infrastructure Deficit 50
1.5 Africa’s High-Cost Infrastructure 50
1.6 10-Year Economic and Social Targets for Investment Needs Estimates, 2006–15 52
1.7 Africa’s Power Needs, 2006–15 53
1.8 Power Spending Needs, 2006–15 54
1.9 Irrigation Spending Needs, 2006–15 55
1.10 Water and Sanitation Spending Needs, 2006–15 55
1.11 Transport Spending Needs, 2006–15 57
1.12 ICT Spending Needs beyond the Purely Market Driven: Investment Only, 2006–15 58
1.13 Overall Infrastructure Spending Needs for Africa, 2006–15 58
2.1 Spending of Most Important Players Traced to Needs (Annualized Flows) 67
Contents xvii
7.2 Investments Needed to Close Gaps in Voice and Broadband Coverage in Sub-Saharan
Africa 172
8.1 Economic Cost of Emergency Power Generation 184
8.2 Power Sector Spending Needs 185
8.3 Financing Flows to the Power Sector 186
8.4 Composition of Power Sector Funding Gap 186
8.5 Cost and Affordability of Monthly Power Bills at Cost-Recovery Prices: Past and
Future 193
10.1 Average Daily Traffic on the Main Road Network 213
10.2 Road Sector Spending Needs 219
10.3 Financing Flows to the Road Sector 219
10.4 Overview of Key Road Freight Parameters on Main International Corridors 220
12.1 Traffic Trends for Container Trade, Sub-Saharan Ports, by Region, 1995–2005 250
12.2 Traffic Trends for General Cargo, 1995–2005 251
12.3 Private Transactions for All Port Sectors, 2000–07 254
12.4 Average Port Delays 255
12.5 Typical Gateway Container and General Cargo Handling Charges in World Markets 255
13.1 Air Traffic Growth in Sub-Saharan Africa, 2001–07 261
13.2 Air Service Liberalization in African Regional Groupings 264
13.3 Runway Quality in Sub-Saharan Africa, 2007 266
14.1 Water Availability and Withdrawal 275
14.2 Capital Investment Needs in Large Multipurpose Hydropower Storage by 2015 281
15.1 Selected Irrigation Investment Indicators for Sub-Saharan Africa 289
15.2 Potential Investment Needs for Large-Scale, Dam-Based, and Complementary Small-Scale
Irrigation in Sub-Saharan Africa 291
15.3 Share of Crops under Irrigation, Irrigation Investment Needs Assessment 293
15.4 Sensitivity of Irrigation Potential to Assumed Investment Cost 293
15.5 Food Price Changes for Various Indicators, 2020 and 2050 294
16.1 Evolution of Water Supply Coverage in Africa, by Source 301
16.2 Services Provided by Utilities in Their Service Areas 301
16.3 Quality of Services Provided by Utilities in Their Service Areas 302
16.4 Average Price for Water Service in 15 Largest Cities, by Type of Provider 303
16.5 Estimated Annual Financing Needed to Meet the Water MDG 303
16.6 Existing Financial Flows to Water Supply and Sanitation 304
16.7 Composition of Water Sector Funding Gap 305
16.8 Overview of Private Sector Participation’s Effect on Utility Performance 312
16.9 Relationship between Rural Reform Index and Success in Expanding Rural Service
Coverage 315
17.1 Patterns of Access to Sanitation in Africa 326
17.2 Cost of Sanitation Facilities in Senegal 332
Preface
This study is part of the Africa Infrastruc- the Democratic Republic of Congo, Ethiopia,
ture Country Diagnostic (AICD), a project Ghana, Kenya, Lesotho, Madagascar, Malawi,
designed to expand the world’s knowledge of Mozambique, Namibia, Niger, Nigeria, Rwanda,
physical infrastructure in Africa. The AICD Senegal, South Africa, Sudan, Tanzania, Uganda,
will provide a baseline against which future and Zambia. Under a second phase of the proj-
improvements in infrastructure services can ect, coverage is expanding to include as many
be measured, making it possible to monitor of the additional African countries as possible.
the results achieved from donor support. It The AICD was commissioned by the
should also provide a more solid empirical Infrastructure Consortium for Africa (ICA)
foundation for prioritizing investments and following the 2005 G8 (Group of Eight) sum-
designing policy reforms in the infrastructure mit at Gleneagles, Scotland, which flagged the
sectors in Africa. importance of scaling up donor finance for
The AICD is based on an unprecedented infrastructure in support of Africa’s develop-
effort to collect detailed economic and techni- ment. The World Bank is implementing the
cal data on the infrastructure sectors in Africa. AICD under the guidance of a steering com-
The project has produced a series of origi- mittee that represents the African Union, the
nal reports on public expenditure, spending New Partnership for Africa’s Development
needs, and sector performance in each of the (NEPAD), Africa’s regional economic com-
main infrastructure sectors, including energy, munities, the African Development Bank
information and communication technologies, (AfDB), the Development Bank of South
irrigation, transport, and water and sanitation. Africa (DBSA), and major infrastructure
This volume synthesizes the most significant donors. Financing for the AICD is provided
findings of those reports. by a multidonor trust fund to which the main
The first phase of the AICD focused on 24 contributors are the United Kingdom’s Depart-
countries that together account for 85 percent ment for International Development (DFID),
of the gross domestic product, population, the Public-Private Infrastructure Advisory
and infrastructure aid flows of Sub-Saharan Facility (PPIAF), Agence Française de Dével-
Africa. The countries are Benin, Burkina Faso, oppement (AFD), the European Commission,
Cameroon, Cape Verde, Chad, Côte d’Ivoire, and Germany’s Entwicklungsbank (KfW).
xix
xx Preface
A group of distinguished peer reviewers from This and other volumes analyzing key infra-
policy-making and academic circles in Africa and structure topics, as well as the underlying data
beyond reviewed all major outputs of the study sources described above, will be available for
to ensure the technical quality of the work. download from http://www.infrastructure
The Sub-Saharan Africa Transport Pol- africa.org. Stand-alone summaries are avail-
icy Program (SSATP) and the Water and able in English and French.
Sanitation Program (WSP) provided technical Inquiries concerning the availability of data
support on data collection and analysis per- sets should be directed to the volume editors
taining to their respective sectors. at the World Bank in Washington, DC.
Acknowledgments
This report was undertaken by the director’s (professor, John F. Kennedy School of Govern-
office of the Department for Sustainable Devel- ment, Harvard University), Cheikh Kane (inde-
opment in the Africa Region of the World Bank. pendent expert on infrastructure finance), and
A number of directors oversaw the implementa- Xinzhu Zhang (professor, Chinese Academy of
tion of the project throughout its life, including Social Sciences, Beijing).
(in chronological order) Michel Wormser, John In order to ensure broad-based par-
Henry Stein (acting), and Inger Andersen. ticipation and consultation of World Bank
The task team leaders for the report were technical practices, a number of internal
Vivien Foster and Cecilia Briceño-Garmendia, peer review groups were formed to provide
and the core team for the project comprised guidance and feedback on earlier drafts of
Aijaz Ahmad, Dominique Akele, Sudeshna the document. The individual groups and
Ghosh Banerjee, Carolina Dominguez Torres, their members are as follows: ICT sector—
Sophie Hans-Moevi, Elvira Morella, Nataliya Mavis Ampah, Philippe Dongier, Clemencia
Pushak, Rupa Ranganathan, Maria Shkaratan, Torres, and Mark Williams; irrigation sector—
and Karlis Smits. Barbara Miller, Stephen Mink, and Ashok
The project team is grateful to a number of Subramanian; power sector—Philippe Benoit,
World Bank colleagues who acted as advisers on David Donaldson, Vijay Iyer, Luiz Maurer,
key cross-cutting aspects of the report. These Rob Mills, Lucio Monari, Kyran O’Sullivan,
include Antonio Estache, Jose Luis Irigoyen, Prasad Tallapragada, Clemencia Torres, and
and Jyoti Shukla, who provided advice on gen- Tjaarda Storm Van Leeuwen; transport sector—
eral infrastructure issues; Sarah Keener, who Pierre Pozzo di Borgo, Michel Luc Donner,
provided advice on social issues; Paul Martin, Michel Iches, Marc Juhel, Cornelis Kruk,
who provided advice on environmental issues; Alain Labeau, Charles Schlumberger, and
and Stephen Mink, who provided advice on Kavita Sethi; water supply and sanitation
rural and agricultural issues. sector—Ventura Bengoechea, Jaime Biderman,
A technical advisory panel provided indepen- Matar Fall, Sarah Keener, Peter Kolsky, Alex
dent, external peer review on the quality of the McPhail, Eustache Ouayoro, Christophe Pre-
background papers on which this report is based. vost, Caroline van den Berg, and Meike van
The panel was cochaired by Shanta Devarajan Ginneken; finance theme—Gerardo Corro-
(chief economist, Africa Region, World Bank) chano, Michael Fuchs, James Leigland, Anand
and Louis Kasekende (chief economist, African Rajaram, Sudhir Shetty, Jyoti Shukla, Clem-
Development Bank), and comprised Adeola encia Torres, Marilou Uy, and Marinus Ver-
Adenikinju (professor, University of Ibadan, hoeven; poverty and inequality theme—Judy
Nigeria), Emmanuelle Auriol (professor, Univer- Baker, Douglas Barnes, Ellen Hamilton, Julian
sity of Toulouse, France), Tony Gomez-Ibanez Lampietti, and Kenneth Simler; institutional
xxi
xxii Acknowledgments
theme—James Leigland and Jyoti Shukla; An editorial team comprising Bruce Ross-
urban theme—Jaime Biderman, Catherine Larson, Steven Kennedy, and Joseph Caponio
Farvacque-Vitkovic, Matthew Glasser, Sumila contributed significantly to improving the
Gulyani, and Uri Raich; and regional integra- quality of the final manuscript submitted to
tion theme—Uwe Deichmann, Jakob Kolster, the World Bank Office of the Publisher for
and Mark Tomlinson. publication.
Abbreviations
Africa’s Infrastructure:
A Time for Transformation
T
he Africa Infrastructure Country Diag- diseconomies of scale in production and
nostic is an unprecedented attempt to high profit margins caused by lack of
collect comprehensive data on the infra- competition.
structure sectors in Africa—covering power, • Power is by far Africa’s largest infrastructure
transport, irrigation, water and sanitation, and challenge, with 30 countries facing regular
information and communication technology power shortages and many paying high pre-
(ICT)—and to provide an integrated analysis miums for emergency power.
of the challenges they face. Based on extensive
fieldwork across Africa, the following main
• The cost of addressing Africa’s infrastruc-
ture needs is around $93 billion a year, about
findings have emerged: one-third of which is for maintenance—
• Infrastructure has been responsible for more than twice the Commission for Afri-
more than half of Africa’s recent improved ca’s (2005) estimate.
growth performance and has the potential • The infrastructure challenge varies greatly
to contribute even more in the future. by country type—fragile states face an
• Africa’s infrastructure networks increas- impossible burden and resource-rich coun-
ingly lag behind those of other developing tries lag despite their wealth.
countries and are characterized by miss- • A large share of Africa’s infrastructure is
ing regional links and stagnant household domestically financed, with the central gov-
access. ernment budget being the main driver of
• Africa’s difficult economic geography pre- infrastructure investment.
sents a particular challenge for the region’s • Even if major potential efficiency gains are
infrastructure development. captured, Africa would still face an infra-
• Africa’s infrastructure services are twice structure funding gap of $31 billion a year,
as expensive as elsewhere, reflecting both mainly in power.
1
2 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
• Africa’s institutional, regulatory, and admin- For most countries, the negative effect of defi-
istrative reforms are only halfway along, cient infrastructure is at least as large as that
but they are already proving their effect on of crime, red tape, corruption, and financial
operational efficiency. market constraints. For one set of countries,
power emerges as the most limiting factor by
far, cited by more than half the firms in more
Finding 1: Infrastructure than half the countries as a major business
Contributed over Half of Africa’s obstacle. For a second set, inefficient function-
Improved Growth Performance ing of ports and associated customs clearance
is equally significant. Deficiencies in transport
Africa’s growth improved markedly in the last and in ICTs are less prevalent but substantial
decade. African countries saw their econo- in some cases.
mies grow at a solid 4 percent a year from Infrastructure not only contributes to eco-
2001 to 2005. Resource-rich countries, which nomic growth, but it is also an important input
have benefited from rising commodity prices, to human development (Fay and others 2005).
demonstrate the highest growth rates. Growth Infrastructure is a key ingredient for achieving
overall still falls short of the 7 percent needed all the MDGs. Safe and convenient water sup-
to achieve substantial poverty reduction and plies save time and arrest the spread of a range
attain the Millennium Development Goals of serious diseases—including diarrhea, a lead-
(MDGs), however. Infrastructure, significant ing cause of infant mortality and malnutrition.
in Africa’s economic turnaround, will need to Electricity powers health and education services
play an even greater role for the continent to and boosts the productivity of small businesses.
reach its development targets. Road networks provide links to global and local
Across Africa, infrastructure contributed markets. ICTs democratize access to informa-
99 basis points to per capita economic growth tion and reduce transport costs by allowing
from 1990 to 2005, compared with 68 basis people to conduct transactions remotely.
points for other structural policies (Calderón
2008). That contribution is almost entirely
attributable to advances in the penetration Finding 2: Africa’s Infrastructure
of telecommunication services. The deterio- Lags Well behind That of Other
ration in the quantity and quality of power Developing Countries
infrastructure over the same period retarded
growth, shaving 11 basis points from per cap- On just about every measure of infrastructure
ita growth for Africa as a whole and as much as coverage, African countries lag behind their
20 basis points for southern Africa. peers in the developing world (Yepes, Pierce,
The growth effects of further improving and Foster 2008). This lag is perceptible for low-
Africa’s infrastructure would be even greater. and middle-income countries in Sub-Saharan
Simulations suggest that if all African coun- Africa relative to other low- and middle-income
tries were to catch up with Mauritius (the countries (table O.1). The differences are par-
regional leader in infrastructure) per capita ticularly large for paved roads, telephone main
growth in the region could increase by 2.2 per- lines, and power generation. For all three, Africa
centage points. Catching up with the Republic has been expanding stocks much more slowly
of Korea would increase per capita growth by than other developing regions; so unless some-
2.6 percentage points a year. In Côte d’Ivoire, thing changes, the gap will continue to widen.
the Democratic Republic of Congo, and Sen- To what extent does Africa’s current deficit
egal, the effect would be even larger. date to a low starting point for infrastructure
In most African countries, particularly stocks? Africa started out with stocks that
the lower-income countries, infrastructure were generally not very different from those
emerges as a major constraint on doing busi- in South or East Asia in the 1960s for roads,
ness, depressing firm productivity by about in the 1970s for telephones, and in the 1980s
40 percent (Escribano, Guasch, and Pena 2008). for power. The comparison with South Asia,
Africa’s Infrastructure: A Time for Transformation 3
which has similar per capita incomes, is par- universal access to these and other household
ticularly striking. In 1970, Sub-Saharan Africa services is more than 50 years away in most
had almost three times the generating capac- African countries (Banerjee, Wodon, and oth-
ity per million people as South Asia. In 2000, ers 2008). Even where infrastructure networks
South Asia had left Sub-Saharan Africa far are in place, a significant percentage of house-
behind—with almost twice the generation holds remains unconnected, suggesting that
capacity per million people. Also in 1970, demand-side barriers exist and that univer-
Sub-Saharan Africa had twice the main-line sal access entails more than physical rollouts
telephone density of South Asia, but by 2000, of networks. As might be expected, access to
the two regions were even. infrastructure in rural areas is only a frac-
Since 1990, coverage of household services tion of that in urban areas, even where urban
has barely improved (figure O.1, panel a). coverage is already low by international stan-
Africa is unlikely to meet the MDGs for water dards (Banerjee, Wodon, and others 2008)
and sanitation. Moreover, on current trends, (figure O.1, panel b).
30 60
% of population
% of population
20 40
10 20
0 0
1990–95 1996–2000 2001–05 piped water electricity flush toilets landline
telephones
piped water electricity
national rural urban
flush toilets landline telephones
Africa’s atomized nation-states are reflected Africa’s water resources are abundant,
in the region’s fragmentary infrastructure but because of an absence of water stor-
networks. Sub-Saharan Africa comprises 48 age and distribution infrastructure, they are
nation-states, many of which are very small. grossly underused. Therefore, water security—
The bulk of those countries have populations eliable water supplies and acceptable risks
of fewer than 20 million and economies smaller from floods and other unpredictable events,
than $10 billion. International frontiers bear including those from climate change—will
little relation either to natural features (such require a significant expansion of water
as river basins) or to artificial features (such as storage capacity from the current 200 cubic
cities and their accessibility to trading chan- meters per capita (Grey and Sadoff 2006). In
nels, such as ports). Intraregional connectiv- other parts of the world, such capacity is in
ity is therefore very low, whether measured in the thousands of cubic meters. The cost of
transcontinental highway links, power inter- expanding water storage is extremely high
connectors, or fiber-optic backbones. Most in relation to the size of Africa’s economies,
continuous transport corridors are concerned suggesting the phasing of investments, with
with providing access to seaports, whereas the initial focus on achieving water security for
intraregional road network is characterized by key growth poles.
major discontinuities. Few cross-border inter- Water also needs to be distributed for agri-
connectors exist to support regional power cultural use. In a handful of countries, only
exchange, even though many countries are too 7 million hectares are equipped for irrigation.
small to produce power economically on their Although the irrigation-equipped area is less
own. Until recently, the whole of East Africa than 5 percent of Africa’s cultivated area, it
lacked access to a global submarine cable to pro- produces 20 percent of the value of agricultural
vide low-cost international communications production. An additional 12 million hectares
and Internet access. The intraregional fiber- could be economically viable for irrigation as
optic network is also incomplete, but growing long as costs are contained (You 2008).
rapidly. Because of their geographic isolation,
landlocked countries in particular suffer from
the lack of regional connectivity. Finding 4: Africa’s Infrastructure
Both the spatial distribution and rapid Services Are Twice as Expensive
migration of Africa’s population create major as Elsewhere
challenges for reaching universal access. In rural
areas, over 20 percent of the population lives Not only are Africa’s infrastructure networks
in dispersed settlements where typical popula- deficient in coverage, but the price of the
tion densities are less than 15 people per square services provided is also exceptionally high
kilometer; hence, the costs of providing infra- by global standards (table O.2). Whether for
structure are comparatively high. In urban areas, power, water, road freight, mobile telephones,
population growth rates averaging 3.6 percent a or Internet services, the tariffs paid in Africa
year are leaving infrastructure service provid- are several multiples of those paid in other
ers severely stretched. As a result, urban service parts of the developing world. The explana-
coverage has actually declined over the last tion for Africa’s higher prices sometimes lies
decade, and lower-cost alternatives are filling the in genuinely higher costs, and sometimes in
resulting gap (Banerjee, Wodon, and others 2008; high profits. The policy prescriptions for the
Morella, Foster, and Banerjee 2008). In addition, two cases are, of course, radically different.
population densities in African cities are rela- Power provides the clearest example of
tively low by global standards and do not benefit infrastructure with costs genuinely higher in
from large economies of agglomeration in the Africa than elsewhere. Many smaller coun-
provision of infrastructure services. As a result, tries have national power systems below the
the costs of providing a basic infrastructure 500-megawatt threshold and therefore often
package can easily be twice as much as in other rely on small diesel generation that can cost
developing cities (Dorosh and others 2008). up to $0.35 per kilowatt-hour to run, about
Africa’s Infrastructure: A Time for Transformation 5
Table O.2 Africa’s High-Cost Infrastructure higher than those without (Minges and oth-
Other ers 2008).
Sub-Saharan developing
Infrastructure sector Africa regions
Power tariffs
($ per kilowatt-hour) 0.02–0.46 0.05–0.10
Finding 5: Power Is Africa’s
Water tariffs
($ per cubic meter) 0.86–6.56 0.03–0.60
Largest Infrastructure Challenge
Road freight tariffs by Far
($ per ton-kilometer) 0.04–0.14 0.01–0.04
Mobile telephony Whether measured in generation capacity,
($ per basket per month) 2.60–21.00 9.90 electricity consumption, or security of sup-
International telephony ply, Africa’s power infrastructure delivers only
($ per 3-minute call to a fraction of the service found elsewhere in
the United States) 0.44–12.50 2.00
the developing world (Eberhard and others
Internet dial-up service
($ per month) 6.70–148.00 11.00 2008). The 48 Sub-Saharan Africa countries
Sources: Authors’ estimates based on Africon 2008; Bannerjee,
(with 800 million people) generate roughly the
Skilling, and others 2008; Eberhard and others 2008; Minges and same power as Spain (with 45 million people).
others 2008; Teravaninthorn and Raballand 2008; Wodon 2008a
and 2008b. Power consumption, at 124 kilowatt-hours
Note: Ranges reflect prices in different countries and various per capita annually and falling, is only 10 per-
consumption levels. Prices for telephony and Internet service
represent all developing regions, including Africa. cent of that found elsewhere in the developing
world, barely enough to power one 100-watt
lightbulb per person for 3 hours a day.
twice the costs faced by larger countries typi- More than 30 African countries experience
cally with coal- or hydropower-based systems power shortages and regular interruptions to
(Eberhard and others 2008). service (figure O.2). The underlying causes
High road freight tariffs in Africa have vary: failures to bring on new capacity to keep
much more to do with high profit margins pace with the demands of economic growth,
than high costs (Teravaninthorn and Rabal- droughts that reduced hydropower in East
land 2008). The costs for Africa’s trucking Africa, oil price hikes that inhibited afford-
operators are not much higher than costs in ability of diesel imports for many West African
other parts of the world, even when informal countries, and conflicts that destroyed power
payments are counted. Profit margins, by con- infrastructure in fragile states. Africa’s firms
trast, are exceptionally high, particularly in report losing 5 percent of their sales because of
Central and West Africa, where they reach 60 frequent power outages—a figure that rises to
to 160 percent. The underlying cause is limited 20 percent for informal firms unable to afford
competition combined with a highly regulated backup generation. Overall, the economic
market based on tour de role principles, which costs of power outages can easily rise to 1–2
allocate freight to transporters through a cen- percent of GDP.
tralized queuing method rather than allowing A common response to the crisis is to ten-
truckers to enter into bilateral contracts with der short-term leases for emergency power. At
customers directly. least 750 megawatts of emergency generation
The high costs of international telephony are operating in Sub-Saharan Africa, which
and Internet services reflect a mixture of cost for some countries constitute a large pro-
and profit factors. Countries without access portion of their national installed capacity.
to a submarine cable must rely on expensive However, emergency generation is expensive
satellite technology for international connec- at costs of $0.20–$0.30 per kilowatt-hour,
tivity and have charges typically twice those and for some countries, the price tag can be
in countries that do enjoy such access. Even as high as 4 percent of GDP. Paying for emer-
when access to a submarine cable is secured, gency leases absorbs significant budgetary
countries with a monopoly on this interna- resources, reducing the funds for longer-term
tional gateway still have tariffs substantially solutions.
6 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
MAURITANIA
CAPE VERDE MALI
NIGER
SE CHAD SUDAN ERITREA
NE
GA
GAMBIA L BURKINA DJIBOUTI
FASO
BENIN
GUINEA-BISSAU GUINEA
NIGERIA
GHANA
D'IVOIRE
CÔTE
SIERRA LEONE ETHIOPIA
CENTRAL
ON
AFRICAN REPUBLIC
RO
LIBERIA
IA
ME
AL
TOGO
DEMOCRATIC
CA
M
SO
GO
EQUATORIAL GUINEA REPUBLIC OF UGANDA
CON
CONGO
GABON KENYA
SAO TOME AND PRINCIPE RWANDA
. OF
ES
LL
REP
BURUNDI HE
C
TANZANIA
SEY
MALAWI
COMOROS
Main cause or trigger MAYOTTE
ANGOLA
natural causes (droughts) ZAMBIA
E
IQU
MB
oil price shock
AR
ZA
ZIMBABWE
ASC
MO
system disrupted by conflict NAMIBIA MAURITIUS
DAG
high growth, low investment/structural issues BOTSWANA
MA
SWAZILAND
SOUTH
AFRICA
LESOTHO
requirements (table O.3; Briceño-Garmendia, inadequate competition for tenders have all
Smits, and Foster 2008). played their role, with the last factor by far
That cost is well over twice the $39 billion of the strongest.
infrastructure spending estimated by the Com- The global financial crisis of 2008 can be
mission for Africa report in 2005. That figure expected to reduce demand for some types of
was based on a cross-country econometric infrastructure, but it would not hugely alter
study, rather than the more detailed country- the estimated spending needs. Planning and
level microeconomic modeling (Estache 2005). social targets rather than economic growth
A more recent update of the cross-country drive a large share of the spending needs, for
model used for the Commission for Africa example, the transport spending needs (which
report came up with revised estimates in the are largely based on connectivity objectives)
range of $80 billion to $90 billion, much closer and the water and sanitation spending needs
to those reported here (Yepes 2007). (which are based on the MDGs). The spending
About 40 percent of the total spending needs with the strongest direct link to economic
needs are associated with power, reflecting growth are those for the power sector. However,
Africa’s particularly large deficits. About one- because of the large investment backlog in the
third of the power investment needs (some sector, the estimated spending needs contain a
$9 billion a year) are associated with multipur- strong component of refurbishment and catch-
pose water storage for hydropower and water up. Thus, even halving economic growth esti-
resource management. After power, water sup- mates for the region would reduce estimated
ply and sanitation and then transport are the power spending needs by only 20 percent.
most significant items. The global recession could also be expected to
Given recent escalations in unit costs, these affect demand for ICT services, as well as trade-
estimates are a lower bound. Although the related infrastructure, such as railways and
investment estimates here are based on the most ports. However, the weight of these infrastruc-
accurate unit-cost data available, develop- tures in the total spending needs is not much
ment agencies are reporting significant cost more than 10 percent.
escalations on projects under implementa-
tion. For road projects, these escalations have
averaged 35 percent but in some cases have
been as high as 50–100 percent. Closer inspec-
Finding 7: The Infrastructure
tion reveals that no single factor explains this Challenge Varies Greatly by
escalation. Domestic inflation, tight construc- Country Type
tion industry conditions, oil price hikes, and
The infrastructure challenge differs mark-
edly across African country groups (Briceño-
Table O.3 Overall Infrastructure Spending Needs for Garmendia, Smits, and Foster 2008). Because
Sub-Saharan Africa
of the widely varying circumstances, distin-
$ billions annually
guishing among middle-income countries
Operation
Infrastructure Capital and Total (like Cape Verde and South Africa), resource-
sector expenditure maintenance spending rich countries with economies heavily reliant
ICT 7.0 2.0 9.0 on petroleum or mineral revenues (like Nige-
Irrigation 2.9 0.6 3.4 ria and Zambia), fragile states emerging from
conflict (like Côte d’Ivoire and the Demo-
Power 26.7 14.1 40.8
cratic Republic of Congo), and the remaining
Transport 8.8 9.4 18.2
low-income countries that are neither fragile
WSS 14.9 7.0 21.9 nor resource rich (like Senegal and Uganda)
Total 60.4 33.0 93.3 is helpful.
Source: Authors’ estimates based on Banerjee, Wodon, and By far the most daunting infrastructure
others 2008; Carruthers, Krishnamani, and Murray 2008; Mayer
and others 2008; Rosnes and Vennemo 2008. challenges are those facing the fragile states
Note: Column totals may not add exactly because of rounding (figure O.3). The recent conflicts affecting these
errors. ICT = information and communication technology; WSS =
water supply and sanitation. countries usually resulted in the destruction
8 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure O.3 Burden of Infrastructure Spending Needs have not tended to do so. Resource-rich coun-
40
tries could meet their infrastructure spending
35 needs for a more manageable price tag of about
30 12 percent of GDP. Moreover, the large roy-
alty payments they received during the recent
% of GDP
25
20 commodity boom provide a ready source of
15 finance. Yet resource rich-countries actually
10 lag nonfragile low-income countries in their
5 infrastructure stocks and spend less on infra-
0 tri e
structure. They have been devoting their added
tri e
tri h
tri e
a
un m
un m
un m
un ric
ric
wealth not to infrastructure development but to
co inco
co co
co inco
es
es
es
es
co rce-
Af
-in
n
-
le-
u
ra
ow
ow
idd
ha
re
el
el
Sa
m
gil
gil
b-
resource-rich environment may thus prevent the
fra
ra
Su
nf
no
finance are substantially higher relative to GDP Figure O.4 Infrastructure Public Spending as a
in the low-income states, typically absorbing Percentage of GDP
5–6 percent of total GDP (figure O.4). In abso- 7.0
lute terms, however, spending remains very
6.0
low, no more than $20–$30 per capita a year
(Briceño-Garmendia, Smits, and Foster 2008). 5.0
Looking only at investment, one finds that
% of GDP
4.0
official development assistance, private partici-
pation in infrastructure, and non-OECD finan- 3.0
ciers together exceed domestically financed
2.0
public investment (Briceño-Garmendia, Smits,
and Foster 2008). The private sector is by far the 1.0
largest source, on a par with domestic public
0
investment. Much smaller, but still significant,
tri e
tri e
es
tri e
a
un m
un m
un m
ric
tri
co inco
co inco
co inco
es
es
es
Af
n
co
le-
ra
ow
ow
idd
ch
ha
el
-ri
Sa
m
gil
gil
ce
b-
ra
fra
Su
nf
so
no
re
Although some of this “overspending” may cost of sound preventive maintenance. For
be justified by phasing or sequencing, at least example, spending $1 on road maintenance
part of these resources could possibly be real- provides a savings of $4 to the economy. So
located to underfunded sectors. A need exists some reallocation of resources from invest-
to monitor infrastructure expenditure more ment to maintenance may be warranted, par-
closely against identified needs and priorities ticularly in low-income countries with very low
and considering expected economic returns. maintenance spending. For roads, an estimated
Second, African countries are typically $1.9 billion of capital spending on rehabilita-
executing only about two-thirds of the budget tion could have been avoided with sound pre-
allocated to public investment in infrastruc- ventive maintenance.
ture (Briceño-Garmendia, Smits, and Foster Fourth, Africa’s power and water utilities
2008). Put differently, public investment could present very high inefficiency in distribution
in theory increase by 30 percent without any losses, undercollection of revenues, and over-
increase in spending, simply by addressing the staffing (figure O.6). Utilities typically collect
institutional bottlenecks that inhibit capital only 70–90 percent of billed revenues, and dis-
budget execution. Changes include better plan- tribution losses can easily be twice the technical
ning of investment projects, earlier completion best practice. According to household surveys,
of feasibility studies, more efficient procure- about 40 percent of those connected to utility
ment processes, and a move to medium-term services do not appear to be paying for them,
multiyear budgeting. Increasing capital budget a share that rises to 65 percent for a significant
execution to 100 percent could capture an addi- minority of countries. Undercollection is also
tional $1.9 billion a year in public investment. a problem for some of Africa’s road funds
Third, on average, about 30 percent of the (Gwilliam and others 2008). State-owned tele-
infrastructure assets of a typical African coun- communication incumbents employ roughly
try need rehabilitation (figure O.5). This share six times the number of employees per con-
is even higher for rural infrastructure and for nection than do privately operated enterprises
countries affected by violent conflict. The reha- in developing countries. For ICT, countries
bilitation backlog reflects a legacy of under- retaining state-owned incumbents are often
funding maintenance, a major waste given incurring significant losses from overstaffing
that the cost of rehabilitating infrastructure that average 0.2 percent of GDP. Similarly,
is several times higher than the cumulative though to a lesser extent, overemployment
in power and water utilities ranges from
20 percent to 80 percent over benchmarks in
other developing areas. Overall, the revenues
Figure O.5 Rehabilitation Backlog
lost through these inefficiencies can easily
50 exceed the current turnover of the utilities by
several multiples. For power, these losses are
% of assets in need of rehabilitation
ge
ge
er
er
ge
ds
ad
ay
tio
at
oa
ra
ra
ra
wa
at
ilw
ro
iga
lw
e
ve
lr
er
av
ain
ra
n
la
ra
irr
ra
en
ba
al
all
ru
ru
ra
m
rg
ur
ur
er
ru
nr
ov
no
po
a. Power b. Water
5 5
4 4
% of GDP
% of GDP
3 3
2 2
1 1
0 0
tri e
es
tri e
tri e
a
tri e
es
tri e
tri e
un m
un m
un m
ric
un m
un m
un m
ric
tri
tri
co inco
co inco
co inco
es
es
es
co inco
co inco
co inco
Af
es
es
es
Af
un
un
n
co
n
le-
-
co
le-
ra
ow
ow
ra
ow
ow
idd
ch
ha
idd
ch
ha
el
el
el
el
-ri
Sa
-ri
Sa
m
m
gil
gil
gil
gil
ce
b-
ce
b-
ra
fra
ra
fra
ur
Su
ur
Su
nf
nf
so
so
no
no
re
re
Table O.5 Finding Resources: The Efficiency Gap and the Funding Gap
$ billions annually
Cross-sector
Item Electricity ICT Irrigation Transport WSS gain Total
Infrastructure spending needs (40.8) (9.0) (3.4) (18.2) (21.9) n.a. (93.3)
Existing spending 11.6 9.0 0.9 16.2 7.6 n.a. 45.3
Efficiency gap 6.0 1.3 0.1 3.8 2.9 3.3 17.4
Gain from raising capital execution 0.2 0.0 0.1 1.3 0.2 n.a. 1.9
Gain from eliminating operational
inefficiencies 3.4 1.2 — 1.9 1.0 n.a. 7.5
Gain from tariff cost recovery 2.3 — — 0.6 1.8 n.a. 4.7
Potential for reallocation n.a. n.a. n.a. n.a. n.a. 3.3 3.3
Funding gap (23.2) 1.3 (2.4) 1.9 (11.4) 3.3 (30.6)
Source: Briceño-Garmendia, Smits, and Foster 2008.
Note: ICT = information and communication technology; n.a. = not applicable; — = not available; WSS = water supply and sanitation.
Parentheses indicate negative values.
2 % of GDP 15
10
1
5
0 0
r
at ly
rt
ch a nd
tri e
nt w-
tri h
tri e
we
un m
un m
tio
un ric
po
nit pp
te mun n a
ou lo
co inco
co inco
ion
log n
es
es
es
co rce-
po
iga
ns
rie
sa su
no tio
e c gile
m tio
y
tra
irr
d er
le-
u
ow
co rma
om fra
so
an at
idd
re
w
el
inc non
o
m
inf
gil
fra
fully exploit efficiency gains. One cannot say Finding 10: Africa’s Institutional,
the same of fragile states, however. They would Regulatory, and Administrative
still require a substantial increase in spending
Reform Process Is Only
to meet the investment targets in any reason-
able time frame, even when inefficiencies are Halfway Along
fully captured.
Another possibility is to adopt lower-cost During the last decade, African states have
technologies to trim investment needs. Sav- made concerted efforts toward institutional
ings of approximately one-third of spending reform in infrastructure. One could probably
requirements in transport and in water and san- fairly say that the institutional reform process
itation are achievable in this way, by adopting is halfway along (Vagliasindi and Nellis 2009).
lower-cost road designs or lower-end solutions They have made progress, but few countries
for water and sanitation (such as standposts have a modern institutional framework for
and improved latrines). Countries face a stark these sectors. Overall, the greatest progress has
trade-off between the level of service provided been in telecommunications, whereas trans-
and the speed with which they can serve their port lags furthest behind (figure O.8). The
entire population. focus also varies. In telecommunications, the
Africa’s Infrastructure: A Time for Transformation 13
emphasis has been on implementing sector businesses delivers cash flows high enough to
reform, and in water on improving the gover- finance investment. However, these arrange-
nance of state-owned enterprises. ments have often (though not always) been
Private participation has varied enormously good for operational performance, even if
(Vagliasindi and Nellis 2009). Since the mid- characterized by renegotiation and premature
1990s, many African countries have experi- cancellation. A growing area of experimenta-
mented with various forms of private participa- tion is the multiyear performance-based road
tion in infrastructure, with very heterogeneous maintenance contract with the private sector,
results (table O.6). which shows promise in safeguarding mainte-
The private sector has proved willing to nance activities and keeping costs down.
invest only in mobile telephones, power plants, Some progress has occurred with gov-
and container terminals. The number of ernance reform of state-owned enterprises,
mobile subscribers and the share of the popu- where incentive-based performance contracts
lation receiving mobile signals increased by a and external auditing seem to be paying off.
factor of 10 in five years, the result of compe- Corporate governance reforms, including the
tition among private operators. Private inves- establishment of a somewhat independent
tors have also provided significant finance for board of directors, are becoming more prev-
thermal power generation (3,000 megawatts) alent across sectors, even if few enterprises
and for container terminals at ports, even if have full corporatization that includes limited
the volumes fall substantially short of require- liability, rate of return, and dividend poli-
ments. Toll-road concessions are confined to cies. Performance contracts with incentives
South Africa; traffic volumes elsewhere are and independent external audits have become
not enough to make such endeavors financially dominant features of the reform process for
self-sustaining. governance of state-owned enterprises, for
In power, water, and railways, the pri- both electricity and water. When combined
vate sector has delivered improvements in with managerial performance incentives, these
operational performance but no new finance. measures seem to be having a material effect
The numerous concessions (and related con- on performance. The introduction of inde-
tractual forms) covering railways, power, and pendent audits has also increased efficiency,
water distribution have not delivered signifi- for both electric and water utilities.
cant investment. Because of a combination Evidence on the links between introduc-
of low tariffs and low volumes, none of these ing an independent regulator and improving
14 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Table O.6 Overview of Private Participation in Infrastructure resulting from excessive discretion and overly
Infrastructure Extent of private Nature of broad objectives (Eberhard 2007). Regulatory
sector participation experience Prospects autonomy remains elusive: in some countries,
ICT turnover among commissioners has been high,
Mobile telephony Over 90 percent of Extremely beneficial Several countries still and the gap between law (or rule) and practice
countries have licensed with exponential have potential to grant has been wide. For water, where the vast major-
multiple mobile operators increase in coverage additional licenses
and penetration ity of service providers are state-owned enter-
Fixed telephony About 60 percent of Controversial in some Several countries still prises, no evidence exists of any benefit from
countries have cases, but have potential to regulation. For power and telecommunications,
divested state-owned has helped improve undertake divestitures some effect is discernible, but it is far from
telecommunication overall sector efficiency
incumbent unambiguous. Weak regulatory autonomy and
Power capacity constraints undermine the credibility
Power generation 34 independent power Few cancellations but Likely to continue,
of independent regulators. Most African regu-
projects provide 3,000 frequent given huge unsatisfied latory agencies are embryonic, lacking funding
MW of new capacity, renegotiations; power demands and limited and in many cases qualified personnel.
investing $2.5 billion purchase agreements public sector capacity
have proved costly for
utilities
Key Recommendations
Power distribution 16 concessions and 17 Problematic and Movement toward hybrid
management or lease controversial; models involving local
contracts in 24 countries one-quarter of private sector in similar Based on these findings, one can make the fol-
contracts cancelled frameworks
before completion
lowing 10 key recommendations:
Transport • Addressing Africa’s infrastructure efficiency
Airports Four airport No cancellations but Limited number of gap is a pressing policy priority with poten-
concessions, investing some lessons learned additional airports tial dividends of $17 billion a year.
less than $0.1 billion, viable for concessions
plus some divestitures • One of the most flagrant inefficiencies is the
Ports 26 container terminal Processes can be Good potential to failure to maintain infrastructure assets—
concessions, investing controversial, but continue maintenance needs to be understood as an
$1.3 billion cancellations have investment in asset preservation.
been few and
results positive • Institutional reform remains essential for
Railroads 14 railroad concessions, Frequent renegotia- Likely to continue but tackling utilities’ operational inefficien-
investing $0.4 billion tions, low traffic, and model needs to be cies, both through private participation
costly public service adapted
obligations keep and through governance reforms for state-
investment below owned enterprises.
expectations
Roads 10 toll-road projects, No cancellations Limited because only
• Institutional reform should also go beyond
almost all in South Africa, reported 8 percent of road utilities to strengthen the planning func-
investing $1.6 billion network meets minimum tions of the line ministries and address seri-
traffic threshold, almost ous deficiencies in the budgetary process.
all in South Africa
Water • Reforms are needed to get full value from
existing infrastructure, where widespread
Water 26 transactions, mainly Problematic and Movement toward hybrid
management or lease controversial; models involving local administrative and regulatory bottlenecks
contracts 40 percent of private sector in similar prevent facilities from being fully used.
contracts cancelled frameworks
before completion • Regional integration can contribute signifi-
Sources: Authors’ elaboration based on Bofinger 2009; Bullock 2009; Eberhard and others 2008; cantly to reducing infrastructure costs, by
Gwilliam and others 2008; Minges and others 2008; Mundy and Penfold 2008; and Svendsen, allowing countries to capture scale econo-
Ewing, and Msangi 2008.
Note: ICT = information and communication technology; MW = megawatts. mies and manage regional public goods
effectively.
performance is currently mixed (Vagliasindi • Development of infrastructure networks
and Nellis 2009). Some critics argue that regu- needs to be strategically informed by the
latory agencies have simply created additional spatial distribution of economic activities
risks because of unpredictable decisions, and by economies of agglomeration.
Africa’s Infrastructure: A Time for Transformation 15
to insulate utilities from political interference A clear example is power generation. Tra-
while closely monitoring enterprises. Improving ditionally, planning and procurement of new
regulatory performance is a long-term process power infrastructure were the province of the
to be pursued where private participation and state-owned utility. With power sector reforms
competitive pressures are significant. The chal- and independent power producers, those func-
lenge of establishing new public institutions in tions were often moved to the ministry of
developing countries is often underestimated. energy or electricity. The transfer of skills was
Independent regulation requires a strong polit- not always simultaneous, however, so plans
ical commitment and competent institutions were not adequately informed by the complex-
and people. Where some or all are lacking, con- ities on the ground. In many cases, planning
sidering complementary or transitional options has collapsed. New plants are rarely timely,
that reduce discretion in regulatory decision thereby opening power gaps that prompt
making through more explicit rules and proce- recourse to temporary power and discour-
dures or by outsourcing regulatory functions to age investors. When procurement is (finally)
advisory regulators and expert panels may be undertaken, the authorities may not take the
wise (Eberhard 2007). trouble to conduct international competitive
bidding. This outcome is unfortunate because
Recommendation 4: Include Line a rigorous bidding process lends credibility
Ministries and Budgetary Processes on and transparency to procurement and results
the Institutional Reform Agenda in more competitively priced power.
Much of the emphasis of recent reforms has Because domestic public spending finances
been on restructuring the service provider or the bulk of Africa’s infrastructure investments,
utility, bringing in private management, applying development partners need a broader view
regulatory oversight, and so on. Little attention of the quality of public spending. Across the
has been given to institutional strengthening of infrastructure sectors, most investments are by
the sector line ministries. These line ministries line ministries through the budgetary process.
have responsibilities, which, if not adequately Shortcomings in the way the rest of the sec-
discharged, can jeopardize the functioning tor budget is allocated and spent may offset
of the sector. They take the lead in sector development finance that focuses too narrowly
planning, participate in the formulation of the on specific project interventions. So donor
public budgets, and execute investments. How- resources are best channeled programmatically
ever, deficiencies exist in all those areas. Unless as budgetary support or through sectorwide
they are tackled head on, the effect of reforms projects, and development partners need to
on service providers will remain limited. take a broader interest in the overall quality of
Stronger sector planning is needed in infra- public spending. Thus, infrastructure interven-
structure line ministries to ensure that the tions must be grounded in a broader under-
construction of critical new assets begins early standing of the public expenditure framework
enough to come on stream when needed. Too in each sector.
often overlooked or debilitated during the Ad hoc political priorities with little or no
course of sector restructuring efforts, plan- economic screening too often characterize the
ning is a critical sector function. It is essential budgetary process. The annual budget cycle
to restore this vital planning capability in the prevents adequate follow-through on the fund-
line ministries and to develop sound techni- ing of multiyear infrastructure projects. When
cal methodologies for identifying and selecting it comes to implementation, many countries
infrastructure projects. More rigorous project have significant problems with budgetary
screening can ensure that infrastructure invest- execution, with procurement bottlenecks pre-
ments are selected according to their expected venting the full budget allocations from mate-
returns and are appropriately sequenced and rializing in actual spending.
synchronized with one another and with Key aspects of the public expenditure
broader development plans to maximize syn- framework need to be addressed. The budget-
ergies and avoid costly bottlenecks. ing process needs to move to a medium-term
Africa’s Infrastructure: A Time for Transformation 19
framework and link sector objectives and Liberalizing the trucking industry can
resource allocations, underpinned by clear reduce the exorbitant road freight costs in
sector plans that go down to specific activities Central and West Africa. The regulation and
and their associated costs. The careful incor- market structures of the road freight industry,
poration of maintenance in medium-term not the quality of road infrastructure, are the
sector-planning tools can prevent the growing binding constraints on international corridors
need for asset rehabilitation. Project appraisal (Teravaninthorn and Raballand 2008). Road
should underpin the budgetary process for freight tariffs, which can reach $0.08–$0.13
public investment to ensure that all invest- per ton-kilometer in Central and West Africa,
ments under political consideration pass at reflect the high profit margins of trucking
least a minimum threshold of economic via- services (60–160 percent). The tour de role
bility. Administrative processes that delay the regulatory framework, based on market shar-
release of budgeted funds must be overhauled, ing and centralized allocations of freight, lim-
and procedures for procurement, disbursement, its vehicle mileage and undermines incentives
financial management, and accountability to improve fleet quality. The alternative is to
must be modernized and streamlined. combine free entry to the market and market
Water provides interesting examples of pricing with regulatory enforcement of rules
how bottlenecks in the budgetary process for quality and operating behavior. Already
can prevent the use of available resources. In practiced in southern Africa, these reforms
West Africa, the binding constraint is not the can reduce road freight tariffs to $0.05 per
availability of budgetary resources in many ton-kilometer. Without such reforms, further
instances but the capacity to disburse them in investments in upgrading road network qual-
a timely fashion (Prevost 2009). In Tanzania, ity will simply lead to higher profit margins for
steep increases in budget allocations to the sec- the trucking industry without lowering trans-
tor followed water’s identification as a priority port costs for consumers.
in the country’s poverty reduction strategy, One-stop border posts are essential to avoid
but disbursements increased at a much slower extensive delays in transit traffic along interna-
pace, thus impeding any immediately discern- tional road corridors. Road conditions along
ible effect on access (Van den Berg 2009). Africa’s major international corridors are good,
Parallel improvements are also needed in the with trucks reaching speeds of 50–60 kilome-
way donor finance is channeled. Given the rele- ters an hour, but long delays at borders slow
vance of external funds, a solid public expendi- effective velocities to little more than 10 kilo-
ture management system for African countries meters an hour. A journey of 2,500 kilometers
requires that donors improve the predictability from Lusaka, Zambia, to the port of Durban
of their support and streamline and harmonize in South Africa takes on average eight days—
their procedures. In that sense, a focus on mul- four days of travel time and four days spent
tidonor initiatives that pool funds to provide at border crossings. Compare that total with
general budgetary support for a sectorwide land border-crossing times of no more than
program of interventions is preferable. half an hour for industrialized countries. The
cost of delays for an eight-axle interlink truck
Recommendation 5: Use Administrative has been estimated at about $300 a day. The
and Regulatory Reforms to Get Full investments to develop one-stop border facili-
Value from Existing Infrastructure ties and to modernize customs procedures are
Africa is failing to get the full development relatively modest and would pay back in barely
potential even from its existing infrastructure a year. Without such reforms, further invest-
networks. Administrative and regulatory fail- ments in the road network will have little effect
ures create bottlenecks and prevent infrastruc- on overall transit times.
ture assets from delivering the services they More reliable interconnection services can
are supposed to. These problems are particu- avoid even longer delays on international rail
larly evident in transport, where high-impact corridors. Locomotives from one country are
reforms are urgently needed. generally not allowed to travel on another
20 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
country’s network, mainly because of the boosted handling rates. Modernizing customs
inability to provide breakdown assistance administration requires modern information
to foreign operators. As a result, rail freight technology and associated database systems.
crossing borders must wait to be picked up Such soft infrastructure has traditionally been
by a different locomotive. These delays can be underfunded, contributing to poor port effi-
extensive. A journey of 3,000 kilometers from ciency. Governance issues may also afflict cus-
Kolwezi on the Democratic Republic of Congo toms administration.
border to the port of Durban in South Africa Port and land distribution infrastructure
takes 38 days—including 9 days of travel time need to be integrated. The lack of an inte-
and 29 days associated primarily with loading grated land distribution system, particularly
and interchange of freight. This delay partly for transit traffic, further impedes container
reflects the lack of reliable, well-maintained traffic. Making the most progress are dry and
locomotives, but it also reflects the absence of liquid bulk exports, where many port facilities
clear contractual incentives to service traffic are privately owned and integrated within a
from a neighboring country’s network. Reduc- comprehensive logistics system. Container-
ing such delays would require total rethinking ized trade, in contrast, is often only skin-deep.
of contractual relationships and access rights Containers are packed and unpacked near the
linking the railways along the corridor. It ports, and the benefits of fully integrated mul-
would also likely require a regional clearing- timodal transport corridors associated with
house to ensure transparency and fairness in container adoption are not secured. As a result,
reciprocal track access rights. little containerized traffic moves into the land-
Slow movement of containers and cargo locked hinterland, and most of those countries’
through Africa’s ports imposes very high eco- imports are transported as general cargo.
nomic costs. Many firms cite bottlenecks at Overall, the transport regulatory and
ports as their most pressing infrastructure administrative framework needs to promote
constraint in countries as diverse as Burkina seamless multimodal transportation networks
Faso, Cameroon, Malawi, Mauritius, and South more consciously. Transport chains can be no
Africa. Container dwell times in East and West stronger than their weakest links, which are
Africa are 12–15 days, twice the international usually the interchanges between different
best practice of 7 days. Most delays are caused modalities—such as road to rail or rail to sea.
by long processing and administration times The weaknesses are partly physical, where no
and poor handling in congested port areas, physical connection exists between the modes
rather than by any real limitations in basic and no infrastructure is available for transship-
quay capacity. These delays can be very costly. ment. However, they are also partly institu-
One extra day in port costs more than $35,000 tional, with responsibility for the interchanges
for a 2,200-TEU (20-foot equivalent unit) ves- not falling clearly to one modal agency or the
sel in 2006 and proportionately more for larger other. Finally, they are partly operational, with
ships. Shipping lines have responded by intro- the government collecting taxes and duties,
ducing “congestion charges”: for a 20-foot or staff collecting bribes, slowing movements,
container in 2006, ranging from $35 a day in and pushing up costs. Even at the sector policy
Dakar, Senegal, to $420 a day in Tema, Ghana. and planning level, Africa’s transport modes
The solution lies in modernizing customs are too often parceled out across separate line
administration and improving efficiency of ministries, thereby preventing a cohesive inter-
cargo handling. The two main bottlenecks modal transport framework from emerging.
within ports are loading and unloading of
cargo and customs administration—both need Recommendation 6: Pursue
to be addressed simultaneously. Inadequate Regional Integration to Reduce
cranes are part of the problem, but new equip- Infrastructure Costs
ment alone will not deliver better performance Regional integration lowers costs across all
unless staff practices are also modernized. aspects of infrastructure. The high cost of infra-
Ports with container terminal concessions have structure services in Africa is partly attributable
Africa’s Infrastructure: A Time for Transformation 21
infrastructure needs. The two go hand in hand, Figure O.9 Access to and Affordability of
and their development should follow the value Household Services
of agricultural land and the spatial proximity a. Access by quintile
to urban markets. ICT has made huge strides 80
% of households
60
rural Africans now in range of a global systems
mobile signal. This platform can contribute to 40
affordability threshold
beyond the 5 percent
80
Although Africa’s infrastructure services are rel-
atively expensive, costs remain even higher than 60
prices, and this lack of cost recovery has major 40
detrimental effects. Underpricing infrastructure
20
services is costing Africa $4.7 billion a year in
forgone revenues. In addition, because of ineq- 0
2 4 6 8 10 12 14 16
uitable access to infrastructure services, these
monthly utility bill (US$)
subsidies are highly regressive, largely bypass-
low-income countries
ing the poor (figure O.9). The underrecovery middle-income countries
of costs impairs the financial health of utilities Sub-Saharan Africa overall
and slows the pace of service expansion.
Concerns about affordability are usually Source: Banerjee, Wodon, and others 2008.
Note: Q1 = first (or poorest) budget quintile; Q2 = second
the pretext for underpricing services but do budget quintile; Q3 = third (or middle) budget quintile;
not bear much scrutiny (figure O.9). A subsis- Q4 = fourth budget quintile; Q5 = top (or richest) budget quintile.
tence-level monthly utility bill priced in cost-
recovery terms typically amounts to $6–$10 a
month. In the middle-income countries, bills The affordability of services depends
of this magnitude do not appear to present not only on prices, but also on the type of
an affordability problem anywhere across the payment arrangements that are made avail-
income spectrum. Nor do bills of this mag- able to consumers. Prepayment (pioneered
nitude pose affordability issues for the more in the mobile telephone sector) can help
affluent groups in low-income countries, the households budget their consumption and
main ones to enjoy access to services. Afford- reduce revenue risks for operators. The same
ability would become a binding constraint in approach is technologically feasible for elec-
low-income countries only when service cov- tricity, and a growing number of power utili-
erage starts to exceed 50 percent. Only in the ties are adopting it.
poorest of countries, and those with excep- Subsidies are important, but subsidy design
tionally high infrastructure costs, does full needs major rethinking, with a sharper focus
cost recovery seem unachievable for today’s on subsidizing connections, which can be
more affluent consumers. Even in these cases, more equitable and effective in expanding
operating cost recovery should be a feasible coverage. The affordability problems with con-
objective, with subsidies limited to capital nection charges are often much more serious
costs. Simulations suggest that raising tariffs to than those with use-of-service charges. More-
cost recovery would have only minimal effects over, the absence of a connection may itself
on poverty rates in most cases. be a good targeting variable for identifying
24 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
disadvantaged households, although less so that Africa can widely and rapidly adopt
in a low-access environment where coverage modern infrastructure services. Low charges
may be far from universal, even among afflu- for initial connection make market entry
ent households. affordable. Prepayment schemes eliminate
An important test of the coherence of a credit risk and give customers full control over
subsidy policy is to see whether it would be their spending. Services are well tailored to
affordable for the country under universal customer demands. Other network services,
access. The existing underpricing of utility notably power and water, have tended to view
services that benefit just a small minority costs access as a matter of simply rolling out new
many African countries as much as 1 percent networks, overlooking the fact that even where
of GDP. As countries move toward universal networks are available, the hookup rates are
access, that subsidy burden would increase relatively low. They need to pay greater atten-
proportionately, rapidly becoming unafford- tion to demand-side issues that prevent cus-
able for the national budget. Countries should tomers from making connections: connection
thus consider how the cost of any proposed charges that are much higher than household
subsidy policy would escalate as coverage incomes, as well as tenure and urban devel-
increases. This test of the fiscal affordability of opment issues. The most cost-effective way
a subsidy is an important reality check that can to increase access for many utilities may be
prevent countries from embarking on poli- through densification programs that increase
cies that are simply not scalable and will keep hookups to existing networks by using greater
coverage low. community outreach to understand better the
demand side of the market.
Recommendation 9: Find Practical Second-best alternatives can be fine-tuned
Ways to Broaden Access to to provide feasible and attractive infrastructure
Infrastructure Services services to those otherwise unserved. The vast
Universal access to infrastructure services majority of those without access to modern
remains distant for most African countries. infrastructure services rely on traditional alter-
The vast majority of African households today natives, such as candles, wells, or unimproved
lack access to modern power, piped water, sew- latrines. Although doing the job, these tradi-
erage, and even all-season roads that service tional alternatives tend to be inconvenient,
their communities. The very slow progress in inferior, or unsafe. Second-best solutions, such
expanding this access since the mid-1990s sug- as street lighting, solar lanterns, standposts,
gests that universal access to infrastructure is and improved latrines, would provide house-
more than 50 years away for most countries holds with superior services at a cost that is
in Africa. somewhat higher than the traditional alterna-
This situation calls for a different approach tives but still falls far short of modern services.
to expanding modern infrastructure services Puzzlingly, these second-best solutions are not
and for greater attention to second-best alter- very prevalent in Africa, and even where they
natives. Business as usual will not bring about exist, they tend to be available primarily to the
the acceleration of infrastructure access that more affluent.
Africa needs. Moreover, even if access can be A key problem seems to be the public-good
accelerated, many people will have to continue nature of many of these solutions (such as
to rely on alternatives to modern infrastruc- standposts and street lighting), which makes it
ture services for many years to come. There- difficult for service providers to recover costs
fore, infrastructure social policies in Africa and greatly complicates the administration of
need to give greater thought to improving and the facilities. Effective institutional arrange-
expanding second-best alternatives. ments must be found to support implemen-
In expanding modern infrastructure tation of these alternatives. Another problem
networks, closer attention should be paid to is that some of these alternatives, although
the demand side of the equation. The mobile cheaper, may simply not be cheap enough to
telephone revolution has clearly demonstrated be widely affordable.
Africa’s Infrastructure: A Time for Transformation 25
Recommendation 10: Close the 2 percent. Growth in Latin America and Asia
Infrastructure Funding Gap was compromised in a “lost decade.”
Notwithstanding the importance of all these Many countries, ranging from China and
efficiency measures, a substantial infrastruc- India to Argentina and Mexico, have used
ture financing gap of $31 billion a year remains. infrastructure-based fiscal stimulus in times of
Such a large shortfall looked daunting even economic crisis. If well targeted to addressing
before the onset of the global financial crisis. key economic bottlenecks and complemented
As of year-end 2007, many factors had by policy reforms, infrastructure investments
converged to bring about rapid and sustained can pave the way for the later resurgence of
increases in all major sources of external economic growth. Furthermore, some kinds
finance for African infrastructure. Following of public works contracts are labor intensive,
the Gleneagles Summit, OECD development creating short-term employment to alleviate
assistance placed greater emphasis on sup- poverty. Although Africa could benefit from
porting African infrastructure. Official devel- such a program, the continent does not have
opment assistance flows almost doubled, from the means to finance it without external sup-
$4.1 billion in 2004 to $8.1 billion in 2007. The port. Estimates suggest that a $50 billion stim-
resurgence of economic growth on the conti- ulus package would be needed to offset the
nent led to an upswing in private participation. impact of the economic crisis on Africa, and
Since the late 1990s, private investment flows that focusing such a package on infrastructure
to Sub-Saharan infrastructure almost tripled, investments would have the largest short-term
going from about $3 billion in 1997 to $9.4 effect on GDP growth, boosting projections for
billion in 2006/07 (about 1.5 percent of regional 2010 to 4 percent, compared with the postcrisis
GDP). In addition, non-OECD countries— 1.7 percent. In the long term, Africa would see
notably China and India—began to take a a permanent increase of 2.5 percent of GDP
growing interest in financing infrastructure (ODI 2009).
within a framework of South-South coop- Any increase in donor finance for African
eration. Their commitments rose from almost infrastructure should pay particular attention
nothing in the early 2000s to finance about to the power sector and to the fragile states.
$2.6 billion of African infrastructure annually Donors have neglected power since the 1990s.
between 2001 and 2006. Although disburse- Although the private sector can contribute to
ments tend to lag commitments by several funding power generation, donors will still
years, if the record commitments of 2007 are need to scale up substantially to address the
fully honored, the disbursements of external current crisis in the sector. This scale-up was
finance for African infrastructure may con- already under way before the onset of the cri-
tinue to increase over the next few years. sis, with donor commitments that first topped
In the absence of any offsetting measures, $1 billion a year in 2005 reaching a peak of
domestic infrastructure spending would likely $2.3 billion in 2007. Fragile states stand out
fall, compromising economic recovery and as receiving less than their fair share of donor
deepening poverty. The existing gap of $31 finance for infrastructure. Given the magni-
billion a year could widen further as public tude of the financing gap that these countries
budgets are squeezed, external capital flows face relative to the size of their economies, as
decline, and consumer ability to pay user well as the importance of infrastructure in
charges is eroded. The ability to construct regenerating their development, a case exists
new infrastructure, address regional bottle- for channeling incremental donor resources in
necks, and maintain existing assets would be their direction.
severely reduced. In Latin America during the Some of Africa’s larger low-income coun-
1990s, some 50 percent of the fiscal compres- tries have the potential to raise a significant
sion to balance the public books came from amount of local finance for infrastructure if
cuts in infrastructure spending. In Indonesia suitable instruments can be developed. In a
following the Asian crisis, public investment in handful of African countries, domestic capi-
infrastructure fell from 7 percent of GDP to tal markets are beginning to look wide and
26 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
deep enough to provide significant volumes of Calderón, César. 2008. “Infrastructure and
infrastructure finance, Nigeria being the most Growth in Africa.” Working Paper 3, Africa
Infrastructure Country Diagnostic, World
salient example (Irving and Manroth 2009).
Bank, Washington, DC.
However, most of this finance takes the form
Carruthers, Robin, Ranga R. Krishnamani, and
of relatively short-maturity commercial bank
Siobhan Murray. 2008. “Improving Connectiv-
lending, often not the best suited for infra- ity: Investing in Transport Infrastructure in Sub-
structure projects. A need exists to further Saharan Africa.” Background Paper 7, Africa
develop corporate bond markets and to create Infrastructure Country Diagnostic, World Bank,
regulatory conditions for greater participation Washington, DC.
by institutional investors in funding infra- Commission for Africa. 2005. Our Common Inter-
structure investments. est: Report of the Commission for Africa. London:
Commission for Africa.
Note Dorosh, Paul, Hyoung-Gun Wang, Liang You,
and Emily Schmidt. 2008. “Crop Production
The authors of this chapter are Vivien Foster
and Road Connectivity in Sub-Saharan Africa:
and Cecilia Briceño-Garmendia. A Spatial Analysis.” Working Paper 19, Africa
Infrastructure Country Diagnostic, World Bank,
References Washington, DC.
Africon. 2008. “Unit Costs of Infrastructure Proj- Eberhard, Anton. 2007. “Matching Regulatory
ects in Sub-Saharan Africa.” Background Paper Design to Country Circumstances: The Poten-
11, Africa Infrastructure Sector Diagnostic, tial of Hybrid and Transitional Models.”
World Bank, Washington, DC. Gridlines, Note 23 (May), Public-Private
Banerjee, Sudeshna, Heather Skilling, Vivien Foster, Infrastructure Advisory Facility, World Bank,
Cecilia Briceño-Garmendia, Elvira Morella, and Washington, DC.
Tarik Chfadi. 2008. “Ebbing Water, Surging Defi- Eberhard, Anton, Vivien Foster, Cecilia Briceño-
cits: Urban Water Supply in Sub-Saharan Africa.” Garmendia, Fatimata Ouedraogo, Daniel
Background Paper 12, Africa Infrastructure Camos, and Maria Shkaratan. 2008. “Under-
Sector Diagnostic, World Bank, Washington, DC. powered: The State of the Power Sector in Sub-
Banerjee, Sudeshna, Quentin Wodon, Amadou Saharan Africa.” Background Paper 6, Africa
Diallo, Taras Pushak, Hellal Uddin, Clarence Infrastructure Sector Diagnostic, World Bank,
Tsimpo, and Vivien Foster. 2008. “Access, Afford- Washington, DC.
ability, and Alternatives: Modern Infrastructure Escribano, Alvaro, J. Luis Guasch, and Jorge Pena.
Services in Africa.” Background Paper 2, Africa 2008. “Impact of Infrastructure Constraints on
Infrastructure Sector Diagnostic, World Bank, Firm Productivity in Africa.” Working Paper 9,
Washington, DC. Africa Infrastructure Sector Diagnostic, World
Bofinger, Heinrich C. 2009. “Air Transport: Chal- Bank, Washington, DC.
lenges to Growth.” Background Paper 16, Africa Estache, Antonio. 2005. “What Do We Know
Infrastructure Sector Diagnostic, World Bank, about Sub-Saharan Africa’s Infrastructure and
Washington, DC. the Impact of Its 1990 Reforms?” World Bank,
Briceño-Garmendia, Cecilia, and Vivien Foster. Washington, DC.
2009a. “Democratic Republic of Congo: Pri- Fay, Marianne, Danny Leipziger, Quentin Wodon,
oritizing Infrastructure Investments—a Spatial and Tito Yepes. 2005. “Achieving Child-Health-
Approach.” World Bank, Washington, DC. Related Millennium Development Goals: The
———. 2009b. “Republic of Congo: Prioritiz- Role of Infrastructure.” World Development 33
ing Infrastructure Investments—a Spatial (8): 1267–84.
Approach.” World Bank, Washington, DC. Grey, David, and Claudia Sadoff. 2006. “Water for
Briceño-Garmendia, Cecilia, Karlis Smits, and Growth and Development: A Framework for
Vivien Foster. 2008. “Financing Public Infra- Analysis.” Theme Document of the 4th World
structure in Sub-Saharan Africa: Patterns, Issues, Water Forum, Mexico City, March.
and Options.” AICD Background Paper 15, Gwilliam, Ken, Vivien Foster, Rodrigo Archondo-
Africa Infrastructure Sector Diagnostic, World Callao, Cecilia Briceño-Garmendia, Alberto
Bank, Washington, DC. Nogales, and Kavita Sethi. 2008. “The Burden
Bullock, Richard. 2009. “Taking Stock of Railway of Maintenance: Roads in Sub-Saharan
Companies in Sub-Saharan Africa.” Background Africa.” Background Paper 14, Africa
Paper 17, Africa Infrastructure Sector Diagnos- Infrastructure Sector Diagnostic,
tic, World Bank, Washington, DC. World Bank, Washington, DC.
Africa’s Infrastructure: A Time for Transformation 27
Irving, Jacqueline, and Astrid Manroth. 2009. Background Paper 4, Africa Infrastructure Sector
“Local Sources of Financing for Infrastructure Diagnostic, World Bank, Washington, DC.
in Africa: A Cross-Country Analysis.” Policy Teravaninthorn, Supee, and Gael Raballand. 2008.
Research Working Paper 4878, World Bank, “Transport Prices and Costs in Africa: A Review
Washington, DC. of the Main International Corridors.” Working
Mayer, Rebecca, Ken Figueredo, Mike Jensen, Tim Paper 14, Africa Infrastructure Sector Diagnos-
Kelly, Richard Green, and Alvaro Federico Barra. tic, World Bank, Washington, DC.
2008. “Costing the Needs for Investment in ICT Torero, Maximo, and Javier Escobal. 2005. “Mea-
Infrastructure in Africa.” Background Paper 3, suring the Impact of Asset Complementarities:
Africa Infrastructure Sector Diagnostic, World The Case of Rural Peru.” Cuadernos de Economia
Bank, Washington, DC. 42 (May): 137–64.
Minges, Michael, Cecilia Briceño-Garmendia, Vagliasindi, Maria, and John Nellis. 2009. “Evalu-
Mark Williams, Mavis Ampah, Daniel Camos, ating Africa’s Experience with Institutional
and Maria Shkratan. 2008. “Information and Reform for the Infrastructure Sectors.” Working
Communications Technology in Sub-Saharan Paper 23, Africa Infrastructure Country Diag-
Africa: A Sector Review.” Background Paper 10, nostic, World Bank, Washington, DC.
Africa Infrastructure Sector Diagnostic, World
Bank, Washington, DC. Van den Berg, Caroline. 2009. “Public Expenditure
Review in the Water Sector: The Case of Tan-
Morella, Elvira, Vivien Foster, and Sudeshna Ghosh zania.” Paper presented at Water Week, World
Banerjee. 2008. “Climbing the Ladder: The State Bank, Washington, DC, February 17–20.
of Sanitation in Sub-Saharan Africa.” Back-
ground Paper 13, Africa Infrastructure Country Wodon, Quentin, ed. 2008a. “Electricity Tariffs
Diagnostic, World Bank, Washington, DC. and the Poor: Case Studies from Sub-Saharan
Africa.” Working Paper 11, Africa Infrastructure
Mundy, Michael, and Andrew Penfold. 2008. Country Diagnostic, World Bank, Washington,
“Beyond the Bottlenecks: Ports in Sub-Saharan DC.
Africa.” Background Paper 8, Africa Infrastructure
Sector Diagnostic, World Bank, Washington, DC. ———. 2008b. “Water Tariffs and the Poor: Case
Studies from Sub-Saharan Africa.” Working
ODI (Overseas Development Institute). 2009. A Paper 12, Africa Infrastructure Country Diag-
Development Charter for the G-20. London: ODI. nostic, World Bank, Washington, DC.
Prevost, Christophe. 2009. “Benin Rural Water Pub- Yepes, Tito. 2007. “New Estimates of Infrastructure
lic Expenditure Review: Findings, Impacts and Expenditure Requirements.” World Bank, Wash-
Lesson Learned.” Paper presented at Water Week, ington, DC.
World Bank, Washington, DC, February 17–20.
Yepes, Tito, Justin Pierce, and Vivien Foster.
Rosnes, Orvika, and Haakon Vennemo. 2008. 2008. “Making Sense of Sub-Saharan Africa’s
“Powering Up: Costing Power Infrastructure Infrastructure Endowment: A Benchmarking
Investment Needs in Southern and Eastern Approach.” Working Paper 1, Africa Infrastruc-
Africa.” Background Paper 5, Africa Infrastruc- ture Country Diagnostic, World Bank, Washing-
ture Sector Diagnostic, World Bank, Washing- ton, DC.
ton, DC. You, Liang Zhi. 2008. “Irrigation Investment Needs
Svendsen, Mark, Mandy Ewing, and Siwa Msangi. in Sub-Saharan Africa.” Background Paper 9,
2008. “Watermarks: Indicators of Irrigation Africa Infrastructure Country Diagnostic, World
Sector Performance in Sub-Saharan Africa.” Bank, Washington, DC.
Par t 1
I
n 2005, the Commission for Africa drew the 1990s, many donors shifted their priorities
public attention to the magnitude and to social interventions focused on poverty alle-
urgency of Africa’s development challenges viation, overlooking the central importance of
and sounded a new appeal to the international economic growth as an engine of poverty reduc-
community to meet them. In its landmark tion. Moreover, private capital flows in the early
report, Our Common Interest, the commission 2000s were weak in the aftermath of the Asian
underscored infrastructure as one of the con- crisis. The commission’s report stated that
tinent’s central challenges: despite its clear benefits, African govern-
Infrastructure is a key component of the ments and their development partners sharply
investment climate, reducing the costs of reduced, over the 1990s, the share of resources
doing business and enabling people to access allocated to infrastructure—reflecting its lower
markets; is crucial to advances in agriculture; is priority in policy discussions. In retrospect,
a key enabler of trade and integration, impor- this was a serious policy mistake, driven by the
tant for offsetting the impact of geographical international community, which undermined
dislocation and sovereign fragmentation, and growth prospects and generated a substantial
critical to enabling Africa to break into world backlog of investment—a backlog that will
markets; and is fundamental to human devel- take strong action, over an extended period
opment, including the delivery of health and of time, to overcome. (Commission for Africa
education services to poor people. Infrastruc- 2005: chap. 7, para. 63, citations omitted)
ture investments also represent an important The report estimated Sub-Saharan Africa’s
untapped potential for the creation of pro-
infrastructure financing needs to be $39 bil-
ductive employment. (Commission for Africa
2005: chap. 7, para. 61, citations omitted)
lion per year, divided almost equally between
capital expenditure ($22 billion) and spending
In the years preceding the commission’s on operation and maintenance ($17 billion).
report, external capital flows for African infra- On this basis, it recommended a doubling of
structure had reached a historic low. During infrastructure spending in the region, to be
31
32 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Neither does the diagnostic consider needs for Table I.1 AICD Background Papers
water storage infrastructure required to protect Number Category and title Authors
countries from droughts and floods beyond Cross-cutting topics
those necessitated by particular downstream BP2 Access, Affordability, and Alternatives: Sudeshna Banerjee, Quentin Wodon, Amadou
uses such as hydropower electricity generation, Modern Infrastructure Services in Africa Diallo, Taras Pushak, Helal Uddin, Clarence
irrigation, and water supply. Tsimpo, and Vivien Foster
The primary unit of analysis for the diag- BP11 Unit Costs of Infrastructure Projects in Willem van Zyl, Lynette Coetzer, and Chris
Sub-Saharan Africa Lombard
nostic is the country. The focus is on Sub-
Saharan Africa, given the genesis of the proj- BP15 Financing Public Infrastructure in Cecilia Briceño-Garmendia, Karlis Smits, and
Sub-Saharan Africa: Patterns, Issues, Vivien Foster
ect as a response to the major infrastructure and Options
deficits in that part of the continent. Owing Spending needs studies
to budgetary and feasibility constraints, the
BP3 Costing the Needs for Spending in ICT Rebecca Mayer, Ken Figueredo, Mike Jensen,
diagnostic was originally limited to 24 of Infrastructure in Africa Tim Kelly, Richard Green, and Alvaro Federico
the 48 countries in the Sub-Saharan region Barra
(figure I.1). This Phase I sample covers almost BP5 Powering Up: Costing Power Orvika Rosnes and Haakon Vennemo
Infrastructure Spending Needs
all of the major countries, which together
account for about 85 percent of the population BP7 Improving Connectivity: Investing Robin Carruthers, Ranga Rajan Krishnamani,
in Transport Infrastructure in and Siobhan Murray
and GDP of the region. They were carefully Sub-Saharan Africa
selected to represent the economic, geographic, BP9 Irrigation Investment Needs in Liang Zhi You
cultural, and political diversity that charac- Sub-Saharan Africa: A Matter of Scale
terizes the region (figure I.2). Therefore, the State-of-the-sector reviews
sample of 24 countries is statistically represen- BP1 Stuck in Traffic: Urban Transport in Africa Ajay Kumar and Fanny Barrett
tative, providing an adequate basis for draw-
BP4 Watermarks: Indicators of Irrigation Mark Svendsen, Mandy Ewing, and Siwa
ing inferences about the overall infrastructure Sector Performance in Sub-Saharan Africa Msangi
situation of Sub-Saharan Africa. BP6 Underpowered: The State of the Power Anton Eberhard, Vivien Foster, Cecilia
Later, the project steering committee rec- Sector in Sub-Saharan Africa Briceño-Garmendia, Fatimata Ouedraogo,
ommended extending the coverage of the Daniel Camos, and Maria Shkaratan
diagnostic to as many of the remaining Afri- BP8 Beyond the Bottlenecks: Ports in Michael Mundy and Andrew Penfold
Sub-Saharan Africa
can countries as possible. Following further
fund-raising, Phase II of the project was ini- BP10 Information and Communications Michael Minges, Cecilia Briceño-Garmendia,
Technology in Sub-Saharan Africa: Mark Williams, Mavis Ampah, Daniel Camos,
tiated in mid-2008. It incorporates 16 more A Sector Review and Maria Shkratan
countries, raising the total to 40. Although BP12 Ebbing Water, Surging Deficits: Urban Sudeshna Banerjee, Heather Skilling, Vivien
the focus remains on Sub-Saharan Africa, Water Supply in Sub-Saharan Africa Foster, Cecilia Briceño-Garmendia, Elvira
Phase II includes greater coverage of North Morella, and Tarik Chfadi
African countries in a number of areas to com- BP13 Climbing the Ladder: The State of Elvira Morella, Vivien Foster, and Sudeshna
Sanitation in Sub-Saharan Africa Ghosh Banerjee
plete the African picture and provide a point of
BP14 The Burden of Maintenance: Roads in Ken Gwilliam, Vivien Foster, Rodrigo
comparison with the Sub-Saharan region. Sub-Saharan Africa Archondo-Callao, Cecilia Briceño-Garmendia,
Alberto Nogales, and Kavita Sethi
BP16 Air Transport: Challenges to Growth Heinrich C. Bofinger
Scope of the Project BP17 Taking Stock of Railway Companies in Richard Bullock
Sub-Saharan Africa
The results of Phase II were not available at the
time of writing, so the results presented in this
volume are based on the analysis of the 24 Phase
I countries. However, all financial aggregates • The spending needs pillar estimates the cost
in this report have been scaled up to cover the of future infrastructure requirements.
whole of Sub-Saharan Africa. Financial estimates • The fiscal costs pillar documents existing
were scaled to reflect the weight of the 24 sample patterns of infrastructure spending.
countries in the overall GDP of the region. • The sector performance pillar clarifies the
The country-level analysis has three pillars, scope for improvement in efficiency as well
each of which is described below: as structural and policy reforms.
34 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure I.1 Country Coverage of the Africa Infrastructure Country Diagnostic of Africa’s spending needs, and most of them
had limited country coverage. The best-known
TUNISIA global cross-country studies estimate spending
MOROCCO needs using econometric techniques and mac-
ALGERIA
LIBYA ARAB REP.
roeconomic panel data (Estache 2005; Fay and
FORMER
SPANISH OF EGYPT
Yepes 2003; Yepes 2007). These studies iden-
SAHARA
tify historical relationships between GDP and
CAPE VERDE MAURITANIA
MALI
NIGER
physical infrastructure stocks to predict infra-
ERITREA
THE GAMBIA
SENEGAL
BURKINA
CHAD
SUDAN structure requirements given current growth
GUINEA-BISSAU DJIBOUTI
GUINEA FASO
BENIN
NIGERIA
forecasts. Unit costs of infrastructure are then
CÔTE ETHIOPIA
SIERRA LEONE
D’IVOIRE GHANA
CENTRAL
AFRICAN REPUBLIC
used to convert these predictions into financial
LIBERIA CAMEROON SOMALIA
TOGO
EQUATORIAL GUINEA
estimates. These types of studies provide inter-
UGANDA
SÃO TOMÉ AND PRÍNCIPE
GABON
CONGO
DEM. REP.
KENYA nationally consistent, first-order approxima-
RWANDA
OF
CONGO BURUNDI
tions of investment needs. They are likely to
TANZANIA SEYCHELLES
underestimate requirements, however, because
COMOROS
they tend to focus on infrastructure quantity
ANGOLA MALAWI
ZAMBIA rather than quality; do not take into account
Phase I
Phase II
repressed demand and social targets; and
ZIMBABWE MOZAMBIQUE MADAGASCAR MAURITIUS
Phase II (partial)
NAMIBIA
use single, global (as opposed to country-
BOTSWANA
specific), unit-cost parameters based on effi-
SWAZILAND cient implementation.
This map was produced by the Map Design Unit of The World Bank. SOUTH
The boundaries, colors, denominations and any other information shown on
this map do not imply, on the part of The World Bank Group, any judgment
AFRICA
LESOTHO Country-specific or sector-specific engi-
on the legal status of any territory, or any endorsement or acceptance of
such boundaries.
neering cost studies existed for particular
infrastructure packages: the West Africa Power
Figure I.2 Representativeness of Phase I Sampled Countries Pool Master Plan, for example, and the African
Development Bank’s study of the Trans-Africa
100
Highway Network, in addition to various
90
country or regional master plans. These stud-
80
ies tend to be accurate and internalize policy-
70 defined targets, but they have a number of dis-
% of countries
ne
idd ome
ed
ed
d
on
on
om
cte
ho
rn
ck
r m inco
ph
ph
inc
inc
ffe
op
dlo
glo
go
nc
lus
ta
low
le
le
lan
an
idd
fra
ly
lic
nf
m
po
co
er
pe
up
BOX I.1
was grounded in an overview of the institu- distinguish to some extent between current
tional framework for delivering infrastructure expenditures, capital expenditures, and vari-
services in each of the countries while aiming ous subcategories.
to identify all of the channels through which Much of the necessary data could be lifted
public resources go into infrastructure. The directly from the budget documents and
work began with a detailed review of the cen- financial statements of the relevant parastatals,
tral government budget. Thereafter, financial although in many cases, careful recoding of the
statements were collected from all the para- data was necessary to align them with the proj-
statals and special funds that had been identi- ect template. Local consultants undertook field-
fied in the institutional review. work that was coordinated centrally to ensure
In countries where infrastructure service quality control and data consistency. The focus
providers are highly decentralized (as in munici- of data collection was on executed expendi-
pal water utilities), financial statements could be tures, but wherever possible, the budgeted and
collected from only the three largest provid- released expenditures were also collected.
ers. Privatized infrastructure service provid- The targeted period for data collection was
ers were included if a majority of their shares 2001–06, although a complete time series was
remained government owned or if they con- not always available. All financial data are pre-
tinued to depend on the state for capital or sented as annual averages over the period, to
operating subsidies. Thus, telecommunication smooth out annual variations and maximize
incumbents were typically included, whereas available data points. All data were denomi-
mobile operators were not. nated in local currency and centrally normal-
In some countries, local governments have ized using exchange rate, GDP, and population
begun to play an increasingly prominent role data taken from the World Development Indi-
in infrastructure service provision, but com- cators database of the World Bank.
prehensive expenditure data at the local gov- Public expenditure data were complemented
ernment level could not always be collected. by financing data from secondary sources to
In some cases, however, the central govern- provide a comprehensive view of financial
ment produces consolidated local government flows to African infrastructure and the rela-
accounts. Otherwise, an alternative source of tive importance of the different players. These
information was the fiscal transfers from cen- secondary sources included the World Bank’s
tral to local governments reported in the budget Private Participation in Infrastructure data-
and on which local governments relied, given base, which documents trends in private capital
limited alternative sources of revenue. In some flows; the Development Assistance Committee
cases, transfers are earmarked for infrastruc- database of the Organisation for Economic
ture spending; in others, the share allocated to Co-operation and Development (OECD), cov-
infrastructure could only be estimated. ering external financial support from bilateral
Data were collected to permit both classi- and multilateral OECD donors; and a new
fication and cross-classification by economic database on non-OECD finance for African
and functional categories. That is, a matrix was infrastructure (Foster and others 2008). To
established so that spending on each functional make these financial flows methodologically
category could be decomposed according to the consistent with those for public expenditure,
economic nature of the expense and vice versa. researchers converted commitments made by
Functional classification followed as closely as external financiers into disbursements using
possible the four-digit category or class level typical disbursement profiles for infrastruc-
of the functional classification (COFOG) pro- ture projects. Every effort was made to avoid
posed in the International Monetary Fund’s double counting between public expenditure
Government Financial Statistics Manual 2001 and external finance.
(IMF 2001), making possible identification of The results of the public expenditure analy-
all major infrastructure subsectors. The eco- sis provide the foundation for chapter 2 of this
nomic classification of expenses also followed report, but they are reported in much greater
the Fund’s framework, making it possible to detail in Background Paper 15 (see table I.1).
38 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
At this time, one can say that the level of pub- provide a snapshot of the situation prevailing
lic expenditure on infrastructure in Africa is in 2006 at the time of data collection. A second
substantially higher than previously thought— block of quantitative indicators documents the
and certainly several times higher than earlier operational, technical, and financial aspects of
estimates. The resulting public expenditure sector performance, with particular focus on
database is now available to the public via the infrastructure service providers such as utilities.
project Web site and can be downloaded by users Wherever possible, the quantitative data cover
for a variety of purposes. The database con- the period 2001 to 2006, and the most recent
tains detailed information about expenditure available year is the one reported.
patterns by institution, sector, and functional For each sector, manuals were developed
category. to guide the data collection for the indicators.
The analysis of public spending patterns The manuals map the rationale and conceptual
was complemented by work on unit costs of structure of the data collection, provide detailed
infrastructure projects and included a review definitions of the indicators, lay out question-
of the costs and cost structures associated with naire formats to assist in eliciting information,
a sample of donor-funded projects covering and map a database structure for coding of the
roads, power, and water supply. The typical data. Such detailed manuals were designed to
outputs of these projects were standardized to guide consultants responsible for data collection
permit the creation of standardized unit costs. and to ensure comparability of indicators across
Data were collected from the bills of quantities countries and ultimately over time, should the
for the public works contracts of these projects process be repeated.
and entered into the standardized template. For some sectors (power, railways, roads,
The overall sample included 115 road projects, water and sanitation), the indicators could
144 water projects, and 58 power projects. The be collected only through detailed in-country
resulting database of unit costs illustrates the fieldwork. For a number of other sectors (air
dispersion that can be experienced in donor- transport, ICT, irrigation, ports), the data
funded infrastructure projects depending on a could be collected remotely through the arm’s-
range of factors. length administration of questionnaires with
telephone follow-up and the compilation of
Understanding Sector Performance data from existing publications and sources.
At the beginning of the project, relatively little The data collection involved contacting several
systematic, comprehensive, and empirically hundred infrastructure institutions around
grounded literature was available on the per- Africa, including more than 16 rail operators,
formance of the five infrastructure sectors. 20 road entities, 30 power utilities, 30 ports,
To develop a comprehensive and detailed 60 airports, 80 water utilities, and 100 ICT
portrait of the infrastructure sectors in Sub- operators, as well as the relevant line ministries
Saharan Africa, the AICD developed a set of in all of the countries.
standardized performance indicators covering The data collection focused on the compi-
both the consumer and the service provider lation of existing information available from
perspectives. These indicators are collected the target institutions through their annual
for the full range of infrastructure subsec- reports, internal databases, and knowledge
tors, including air transport, ICT, irrigation, of their managers. Thus, the coverage of the
ports, power, railways, roads, and water and databases reflects the state of self-knowledge of
sanitation. In each case, a common conceptual the institutions. The project did not have the
structure was adopted. resources to undertake primary survey work to
A first block of qualitative indicators was gen- obtain data on missing indicators.
erated through a substantial questionnaire that The resulting data were centralized, and two
documents the details of the legal, institutional, forms of quality control were conducted. The
and regulatory framework, which are summa- first was a review by specialists knowledgeable
rized in a series of specially developed indexes about the countries in question. The second
(see chapter 4 in this volume). Qualitative data, consisted of logic and consistency checks on
The Africa Infrastructure Country Diagnostic 39
the database as a whole by examining data pat- and up-to-date information on the status of
terns and outliers. ICT in Africa not available from any other
The survey of infrastructure service provid- single source.
ers was complemented by work on patterns of In transportation, the Sub-Saharan Africa
household access to and expenditure on infra- Transport Policy Program (SSATP) has played
structure services aimed to integrate all exist- an important role in developing the knowl-
ing household surveys conducted in Africa edge base through an abundant literature of
from 1990 to 2005. These sources included case studies and policy reports. The SSATP
67 demographic and health surveys (DHSs) has made some efforts to move toward a set of
and multi-indicator cluster surveys contain- quantitative indicators for the transport sector,
ing detailed information on household access although these remain limited in scope. The
patterns, and 30 budget surveys containing program has played a leading role in develop-
detailed information on household expendi- ing road sector modeling tools, most notably
ture patterns. Data from all of these surveys RONET, that enable road maintenance costs
were standardized (based on a careful com- to be estimated based on a detailed physical
parison of questionnaires) and integrated into specification of the road network. In addition,
a single meta-database, making consistent there have been important contributions to the
analysis possible of time trends within coun- understanding of institutional reform in the
tries and diverging patterns across countries. road sector (Benmaamar 2006) and some work
A standardized approach was used to group on the performance of African rail concessions.
households socioeconomically according to At the outset of the AICD, however, no unified
asset quintiles in the case of the DHSs and database existed on road type, condition, and
expenditure quintiles in the case of the budget traffic. These data were collected on a georefer-
surveys. The meta-database covers 39 countries enced, link-by-link basis that allows the infor-
in Africa; time trends are available for 23 of the mation to be presented graphically in a map
countries. and that underpins detailed financial analysis
The main source of telecommunications of the road network using the RONET model.
data for Africa is the International Telecom- At the outset of the project, relatively little
munication Union, which compiles time- continent-wide analysis of the African water
series data on a number of indicators and utilities existed (Estache and Gassner 2004b).
publishes information on telecommunications The starting point for water utility data col-
regulation. In addition, a number of one-off lection was the databases developed by the
reports had been written on the African tele- Water Utilities Partnership and the Interna-
communication sector. These reports quickly tional Benchmarking Network (IBNET). Both
become outdated and are often limited to cer- sources were sparse in their country coverage
tain groups of countries. The AICD project and focused primarily on utility operational
has improved the timeliness, detail, and scope performance without covering the institutional
of these data sets, including compiling more framework in any great depth. Both initia-
recent data than are available from intergov- tives informed the development of indicators
ernmental sources, verifying the accuracy of under the AICD, which aimed to be consistent
existing information, widening and completing with them in areas of overlap. The data col-
coverage for all African countries, and enhanc- lection process was coordinated with IBNET
ing data to incorporate more detailed and spe- to increase African country coverage for both
cific indicators for tariffs, regulations, market projects. Generally speaking, the AICD opera-
structure, and the user’s perspective, among tional and financial performance indicators
others. The project has also structured the data are a subset of those collected by IBNET, but
into analytical categories and compiled several the qualitative indicators and tariff schedules
indexes to facilitate comprehension of the vast collected through AICD go much further than
amount of data. In summary, the information anything done before. Five modules of qualita-
in the AICD data set for ICT provides a struc- tive data were collected for each country, cover-
tured framework of comprehensive, inclusive, ing the institutional and regulatory framework
40 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
for water provision, governance arrangements indicators are now available to the public on the
for specific water utilities, the status of the sani- project Web site, http://www.infrastructureafrica
tation sector, the status of the rural water sector .org, and through the Development Data Plat-
in each country, and the prevalence and charac- form of the World Bank. Containing detailed
teristics of small-scale service providers in the information about institutional, operational,
largest city in each country. The quantitative technical, and financial indicators relating to
indicators aim to capture the operational and each of the sectors covered, the database can be
financial performance of utilities from 2001 to downloaded for a variety of purposes.
2006, together with their tariff schedules. In The work on these three pillars and cross-
countries where service provision is decentral- cutting issues resulted in the creation of 17
ized, comprehensive data could not be captured original background papers on which this
on all utilities, but efforts were made to cover “Flagship Report” is based (table I.1). The
the three largest utilities in each country. main findings that follow in this report refer
In the case of irrigation, limited data were to these background papers. Readers seeking
available at the country level. Sub-Saharan further technical details on any of these issues
Africa has little experience with irrigated can find these papers through the project Web
agriculture. Most performance indicators are site (http://www.infrastructureafrica.org). In
limited to specific irrigation systems. The best due course, the background papers will be
single source of data on comparable cross- repackaged as four sectoral volumes on ICT,
country indicators was the global databases of power, transport, and water and sanitation,
the Food and Agriculture Organization. They which will be technical companions to this
were complemented where needed by data Flagship Report.
from the World Bank and the International In addition to these three central pillars
Food Policy Research Institute. of the data collection effort, more than 20
Power was undoubtedly the least docu- working papers were commissioned, cover-
mented of Africa’s infrastructure sectors at ing a range of ad hoc topics of relevance to
the outset of the project (Estache and Gassner African infrastructure (table I.2). The top-
2004a). Some basic indicators on overall energy ics include linkages between infrastructure,
balance and national power generation port- growth and fiscal sustainability, welfare effects
folios were available from the International of infrastructure reforms, utility tariffs and
Energy Agency and others, but coverage of subsidies, urban infrastructure services, local
African countries remained quite limited, and private finance of infrastructure, impact of
the available indicators did not provide any real inadequate power supply on firms, and the
picture of power utility performance. Although role of small, independent suppliers of water.
the Union of Producers, Transporters and Dis- The working papers are also available on the
tributors of Electric Power in Africa (the Afri- project Web site.
can power utilities association) has developed Beyond the initial data baseline estab-
its own database of performance indicators, it lished here, the AICD project aims to estab-
is not available to the general public. The Africa lish a sustainable basis for ongoing data
Energy Commission is also developing a data- collection on Africa’s infrastructure sectors.
base of energy indicators for the continent, but This Flagship Report presents and analyzes
it was not available in time for this project. the baseline information on the African
The results of the various sector reviews infrastructure sectors collected because of
provide the foundation for the corresponding this project. The long-term value of the effort
sector chapters contained in part 2 (chapters depends on the sustainability of data collec-
7–17) of this report. In addition, the results of tion efforts to ensure that key infrastructure
the household survey analysis are reported in trends on the continent can be tracked over
chapter 3 on poverty and inequality, while the time and progress against this benchmark
overall findings of the institutional analysis can be accurately measured. Plans are under
are summarized in chapter 4 on institutions. way for the Statistical Department of the
The resulting databases of sector performance African Development Bank to take over the
The Africa Infrastructure Country Diagnostic 41
long-term data collection effort based on the Briceño-Garmendia, Cecilia. 2007. “Fiscal Costs of
methodological framework developed under Infrastructure Provision: A Practitioner’s Guide.”
Working Paper 17, Africa Infrastructure Coun-
the AICD project. The sponsors of the AICD
try Diagnostic, World Bank, Washington, DC.
project remain firmly committed to ensuring
Commission for Africa. 2005. Our Common Inter-
the sustainability of the data collection effort. est: Report of the Commission for Africa. London:
Commission for Africa.
Note Estache, Antonio. 2005. “What Do We Know
The authors of this chapter are Vivien Foster about Sub-Saharan Africa’s Infrastructure and
the Impact of Its 1990 Reforms?” World Bank,
and Cecilia Briceño-Garmendia.
Washington, DC.
Estache, Antonio, and Katharina Gassner. 2004a.
References “The Electricity Sector of Sub-Saharan Africa:
Benmaamar, Mustapha. 2006. “Financing of Road Basic Facts and Emerging Issues.” World Bank,
Maintenance in Sub-Saharan Africa: Reforms Washington, DC.
and Progress towards Second Generation Road ———. 2004b. “Recent Economic Developments
Funds.” Discussion Paper 6, Road Management in the Water and Sanitation Sectors of Selected
and Financing Series, Sub-Saharan Africa Sub-Saharan African Countries: Overview of
Transport Policy Program, World Bank, Basic Facts and Emerging Issues.” World Bank,
Washington, DC. Washington, DC.
42 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Estache, Antonio, Marianela Gonzalez, and Mehta, Meera, Thomas Fugelsnes, and Kameel
Lourdes Trujillo. 2007. “Government Virjee. 2005. “Financing the Millennium Devel-
Expenditures on Health, Education and opment Goals for Water and Sanitation: What
Infrastructure: A Naïve Look at Levels, Out- Will It Take?” International Journal of Water
comes and Efficiency.” Policy Research Working Resources Development 21 (2): 239–52.
Paper 4219, World Bank, Washington, DC. Water and Sanitation Program. 2006. Getting
Fay, Marianne, and Tito Yepes. 2003. “Investing Africa on Track to Meet the MDGs for Water and
in Infrastructure: What Is Needed from 2000 Sanitation: A Status Overview of Sixteen African
to 2010?” Policy Research Working Paper 3102, Countries. Joint Report of African Ministers
World Bank, Washington, DC. http://ssrn.com/ Council on Water, African Development Bank,
abstract=636464. European Union Water Initiative, and Water and
Foster, Vivien, William Butterfield, Chuan Chen, Sanitation Program. Nairobi: Water and Sanita-
and Nataliya Pushak. 2008. Building Bridges: tion Program–Africa, World Bank.
China’s Growing Role as Infrastructure Finan- WHO and UNICEF (World Health Organization
cier for Sub-Saharan Africa. Trends and Policy and United Nations Children’s Fund). 2006.
Options no. 5. Washington, DC: Public-Private Meeting the MDG Drinking Water and Sanitation
Infrastructure Advisory Facility, World Bank. Target: The Urban and Rural Challenge of the
IMF (International Monetary Fund). 2001. Gov- Decade. Geneva: WHO and UNICEF.
ernment Financial Statistics Manual 2001 Yepes, Tito. 2007. “New Estimates of Infrastructure
(GFSM 2001). Washington, DC: IMF Statistics Expenditure Requirements.” World Bank,
Department. Washington, DC.
Chapter 1
I
nfrastructure is central to Africa’s develop- The estimated spending needs are $93 billion
ment.1 Major improvements in information a year (15 percent of the region’s GDP)—more
and communication technology (ICT), for than twice the 2005 estimate by the Commission
example, added as much as 1 percentage point for Africa.3 Total spending estimates divide fairly
to Africa’s per capita growth rate during the evenly among the middle-income countries,
last decade, since the mid-1990s. However, the resource-rich countries, and low-income
deficiencies in infrastructure are holding back nonfragile states (in the neighborhood of $28
the continent by at least 1 percentage point in billion–$30 billion a year), with low-income frag-
per capita growth. In many countries, infra- ile states accounting for a smaller share of total
structure limitations, particularly in power, needs (about $14 billion a year). The burden on
depress productivity at least as much as red their economies varies dramatically per income
tape, corruption, and lack of finance—the group, ranging from 10–12 percent of GDP for
usual suspects in many people’s minds when middle-income and resource-rich countries to
they think of constraints on growth. 25 percent of GDP for low-income nonfragile
In density of paved roads, capacity to gener- states and 36 percent for fragile states. The total
ate power, and coverage of telephone main lines, cost splits two to one between capital investment
both low-income and middle-income African and operation and maintenance expenses.
countries lag behind their peers elsewhere in Over 40 percent of the expenditure needed
the developing world.2 A few decades ago, in the is in the power sector, which must install
1960s to 1980s, Africa’s infrastructure endow- 7,000 megawatts of new generation capac-
ments were similar to those in East and South ity each year just to keep pace with demand.
Asia, but those regions have since expanded their Slightly more than 20 percent is associated with
infrastructure stocks more rapidly, surpassing achievement of the Millennium Development
Africa’s position. Meeting Africa’s infrastructure Goals (MDGs) for water supply and sanitation.
needs and developing cost-effective modes of A further 20 percent of the spending require-
infrastructure service delivery will entail a sub- ment is associated with the transport sector to
stantial program of infrastructure investment. In achieve a reasonable level of regional, national,
addition to building new infrastructure, existing rural, and urban connectivity and to maintain
facilities must be rehabilitated and maintained. existing assets.
43
44 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Infrastructure: The Key to Africa’s allow isolation of the first of these effects with
Faster Growth some precision. The estimated effect of rais-
ing Africa’s infrastructure to some regional or
African economies have grown at a solid 4 international benchmark shows considerable
percent annual average in recent years. The consistency of 1 or 2 percentage points in per
fastest growth has been in resource-rich capita growth.
countries, which have benefited from rising A key question for policy makers is how
commodity prices. In almost all cases, how- much infrastructure development contributes
ever, that performance still falls short of the to growth relative to other policy parameters.
7 percent growth needed to achieve substan- One study finds that expanding and improving
tial poverty reduction and attain the MDGs. infrastructure contributed almost 1 percent-
Although infrastructure has contributed to age point to per capita economic growth from
Africa’s recent economic turnaround, it will 1990 to 2005, compared with only 0.8 percent-
need to do even more to reach the continent’s age point for macroeconomic stabilization and
development targets. structural policies (Calderón 2008). Stabiliza-
Inadequate infrastructure impedes faster tion policies include measures to control price
growth in Africa. This view, highlighted by the inflation and rein in fiscal deficits, while struc-
Commission for Africa (2005), is supported tural policies include measures to enhance
by considerable economic research (table 1.1). human capital, increase financial depth, pro-
Based on a cross-country econometric analysis mote trade openness, and improve governance.
and a handful of country studies, the research Central Africa is the region where infrastruc-
confirms a strong and significant connection ture improvements have made the largest
between infrastructure stocks and economic contribution to recent growth, totaling 1.1
growth. Although the relationship undoubt- percentage points. Only in West Africa did the
edly runs in both directions—infrastructure effect of macroeconomic policies on growth
supporting growth and growth promoting exceed that of infrastructure. Over the same
infrastructure—modern research techniques period, infrastructure in East Asia contributed
Table 1.1 Links between Infrastructure and Growth in Africa: What the Research Says
Study Method Scope Sector Conclusions
Easterly and Levine 1997 Multicountry Africa Telecommunications, Infrastructure is strongly and significantly correlated with growth.
power
Esfahani and Ramirez 2003 Multicountry Africa Telecommunications, Africa’s growth per capita would be 0.9 point higher with East
power Asia’s infrastructure.
Calderón and Servén 2008 Multicountry Africa Telecommunications, Africa’s growth per capita would be 1.0 point higher with the
power, roads Republic of Korea’s infrastructure.
Estache, Speciale, and Veredas 2005 Multicountry Africa Various Confirms earlier work and underscores equal relevance for
coastal and landlocked countries.
Calderón 2008 Multicountry Africa Telecommunications, Africa’s growth per capita would be 2.3 points higher with
power, roads Mauritius’s infrastructure.
Calderón and Servén 2008 Multicountry Africa Telecommunications, Extends earlier results to show infrastructure also has a negative
power, roads effect on inequality.
Fedderke and Bogetic 2006 Country study South Africa Various Finds long-term relationship between infrastructure and growth
based on robust econometric techniques.
Ayogu 1999 Production function Nigeria Various Finds strong association between infrastructure and output in
panel data.
Kamara 2008 Production function Various Africa Various Finds strong association between infrastructure and output in
panel data.
Reinikka and Svensson 1999a Enterprise surveys Uganda Power Unreliable power is a significant deterrent to private sector
investment.
Escribano, Guasch, and Pena 2008 Enterprise surveys Africa Various Infrastructure has a substantial effect on total factor productivity.
Source: Authors’ elaboration.
Meeting Africa’s Infrastructure Needs 45
1.2 percentage points to per capita growth Figure 1.1 Changes in Growth per Capita Caused by Changes in Growth Fundamentals
(figure 1.1). 2.5
economic growth
attributable to greater penetration of telecom- 1.5
munications (figure 1.2). In contrast, the defi- 1.0
cient infrastructure of the power sector has 0.5
retarded growth, reducing per capita growth
0
for Africa as a whole by 0.11 percentage point
and for southern Africa by as much as 0.2 per- –0.5
ca
a
er
op
ric
ric
ric
ric
ric
centage point. The effect of road infrastructure
fri
Tig
Af
Af
Af
Af
Af
r
lA
Eu
rth
st
rn
ian
ra
es
rn
is generally positive, if rather small, perhaps
Ea
he
nt
No
W
As
te
ut
Ce
es
st
So
W
because of the absence of a widely available
Ea
measure of road quality, which is the critical stabilization policies structural policies infrastructure
variable affecting transport costs.
Source: Calderón 2008.
More detailed microeconomic work on the
relationship between infrastructure and the per-
formance of firms (see table 1.1) supports these Figure 1.2 Changes in Growth per Capita Caused by Changes in Different Kinds of
macroeconomic findings. The data consistently Infrastructure
show a strong relationship between infrastruc- 2.5
percentage points of per capita
ca
ca
a
tivity by about 40 percent. The study first looked
ric
ric
ric
ric
fri
fri
Af
Af
Af
Af
tA
lA
rth
st
rn
ra
at the relative contribution of infrastructure
es
Ea
he
nt
No
ut
Ce
So
Botswana Ethiopia
Swaziland Botswana
Mauritius Swaziland
Kenya Mali
Madagascar Uganda
Tanzania Zambia
Niger Kenya
Mauritania Madagascar
Mali Malawi
Eritrea Mauritania
Zambia Niger
Uganda Cameroon
Senegal Benin
Benin Tanzania
Malawi Mauritius
0 20 40 60 80 100 0 20 40 60 80 100
% contribution to total factor productivity % contribution to total factor productivity
infrastructure others electricity customs clearance transportation
information and cummunication water
technology
children, face a substantial opportunity cost in 1990; Venkataraman 1990). Similarly, better
travel time when they have to fetch water. More access to electricity lowers costs for businesses
than 20 percent of the population in Cameroon, and increases investment, driving economic
Ghana, Mauritania, Niger, and Tanzania must growth (Reinikka and Svenson 1999b).
travel more than 2 kilometers to their primary Improved transportation networks enable
water supply. Rural dwellers tend to travel far- isolated rural communities to move into com-
ther than urban dwellers (Blackden and Wodon mercial agriculture, thereby increasing their
2005; Wodon 2008). income, and to use health and education ser-
Better provision of electricity has impor- vices some distance away (Barwell 1996; Calvo
tant benefits for health because vaccines and and others 2001; Davis, Lucas, and Rikard
medications can be safely stored in hospitals 1996; Ellis and Hine 1998; World Bank 1996).
and food can be preserved at home (Jimenez By reducing the time and money it takes to
and Olson 1998). Electricity also improves move goods, better transportation improves
literacy and primary school completion rates competitiveness, helping create more jobs and
because students can read and study after boost incomes (Limão and Venables 1999;
sundown (Barnes 1988; Brodman 1982; Foley World Bank 2000, 2001).
Meeting Africa’s Infrastructure Needs 47
The expansion of ICT networks democra- The differences are particularly large for paved-
tizes access to information. It can be particu- road density, telephone main lines, and power
larly critical for rural populations otherwise cut generation. The gap exists for both low-income
off from important technological know-how or and middle-income groups.
critical information about market prices (Kenny Was Africa’s current infrastructure defi-
2002; Saunders, Warford, and Wellenius 1994). cit caused by a low historic starting point?
In many cases, telecommunication improve- Has it always been worse-off than the rest of
ments also reduce transportation spending by the world? In the 1960s (roads), 1970s (tele-
allowing people to avoid fruitless journeys or to phones), and 1980s (power), Africa’s stocks
perform transactions remotely (Telecommuni- were quite similar to those of South or East
cation Development Bureau 1999). Asia. (The one exception was paved-road den-
sity, in which South Asia already enjoyed a
huge advantage over both Africa and East Asia
Africa’s Infrastructure Deficit as far back as the 1960s. For household cov-
erage of electricity, both South and East Asia
By just about every measure of infrastructure were already far ahead of Africa in the early
coverage, African countries lag behind their 1990s, and this gap has widened over time.)
peers in other parts of the developing world Africa expanded its infrastructure stocks
(see table 1.3; Yepes, Pierce, and Foster 2008). more slowly than other developing regions,
48 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
opening a gap between Africa and Asia (figure sanitation, the differences between subregions
1.4). The comparison with South Asia—with have been relatively small. Today, the South-
a similar per capita income—is particularly ern African Development Community region
striking. In 1970, Africa had almost three times has a strong lead over all other subregions on
more electricity-generating capacity per million almost every aspect of infrastructure. The weak-
people than did South Asia. By 2000, South est infrastructure endowments are in Central
Asia had left Africa far behind—it now has Africa (for roads, water, and sanitation) and in
almost twice the generating capacity per mil- East Africa (for ICT and power) (table 1.4).
lion people. Similarly, in 1970 Africa had twice To better portray the diversity that exists
the main-line telephone density of South Asia, across Africa, this report classifies countries into
but by 2000, South Asia had drawn even. And in four types: (a) middle-income countries, (b)
the case of mobile density, low-income African resource-rich countries, (c) fragile states, and
countries are actually ahead of South Asia. (d) other low-income countries. (See box 1.1 for
China and India have largely driven the full definitions.) These categories were chosen
rapid infrastructure expansion in South and because they capture differences in financing
East Asia. In particular, China has pursued a capacity and institutional strength that are rele-
conscious strategy of infrastructure-led growth vant in understanding infrastructure outcomes.
since the 1990s, committing more than 14 per- Outcomes across these different types of
cent of GDP to infrastructure investment in countries are strikingly diverse. The difference in
2006 (Lall, Anand, and Rastogi 2008). infrastructure stocks between African middle-
At independence, substantial variations in income countries and other African countries is
infrastructure existed across different subre- to be expected, although African middle-income
gions in Africa. Southern Africa started with countries have only a narrow edge over low-
relatively high infrastructure endowments and income countries elsewhere in the developing
achieved some of the highest annual growth world. The lags associated with fragile states are
rates in infrastructure stocks over the last four readily understandable, given the disruption of
decades. In 1980, the subregion had more than conflict.
three times the generating capacity per million Especially striking is the extent to which
people of other subregions; in 1970, it had five resource-rich countries lag behind others in
times the telecommunication density of the their infrastructure endowment, despite their
other subregions. With regard to roads, West greater wealth. In recent years, resource-
Africa was in a much stronger position than the rich countries have devoted their additional
other subregions in the 1960s but was overtaken wealth not to infrastructure development
by southern Africa by the 1980s. In water and but to paying off their debt. The governance
Meeting Africa’s Infrastructure Needs 49
20 4
15 3
index
index
10 2
5 1
0 0
1960 1970 1980 1990 2000 1980 1990 2000
50
2 40
30
1 20
10
0 0
1960 1970 1980 1990 2000 1990–95 1996–2000 2001–05
20 80
percent
15 60
index
10 40
5 20
0 0
1970 1980 1990 2000 1990–95 1996–2000 2001–05
Sources: Banerjee and others 2008; Yepes, Pierce, and Foster 2008.
Note: Road density is measured in kilometers per 100 square kilometers of arable land; telephone density, in lines per 1,000 people;
generation capacity, in megawatts per 1 million people.
Power provides the clearest example of a margins are exceptionally high, particularly in
sector with genuinely higher costs in Africa Central and West Africa where they reach lev-
than elsewhere. Many small countries rely on els of 60 to 160 percent. The underlying cause
small-scale diesel generation that can cost up is the limited competition in the sector, com-
to $0.40 per kilowatt-hour in operating costs bined with a highly regulated market based
alone—about three times higher than coun- on tour de role principles, whereby freight is
tries with larger power systems (over 500 allocated to transporters through a central-
megawatts), which are typically hydropower ized queuing method rather than by allowing
based (Eberhard and others 2008). truckers to enter into bilateral contracts with
In contrast, high road freight tariffs in Africa customers directly.
are caused more by excessive profit margins than The high prices for international tele-
by high costs (Teravaninthorn and Raballand phone and Internet service in Africa reflect a
2008). The costs that Africa’s trucking operators mixture of cost and profit. In countries that
face are not significantly higher than in other have no access to a submarine cable and are
parts of the world, even when informal pay- forced to rely on expensive satellite technol-
ments are taken into account. However, profit ogy, charges are typically twice as high as in
Meeting Africa’s Infrastructure Needs 51
BOX 1.1
MAURITANIA
CAPE MALI NIGER
VERDE CHAD ERITREA
SENEGAL SUDAN
GAMBIA
GUINEA-BISSAU BURKINA FASO
GUINEA BENIN NIGERIA SOMALIA
SIERRA LEONE CÔTE GHANA ETHIOPIA
D'IVOIRE TOGO CENTRAL AFRICAN
LIBERIA
CAMEROON REPUBLIC
EQUATORIAL GUINEA
UGANDA
KENYA
O
GABON
NG
CONGO, RWANDA
CO
TANZANIA
MALAWI
ANGOLA
ZAMBIA
MOZAMBIQUE
ZIMBABWE MADAGASCAR MAURITIUS
countries that enjoy cable access. Even when Power Spending Needs Are by Far
access to a submarine cable is obtained, coun- the Largest
tries with a monopoly on this international Africa’s largest infrastructure needs are in the
gateway have tariffs that are substantially power sector. Whether measured in generat-
higher than those without a monopoly (Min- ing capacity, electricity consumption, or secu-
ges and others 2008). rity of supply, Africa’s power infrastructure
delivers only a fraction of the service found
elsewhere in the developing world (Eberhard
How Much Does Africa Need to and others 2008). The 48 countries of Africa
Spend on Infrastructure? (with a combined population of 800 million)
generate roughly the same amount of power as
Meeting Africa’s infrastructure needs and Spain (with a population of 45 million). Power
developing cost-effective modes of infra- consumption, which is 124 kilowatt-hours per
structure service delivery call for a substantial capita per year and falling, is only 10 percent of
program of investment, rehabilitation, and dis- that found elsewhere in the developing world,
ciplined maintenance combined. The physical barely enough to power one 100-watt light-
infrastructure requirements are the grounds bulb per person for 3 hours a day. Africa’s firms
for a new set of estimates for spending require- report that frequent power outages cause them
ments that are the foundation of this report. to lose 5 percent of their sales; this figure rises to
In all cases, the estimated spending takes 20 percent for firms in the informal sector that
into account both growth-related and social are unable to afford backup generators. Chap-
demands for infrastructure, and it incorporates ter 8 in this volume contains a more detailed
the costs of maintenance and rehabilitation as discussion of Africa’s power challenges.
well as new investment. Addressing this power shortage will require
The time horizon for estimating spend- enormous investments in infrastructure over
ing needs is a decade. The assumption is that the next decade. Based on four economic mod-
over a period of 10 years running up to 2015, els, covering the Central, East, Southern, and
the continent should be expected to address West African Power Pools, potential generation
its infrastructure backlog, keep pace with the projects in each power pool are identified and
demands of economic growth, and attain a ranked according to cost-effectiveness. These
number of key social targets for broader infra- models make possible estimating the cost
structure access (table 1.6). of meeting power demand under a range of
Table 1.6 10-Year Economic and Social Targets for Investment Needs Estimates, 2006–15
Sector Economic target Social target
Information and communication Complete submarine cable loop around Africa and 36,000-kilometer Extend GSM voice signal and public access broadband to
technology fiber-optic backbone network interconnecting national capitals to 100 percent of the rural population.
each other and to submarine cable loop.
Irrigation Develop all financially viable opportunities for large- and small-scale n.a.
irrigation, potentially some 12 million hectares.
Power Attain demand-supply balance in power production, developing Raise household electrification rate by about 10 percentage points
7,000 megawatts of new generation capacity annually within a over current levels, entailing an additional 57 million new house-
regional framework entailing 22,000 megawatts of new cross-border hold connections.
interconnections.
Transport Attain 250,000 kilometers of good-quality road networks supporting Raise the Rural Access Index from the current level of 34 percent
regional and national connectivity goals. nationally to 100 percent in highest-value agricultural areas.
Place entire urban population within 500 meters of road
supporting motorized access.
Water and sanitation n.a. Meet the Millennium Development Goals for water and sanitation.
Sources: Banerjee and others 2008; Carruthers, Krishnamani and Murray 2008; Mayer and others 2008; Rosnes and Vennemo 2008; You 2008.
Note: GSM = global systems mobile. n.a. = not applicable.
Meeting Africa’s Infrastructure Needs 53
alternative scenarios that consider access tar- next decade (to 2015), to allow power to flow
gets, fuel prices, unit costs of investment, and freely from country to country. The financial
feasibility of cross-border trade (Vennemo and returns on these interconnectors can be as high
Rosnes 2008). as 120 percent in the Southern African Power
Demand for power is almost directly pro- Pool; it is typically 20–30 percent in the other
portional to economic growth. Installed capac- pools. Regional trade can also put Africa on a
ity will need to grow by more than 10 percent path to cleaner development, because it would
annually—or more than 7,000 megawatts a increase hydropower’s share of the continent’s
year—just to meet Africa’s suppressed demand, generation portfolio from 36 percent to 48 per-
keep pace with projected economic growth, cent, displacing 20,000 megawatts of thermal
and provide additional capacity to support the plant in the process and saving 70 million tons
rollout of electrification. Since 1995, expan- of carbon emissions each year. Finally, raising
sion of the sector has averaged barely 1 per- electrification rates will require extending dis-
cent annually, or less than 1,000 megawatts tribution networks to reach almost 6 million
a year. Most of that power would go to meet additional households a year over the next dec-
nonresidential demands from the commercial ade (to 2015).
and industrial sectors. The overall costs for the power sector in
The most cost-effective way to expand Afri- Africa are a staggering $41 billion a year—$27
ca’s power generation is through regional trade billion for investment and $14 billion for
that allows countries to pool the most attrac- operation and maintenance (table 1.8). About
tive primary energy resources across national half the investment costs are for development
boundaries. Regional trade shaves around of new generating capacity. Approximately 15
$0.01 per kilowatt-hour off the marginal cost percent is earmarked for rehabilitation of exist-
of power generation in each of the power pools ing generation and transmission assets. About
(and as much as $0.02 to $0.04 per kilowatt- 40 percent of the costs are for the Southern
hour for some countries), leading to savings of Africa Power Pool alone.
about $2 billion a year in the costs of develop-
ing and operating the power system. Mobilizing Achieving Water Security Remains an
the benefits of regional trade depends on devel- Unquantified Challenge
oping major untapped hydropower projects One important infrastructure requirement
in the Democratic Republic of Congo, Ethio- not explicitly estimated in the investment costs
pia, and Guinea, which would become major is water storage capacity, which is required to
exporters in the Southern, East, and West Afri- reach water security. Africa experiences huge
can Power Pools, respectively (table 1.7). It also swings in precipitation across areas, across
hinges on establishing some 22,000 megawatts seasons, and over time (Grey and Sadoff 2006).
of interconnectors that will be needed over the Climate change will only exacerbate this
improved sanitation, it is 63 percent. As of 2006, Table 1.10 Water and Sanitation Spending Needs,
the last year for which official data have been 2006–15
$ billions annually
published, the figures for Africa were 58 per-
cent and 31 percent, respectively. To meet the Sector Investment Maintenance Total
MDG goal, the number of people with access Water 11.0 5.5 16.5
to safe water would need to increase from 411 Sanitation 3.9 1.4 5.4
million to 701 million by 2015—an increase of Total 14.9 7.0 21.9
29 million a year compared with recent prog- Source: Banerjee and others 2008.
ress of only 11 million per year. To meet the
MDG sanitation goal, the number of people Transport Needs Are Substantial
with access to improved service would need to Africa’s road density seems sparse compared
increase from 272 million in 2006 to 617 mil- with the vastness of the continent, but it is not
lion by 2015—an increase of 35 million a year unreasonable relative to the continent’s popu-
compared with recent progress of only 7 mil- lation and income. A more detailed discus-
lion a year. Chapters 16 and 17 in this volume sion of Africa’s transport challenges appears
offer more detailed discussions of Africa’s water in chapters 9–13 in this volume. The adequacy
supply and sanitation challenges, respectively. of Africa’s current transport network can best
The overall price tag for reaching the water be assessed by examining whether it provides
and sanitation MDG access is estimated at $22 an adequate level of connectivity to facilitate
billion (roughly 3.3 percent of Africa’s GDP), the movement of people and goods between
with water accounting for more than two- regions, within nations, out of rural areas,
thirds (table 1.10). Capital investment needs and across cities. Using a spatial model, one
can be conservatively estimated at $15 billion can assess the cost of linking economic and
a year (2.2 percent of the region’s GDP). These demographic nodes through transport infra-
needs include both new infrastructure and structures so as to achieve regional, national,
rehabilitation of existing assets. Estimates are urban, and rural connectivity.
based on minimum acceptable asset standards. Regional connectivity within the African
It is assumed that access patterns (or relative continent requires a network that links all
prevalence of water and sanitation modali- capital cities and cities with over 1 million
ties) remain broadly the same between 2006 inhabitants to deep-sea ports and interna-
and 2015 and that services are upgraded for tional borders. This objective can be achieved
only a minimum number of customers. The with a two-lane network of a little over 100,000
maintenance requirements stand at $7 bil- kilometers maintained in good condition.
lion a year (1.1 percent of the region’s GDP). About 70 percent of this network is already in
Operation and maintenance of network and place, but about one-quarter of it needs to be
non-network services, respectively, amount widened from one lane to two lanes, and about
to 3 percent and 1.5 percent of the replace- three-quarters of it needs to be improved to
ment value of installed infrastructure. Reha- good quality. The overall cost of meeting this
bilitation costs have been estimated based on target amounts to $2.7 billion a year, or barely
a model that takes into account the mainte- 15 percent of total spending needs for the
nance backlog of network infrastructure in transport sector. The bulk of this expenditure
each country. is for investment.
56 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
facilities. The ICT revolution has been accom- Table 1.11 Transport Spending Needs, 2006–15
plished largely through market liberalization $ billions annually
and private sector investment, which will Investment
continue to be the main driving force behind Improve Upgrade Add Total Total Overall
future investments. The state will need to con- Sector/area condition category capacity investment maintenance total
tinue investing in a few critical areas, however. Regional
connectivity 0.5 1.1 0.2 1.8 0.9 2.7
Chapter 7 in this volume contains a more
detailed discussion of Africa’s ICT challenges. National
connectivity 0.5 1.2 0.2 1.9 1.0 2.9
The private sector will undertake the major
Rural
expenditures in this sector to service growth in connectivity 0.8 0.4 0.1 1.3 1.2 2.5
market demand. The urban market for ICT ser-
Urban
vices is well established and profitable. Demand connectivity 0.3 0.4 0.4 1.1 0.5 1.6
for voice services in this market is expected to Railways, ports,
grow as penetration rates continue to rise from and airports 0.2 0.6 1.9 2.7 5.9 8.6
20 to 46 lines per 100 inhabitants. In addition, Total 2.2 3.7 2.7 8.6 9.6 18.2
incipient markets for broadband services are Source: Carruthers, Krishnamani, and Murray 2008.
expected to expand from 0.04 to 2.54 lines per Note: Railways, ports, and airports include investments by South Africa’s Transnet and other
demand-driven transport investment needs covered by the private sector.
100 inhabitants. These demands can be met Column totals may not add exactly because of rounding errors.
entirely by private sector investment.
Spatial models are used to simulate the com-
mercial viability of further expanding cover- would be required. Private finance would likely
age of voice and broadband signals into rural be forthcoming for the highest-traffic seg-
areas using global systems mobile and WiMAX ments. However, the more ambitious the aspi-
(Worldwide Interoperability for Microwave rations for extending connectivity, the larger
Access) technologies (Mayer and others 2008). the component of public finance that would
The models consider the cost of network rollout be required.
based on topographical factors and local avail- A modest level of broadband service could
ability of power. They also estimate local revenue be provided using WiMAX technology to
potential based on demographic densities, per provide low-volume connectivity to a lim-
capita incomes, and estimated subscriber rates. ited number of institutions and public access
With no market barriers, the private sector telecenters in rural areas. Using this approach,
alone could profitably extend global systems and again in the absence of market barri-
mobile signal coverage to about 95 percent of ers, the private sector alone could profitably
Africa’s population (Mayer and others 2008). extend WiMAX coverage to about 89 percent
The remaining 5 percent, living in isolated of Africa’s population (Mayer and others
rural communities, is not commercially viable 2008). The remaining 11 percent, living in
and would require a significant state subsidy to isolated rural communities, are not commer-
connect. The percentage of the population that cially viable and would require a significant
is not commercially viable varies substantially state subsidy to support network rollout. As
across countries, from less than 1 percent in with voice, the percentage of the population
Nigeria to more than 20 percent in the Demo- that is not commercially viable to cover varies
cratic Republic of Congo. substantially across countries, from less than
Broadband service, by contrast, is still in 1 percent in Nigeria to more than 70 percent
its infancy and will expand only if significant in the Democratic Republic of Congo.
investments are made in rolling out high- Finally, Africa is in the process of complet-
capacity fiber-optic cable across the continent. ing a network of submarine cables that links it
Just interconnecting all Africa’s capitals would to the global intercontinental network. Several
require a network of 36,000 kilometers of projects are already under way to close the loop
fiber-optic cable. If the network were extended around the eastern side of the continent. Some
to cover all cities with 500,000 or more inhab- strengthening of the West African submarine
itants, more than 100,000 kilometers of cable system is also needed, plus cable links to service
58 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
outlying islands, such as the Comoros, Mada- Comparison with the Commission
gascar, and the Seychelles. The private sector is for Africa
showing considerable appetite to take on this The $93 billion estimate is more than twice the
kind of investment. estimate of the Commission for Africa in 2005,
The investment costs of this additional ICT which was based on cross-country econo-
infrastructure, beyond what would be purely metric studies, rather than the more detailed
driven by market demand, are relatively mod- country-level microeconomic modeling of
est when compared with other infrastructure the Africa Infrastructure Country Diagnostic
sectors. Achieving universal rural access for (Estache 2006). A recent update of the cross-
both voice service and limited broadband ser- country model used for the Commission for
vice based on WiMAX technology could be Africa report came up with a revised estimate
accomplished for an investment of $1.7 billion of $80 billion to $90 billion (Yepes 2007).
a year, the bulk of which could come from the Some 40 percent of the total is for the power
private sector, with additional public fund- sector, which requires about $41 billion each year
ing amounting to no more than $0.4 billion a (6 percent of African GDP; Rosnes and Vennemo
year. Completing the submarine and intrare- 2008). A significant share of the spending for
gional fiber-optic backbone would entail an power is for investment in multipurpose water
annual (private sector) investment of less than storage schemes and thus makes an important
$0.2 billion, although this sum would more contribution to water resources management.
than double if a more ambitious network con- The second-largest component is the cost of
necting all cities with over 500,000 inhabitants meeting the MDGs for water and sanitation—
were envisaged (table 1.12). Factoring in the about $22 billion (3 percent of regional GDP).
market-driven investments needed to keep pace The third-largest price tag is associated with the
with demand in established urban markets, transport sector, which comes in at just over $18
the estimated ICT sector annual investment need billion (3.6 percent of GDP).
rises to $7 billion a year, plus another $2 billion
annually for operation and maintenance. Distribution of Spending among
Countries
Three groups of countries—the middle-
Overall Price Tag income countries, the resource-rich countries,
and the low-income nonfragile states—share
Africa’s overall cost to build new infrastruc- roughly equally in the bulk of total spending.
ture, refurbish dilapidated assets, and operate Each of these groups needs to spend around
and maintain all existing and new installa-
tions is estimated at almost $93 billion a year
for 2006 through 2015 (15 percent of African Table 1.13 Overall Infrastructure Spending Needs for
GDP; table 1.13 and figure 1.5). Africa, 2006–15
$ billions annually
Capital Operation and Total
Sector expenditure maintenance needs
Table 1.12 ICT Spending Needs beyond the Purely Market Driven: Investment Only,
2006–15 ICT 7.0 2.0 9.0
$ billions annually Irrigation 2.7 0.6 3.3
Universal Universal access Fiber-optic Power 26.7 14.1 40.8
Type of access to voice to broadband backbone linking Submarine
investment signal platform capital cities cables Transport 8.8 9.4 18.2
Total investment 0.78 0.91 0.03 0.18 Sources: Authors’ calculations based on Banerjee and others
2008; Carruthers, Krishnamani, and Murray 2008; Mayer and
Source: Mayer and others 2008. others 2008; Rosnes and Vennemo 2008.
Note: In contrast to the preceding tables, the expenditure for operation and maintenance is excluded Note: ICT = information and communication technology;
because of the difficulty of apportioning it across the different subcategories presented. WSS = water supply and sanitation.
— Not available. Row totals may not add exactly because of rounding errors.
Meeting Africa’s Infrastructure Needs 59
Sources: Authors’ calculations based on Banerjee and others 2008; Carruthers, Krishnamani, and Murray 2008; Mayer and others 2008; Rosnes and Vennemo 2008.
$28 billion to $30 billion to meet its infrastruc- amounting to no more than 10 percent to
ture needs. The price tag for the fragile states 13 percent of their respective GDPs. For
is only about half as much at $13 billion. The low-income countries, however, as much as
largest spending needs for an individual coun- 25 percent of GDP would be needed, rising to
try by far are in South Africa, which requires an implausible 37 percent for the low-income
$27 billion a year. fragile states. Ethiopia, Madagascar, Niger, and
The burden of spending relative to the above all, the Democratic Republic of Congo
countries’ GDPs is very different across face an impossible challenge—their infrastruc-
groups. For middle-income and resource-rich ture needs range from 26 to over 70 percent of
countries, the burden appears manageable, GDP (see figure 1.5, panel a).
60 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Ayogu, Melvin. 1999. “Before Prebendalism: A Collier, Paul, and Stephen O’Connell. 2006.
Positive Analysis of Core Infrastructure Invest- “Opportunities and Choices.” Draft chapter for
ment in a Developing Fiscal Federalism.” African the synthesis volume of the Explaining African
Development Review 11 (2): 169–98. Economic Growth PRoject of Oxford University
Banerjee, Sudeshna, Quentin Wodon, Amadou and Swarthmore College, Centre for Study of
Diallo, Taras Pushak, Hellal Uddin, Clarence African Economies.
Tsimpo, and Vivien Foster. 2008. “Access, Afford- Commission for Africa. 2005. Our Common Inter-
ability, and Alternatives: Modern Infrastructure est: Report of the Commission for Africa. London:
Services in Africa.” Background Paper 2, Africa Commission for Africa.
Infrastructure Sector Diagnostic, World Bank,
Washington, DC. Croppenstedt, Andre, and Mulat Demeke. 1996.
“Determinants of Adoption and Levels of
Barnes, Douglas F. 1988. Electric Power for Rural Demand for Fertiliser for Cereal Growing
Growth: How Electricity Affects Rural Life in Devel- Farmers in Ethiopia.” Working Paper 96-3,
oping Countries. Boulder, CO: Westview Press. Centre for the Study of African Economies,
Barwell, Ian. 1996. “Transport and the Village: Oxford, U.K.
Findings from African Village-Level Travel Davis, Tony, Kimberly Lucas, and Ken Rikard.
and Transport Surveys and Related Studies.” 1996. Agricultural Transport Assistance Program:
Discussion Paper 344, World Bank, Impact Study. Washington, DC: United States
Washington, DC. Agency for International Development.
Blackden, Mark, and Quentin Wodon, eds. 2005. Doumani, Fadi M., and James A. Listorti. 2001.
“Gender, Time Use and Poverty in Sub-Saharan “Environmental Health: Bridging the Gaps.”
Africa.” Working Series Paper 73, World Bank, World Bank Discussion Paper 422, World Bank,
Washington, DC. Washington, DC.
Briceño-Garmendia, Cecilia, Karlis Smits, and Easterly, William, and Ross Levine. 1997. “Africa’s
Vivien Foster. 2008. “Financing Public Infra- Growth Tragedy: Policies and Ethnic Divisions.”
structure in Sub-Saharan Africa: Patterns, Issues, Quarterly Journal of Economics 112 (4): 1203–50.
and Options.” AICD Background Paper 15, http://ssrn.com/abstract=88828.
Africa Infrastructure Sector Diagnostic, World
Bank, Washington, DC. Eberhard, Anton, Vivien Foster, Cecilia Briceño-
Garmendia, Fatimata Ouedraogo, Daniel
Brodman, Janice. 1982. “Rural Electrification and
Camos, and Maria Shkaratan. 2008. “Under-
the Commercial Sector in Indonesia.” Discus-
powered: The State of the Power Sector in
sion Paper D-73L, Resources for the Future,
Sub-Saharan Africa.” Background Paper 6,
Washington, DC.
Africa Infrastructure Sector Diagnostic, World
Calderón, César. 2008. “Infrastructure and Growth Bank, Washington, DC.
in Africa.” Working Paper 3, Africa Infrastruc-
ture Country Diagnostic, World Bank, Eberhard, Anton, and Clive Van Horen. 1995.
Washington, DC. Poverty and Power: Energy and the South African
State. London: UCT and Pluto.
Calderón, César, and Luis Servén. 2008. “Infra-
structure and Economic Development in Sub- Ellis, Simon, and John Hine. 1998. “The Provi-
Saharan Africa.” Policy Research Working Paper sion of Rural Transport Services: Approach
4712, World Bank, Washington, DC. Paper.” Sub-Saharan Africa Transport Policy
Program Working Paper 37, World Bank,
Calvo, Christina Malmberg. 1994. “Case Study on Washington, DC.
the Role of Women in Rural Transport: Access
of Women to Domestic Facilities.” Sub-Saharan Escribano, Alvaro, J. Luis Guasch, and Jorge Pena.
Africa Transport Policy Program Working Paper 2008. “Impact of Infrastructure Constraints on
11, World Bank, Washington, DC. Firm Productivity in Africa.” Working Paper 9,
Africa Infrastructure Sector Diagnostic, World
Calvo, Christina Malmberg, Colin Gannon, Bank, Washington, DC.
Kenneth M. Gwilliam, and Zhi Liu. 2001.
“Transport: Infrastructure and Services— Esfahani, Hadi Salehi, and Maria Teresa
Technical Notes.” PRSP Toolkit, World Bank, Ramirez. 2003. “Institutions, Infrastructure,
Washington, DC. and Economic Growth.” Journal of Develop-
ment Economics 70: 443–77. http://ssrn.com/
Carruthers, Robin, Ranga R. Krishnamani, and
abstract=310779.
Siobhan Murray. 2008. “Improving Connectiv-
ity: Investing in Transport Infrastructure in Sub- Esrey, S. A., J. B. Potash, L. Roberts, and C. Shiff.
Saharan Africa.” Background Paper 7, Africa 1991. “Effects of Improved Water Supply and
Infrastructure Country Diagnostic, World Bank, Sanitation on Ascariasis, Diarrhea, Dracunculia-
Washington, DC. sis, Hookworm Infection, Schistosomiasis and
62 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Trachoma.” Bulletin of the World Health Organi- ———. 2003b. “Linkages between ICT and Pov-
zation 89 (5): 609–21. erty Alleviation.” World Bank, Washington,
Estache, Antonio. 2006. “Africa’s Infrastructure: DC.
Challenges and Opportunities.” Paper pre- ———. 2003c. “Linkages between Transport and
sented at “Realizing the Potential for Profitable Poverty Alleviation.” World Bank, Washington,
Investment in Africa” seminar, IMF Institute DC.
and Joint Africa Institute, Tunis, February ———. 2003d. “Linkages between Water and Sani-
28–March 1. tation and Poverty Alleviation.” World Bank,
Estache, Antonio, Biagio Speciale, and David Vere- Washington, DC.
das. 2005. “How Much Does Infrastructure Mat- Lall, Rajiv, Ritu Anand, and Anupam Rastogi. 2008.
ter for Growth in Sub-Saharan Africa?” World “Developing Physical Infrastructure: A Com-
Bank, Washington, DC. parative Perspective on the Experience of the
Fay, Marianne, Danny Leipziger, Quentin Wodon, People’s Republic of China and India.” Asian
and Tito Yepes. 2005. “Achieving Child-Health- Development Bank, Manila.
Related Millennium Development Goals: The Lanjouw, Peter, Jamie Quizon, and Robert Spar-
Role of Infrastructure.” World Development 33 row. 2001. “Non-Agricultural Earnings in
(8): 1267–84. Peri-Urban Areas of Tanzania: Evidence from
Fedderke, Johannes W., and Zeljko Bogetic. 2006. Household Survey Data.” Food Policy 26:
“Infrastructure and Growth in South Africa: 385–403.
Direct and Indirect Productivity Impacts of 19 Limão, Nuno, and Anthony Venables. 1999. “Infra-
Infrastructure Measures.” Policy Research Work- structure, Geographical Disadvantage and
ing Paper 3989, World Bank, Washington, DC. Transportation Costs.” Policy Research Working
Foley, Gerald. 1990. Electricity for Rural People. Paper 2257, World Bank, Washington, DC.
London: Panos Institute.
Mayer, Rebecca, Ken Figueredo, Mike Jensen, Tim
Grey, David, and Claudia Sadoff. 2006 “Water for Kelly, Richard Green, and Alvaro Federico Barra.
Growth and Development: A Framework for 2008. “Costing the Needs for Investment in ICT
Analysis.” Theme Document of the 4th World Infrastructure in Africa.” Background Paper 3,
Water Forum, Mexico City, March. Africa Infrastructure Sector Diagnostic, World
Hutton, Guy. 2000. Considerations in Evaluating the Bank, Washington, DC.
Cost-Effectiveness of Environmental Health Inter- McCarthy, F. Desmond, and Holger Wolf. 2001.
ventions. Geneva: World Health Organization. “Comparative Life Expectancy in Africa.” Policy
Hutton, Guy, and Laurence Haller. 2004. Evalu- Research Working Paper 2668, World Bank,
ation of the Costs and Benefits of Water and Washington, DC.
Sanitation Improvements at the Global Level: Minges, Michael, Cecila Briceño-Garmendia, Mark
Water, Sanitation and Health Protection of the Williams, Mavis Ampah, Daniel Camos, and
Human Environment. Geneva: World Health Maria Shkratan. 2008. “Information and Com-
Organization. munications Technology in Sub-Saharan Africa:
A Sector Review.” Background Paper 10, Africa
IMF (International Monetary Fund). 2007.
Infrastructure Sector Diagnostic, World Bank,
Regional Economic Outlook: Sub-Saharan Africa.
Washington, DC.
Washington, DC: IMF.
Reinikka, Ritva, and Jakob Svensson. 1999a.
Jimenez, Antonio, and Ken Olson. 1998. Renew-
“Confronting Competition: Firms’ Investment
able Energy for Rural Health Clinics. Golden, CO:
Response and Constraints in Uganda.” In
National Renewable Energy Laboratory. http://
Assessing an African Success: Farms, Firms, and
www.nrel.gov/docs/legosti/fy98/25233.pdf.
Government in Uganda’s Recovery, ed. Paul Col-
Kamara, Samura. 2008. “Restoring Economic lier and Ritva Reinikka. Washington, DC: World
Growth and Stability in Fragile and Post- Bank.
Conflict Low Income States: Views from the ——. 1999b. “How Inadequate Provision of Pub-
Trench.” Paper presented at World Bank Poverty lic Infrastructure and Services Affects Private
Reduction and Economic Management confer- Investment.” Policy Research Working Paper
ence, Washington, DC, April 15–16. 2262, World Bank, Washington, DC.
Kenny, Charles. 2002. “The Costs and Benefits of Rosnes, Orvika, and Haakon Vennemo. 2008.
ICTs for Direct Poverty Alleviation.” Develop- “Powering Up: Costing Power Infrastructure
ment Policy Review 20: 141–57. Investment Needs in Southern and Eastern
Kerf, Michel. 2003a. “Linkages between Energy Africa.” Background Paper 5, Africa Infra-
and Poverty Alleviation.” World Bank, structure Sector Diagnostic, World Bank,
Washington, DC. Washington, DC.
Meeting Africa’s Infrastructure Needs 63
Saunders, Robert J., Jeremy Warford, and Bjorn World Bank. 1996. “Socio-Economic Impact of
Wellenius. 1994. Telecommunications and Rural Roads: Impact Evaluation Report for
Economic Development. 2nd ed. Baltimore: the Fourth Highway Project in Morocco.”
Johns Hopkins University Press for the OED Report 15808-MOR, World Bank,
World Bank. Washington, DC.
Telecommunication Development Bureau. 1999. ———. 2000. “Ghana: Building a Stronger
Telemedicine and Developing Countries—Lessons Transportation System.” Precis 189,
Learned. Geneva: International Telecommunica- Washington, DC.
tion Union. ———. 2001. Cities on the Move: A World Bank
Teravaninthorn, Supee, and Gaël Raballand. Urban Transport Strategy Review. Washington,
2008. “Transport Prices and Costs in Africa: DC: World Bank.
A Review of the Main International Corri-
———. 2005. Global Monitoring Report 2005.
dors.” Working Paper 14, Africa Infrastructure
Washington, DC: World Bank.
Sector Diagnostic, World Bank,
Washington, DC. ———. 2007. DEPweb glossary. Development
Education Program, World Bank. http://www
Venkataraman, K. 1990. “Rural Electrification in
.worldbank.org/depweb/english/modules/
the Asian and Pacific Region.” In Power Systems
glossary.html#middle-income.
in Asia and the Pacific, with Emphasis on Rural
Electrification, ed. Economic and Social Com- Yepes, Tito. 2007. “New Estimates of Infrastructure
mission for Asia and the Pacific, 310–32. New Expenditure Requirements.” World Bank,
York: United Nations. Washington, DC.
Vennemo, Haakon, and Ornica Rosnes. 2008. Yepes, Tito, Justin Pierce, and Vivien Foster.
“Powering-Up: Costing Power Infrastructure 2008. “Making Sense of Sub-Saharan Africa’s
Investment Needs in Africa.” Background Paper Infrastructure Endowment: A Benchmarking
5, Africa Infrastructure Country Diagnostic, Approach.” Working Paper 1, Africa Infra-
World Bank, Washington, DC. structure Country Diagnostic, World Bank,
Wodon, Quentin, ed. 2008. “Electricity Tariffs Washington, DC.
and the Poor: Case Studies from Sub-Saharan You, Liang Zhi. 2008. “Irrigation Investment Needs
Africa.” Working Paper 11, Africa Infrastruc- in Sub-Saharan Africa.” Background Paper 9,
ture Country Diagnostic, World Bank, Africa Infrastructure Country Diagnostic, World
Washington, DC. Bank, Washington, DC.
Chapter 2
T
he cost of addressing Africa’s infrastruc- Substantial evidence indicates that a lot
ture needs is estimated at $93 billion, more can be done within Africa’s existing
some 15 percent of Africa’s GDP—about resource envelope. Inefficiencies of various
two-thirds for investment and one-third for kinds total about $17 billion a year. If appro-
maintenance. The burden varies greatly by priately tackled, fixing these inefficiencies
country type. About half of the capital invest- could expand the existing resource envelope
ment needs are for power, reflecting the par- by 40 percent.
ticularly large physical deficits in that area. First, countries and development institu-
Existing spending is higher than previously tions allocate $3.3 billion in infrastructure
thought. African governments, infrastructure spending to areas that appear surplus to the
users, the private sector, and external sources basic infrastructure requirements (as defined in
together already contribute about $45 billion chapter 1 of this volume), which suggests that
to directly address the infrastructure needs public and aid flows can be redirected toward
previously identified. About one-third of this areas of greater impact on development.
amount is spent by middle-income countries, Second, because only three-fourths of the
whereas fragile states barely account for 5 per- capital budgets allocated to infrastructure are
cent of it (about $2 billion in total), mirroring actually executed, about $2 billion in public
the weakness of their economies and the investment is being lost.
enormous disparity in terms of financing and Third, underspending on infrastructure
institutional capabilities across Sub-Saharan asset maintenance is another major waste of
African countries. About two-thirds of the resources because the cost of rehabilitating
existing spending is domestically sourced, infrastructure assets is several times higher
from taxes or user charges, and channeled than the cumulative cost of sound preven-
through public institutions, making the public tive maintenance. In the road sector alone,
sector—governments and nonfinancial pub- addressing undermaintenance can save $1.9
lic enterprises together—the most important billion a year in rehabilitation, or spending $1
financier of capital investment, funding more on maintenance can be a savings of about $4
than half of total investment. to the economy.
65
66 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Fourth, Africa’s power and water utilities and the countries benefit fully from the additional
state-owned telecommunication incumbents finance. Otherwise, what is the use of pouring
waste about $6 billion a year on inefficiencies water into a leaking bucket?
such as overstaffing, revenue undercollection,
and distribution losses.
Fifth, underpricing infrastructure services Spending Allocated to Address
accounts for $4.7 billion a year in lost revenues. Infrastructure Needs
In all, with existing allocation patterns
and even if potential efficiency gains are fully Africa is spending $45 billion a year to address
captured, a funding gap of $31 billion a year its infrastructure needs. Existing spending on
remains: three-quarters for capital and one- infrastructure in Africa is higher than previ-
quarter for maintenance. About $23 billion of ously thought when budget and off-budget
this gap relates to power and a further $11 bil- spending (including state-owned enterprises
lion to water supply and sanitation (WSS). For and extrabudgetary funds) and external financ-
fragile states, the funding gap is an implausible ing (comprising official development assistance
25 percent of GDP on average, almost equally [ODA], financiers from outside the Organisation
divided among energy, water, and transport. for Economic Co-operation and Development
How can Africa close such a sizable funding [OECD], and private participation in infra-
gap, equivalent to one-third of the estimated structure [PPI]) are taken into account. This
infrastructure needs? Additional funds will be level of spending is associated with allocations
required, and in a few countries—mainly the directly targeted to cover the needs identified
fragile ones—the magnitude of the funding in chapter 1. In practice, however, some coun-
gap calls for considering taking more time to tries spend more on some infrastructure subsec-
attain targets or using lower-cost technologies. tors than the estimated benchmark requirements
Historical trends do not suggest much pros- while incurring funding gaps in other subsec-
pect of increasing allocations from the public tors. This existing spending with potential for
budget: even when fiscal surpluses existed, they reallocation is not counted here but is consid-
did not visibly favor infrastructure. External ered later in this chapter.
finance has been buoyant in recent years, and The four-way country typology introduced
disbursements will likely continue to grow as in chapter 1 of this volume—comprising mid-
projects committed move to the implementa- dle-income countries, resource-rich coun-
tion stage. In light of today’s financial crisis, tries, fragile states, and other low-income
however, prospects for new commitments do countries—serves as a basis for summarizing
not look good. Private capital flows, in particu- the diversity of infrastructure financing chal-
lar, can be expected to decline. Fiscal pressure lenges (see box 1.1). Expressed as a percentage
is growing in donor countries, and to judge by of GDP, infrastructure spending is comparable
previous crises, foreign aid is likely to slow. across the different country types, at around
By delaying investment timetables, and 5–6 percent of GDP, with the exception of
assuming that efficiency gains are fully cap- nonfragile low-income countries, which spend
tured, many countries could even attain the at 10 percent of their GDP. In absolute dollar
infrastructure targets without increasing their terms, the middle-income countries spend the
spending envelopes. Targeting a high level of most (roughly $16 billion), reflecting their
service might not always work in the best inter- much larger purchasing power. Fragile states,
est of a country. Lower-cost technologies can by contrast, account for a tiny amount of over-
permit broadening the portion of the popula- all spending (about $2 billion), reflecting the
tion with access to some level of service. weakness of their economies (table 2.1).
Closing Africa’s funding gap inevitably The public sector, with the lion’s share of
requires undertaking needed reforms to reduce spending, is by far the most important finan-
or eliminate the inefficiencies of the system. cier. In the middle-income countries, domestic
Only then can the infrastructure sectors become public sector resources (comprising tax reve-
attractive to a broader array of investors and nues and user charges raised by state-owned
Closing Africa’s Funding Gap 67
Table 2.1 Spending of Most Important Players Traced to Needs (Annualized Flows)
Percentage of GDP $ billions
O&M Capital expenditure O&M Capital expenditure
Country Public Public Non-OECD Total capital Public Public Non-OECD Total capital
type sector sector ODA financiers Private expenditures Total sector sector ODA financiers Private expenditures Total
Middle
income 3.7 1.2 0.1 0.0 0.8 2.1 5.8 10.0 3.1 0.2 0.0 2.3 5.7 15.7
Resource rich 1.1 1.5 0.2 0.7 1.7 4.1 5.2 2.5 3.4 0.5 1.4 3.8 9.1 11.7
Low-income
nonfragile 4.0 1.5 2.2 0.5 1.9 6.1 10.1 4.4 1.6 2.5 0.6 2.1 6.7 11.1
Low-income
fragile 2.0 0.6 1.0 0.8 1.2 3.6 5.6 0.8 0.2 0.4 0.3 0.5 1.4 2.1
Africa 3.2 1.5 0.6 0.4 1.5 3.9 7.1 20.4 9.4 3.6 2.5 9.4 24.9 45.3
Sources: Africa Infrastructure Country Diagnostic (AICD); Briceño-Garmendia, Smits, and Foster 2008 for public spending; PPIAF 2008 for private flows; Foster and others 2008
for non-OECD financiers.
Note: Aggregate public sector covers general government and nonfinancial enterprises. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1.
Totals may not add exactly because of rounding errors. O&M = operation and maintenance; ODA = official development assistance; OECD = Organisation for Economic
Co-operation and Development; Private = private participation in infrastructure and household self-finance of sanitation facilities.
enterprises) account for the bulk of spending Patterns of specialization are clear across
across all infrastructure subsectors. Across the the different sources of external finance (figure
other country typologies, domestic public sec- 2.1). Across sectors, PPI is strongly concen-
tor resources contribute approximately half trated on information and communication
of total spending. One-third of this aggregate technology (ICT), which shows the highest
public sector spending (or an equivalent of commercial returns. ODA has tended to focus
1.5 percent of GDP) can be traced exclusively on public goods with high social returns, nota-
to capital investments. bly roads and water. Much non-OECD finance
This level of effort by African governments has gone to energy and, to a lesser extent, to
to develop their infrastructure pales when com- railways, both sectors with strong links to
pared with what East Asian countries have done industry and mining. Across countries, PPI has
in recent decades. China, for example, adopted a tended to go to middle-income and resource-
determined and clear strategy to increase infra- rich countries, which have the greatest ability
structure investment (publicly and privately to pay for services. Non-OECD finance has
financed) as a means of achieving accelerated shown a preference for resource-rich countries,
economic growth. Fixed capital formation in with a strong pattern linking infrastructure
Chinese infrastructure more than doubled investment and natural resource extraction,
between 1998 and 2005. By 2006, only infra- and ODA has preferred nonfragile low-income
structure investment was higher than 14 percent states with limited domestic resources but ade-
of GDP, perhaps the highest in the world. quate institutional capacity. The fragile states
Excluding middle-income countries, external do not seem to have captured their fair share
financiers contribute roughly one-half of Afri- from any of the external sources.
ca’s total spending on infrastructure. External
sources include ODA from the OECD countries,
official finance from non-OECD countries (such How Much More Can Be Done
as China, India, and the Arab funds), and PPI. within the Existing Resource
External finance is primarily for investment— Envelope?
broadly defined to include asset rehabilitation
and reconstruction—and in most cases does not Africa is losing about $17 billion per year to
provide for O&M. Since the late 1990s, PPI has various inefficiencies in infrastructure opera-
been the largest source of external finance, fol- tions or spending. In this context, four distinct
lowed by ODA and non-OECD finance, which opportunities can be identified for efficiency
are broadly comparable in magnitude. gains. First, improving budget execution rates
68 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 2.1 Sources of Financing for Infrastructure Capital Investment, by Sector and Country Type
US$ billions
US$ billions
6 6
5
4 4
3
2 2
1
0 0
information and power transport water supply middle income resource rich nonfragile fragile
communication and sanitation low income low income
technology
Sources: Africa Infrastructure Country Diagnostic (AICD); Briceño-Garmendia, Smits, and Foster 2008 for public spending; PPIAF 2008 for
private flows; Foster and others 2008 for non-OECD financiers.
would increase the potential of fully using terms, however, middle-income countries
resources allocated to public investment. Sec- have a much larger infrastructure budget, with
ond, reallocating existing spending toward spending per capita at $150–$200, compared
subsectors in greatest need, therefore with with $20–$40 in low-income countries. In
highest economic returns, would allow the other words, per capita budgetary spending
existing budget envelope to better cover exist- on infrastructure by middle-income coun-
ing needs. Third, raising user charges closer to tries is about five times that of low-income
cost-recovery levels would provide more effi- countries.
cient price signals and help capture lost rev- Overall, spending on transport (notably
enues. Fourth, reducing operating efficiencies roads) is the single-largest infrastructure item
of utilities and other service providers would in general government accounts. It ranges from
prevent waste of significant resources, support about half of all general government spending
healthier utilities, and improve service quality. on infrastructure in middle-income countries
to 60 percent in low-income countries. Water
Raising Capital Budget Execution and sanitation spending is the second-largest
African central governments alone allocate, on category, particularly in the middle-income
average, 1.5 percent of GDP, or 6–8 percent of countries. Energy spending features heavily in
their national budgets, to support the provi- resource-rich countries.
sion of infrastructure (table 2.2). For Africa, From a functional perspective, more than 80
this effort translates into about $300 million percent of budgetary spending goes to invest-
a year for an average country, which would ment. With the exception of middle-income
not take many African countries a long way. countries and the ICT sector, the central
To put this figure in perspective, an invest- government makes the bulk of public invest-
ment of $100 million can purchase about 100 ment, even in sectors in which state-owned
megawatts of electricity generation, 100,000 enterprises provide most services. Strikingly,
new household connections to water and sew- relative to central government, nonfinancial
erage, or 300 kilometers of a two-lane paved public institutions, such as utilities and other
road. It runs well short of covering the invest- service providers, make little infrastructure
ment needs estimated in chapter 1 of this vol- investment (figure 2.2). The state-owned
ume (see chapter 1 for details). enterprises are essentially asset administrators.
As a percentage of GDP, budget spending This spending pattern reflects government
on infrastructure is comparable across low- control of some of the main sources of invest-
and middle-income countries. In absolute ment finance, be they royalty payments (in
Closing Africa’s Funding Gap 69
resource-rich countries) or external develop- Figure 2.2 Split Investment Responsibilities between Governments and Public
ment funds (in low-income countries). It also Enterprises, by Type of Country and Sector
reflects to some extent the limited capability of 1.0
state-owned enterprises to fund their capital 0.9
0.8
investment through user fees. 0.7
% of GDP
in e
e
r e
gi w i ich
in e
e
r e
gi w i ich
in e
e
r e
gi w i ich
in e
e
ra reso om
lo om
ra so om
lo om
ra so om
lo om
ra so om
lo om
m
gets. African countries are, on average, unable
fra e lo ce r
fra e lo ce r
fra e lo ce r
fra e lo ce r
co
co
co
co
nc
le nc
nf re nc
le nc
nf re nc
le nc
nf re nc
le nc
ei
ei
ei
gi u
gi u
gi u
gi u
w
w
dl
dl
dl
dl
to spend as much as one-quarter of their capi-
id
id
id
id
m
m
l
l
tal budgets and one-third of their recurrent
nf
no
no
no
no
budgets in the corresponding fiscal year (table information and power transport water
2.3). The poor timing of project appraisals and communication
technology
late releases of budgeted funds because of pro-
curement problems often prevent the use of government public enterprises
leakages of public capital spending as high as the highest, averaging for the continent more
92 percent (see Pritchett 1996; Rajkumar and than a 100 percent economic rate of return
Swaroop 2002; Reinikka and Svensson 2002, and well above returns for rehabilitation and
2003; Warlters and Auriol 2005; and references new construction (table 2.5). By favoring
cited therein). investment over maintenance, African govern-
ments have been implicitly equating public
Reallocating Existing Spending to investment with productive expenditure,
Subsectors in Need even though not all investment is productive
About $3.3 billion a year is spent above the and not all current spending is wasteful.1 The
estimated requirements to meet the identified maintenance of public goods under the juris-
infrastructure needs (see chapter 1 in this vol- diction of general governments is essential
ume). This spending—funded by (or through) to harness the economic returns to capital and
public budgets—includes domestically raised to avoid costly rehabilitation. Highest returns to
funds and international aid (OECD and non- maintenance are seen in networks already well
OECD sources). Most of this apparent over- developed, particularly in middle-income coun-
spending is in telecommunications in countries tries and nonfragile low-income countries.
that have maintained state ownership of the From a sectoral perspective, economic
fixed-line incumbent. State ownership not only returns to railway investments are the lowest
uses expensive public resources in activities that among infrastructure interventions. Railway
the already competitive telecommunication rehabilitation interventions are justified only
market can provide but also forgoes future tax for a few higher traffic systems. Investment in
revenues from expanded business activity. To a water supply and irrigation would bring very
much lesser extent and only in middle-income solid returns in health benefits and productiv-
countries, the other sector showing potential ity, but returns to power generation need to
for reallocation is transport. The overspending be compounded by coordinated investment in
in this case is driven by apparent overinvest- transmission and distribution networks.
ment in road networks that, as will be seen
later, paradoxically coexists with undermainte- Improving Cost Recovery from
nance (table 2.4). User Charges
How much of that spending in ‘‘excess’’ of Two-thirds of African power and water utili-
infrastructure needs is influenced by politi- ties apply tariffs that comfortably cover oper-
cal factors? How far are these politics-tainted ating costs, but only one-fifth of those utilities
decisions from economic optimization? set tariffs high enough to recover full capital
How should these resources be reallocated? costs. Achieving recovery of only operating
Estimates of the economic rates of return to costs across all African power and water utili-
key infrastructure interventions can provide ties would raise $2.5 billion a year (0.4 percent
some answers. of the region’s GDP). Revising tariffs to make
Across infrastructure interventions in Africa, them equal to long-term marginal costs, and
the rates of return to road maintenance are thereby enabling all African power and water
Closing Africa’s Funding Gap 71
Table 2.4 Existing Disbursements above Those Directed to Infrastructure Needs, Annualized Flows
Percentage of GDP $ billions
Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total
Middle income — 1.4 — 0.0 0.1 1.5 — 3.7 — 0.0 0.3 4.1
utilities to recover capital costs also, would Underpricing user charges for roads costs the
increase the potential for efficiency gains to region some $0.6 billion a year (0.1 percent
$4.2 billion a year (0.7 percent of the region’s of GDP).
GDP; table 2.6). Although underpricing is
equally prevalent in power and water utilities, Reducing the Operating Inefficiencies
the value of the lost GDP revenues is slightly of Utilities
higher for power (at 0.4 percent of GDP) than African state-owned enterprises are charac-
for water (at 0.3 percent). terized by low investment and high operating
Raising tariffs to cost-recovery levels is evi- inefficiency. State-owned enterprises account
dently easier said than done and entails a host for between 80 percent (energy) and 40 per-
of social and political challenges. Chapter 3 in cent (water) of total public expenditures
this volume examines these issues in greater (general government and nonfinancial enter-
depth and provides a realistic appraisal of the prises). Despite their large resource base, they
feasibility of improving cost recovery for util- invest comparatively little—on average, an
ity services in Africa. equivalent of between 15 percent (energy) and
In the road sector, a widespread movement 18 percent (water) of the government resource
exists for using fuel levies and taxes as indirect envelope. As a result, governments are typi-
user charges (see chapter 10 in this volume). cally required to step in to assume most of the
For this system to work, fuel levies need to investment responsibilities of state-owned
be set high enough to cover the maintenance enterprises, which are relegated to undertak-
costs imposed by the use of the road network. ing daily operation and maintenance. In many
Comparing existing fuel levies with the levels cases, investment is unaffordable because of
needed to secure road maintenance makes it the significant underpricing of services, which
possible to estimate the underpricing in roads. barely allows the recovery of operating costs.
72 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
twice the best practice of 10 percent. Similarly, Figure 2.3 Rehabilitation Backlog
African water utilities typically lose 35 percent 50
of their water in distribution losses, nearly
twice the 20 percent benchmark. The finan- % of assets in need of rehabilitation
40
cial value of those distribution losses is much
higher for power at $1.3 billion per year than
30
for water at $0.5 billion per year.
20
Undercollection of Bills. The undercollection
of bills amounts to $2.9 billion a year (0.5 per-
10
cent of GDP). African power and water utili-
ties manage to collect about 90 percent of the
0
bills owed to them by their customers, short of
ion
er
ge
ds
te
ad
ay
ag
ag
tio
at
a best practice of 100 percent. Again, although
oa
ra
wa
at
ilw
ro
er
er
iga
lw
ve
lr
er
av
av
ain
ra
n
la
ra
irr
ra
en
ba
water utilities perform worse than power utili-
al
all
ru
ru
ra
m
rg
ur
ur
er
ru
we
nr
ov
no
of the losses is much greater for power. In Source: Briceño-Garmendia, Smits, and Foster 2008.
many African countries, public institutions
are among the worst offenders in failing to
pay for utility services. The undercollection of On average, 30 percent of African infra-
fuel levies for road sector maintenance is also structure assets need rehabilitation (figure 2.3).
an issue, although the absolute values for this Although documenting the exact magnitude
inefficiency are smaller than expected. of undermaintenance is difficult, the share of
today’s assets in need of rehabilitation provides
Undermaintenance. Deferring maintenance a good indicator of past neglect. In general,
expenditures is perhaps the most perverse ineffi- the state of rural infrastructure is substantially
ciency and the hardest to quantify. Given the pre- worse than the rest, with 35 percent of assets in
carious financing position of the infrastructure need of rehabilitation, compared with 25 per-
sectors, cutting back on maintenance is often cent elsewhere and 40 percent of roads. Wide
the only way to make ends meet, but spending differences exist across countries. In the best
too little on maintenance is a false economy. cases (Burkina Faso and South Africa), little
Rehabilitating or replacing poorly maintained more than 10 percent of assets need rehabilita-
assets is much more costly than keeping them up tion, and in the worst cases (the Democratic
with sound preventive maintenance. Moreover, Republic of Congo, Nigeria, Rwanda, and
consumers end up suffering as service quality Uganda), more than 40 percent do.
gradually declines. Indeed, not providing main- For roads alone, undermaintenance over
tenance and replacement investment is the most time leads to additional capital spending
costly way of financing today’s operations. of $1.4 billion a year (0.2 percent of GDP).
74 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Although undermaintenance affects all infra- transferring resources from areas that seem to
structure sectors, only for roads were suffi- be overfunded to areas that are clearly under-
cient data available to quantify the cost. Every funded. The next-largest potential gain of
$1 that goes unspent on road maintenance $4.7 billion a year would come from raising
leads to a $4 liability to the economy (Nogales user charges for infrastructure services. Again,
2009). Therefore, capital spending on roads is better pricing of power produces the greatest
much higher than it would otherwise need to dividends. Finally, raising budget execution
be—with continual reconstructing of the same ratios through improvements in the public
assets rather than creating new ones. The vast expenditure framework could capture an addi-
majority of Sub-Saharan African countries do tional $2 billion a year.
not cover road maintenance costs; more than Addressing some of the operational defi-
half of the countries have shortfalls of over ciencies may require substantial investments
40 percent of maintenance needs. However, in network rehabilitation or system upgrades.
institutions seem to have an important role to Reallocating resources, raising user charges, and
play. Countries with well-designed second- reducing overemployment all carry significant
generation road funds seem to do much bet- political costs, which complicate their imple-
ter in meeting their maintenance needs (see mentation. Therefore, expecting that all these
chapter 10 in this volume). efficiency gains could be fully captured is unre-
alistic. Given the magnitude of the needs, cap-
Closing the Efficiency Gap by turing only half of them would much improve
Promoting Reforms the financing and the perspectives for new
In sum, $17.4 billion could be captured financing in the African infrastructure sectors.
through improvements in infrastructure man- Even if all these efficiency gains could be
agement and institutions. The largest potential fully realized, a sizable funding gap would
gains of $7.5 billion a year come from address- remain. Chapter 1 of this volume identified
ing operating inefficiencies. Some of the most spending requirements of $93 billion a year to
pressing and most rewarding would be resolv- address Africa’s infrastructure needs. Setting
ing undermaintenance of roads and increas- these requirements against the $45 billion of
ing the efficiency of the power utilities. The existing spending directly traced to these needs
second-largest potential gains of $3.3 billion and the $17 billion of potential efficiency gains
a year come from improving the allocation of still leaves an annual infrastructure funding
existing resources across sectors, essentially gap of $31 billion (table 2.8).
Table 2.8 Finding Resources: The Efficiency Gap and the Funding Gap
$ billions annually
Cross-sector
Item Electricity ICT Irrigation Transport WSS gain Total
Infrastructure spending needs (40.8) (9.0) (3.4) (18.2) (21.9) n.a. (93.3)
Existing spending 11.6 9.0 0.9 16.2 7.6 n.a. 45.3
Efficiency gap 6.0 1.3 0.1 3.8 2.9 3.3 17.4
Gain from raising capital execution 0.2 0.0 0.1 1.3 0.2 n.a. 1.9
Gain from eliminating operational
inefficiencies 3.4 1.2 — 1.9 1.0 n.a. 7.5
Gain from tariff cost recovery 2.3 — — 0.6 1.8 n.a. 4.7
Potential for reallocation n.a. n.a. n.a. n.a. n.a. 3.3 3.3
Funding gap (23.2) 1.3 (2.4) 1.9 (11.4) 3.3 (30.6)
Source: Briceño-Garmendia, Smits, and Foster 2008.
Note: ICT = information and communication technology; n.a. = not applicable; — = not available; WSS = water supply and sanitation.
Parentheses indicate negative values.
Closing Africa’s Funding Gap 75
in recent years in line with political pledges, but with infrastructure investment bearing much
this assistance could slow down if countercycli- of that, falling by almost 1.5 percent of GDP. In
cal assistance is put in place. Third, non-OECD middle-income countries, budgetary spend-
finance has been rising steeply, but its future ing increased by almost 4.1 percent of GDP,
is now unclear. Fourth, private participation, but the effect on infrastructure spending was
also very buoyant during Africa’s recent growth almost negligible, and the additional resources
upswing, will be particularly vulnerable to the went primarily to current social sector spend-
downturn in global markets. Finally, local ing. Only in the low-income countries did the
capital markets have so far contributed little to overall increases in budgetary expenditure
infrastructure finance outside South Africa, but have some effect on infrastructure spend-
they could eventually become more important ing. Even there, however, the effect was fairly
in some of the region’s larger economies. modest and confined to capital spending. The
nonfragile low-income countries have allo-
Little Scope for Raising More cated 30 percent of the budgetary increase
Domestic Finance to infrastructure investments. The fragile
A key question is the extent to which coun- states, despite seeing their overall budgetary
tries may be willing to allocate additional fiscal expenditures increase by about 3.9 percent
resources to infrastructure. In the run-up to the of GDP, have allocated only 6 percent of the
current financial crisis, the fiscal situation in increase to infrastructure.
Sub-Saharan Africa was favorable. Rapid eco- Compared with other developing regions,
nomic growth averaging 4 percent a year from public financing capabilities in Sub-Saharan
2001 to 2005 translated into increased domestic Africa are characterized by weak tax revenue
fiscal revenues of just over 3 percent of GDP on collection. Domestic revenue generation around
average. In resource-rich countries, burgeoning 23 percent of GDP trails averages for other
resource royalties added 7.7 percent of GDP to developing countries and is lowest for low-
the public budget. In low-income countries, income countries (less than 15 percent of GDP
substantial debt relief increased external grants a year). Despite the high growth rates in the last
by almost 2 percent of GDP. decade, domestically raised revenues grew by less
To what extent were the additional resources than 1.2 percent of GDP. This finding suggests
available during the recent growth surge allo- that increasing domestic revenues above what
cated to infrastructure? The answer is surpris- is currently raised would require undertaking
ingly little (table 2.10). The most extreme case challenging institutional reforms to increase the
is that of the resource-rich countries, particu- effectiveness of revenue collection and broaden
larly Nigeria. Huge debt repayments more the tax base. Without such reforms, domestic
than fully absorbed the fiscal windfalls in revenue generation will remain weak.
these countries. As a result, budgetary spend- The borrowing capacity from domestic
ing actually contracted by 3.7 percent of GDP, and external sources is also limited. Domestic
Table 2.10 Net Change in Central Government Budgets, by Economic Use, 1995–2004
percentage of GDP
Sub-Saharan Low-income Low-income
Use Africa Middle income Resource rich nonfragile fragile
Net expenditure budget 1.89 4.08 (3.73) 1.69 3.85
Current infrastructure
spending as a share
of expenditures 0.00 0.02 0.03 0.00 0.09
Capital infrastructure
spending as a share
of expenditures (0.14) 0.04 (1.46) 0.54 0.22
Source: Africa Infrastructure Country Diagnostic, adapted from Briceño-Garmendia, Smits, and Foster 2008.
Note: Based on annualized averages for 2001–06. Averages weighted by country GDP. Totals are extrapolations based on the 24-country
sample as covered in AICD Phase 1.
Closing Africa’s Funding Gap 77
borrowing is often very expensive, with inter- example, Indonesia’s total public investment
est rates far exceeding those on concessional in infrastructure dropped from 6–7 percent of
external loans. Particularly for the poorest GDP in 1995–97 to 2 percent in 2000. Given
countries, the scarcity of private domestic sav- recent spending patterns, every reason exists
ings means that public domestic borrowing to expect that, in Africa, changes in the over-
tends to precipitate sharp increases in interest all budget envelope will affect infrastructure
rates, building up a vicious circle. For many investment in a similar pro-cyclical manner.
Sub-Saharan countries, the ratios of debt ser-
vice to GDP are more than 6 percent. Official Development Assistance—
The global financial crisis can be expected to Sustaining the Scale-Up
reduce fiscal receipts because of lower revenues For most of the 1990s and early 2000s, ODA to
from taxes, royalties, and user charges. Africa is infrastructure in Sub-Saharan Africa remained
not exempt from its impact. Growth projections steady at a meager $2 billion a year. The launch
for the coming years have been revised down- of the Commission for Africa Report in 2004
ward from 5.1 to 3.5 percent, which will reduce was followed by the Group of Eight Gleneagles
tax revenues and likely depress the demand
and willingness to pay for infrastructure ser-
vices. Commodity prices have fallen to levels of
the early 2000s. The effect on royalty revenues, BOX 2.1
however, will depend on the saving regime in
each country. A number of oil producers have
been saving royalty revenues in excess of $60 a
Does Deficit-Financed Public Investment in
barrel, so the current downturn will affect sav- Infrastructure Pay for Itself?
ings accounts more than budgets. Overall, this Underinvestment in infrastructure, health, and education during much
adverse situation created by the global finan- of the 1990s has ignited a lively debate on whether some countries
cial crisis will put substantial pressure on public could tolerate a larger public deficit if the additional resources were
sector budgets. In addition, many African coun- invested in growth-enhancing sectors. The analysis undertaken by
tries are devaluing their currency, reducing the the International Monetary Fund does not explicitly take into account
purchasing power of domestic resources. the potential link between public investment and growth—only its
short-term costs. Nevertheless, running a short-term deficit now may
Based on recent global experience, fiscal
help produce the growth that will balance the budget later.
adjustment episodes tend to fall dispropor-
By incorporating this long-run growth effect into the standard
tionately on public investment—and infra- models used to assess fiscal sustainability, one can see whether taking
structure in particular.2 Experience from a longer-term perspective would lead to a more favorable stance for
earlier crises in East Asia and Latin America deficit-financed infrastructure. The results turn out to be very country
indicates that infrastructure spending is par- specific, underscoring the difficulty of generalizing in this area.
ticularly vulnerable to budget cutbacks dur- In Uganda, investment in infrastructure leads to higher output,
ing crisis periods. Based on averages for eight but also—because of its relatively low productivity—worsens the
Latin American countries, cuts in infrastruc- debt ratio. A better way to finance infrastructure may be to improve
ture investment amounted to about 40 percent the existing capital stock by prioritizing O&M expenditure over new
of the observed fiscal adjustment between the investments. Although increased public expenditure on health and
education also leads to higher output, the effect is not as large as for
early 1980s and late 1990s (Calderón and
infrastructure.
Servén 2004). This reduction was remark-
In Senegal, by contrast, public investment in infrastructure does
able because public infrastructure investment not seem to be as effective in boosting growth. Both O&M spending
already represented less than 25 percent of on infrastructure and public investment in other sectors such as health
overall public investment in Latin American and education seem to have a stronger effect on growth. However,
countries. These infrastructure investment no matter how spending is allocated, it seems to worsen the debt-
cuts were later identified as the underlying to-GDP ratio, reflecting the low productivity of public expenditure in
problem holding back economic growth in the this case.
whole region during the 2000s (box 2.1). Sim-
Source: Estache and Muñoz 2008.
ilar patterns were observed in East Asia dur-
ing the financial crisis of the mid-1990s. For
78 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Summit in July 2005, where the Infrastructure ODA commitments were also set to increase
Consortium for Africa was created to focus further before the crisis, but prospects no longer
on scaling up donor finance to meet Africa’s look so good. The three multilateral agencies—
infrastructure needs. The main bilateral and the African Development Bank, the European
multilateral donors committed to double by Commission, and the World Bank—secured
2010 the (already higher) 2004 flows to reach record replenishments for their concessional
$10 billion a year, about 1.6 percent of Africa’s funding windows for the three to four years
GDP at that time. Donors have so far lived up beginning in 2008. In principle, funding allo-
to their promises, and ODA flows to African cations to African infrastructure totaling $5.2
infrastructure almost doubled from $4.1 bil- billion a year could come from the multilateral
lion in 2004 to $8.1 billion in 2007. Close to agencies alone in the near future. In practice,
three-quarters of ODA comes from multi- however, the crisis may divert multilateral
lateral donors (African Development Bank, resources from infrastructure projects and
European Community, and International toward emergency fiscal support. Bilateral sup-
Development Association [IDA]), while Japan port, based on annual budget determinations,
and the United States drive the doubling of may be more sensitive to the fiscal squeeze
bilateral commitments. in OECD countries, and some decline can be
A significant lag occurs between ODA com- anticipated. Historical trends suggest that ODA
mitments and their disbursement, suggesting has tended to be pro-cyclical rather than coun-
that disbursements should continue to increase tercyclical (IMF 2009; ODI 2009; UBS Invest-
in the coming years. The commitments just ment Research 2008; World Economic Outlook
reported are significantly higher than the 2008; and references cited therein).
estimated ODA disbursements of $3.8 billion
(table 2.11). This gap reflects the normal Non-OECD Financiers—Will
lags associated with project implementation. Growth Continue?
Because ODA is channeled through the gov- Non-OECD countries financed about $2.6 bil-
ernment budget, the execution of funds faces lion of African infrastructure annually between
some of the same problems affecting domes- 2001 and 2006 (table 2.12).3 This sum is not
tically financed public investment, including far short of the volumes from ODA; however,
procurement delays and low country capacity the focus of the finance is very different. Non-
to execute funds. Divergences between donor OECD financiers have been active primarily in
and country financial systems, as well as unpre- oil-exporting countries (Angola, Nigeria, and
dictability in the release of funds, may further Sudan). The bulk of their resources have gone
retard the disbursement of donor resources. to power and to transport. In the power sec-
Bearing all this in mind, if all commitments tor, primarily hydroelectric schemes received
up to 2007 are fully honored, ODA disburse- $1 billion per year, and in the transport sec-
ments could be expected to rise significantly tor, railways received nearly $1 billion a year.
(IMF 2009; World Economic Outlook 2008). For electricity, that amounts to 0.17 percent
Table 2.12 Historic Annualized Investment Flows from China, India, and Arab Countries
Percentage of GDP $ billions
Country type Electricity ICT Irrigation Transport WSS Total Electricity ICT Irrigation Transport WSS Total
Middle income 0.00 0.01 0.00 0.01 0.00 0.02 0.00 0.02 0.00 0.02 0.01 0.05
Resource rich 0.33 0.06 0.00 0.34 0.04 0.76 0.74 0.13 0.00 0.75 0.08 1.69
Low-income nonfragile 0.12 0.15 0.00 0.22 0.05 0.54 0.13 0.17 0.00 0.24 0.05 0.59
Low-income fragile 0.58 0.07 0.00 0.11 0.06 0.82 0.21 0.03 0.00 0.04 0.02 0.30
Africa 0.17 0.05 0.00 0.16 0.03 0.41 1.08 0.34 0.00 1.06 0.16 2.64
Source: Africa Infrastructure Country Diagnostic, adapted from Foster and others 2008.
Note: Based on annualized averages for 2001–06. Averages weighted by country GDP. Figures are extrapolations based on the 24-country sample covered in AICD Phase 1.
Totals may not add exactly because of rounding errors. ICT = information and communication technology; WSS = water supply and sanitation.
of African GDP, significantly larger than the For the three major sources of external finance,
0.11 percent coming from ODA. significant complementarity exists, despite
China’s official economic assistance qua- some overlap. PPI seeks the most commercially
drupled between 2001 and 2005, reaching lucrative opportunities in telecommunications.
more than 35 Sub-Saharan countries. Most Non-OECD financiers focus on productive
of the inflows have gone to resource-rich infrastructure (primarily power generation and
countries, in some cases making use of barter railroads). Traditional ODA focuses on financ-
arrangements under the “Angola mode.” 4 This ing public goods (such as roads and water sup-
type of South-South cooperation builds on ply) and plays a broader role in power system
economic complementarities between China development and electrification.
and Africa. China takes a strategic interest in A similar pattern of specialization emerges
Africa’s natural resource sector, while Africa geographically, with different countries rely-
harnesses China’s strengths in construction to ing to differing degrees on the various sources
develop its economic infrastructure. of finance. The countries most heavily reliant
India has become a significant financier on PPI are Kenya and Nigeria, supplemented
of energy projects in Africa. India’s financial by ODA in Kenya and by Chinese financing in
assistance focused initially on export credits to Nigeria. The countries that rely predominantly
facilitate the purchase of Indian goods. However, on non-OECD financiers are often oil produc-
India has signaled a bold commitment to sup- ers (Angola, Gabon, Guinea, Mauritania, and
port big infrastructure projects, predominantly Sudan). Most of the remaining countries rely
in energy, with up to $1 billion in Nigeria primarily on traditional ODA (Burundi, Mali,
(including a 9-million-ton per year refinery, Niger, Rwanda, and Tanzania). Other coun-
a 200-megawatt power plant, and a 1,000- tries (the Democratic Republic of Congo and
kilometer cross-country railway) and close to Guinea) draw on a mixture of OECD and non-
$100 million a year in Sudan (for a 700-kilometer OECD sources.
oil pipeline from Khartoum to Port Sudan and The implementation process for ODA and
four 125-kilowatt power plants). non-OECD finance is completely different. A
The Gulf States, through their various devel- key difference between Chinese finance and
opment agencies, have been funding African ODA is that whereas the latter is channeled
infrastructure for some time. Infrastructure through the government budget, the former
projects of a smaller scale than those funded tends to be executed directly by China, often
by the Chinese and Indian governments char- with associated imports of human resources.
acterize their portfolio, with strong support to Although this approach raises significant chal-
such countries as Mali, Mauritania, Senegal, lenges, it does at least offer the possibility of
and Sudan. Resources from the Gulf States circumventing some of the capital budget
have been distributed almost equally among execution problems typically associated with
water, roads, and small energy projects. public investment.
80 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Non-OECD finance also raises concerns sectors, such as WSS, attracted almost no
about sustainability. The non-OECD financiers private activity. The same is true of longer-
from China, India, and the Gulf States follow term and higher-risk projects. Through 2004,
sectors, countries, and circumstances aligned greenfield and small projects accounted for
with their national business interests. They offer 70 percent of all PPI, while concessions and
realistic financing options for power and trans- divestitures of incumbent utilities accounted
port and for postconflict countries with natural for less than 10 percent. Greenfield transac-
resources. However, nongovernmental organiza- tions, with no long-term risks and little or no
tions are voicing concerns about the associated investment, are much more prominent than in
social and environmental standards. Non-OECD other regions, and they tend to be small.
financiers also provide investment finance Africa’s resource-rich countries have been
without associated support on the operational, capturing the largest volume of private partici-
institutional, and policy sides, raising questions pation. Relative to their GDP, Africa’s middle-
about the sustainability of the new assets. income countries are not doing that well, while
How the current economic downturn will low-income countries—even fragile states—
affect non-OECD finance is difficult to pre- are capturing flows worth well over 1 percent
dict because of the relatively recent nature of of GDP.
these capital inflows. Coming from fiscal and Since the mid-1990s, a shift has occurred
royalty resources in their countries of origin, toward projects with longer horizons. Conces-
they will likely suffer from budgetary cutbacks. sions and existing assets increased to 20 percent
The downturn in global commodity prices of the private partnerships in infrastructure.
may also affect the motivation for some of the Sectors other than ICT have increased; the most
Chinese infrastructure finance linked to natu- important recorded transactions are in trans-
ral resource development. port, such as the concessions in Sudan for the
Juba Port ($30 million) and Uganda’s Rift Val-
Private Investors—over the Hill ley railways ($400 million concession). More-
Since the late 1990s, private investment flows over, larger greenfield power projects, beyond
to Sub-Saharan African infrastructure tripled, concessions and management contracts, are
increasing from about $3 billion in 1997 to starting to emerge.
$9.4 billion in 2006/07. That is about 1.5 per- Private capital flows, in particular, are likely
cent of regional GDP for all sectors, more than to be affected by the global financial crisis. In
recent ODA flows (0.6 percent of GDP, or $3.7 the aftermath of the Asian financial crisis, pri-
billion a year) but still less than half of general vate participation in developing countries fell
government spending (table 2.13). by about one-half over a period of five years,
Close to two-thirds of cumulative private following its peak in 1997. Existing transac-
commitments between 1990 and 2006 were in tions are also coming under stress as they
ICT-related projects (Leigland and Butterfield encounter difficulties refinancing short- and
2006). Power was second. Socially challenging medium-term debt.
Local Sources of Finance—a Possibility local financing was in corporate bonds, all to
in the Medium Term finance transport.
Local capital markets are a major source of Only 10 percent of outstanding bank loans
infrastructure finance in South Africa, but are for financing infrastructure investments. At
not yet elsewhere. Local infrastructure finance about $5 billion, this sum is a little less than the
consists primarily of commercial bank lend- total for Malaysia alone.
ing, some corporate bond and stock exchange However, a recent trend indicates new issu-
issues, and a nascent entry of institutional ers (particularly of corporate bonds) are com-
investors. Because the scale of local financ- ing to market in several countries, in some cases
ing in South Africa and its advanced state of with a debut issue. More than half (52 percent)
evolution are so far ahead of those elsewhere, of the corporate bonds listed on the markets at
attention here focuses on prospects elsewhere year-end 2006 were by infrastructure compa-
in the region. nies. The share of corporate bonds outstand-
Outside South Africa, the stock of outstand- ing at year-end 2006 that had been issued to
ing local infrastructure finance amounts to finance infrastructure exceeded half in 7 of the
$13.5 billion (table 2.14). This figure comprises 11 countries with bond markets reporting these
transport, the first-ranking sector of all local data. West Africa’s regional exchange, the BRVM
infrastructure financing, attracting 47 percent (Bourse Régionale des Valeurs Mobilières), had
of the total, followed by ICT at 32 percent.5 the highest share of issues financing infrastruc-
The low-income nonfragile countries were ture (more than 90 percent). The amount of
the destination for 55 percent of all local infra- financing is still small, however.
structure financing identified in this study. The Local financial markets remain underdevel-
two low-income fragile countries (Côte d’Ivoire oped, shallow, and small. Long-term financing
and the Democratic Republic of Congo) with maturities commensurate with infra-
attracted just 3.5 percent ($474 million), nearly structure projects is scarce.6 The capacity of
three-quarters of it in bank financing and the local banking systems remains too small and
remainder in equity issues by companies in constrained by structural impediments to
Côte d’Ivoire. For the resource-rich countries, finance infrastructure. Most countries’ banks
the $4.9 billion in local infrastructure financ- have significant asset-liability maturity mis-
ing was a nearly equal mix of bank and equity matches for infrastructure financing. Bank
financing. For the three middle-income coun- deposits and other liabilities still have largely
tries, more than half of the $544 million in short-term tenors. More potential may exist
for syndicated lending with local bank par- (figure 2.4). For public funds, raising taxes is
ticipation—though the increase in new loans not a costless exercise. Each dollar raised and
over 2000–06 occurred in a favorable external spent by a Sub-Saharan African government
financing environment. has a social value premium (or marginal cost
Harnessing the significant potential for of public funds) of almost 20 percent. That
local capital markets to finance infrastructure, premium captures the incidence of that tax
particularly local bond markets, is contingent on the society’s welfare (caused by changes
on their further development. It is also contin- in consumption patterns and administrative
gent on further reforms, including those that costs, among other things).9 To allow ready
would deepen the local institutional inves- comparisons across financing sources, this
tor base. Well-functioning and appropriately study standardized the financial terms as the
regulated local institutional investors (pen- present value of a dollar raised through each
sion funds and insurance companies) would of the different sources. In doing so, it recog-
be natural sources of long-term financing for nized that all loans must ultimately be repaid
infrastructure because their liabilities would with tax dollars, each of which attracts the
better match the longer terms of infrastructure 20 percent cost premium.
projects. Private pension providers have begun Wide variation exists in lending terms.
to emerge with a shift from defined-benefit to The most concessional IDA loans charge zero
defined-contribution schemes, viewed as less interest (0.75 percent service charge) with
costly, more transparent, and easier to man- 10 years of grace. India, China, and the Gulf
age. Moreover, local institutional investors are States, respectively, charge 4 percent, 3.6 per-
taking a more diversified portfolio approach to cent, and 1.5 percent interest and grant four
asset allocation. years of grace.10
Regional integration of financial mar- The cost of non-OECD finance is some-
kets could achieve greater scale and liquidity. where between that of public funds and ODA.
More cross-border intraregional listings—of The subsidy factor for Indian and Chinese
both corporate bonds and equity issues—and funds is about 25 percent and for the Arab
more cross-border intraregional investment funds, 50 percent. ODA typically provides a
(particularly by local institutional investors) subsidy factor of 60 percent, rising to 75 per-
could help overcome national capital markets’ cent for IDA resources. In addition to the cost
impediments of small size, illiquidity, and of capital, the different sources of finance dif-
inadequate market infrastructure. They could fer in the transaction costs associated with
also facilitate the ability of companies and gov- their use, which may offset or accentuate some
ernments to raise financing for infrastructure.7 of the differences.
So far, this intraregional approach to rais-
ing infrastructure financing remains largely
untapped.8 Most Promising Ways to
The African banking system did not feel the Increase Funds
effects of the global financial crisis at first, but
the crisis is slowly but surely affecting finan- Given this setting, what are the best ways of
cial systems around the region, adding to the increasing the availability of funds for infra-
already enormous challenge of developing structure development? The place to start is
local financial markets. clearly to get the most from existing budget
envelopes, which can provide up to $17.4 bil-
lion a year of additional resources internally.
Costs of Capital from Different Beyond that, a substantial funding gap still
Sources remains. Before the financial crisis, the pros-
pects for reducing—if not closing—this gap
The various sources of infrastructure finance were reasonably good. Resource royalties were
reviewed in the previous sections differ at record highs, and all sources of external
greatly in their associated cost of capital finance were buoyant and promising further
Closing Africa’s Funding Gap 83
growth. With the onset of the global financial Figure 2.4 Costs of Capital by Funding Source
crisis, that situation has changed significantly
and in ways that are not yet entirely foresee- public funds 1.17
able. The possibility exists across the board that private sector 1.1
all sources of infrastructure finance in Africa India 0.91
may fall rather than increase, further widening China 0.87
the funding gap. Only resource-rich countries Gulf States 0.65
have the possibility of using natural resource 0.51
official development assistance
savings accounts to provide a source of financ-
International Development Association 0.33
ing for infrastructure, but only if macroeco-
nomic conditions allow. One of the few things 0 0.20 0.40 0.60 0.80 1.00 1.20
cost of raising $1 of financing (US$)
that could reverse this overall situation would
be the agreement upon a major stimulus pack- Sources: Average marginal cost of public funds: as estimated by Warlters and Auriol 2005; cost of
equity for private sector: as in Estache and Pinglo 2004 and Sirtaine and others 2005; authors’
age for Africa by the international community, calculations.
with a focus on infrastructure as part of the
effort to rekindle economic growth and safe-
guard employment. conclusion assumes they have first fully cap-
tured efficiency gains. Without such efficiency
gains, the targets could not be met even over 30
What Else Can Be Done? years without increasing spending above cur-
rent levels (figure 2.5, panel b).
Most of the low-income countries, and in Low-income nonfragile and resource-rich
particular the fragile states, face a substantial countries would need to delay an additional
funding gap even if all the existing sources decade to meet targets with existing spending
of funds—including efficiency gains—are levels. By spreading the investment needs over
tapped. What other options do these countries 20 rather than 10 years, these countries could
have? Realistically, they need either to defer the achieve the proposed targets within existing
attainment of the infrastructure targets pro- spending envelopes (figure 2.5, panel a). Again,
posed here or to try to achieve them by using this outcome would be possible only if efficiency
lower-cost technologies. gains are fully exploited. Otherwise, they would
need more than 30 years to reach the target with
Taking More Time existing resources (figure 2.5, panel b).
The investment needs presented in this book Low-income fragile states would need to
are based on the objective of addressing Africa’s delay by more than two decades to meet infra-
infrastructure backlog within 10 years. To meet structure targets within existing spending lev-
this target, middle-income, resource-rich, and els. By spreading the investment needs over 30
low-income nonfragile states would need to rather than 10 years, low-income fragile states
increase their existing infrastructure spending could achieve the proposed targets within
by 50 to 100 percent, while low-income fragile existing spending envelopes (figure 2.5, panel
states would need to increase their infrastructure a). However, without efficiency gains, these
spending by an impossible 350 percent. Extend- countries would take much longer than 30
ing the time horizon for the achievement of years to meet the associated targets or alter-
these goals should make the targets more afford- natively would still need to double their exist-
able. But how long a delay would be needed to ing spending to reach the target in 30 years
make the infrastructure targets attainable with- (figure 2.5, panel b).
out increasing existing spending envelopes?
By delaying only three years, spreading Using Lower-Cost Technologies
the investment needs over 13 rather than 10 Many possible alternative technological solu-
years, middle-income countries could achieve tions exist for meeting a given infrastructure
the proposed targets within existing spending target, and each offers a particular combina-
envelopes (figure 2.5, panel a). However, this tion of financial cost and quality of service.
84 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
500 500
400 400
350 350
300 300
250 250
200 200
150 150
100 100
50 50
0 0
–50 –50
10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
number of years needed to attain infrastructure targets number of years needed to attain infrastructure targets
fragile low-income countries nonfragile low-income countries
middle-income countries resource-rich countries
Where budgets are constrained, policy mak- unimproved latrines, remains the same as it is
ers face a choice between providing a high today (see chapters 16 and 17 in this volume).
level of service to a few people or a lower level Thus, as population grows, the number of peo-
of service to a broader cross-section of the ple accessing high-level services will be larger
population. Critical trade-offs must be con- in absolute terms. If instead, all additional
sidered; thus, one cannot jump to the conclu- people served use cheaper solutions, such as
sion that a high level of service is always in a standposts and improved latrines in urban
country’s best interest. The extent to which areas or boreholes and unimproved latrines
cost-saving technologies are available var- in rural areas, the overall cost of meeting the
ies considerably across sectors. Two of the MDGs would fall by 30 percent.
clearest cases are water and roads, which are
discussed in detail next. Unfortunately, the Using Alternative Technologies in Roads. In the
power sector (which has by far the largest case of roads, the cost of reaching regional and
associated investment tag) does not present national connectivity targets can be reduced
many technological alternatives for reducing by 30 percent by adopting lower standards for
the cost of electricity generation. trunk roads. Road connectivity targets can be
attained by using different engineering stan-
Using Alternative Technologies in WSS. In the dards. The scenario considered here is one in
case of water and sanitation, the cost of achiev- which regional and national connectivity are
ing the MDGs drops by 30 percent with greater achieved by a good-condition asphalt road
reliance on lower-cost technologies. The network with at least two lanes for regional
MDGs can be achieved using either higher- and at least one lane for national connectivity.
end solutions, such as piped water and septic The same connectivity could be achieved at a
tanks, or cheaper solutions, such as standposts cost reduction of 30 percent if a single-surface-
and improved latrines. The scenario consid- treatment road in fair condition is substituted
ered here is one where the MDGs are met by for an asphalt road in good condition.
preserving the prevalent mix of high-end and
Notes
lower-end technologies. That is, the relative
The authors of this chapter are Cecilia Briceño-
share of the population enjoying access to a Garmendia and Nataliya Pushak, who drew
direct water connection, sewers, or a septic on background material and contributions
tank—all regarded as high-level services— from William Butterfield, Chuan Chen, Vivien
compared to the share of people with access Foster, Jacqueline Irving, Astrid Manroth, Afua
to lower-end solutions, such as standposts and Sarkodie, and Karlis Smits.
Closing Africa’s Funding Gap 85
1. In particular, maintenance is essential to har- Calderón, César, and Luis Servén. 2004. “Trends
ness the economic returns of capital, but good- in Infrastructure in Latin America, 1980–2001.”
quality data on how much of current expendi- Policy Research Working Paper 3401, World
tures go to maintenance is hard to track. Bank, Washington, DC.
2. Servén (2005) and Hicks (1991) summarize the De Haan, Jakob, Jan Sturm, and Bernd Sikken.
facts on Latin American and other developing 1996. “Government Capital Formation: Explain-
countries. For industrialized countries, see also ing the Decline.” Weltwirtschaftliches Archiv
Roubini and others (1989); De Haan, Sturm, 132 (1): 55–74.
and Sikken (1996) document the experience of Estache, Antonio, and Rafael Muñoz. 2008. “Build-
industrialized countries. ing Sector Concerns into Macro-Economic
3. This section draws heavily on Foster and others Financial Programming: Lessons from Senegal
(2008). and Uganda.” Working Paper 6, Africa Infra-
4. Essentially, the Angola mode was devised to structure Country Diagnostic, World Bank,
Washington, DC.
enable African nations to pay for infrastructure
with natural resources. In a single transaction, Estache, Antonio, and Maria Elena Pinglo. 2004.
China bundles development-type assistance “Are Returns to Private Infrastructure in
Developing Countries Consistent with Risks
with commercial-type trade finance. A Chi-
since the Asian Crisis?” Policy Research Working
nese resource company makes repayments in Paper 3373. World Bank, Washington, DC.
exchange for the oil or mineral rights. The China
Export-Import Bank acts as a broker, receiving Foster, Vivien, William Butterfield, Chuan Chen,
and Nataliya Pushak. 2008. Building Bridges:
money for the sale and paying the contractor
China’s Growing Role as Infrastructure Finan-
for providing the infrastructure. This arrange- cier for Sub-Saharan Africa. Trends and Policy
ment safeguards against currency inconvertibil- Options no. 5. Washington, DC: Public-Private
ity, political instability, and expropriation. Infrastructure Advisory Facility, World Bank.
5. Data are as of year-end 2006, or most recent
Hicks, Norman. 1991. “Expenditure Reductions
available, for the sampled countries, excluding in Developing Countries Revisited.” Journal of
South Africa. International Development 3 (1): 29–37.
6. Because South Africa’s financial markets are so
Irving, Jacqueline, and Astrid Manroth. 2009.
much more developed than any of those of the
“Local Sources of Financing for Infrastructure
other 23 focus countries, this section excludes in Africa: A Cross-Country Analysis.” Policy
South Africa. Research Working Paper 4878, World Bank,
7. One new initiative is the Pan-African Infra- Washington, DC.
structure Development Fund, a 15-year regional IMF (International Monetary Fund). 2009. The State
fund for raising finance for commercially viable of Public Finances: Outlook and Medium-Term Pol-
infrastructure projects in Africa, which raised icies after the 2008 Crisis. Washington, DC: IMF.
$625 million in its first close in 2007, including
Leigland, James, and William Butterfield. 2006.
funds from Ghanaian and South African insti- “Reform, Private Capital Needed to Develop
tutional investors. Infrastructure in Africa: Problems and Pros-
8. In addition, the lack of a benchmark yield curve in pects for Private Participation.” Gridlines, Note
the vast majority of those African countries that 8 (May), Public-Private Infrastructure Advisory
have an organized bond market has limited cor- Facility, World Bank, Washington, DC.
porate bond issuance, as has the general absence Nogales, Alberto. 2009. “The Cost of Postponing
of credit ratings agencies and a lack of awareness Roads Maintenance.” World Bank, Washington,
among prospective issuers as well as investors. DC.
9. The marginal cost of public funds measures
ODI (Overseas Development Institute). 2009. A
the change in welfare associated with raising Development Charter for the G-20. London: ODI.
an additional unit of tax revenue (Warlters and
Auriol 2005). PPIAF (Public-Private Infrastructure Advisory
Facility). 2008. Private Participation in Infrastruc-
10. See Foster and others (2008) for further details.
ture Project Database, http://ppi.worldbank .org/.
Pritchett, Lant. 1996. “Mind Your P’s and Q’s. The
References Cost of Public Investment Is Not the Value of
Briceño-Garmendia, Cecilia, Karlis Smits, and Vivien Public Capital.” Policy Research Working Paper
Foster. 2008. “Fiscal Costs of Infrastructure in 1660, World Bank, Washington, DC.
Sub-Saharan Africa.” Africa Infrastructure Country Rajkumar, Andrew, and Vinaya Swaroop. 2002. “Pub-
Diagnostic. Washington, DC: World Bank. lic Spending and Outcomes: Does Governance
86 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Matter?” Policy Research Working Paper 2840, Sirtaine, Sophie, Maria Elena Pinglo, J. Luis
World Bank, Washington, DC. Guasch, and Viven Foster. 2005. How Profitable
Reinikka, Ritva, and Jakob Svensson. 2002. Are Infrastructure Concessions in Latin America?
“Explaining Leakage of Public Funds.” Empirical Evidence and Regulatory Implications.
Discussion Paper 3227, Centre for Economic Trends and Policy Options no. 2. Washington,
Policy Research, London, U.K. DC: Public-Private Infrastructure Advisory
Facility, World Bank.
———. 2003. “The Power of Information: Evi-
dence from a Newspaper Campaign to Reduce UBS Investment Research. 2008. “Global
Capture.” Policy Research Working Paper 3239, Economic Perspectives: The Global Impact
World Bank, Washington, DC. of Fiscal Policy.”
Roubini, Nouriel, Jeffrey Sachs, Seppo Honkapohja, Warlters, Michael, and Emmanuelle Auriol. 2005.
and Daniel Cohen. 1989. “Government Spending “The Marginal Cost of Public Funds in Africa.”
and Budget Deficits in the Industrial Countries.” Policy Research Working Paper 3679, World
Economic Policy 8 (4): 99–132. Bank, Washington, DC.
Servén, Luis. 2005. “Fiscal Discipline, Public World Economic Outlook. 2008. “Estimating the
Investment and Growth.” World Bank, Size of the European Stimulus Packages for
Washington, DC. 2009.”
Chapter 3
C
overage of modern infrastructure ser- In the absence of modern infrastructure
vices has been stagnant since the mid- services, the next best option would be to
1990s and remains strongly skewed reach households with lower-cost, second-best
toward more affluent households. In urban solutions, such as standposts, improved
areas, those who fail to hook up to nearby net- latrines, or street lighting. However, the preva-
works form a significant share of the unserved lence of these second-best solutions is surpris-
population, suggesting that demand-side barri- ingly low in Africa, and those that exist tend
ers are also at work. In these circumstances, the to cater more to the higher-income groups
key questions are whether African households than to the middle of the income distribu-
can afford to pay for modern infrastructure tion. The majority of Africans resort instead
services, and if not, whether African govern- to traditional alternatives, such as wells, unim-
ments can afford to subsidize them. proved latrines, or kerosene lamps. Significant
A subsistence power or water bill ranges challenges exist in increasing the coverage of
between $2 and $8 a month. This cost is well second-best alternatives, particularly because
within the affordable range for most house- their public good nature makes some of these
holds in Africa’s middle-income countries technologies more difficult for service provid-
and for the more affluent segments that cur- ers to operate on a commercial basis.
rently enjoy access to utilities in low-income The business-as-usual approach to expand-
countries. However, affordability would defi- ing service coverage in Africa does not appear
nitely become an issue for most people in the to be working. Turning this situation around
poorest low-income countries should access be will require rethinking the approach to service
broadened. expansion in four ways. First, coverage expan-
African governments already spend $4.1 sion is not only about network rollout, but also
billion a year (0.7 percent of GDP) on power about a need to address demand-side barriers
and water subsidies that benefit mainly a small such as high connection charges or legal tenure.
group of affluent customers. Expanding these Second, cost recovery for household services
levels of subsidy to the entire population would needs to be improved to ensure that utilities
be fiscally unsustainable for most countries. have the financial basis to invest in service
87
88 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
expansion. Third, rethinking the design of util- Figure 3.1 Coverage of Network Infrastructure
ity subsidies to better target them and to accel- Services, 1990–2005
erate service expansion is desirable. Fourth, 35
any approach must consider the actual level of 30
% of households
service that households can afford to pay and 25
that governments can afford to subsidize and 20
put greater emphasis on second-best alterna- 15
tives to modern infrastructure services. 10
5
0
1990–95 1996–2000 2001–05
piped water electricity
Access to Modern Infrastructure
flush toilets landline telephone
Services—Stagnant and
Inequitable Source: Banerjee, Wodon, and others 2008.
Among the poorest 60 percent of the population, Figure 3.2 Expansion of Access to Infrastructure Services Each Year,
coverage of almost all modern infrastructure Mid-1990s to Mid-2000s
services is well below 10 percent.Conversely, 2.0
the vast majority of households with coverage
% of population
1.5
belong to the more affluent 40 percent of the
population. In most countries, inequality of 1.0
access has increased over time, suggesting that
0.5
most new connections have gone to more afflu-
ent households (Diallo and Wodon 2005). This 0
es
ts
les
ty
es
situation is not entirely surprising, given that,
te
ile
ne
ne
os
ici
rin
rin
wa
ho
to
ho
ho
dp
ctr
lat
lat
re
sh
lep
ll p
ed
ele
even among households with greater purchas-
bo
sta
ed
al
flu
pip
te
ce
ion
ov
ne
pr
dit
ing power, coverage is far from universal, and
dli
im
tra
lan
well under 50 percent in most cases. Relative
to the other modern infrastructure services, Source: Banerjee, Wodon, and others 2008.
electricity coverage is somewhat higher across
the spectrum. Figure 3.3 Projected Universal Access for Piped Water
Low coverage rates can reflect both sup- for Sub-Saharan African Countries, 2050 and Beyond
ply and demand factors. On the one hand, 100
the household may be physically distant from
% of countries
80
an infrastructure network (and thus face an 60
absence of supply). On the other hand, the 40
household may choose not to connect to a 20
network even when it is nearby (and thus 0
piped water flush toilets electricity landline
express a lack of demand). Understanding telephones
this difference is important because the policy by 2050 2050–2100 2100–2200 after 2200
implications differ radically. By exploiting
the spatial distribution of household survey Source: Banerjee, Wodon, and others 2008.
samples in urban areas, one can quantify
the relative importance of these supply and Figure 3.4 Coverage of Modern Infrastructure
demand factors in accounting for low ser- Services, by Budget Quintile
vice coverage. Using this approach, one can
100
distinguish between the percentage of popula-
% of households
Table 3.1 Understanding Coverage of Infrastructure Services: The Role of The tenure status of households may also
Supply and Demand Factors significantly impede hookup to modern infra-
percent (population-weighted average)
structure services. A study of slum households
Decomposition of coverage Unserved due to factors in Dakar and Nairobi finds that coverage of
Infrastructure Access Hookup Coverage Supply Demand piped water and electricity is more than twice
Piped water as high among owner-occupiers as among ten-
Low-income countries 68 42 31 86 14 ants (Gulyani, Talukdar, and Jack 2008). Even
Middle-income countries 91 74 69 64 36 among owner-occupiers, lack of formal legal
titles can affect hookup to services.
Overall 73 48 38 81 19
Electricity
Low-income countries 93 73 69 50 50 Affordability of Modern
Middle-income countries 95 86 81 39 61 Infrastructure Services—
Overall 93 75 71 48 52 Subsidizing the Better Off
Source: Banerjee, Wodon, and others 2008.
Note: Access is defined as the percentage of the population that lives physically close to infrastructure. African households exist on very limited
Hookup is defined as the percentage of the population that connects to infrastructure when it is avail-
able. Coverage is defined as the percentage of the population that has the infrastructure service; it household budgets. The average African house-
is essentially the product of access and hookup.. In calculating the proportion of gap attributable to
demand and supply factors, one considers the hookup rate of the top budget quintile in each geo- hold of five persons has a monthly budget of
graphical area to be an upper bound on potential hookup, absent demand-side constraints. less than $180, ranging from about $60 in the
poorest quintile to $340 in the richest quintile
(table 3.2). Thus, purchasing power—even in
factors is higher in middle-income countries Africa’s most affluent households—is modest
than in low-income countries, reflecting the in absolute terms. Across the spectrum, house-
fact that infrastructure networks are more hold budgets in middle-income countries are
highly developed in the former and have a roughly twice those in low-income countries.
broader geographical reach. Most African households spend more than
It may appear paradoxical that households half their modest budgets on food, with little left
do not universally take up connections to over for other items. Spending on infrastructure
modern infrastructure services as networks services (including utilities, energy, and trans-
become physically available. Clear economic port) averages about 7 percent of a household’s
reasons exist, however, why this might be so. budget, though this can be 15–25 percent in
In some cases, households may have access to some countries and represents overall a signifi-
cheaper substitutes, such as boreholes. More cant share of the nonfood budget. As household
substitutes are available for piped water than budgets increase, infrastructure services absorb
for electricity, which may explain the much a growing share, rising from less than 4 percent
lower hookup rates for the former. In other among the poorest quintile to more than 8 per-
cases, utility connection charges are set pro- cent among the richest (figure 3.5). In terms
hibitively high for low-income households. For of absolute expenditure, this difference is even
example, 60 percent of the water utilities sur- more pronounced: whereas households in the
veyed for this study apply connection charges poorest quintile spend, on average, no more
in excess of $100. Charges range from about than $2 per month on all infrastructure services,
$6 in the Upper Nile in Sudan to more than households in the richest quintile spend almost
$240 in Côte d’Ivoire, Mozambique, Niger, and $40 per month.
South Africa. The average connection charge Given such low household budgets, a key
across the region is 28 percent of gross national question is whether households can afford to
income (GNI) per capita. For Niger, the charge pay for modern infrastructure services. One
is more than 100 percent of GNI per capita. measure of affordability is nonpayment for
Similarly, the five water utilities in Mozambique infrastructure services. Nonpayment directly
charge more than 75 percent of GNI per capita. limits the ability of utilities and service provid-
These comparisons illustrate how high connec- ers to expand networks and improve services by
tion charges present a barrier to affordability. undermining their financial strength. Based on
Dealing with Poverty and Inequality 91
Figure 3.5 Share of Household Budgets Spent on Figure 3.6 Population with Service Connections Who
Infrastructure Services, by Budget Quintile Do Not Pay for Service
9 40 100
8 35 90
expenditure (US$/month)
7 30 80
budget share (%)
% of population
6 70
25
5 60
20
4 50
15
3 40
2 10 30
1 5 20
0 0 10
poorest second third fourth richest
0
quintile poorest second third fourth richest
budget share (left axis) expenditure (right axis) quintile
electricity piped water
Source: Banerjee, Wodon, and others 2008.
Source: Banerjee, Wodon, and others 2008.
more typical modest household consumption few poor consumers are connected to the ser-
of 10 cubic meters). The cost of a monthly vice (Boccanfuso, Estache, and Savard 2008a,
subsistence consumption of power can range 2008b, 2008c). As the consequences of higher
between $2 (based on a low-cost country tar- power or water prices work their way through
iff of $0.08 per kilowatt-hour and an absolute the economy, however, broader second-order
minimum consumption of 25 kilowatt-hours) effects on wages and prices of goods in the
and $8 (based on a high-cost country tar- economy as a whole can lead to more substan-
iff of $0.16 per kilowatt-hour and a more tial effects on poverty (Boccanfuso, Estache,
typical modest household consumption of and Savard 2008a, 2008b, 2008c).
50 kilowatt-hours). Notwithstanding these findings, most Afri-
An affordability threshold of 5 percent can countries heavily subsidize tariffs for power
of household budgets provides a gauge for and water services. On average, power tariffs
measuring which utility bills might be afford- recover only 75 percent of full costs, and water
able to African households. By looking at the tariffs only 64 percent. The resulting implicit
distribution of household budgets, one can cal- service subsidies amount to as much as $4.1
culate the percentage of households for which billion a year (0.7 percent of Africa’s GDP),
such bills would absorb more than 5 percent divided evenly between power and water (see
of their budgets and thus prove unaffordable. chapter 2 of this volume).
Monthly bills of $2 are affordable to almost the Because electricity and water subsidies are
entire African population. Monthly bills of $8 typically justified by the need to make ser-
would remain affordable for the entire popula- vices affordable to low-income households, a
tion of the middle-income African countries, key question is whether subsidies reach such
indicating that cost recovery should not be a households. Results across a wide range of
major problem for these countries. African countries, for both power and water
Cost-recovery tariffs would also be afford- sectors, show that the share of subsidies going
able for those currently enjoying access in low- to the poor is less than half of the share of the
income countries, but not for the remaining poor in the overall population, indicating a very
population. In low-income countries, monthly regressive distribution (figure 3.8). This result
bills of $8 would remain perfectly affordable is hardly surprising given that connections to
for the most affluent 20–40 percent of the power and water services are already highly
population, the only portion enjoying access. skewed toward more affluent households. To
However, such bills would not be affordable to put these results in perspective, one must com-
the poorest 60–80 percent that currently lack pare them with the targeting achieved by other
access even if services were extended to them. forms of social policy. Estimates for Cameroon,
The affordability problems associated with a Gabon, and Guinea indicate that expenditures
universal access policy would be particularly on primary education and basic health care
large for a handful of the poorest low-income reach the poor better than do power and water
countries—Burundi, the Democratic Repub- subsidies (Wodon 2008a, 2008b).
lic of Congo, Ethiopia, Guinea-Bissau, Malawi, Can African governments afford to further
Niger, Tanzania, and Uganda—where as much expand today’s subsidy model to achieve uni-
as 80 percent of the population could not versal access? Little justification currently exists
afford a monthly bill of $8. for utility subsidies, given that they typically
The immediate poverty-related effect of do not reach unconnected low-income house-
raising tariffs to cost-recovery levels is gener- holds but rather favor more affluent, connected
ally quite small, although it may have second- households that do not really need subsidies
order effects. Detailed analysis of the effect to afford the service. However, the preced-
of significant tariff increases of the order of ing analysis indicates that affordability would
40 percent for power and water services in become a major issue to the extent that Africa’s
Senegal and power services in Mali confirms low-income countries move aggressively toward
that the immediate poverty-related effect on universal access. Given the very high macro-
consumers is small, essentially because very economic cost today of subsidizing even the
Dealing with Poverty and Inequality 93
Figure 3.8 Extent to Which Electricity and Water Subsidies Reach the Poor, by Country
a. Electricity b. Water
Nigeria Senegal
progressive
progressive
regressive
regressive
Mozambique Niger
Ghana Chad
minority of the population with access to power unconnected households over 20 years would
and water, a legitimate question is whether be substantially low, only 0.35 percent of GDP
African governments can afford to scale up this for power and about 0.25 percent of GDP for
subsidy-based model to the remainder of their piped water. A key difference is that the cost of
populations. this one-time subsidy would disappear at the
Providing universal use-of-service sub- end of the decade, whereas the use-of-service
sidies for power and water would absorb an subsidy would continue indefinitely.
unaffordable 1.6 percent of GDP above exist- The welfare case is quite strong for one-
ing spending, about two-thirds for power and time capital subsidies to support universal
one-third for water. These values are high in connection. Households without access to
relation to existing operation and maintenance utility services eventually pay much higher
expenditure, so it is difficult to believe that prices and, as a result, limit their consumption
they would be fiscally affordable for govern- to very low levels. Small-scale piped-network
ments (figure 3.9). operators charge 1.5 times the formal network
One-time capital subsidies could be pro- price, point sources charge 4.5 times the for-
vided at a lower cost and if spread over 20 mal network price, and mobile distributors
years, might just be affordable. The cost of can charge up to 12 times the formal utility
providing a one-time capital subsidy of $200 tariff (Kariuki and Schwartz 2005). A recent
to cover network connection costs for all survey of Accra, Dar es Salaam, and Nairobi
94 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 3.9 Amount of Subsidy Needed to Maintain Affordability of Water and Electricity Service, 2005
2.0 2.0
% of GDP
% of GDP
1.5 1.5
1.0 1.0
0.5 0.5
0 0
electricity water electricity water
operating subsidy needed to maintain affordability capital subsidy needed to maintain affordability
operation and maintenance spending (2005) capital spending (2005)
found that the price of utility-piped water Box 3.2 details some of the access and afford-
ranges from $0.5 to $1.5 a cubic meter, whereas ability challenges arising for urban transpor-
small water enterprises charge between $4 and tation in Africa’s burgeoning cities.
$6 (McGranahan and others 2006). Similarly,
for electricity, the cost of providing basic illu-
mination through candles or kilowatt-hours Alternatives to Modern
is an order of magnitude greater than that for Infrastructure Services—the
electricity per effective unit of lighting (Foster Missing Middle
and Tré 2003).
Interestingly, even though nonutility water Even with renewed efforts, most African coun-
vendors charge higher unit prices, those pur- tries will not realize universal access to modern
chasing water from vendors do not necessar- infrastructure services for some decades. In
ily spend more than those purchasing water the meantime, most households will continue
from the public utility—they simply adjust the to rely on alternative ways of meeting their
quantity consumed. water and energy requirements (figure 3.10).
Nonmonetary benefits of connection can For the most part, these methods comprise
also be very significant. Beyond the potential traditional alternatives such as wells, unim-
monetary savings, piped water and electricity proved latrines, and kerosene lamps. However,
are associated with a wide range of health, edu- second-best alternatives also exist that provide
cation, and productivity benefits (see chapter a significantly higher level of service than the
1 of this volume). Better water and sanitation traditional solutions but at a substantially
service is associated with less malnutrition and lower cost than full-blown piped water or
stunting, and it liberates women from the time- power connections. Examples include stand-
consuming chore of collecting water, leaving posts for water supply, improved latrines for
more time for income-generating activities sanitation, and street lighting for basic neigh-
(box 3.1). Better electricity provision improves borhood illumination.
literacy and primary school completion rates, Although traditional alternatives are in
because better quality of light allows students widespread use, second-best alternatives have
to read and study without sunlight. not yet become popular in most countries. The
Finally, affordability concerns also exist for water and sanitation sectors illustrate this point
urban transportation services. Transportation very clearly. In both cases, the traditional alter-
represents one of the largest household bud- natives (whether wells or unimproved latrines)
get expenditure shares among infrastructure provide the largest share of service across the
services, absorbing 4–6 percent of the budgets income spectrum. However, the second-best
of those households reporting expenditures. alternatives (whether standposts or improved
Dealing with Poverty and Inequality 95
BOX 3.1
latrines) have coverage rates comparable to or conspire to limit the extension of the second-
even lower than the best alternatives (whether best alternatives.
piped water or flush toilets) despite their sig- On the demand side, the costs of the
nificant cost advantages. Moreover, coverage second-best alternatives may still be relatively
of the second-best alternatives is just as regres- high, given limited household budgets. Water
sive as that of the best alternatives. Neverthe- from standposts, though relatively cheap to
less, some countries have made significant provide, is often retailed by intermediaries
progress in expanding coverage of second-best charging substantial markups that outweigh
alternatives—such as the Democratic Republic the underlying advantages in construction
of Congo, Mozambique, Tanzania, and Uganda costs. Improved latrines, though cheaper than
for standposts; and Burkina Faso, Cameroon, flush toilets, are nonetheless substantially
Ghana, Madagascar, and Rwanda for improved more expensive than unimproved latrines, and
latrines. Although data are not available to uneducated households may not be aware of
make a similar comparison for lighting, it is well the health benefits.
known that coverage of street lighting lags. On the supply side, their public good
The capital costs of the second-best alter- nature greatly complicates the implementa-
natives are still only a fraction of those asso- tion of second-best alternatives. The provision
ciated with the best alternatives, even if they of standposts and street lighting is unattractive
are also significantly more expensive than the to utilities because of their limited scope for
traditional alternatives (table 3.3). Thus, the revenue collection as well as the greater poten-
second-best alternatives provide an opportu- tial for revenue loss from increased clandestine
nity to make limited investment budgets go connections once networks are provided. For
further and accelerate the expansion of service improved latrines, the limited experience of
improvements. the local construction sector restricts the avail-
Therefore, understanding the factors that ability of such designs and may keep their costs
lie behind this “missing middle” is important. higher than they would be in a mass market
Once again, both demand and supply issues (see chapter 17 of this volume).
96 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
BOX 3.2
Figure 3.10 Access to Alternative Water and Sanitation Services across All Income Levels
50 50
% of households
% of households
40 40
30 30
20 20
10 10
0 0
Q1 Q2 Q3 Q4 Q5 Q1 Q2 Q3 Q4 Q5
budget quintile budget quintile
BOX 3.3
subsidies created by underpricing are extremely Various tactical measures can improve the
regressive in their distributional incidence. acceptability of tariff increases, but ensuring
In all but the poorest African countries, ser- their sustainability is most important. Tariff
vice coverage could be substantially increased increases can be phased in either gradually or
before any real affordability problems would be instantly through a one-time adjustment. Both
encountered. In the poorest of the low-income approaches have advantages and disadvantages.
countries, affordability is a legitimate concern The public acceptability of tariff increases can
for the bulk of the population and would con- be enhanced if they are part of wider measures
strain universal coverage. Even in the poorest that include service quality improvements.
countries, however, recovering operating costs One method to strengthen social accountabil-
should be feasible, with subsidies limited to ity is adopting communication strategies that
capital costs. link tariffs with service delivery standards and
How would removal of utility subsidies suggest conservation measures to contain the
affect poverty reduction? For most countries, overall bills. In any event, ensuring that the
electricity and water spending accounts for realignment of tariffs and costs is sustained
only a tiny fraction of total consumption. At by providing for automatic indexation and
the national level, a 50 percent increase in tar- periodic revisions of tariffs is perhaps most
iffs or even a doubling of tariffs has a marginal important.
effect, with the share of the population living Countries have pursued different paths
in poverty increasing barely 0.1 of a percentage to increase tariffs critical for operational and
point. Among households with a connection to financial sustainability. In Niger, the standpost
the network, the effect is larger but still limited. and low-volume tariffs have barely increased
Indeed, an increase in the share of households since 2000, but the industrial and commer-
in poverty larger than 1 or 2 percentage points cial tariffs have grown 6–7 percent in nominal
rarely occurs. Because the households that terms (figure 3.11). In Lilongwe, Malawi, the
benefit from a connection also tend to be bet- same increases have been applied across all tar-
ter off than other households, the increase in iff categories. In addition, Botswana, Namibia,
poverty starts from a low base. Thus, the small South Africa (Eskom), and Tanzania recently
effect of a tariff increase on poverty could be increased electricity tariffs as a result of oil
offset by reallocating utility subsidies to other price shocks in 2007–08.
areas of public expenditure with a stronger A danger always exists that higher tariffs will
pro-poor incidence. simply lead to lower revenue collection, but
Figure 3.11 Increased Industrial and Commercial Tariffs, Niger and Malawi
350
80
300
250 60
200
150 40
100 20
50
0 0
2000 2001 2002 2003 2004 2005 2006 1999 2000 2001 2002 2003 2004 2005 2006
government commerce September
outcome would remain regressive. Providing a Table 3.4 Potential Targeting Performance of Connection Subsidies under Various
connection subsidy equally likely to reach all Scenarios
percentage of total poor getting connection subsidies relative to percentage of
unconnected households would ensure that poor in the population
the percentage going to the poor exceeds their
New connections Only households beyond All unconnected
share of the population by 112–118 percent— mirror pattern of reach of existing network households receive
finally, a progressive result. Improving the Utility existing connections receive connection subsidies subsidy
distributional incidence beyond this modest Electricity 37 95 118
level would require connection subsidies to be Water 35 74 112
accompanied by other socioeconomic screens. Sources: Banerjee, Wodon, and others 2008; Wodon 2008a, 2008b.
These findings illustrate that in the low-access
environment in most African countries, the
absence of a connection remains a relatively
weak targeting variable. support the classification of beneficiaries, as
Can anything be done to improve targeting well as a significant amount of administrative
of use-of-service subsidies? The poor perfor- capacity. Both factors are often absent in Africa,
mance of existing utility subsidies is explained particularly in the low-income countries.
by pro-rich coverage and by the widespread An important test of the coherence of a
use of poorly designed IBTs. Common design subsidy policy is whether the country could
failures in water IBTs include high fixed afford the policy if it were scaled up to univer-
charges and minimum consumption levels sal access. The underpricing of utility services
that penalize small consumers, as well as the that benefit just a small minority of the popu-
large size and universal applicability of the first lation costs many African countries as much as
subsidized block (Banerjee, Foster, and oth- 1 percent of GDP. As countries move toward
ers 2008). Common design failures in power universal access, that subsidy burden would
IBTs include large subsistence thresholds increase proportionately, rapidly becoming
that allow only consumers with exceptionally unaffordable for the national budget. Thus,
high consumption to contribute fully to cost countries should consider how the cost of any
recovery (Briceño-Garmendia and Shkaratan proposed subsidy policy would escalate as cov-
2008). Achieving major improvements in the erage improves. This test of a subsidy’s fiscal
targeting of use-of-service subsidies by over- affordability is an important consideration to
hauling the design of increasing block tariffs help countries avoid embarking on policies
(IBTs) is difficult. Some improvements in tar- that are simply not scalable.
geting could be achieved by eliminating fixed Another potentially effective method of
charges, reducing the size of first blocks to targeting is to limit the allocation of subsidies
cover only genuinely subsistence consump- to lower-cost and lower-quality alternatives
tion, and changing from an IBT to a volume- that encourage self-selection. For services such
differentiated tariff in which those consuming as water, for which different modes of service
beyond a certain level forfeit the subsidized provision exist, subsidies could possibly be
first-block tariff completely. Even with these concentrated on second-best alternatives such
modifications, however, the targeting of such as standposts while requiring full cost recov-
tariffs would improve only marginally and not ery from private piped-water connections. The
become strongly pro-poor in absolute terms. theory is that more affluent customers will
Global experience suggests that the target- eschew second-best services and automatically
ing of utility subsidies can be improved and opt to pay the full cost of the best alternative,
become reasonably progressive, if some form of thus identifying themselves and leaving the
geographical or socioeconomic targeting vari- subsidized service to less affluent customers. In
ables can be used beyond the level of consump- Africa, however, the use of self-selection may be
tion (Komives and others 2005). However, such less effective, because coverage of second-best
targeting schemes hinge on the existence of alternatives such as standposts and improved
household registers or property cadastres that latrines is just as regressive as coverage of best
102 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
alternatives such as piped water and flush alternative to standposts, although it is often
toilets. not legally recognized. In addition, yard taps
that serve four or five households—not several
Giving More Consideration to hundred—can reduce costs while avoiding some
Second-Best Solutions of the most serious public good problems.
Second-best solutions appear to provide a
happy compromise but face many implemen- Notes
tation challenges. As noted earlier, second-best The authors of this chapter are Sudeshna Ghosh
approaches provide modern services that are far Banerjee, Quentin Wodon, and Vivien Foster,
preferable in welfare terms to their traditional who drew on background material and contribu-
alternatives and are still much less costly than tions from Tarik Chfadi, Amadou Diallo, Sarah
Keener, Taras Pushak, Maria Shkaratan, Clarence
best modern infrastructure services. So why are
Tsimpo, Helal Uddin, and Yvonne Ying.
these second-best services fairly rare in Africa—
1. The cross-regional figures for infrastructure
and skewed toward more affluent households? coverage are unweighted simple averages.
A key problem of many second-best solu-
tions is that their public good nature compli-
References
cates their adoption. Both public standposts and Banerjee, Sudeshna, Vivien Foster, Yvonne Ying,
street lighting are essentially public goods. This Heather Skilling, and Quentin Wodon. 2008.
characterization makes it difficult for utilities to “Achieving Cost Recovery, Equity and Efficiency
recover the costs of these services and exposes in Water Tariffs: Evidence from African Utilities.”
Working Paper 7, Africa Infrastructure Country
them to theft by extending the network’s reach Diagnostic, World Bank, Washington, DC.
into lower-income areas. Thus, utilities have
Banerjee, Sudeshna, Quentin Wodon, Amadou
no real incentive to provide such loss-prone Diallo, Taras Pushak, Hellal Uddin, Clarence
services. In addition, facilities are vulnerable to Tsimpo, and Vivien Foster. 2008. “Access, Afford-
maintenance issues, because nobody is respon- ability, and Alternatives: Modern Infrastructure
sible for preventing, reporting, or addressing Services in Africa.” Background Paper 2, Africa
problems. One solution is to introduce an agent Infrastructure Sector Diagnostic, World Bank,
Washington, DC.
responsible for managing the facility, charging
for service, and soliciting maintenance activi- Bardasi, Elena, and Quentin Wodon. 2009. “Work-
ing Long Hours and Having No Choice: Time
ties. However, covering the agent’s salary adds Poverty in Guinea.” Policy Research Working
significantly to the cost of the second-best alter- Paper 4961, World Bank, Washington, DC.
native, and agents often exploit their control- Boccanfuso, Dorothée, Antonio Estache, and Luc
ling position to charge excessive rates. Savard. 2008a. “Electricity Reforms in Mali: A
The African experience with standposts pro- Micro-Macro Analysis of the Effects on Pov-
vides pointers for improving the performance erty and Distribution.” Working Paper 4, Africa
of such public facilities (Keener, Luengo, and Infrastructure Country Diagnostic, World Bank,
Washington, DC.
Banerjee 2008). Where standposts are admin-
istered by local agents, the management model ———. 2008b. “Electricity Reforms in Senegal:
A Micro-Macro Analysis of the Effects on Pov-
should be grounded in the prevailing culture erty and Distribution.” Working Paper 5, Africa
of the beneficiary community. Checks and bal- Infrastructure Country Diagnostic, World Bank,
ances are needed to ensure that the delegated Washington, DC.
manager behaves responsibly. The utility also ———. 2008c. “Water Reforms in Senegal: A
must be closely involved in monitoring the sta- Micro-Macro Analysis of the Effects on Poverty
tus of the standposts, regularly collecting the and Distribution.” Working Paper 16, Africa
water revenues, and providing technical assis- Infrastructure Country Diagnostic, World Bank,
Washington, DC.
tance to the standpost operators. Therefore,
defining a useful set of incentives that bolster Briceño-Garmendia, Cecilia, and Maria Shkaratan.
2008. “Achieving Cost Recovery, Equity, and
the utility’s growing interest in participating Efficiency in Power Tariffs: Evidence from
in the standpost business is essential. In some African Utilities.” Working Paper 21, Africa
environments, resale of water by households Infrastructure Country Diagnostic, World Bank,
with private connections can be a practical Washington, DC.
Dealing with Poverty and Inequality 103
Diallo, Amadou, and Quentin Wodon. 2005. “A Africa Infrastructure Country Diagnostic, World
Note on Access to Network-Based Infrastructure Bank, Washington, DC.
Services in Africa: Benefit and Marginal Inci- Komives, Kristin, Vivien Foster, Jonathan Halpern,
dence Analysis.” World Bank, Washington, DC. and Quentin Wodon. 2005. Water, Electricity,
Eberhard, Anton, Vivien Foster, Cecilia Briceño- and the Poor: Who Benefits from Utility Subsidies?
Garmendia, Fatimata Ouedraogo, Daniel Washington, DC: World Bank.
Camos, and Maria Shkaratan. 2008. “Under- Kumar, Ajay, and Fanny Barrett. 2008. “Stuck in
powered: The State of the Power Sector in Sub- Traffic: Urban Transport in Africa.” Background
Saharan Africa.” Background Paper 6, Africa Paper 1, Africa Infrastructure Country Diagnos-
Infrastructure Sector Diagnostic, World Bank, tic, World Bank, Washington, DC.
Washington, DC.
McGranahan, Gordon, Cyrus Njiru, Mike Albu,
Estache, Antonio, and Quentin Wodon. 2007. Mike Smith, and Diana Mitlin. 2006. How Small
Infrastructure and Poverty in Sub-Saharan Africa. Water Enterprises Can Contribute to MDGs: Evi-
Directions in Development Series. Washington, dence from Dar es Salaam, Nairobi, Khartoum,
DC: World Bank. and Accra. Leicestershire, U.K.: Water, Engineer-
Foster, Vivien, and Jean-Philippe Tré. 2003. “Mea- ing and Development Centre, Loughborough
suring the Impact of Energy Interventions on University.
the Poor—An Illustration from Guatemala.” In Tsimpo, Clarence, and Quentin Wodon. 2009.
Infrastructure for Poor People: Public Policy for “Who Benefits from Electricity Consumption
Private Provision, ed. Penelope Brook and Tim versus Connection Subsidies? Evidence from
Irwin, 125–78. Washington, DC: World Bank. Niger.” Development Dialogue on Values and
Gulyani, Sumila, Debabrata Talukdar, and Darby Ethics, World Bank, Washington, DC.
Jack. 2008. “A Tale of Three Cities: Understand- Wodon, Quentin, ed. 2008a. “Electricity Tariffs
ing Differences in Provision of Modern Services.” and the Poor: Case Studies from Sub-Saharan
Working Paper 10, Africa Infrastructure Country Africa.” Working Paper 11, Africa Infrastructure
Diagnostic, World Bank, Washington, DC. Country Diagnostic, World Bank, Washington,
Kariuki, Mukami, and Jordan Schwartz. 2005. DC.
“Small-Scale Private Service Providers of Water ———. 2008b. “Water Tariffs and the Poor:
and Electricity: A Review of Incidence, Struc- Case Studies from Sub-Saharan Africa.”
ture, Pricing, and Operating Characteristics.” Working Paper 12, Africa Infrastructure
Policy Research Working Paper 3727, World Country Diagnostic, World Bank,
Bank, Washington, DC. Washington, DC.
Keener, Sarah, Manuel Luengo, and Sudeshna G. Wodon, Quentin, and Yvonne Ying. 2009. “The
Banerjee. 2008. “Provision of Water to the Poor Determinants of Domestic Work Time in Sierra
in Africa: Informal Water Markets and Experi- Leone.” Development Dialogue on Values and
ence with Water Standposts.” Working Paper 13, Ethics, World Bank, Washington, DC.
Chapter 4
I
nstitutional competence and capacity are providers; outcomes below expectations; and
important determinants of the performance a high degree of official and public skepticism
of infrastructure providers in every sector. about whether the application of the standard
That seems obvious, but systematic analysis has package is producing (or even could produce)
been lacking on the nature and extent of the the desired results. A large part of the explana-
links between stronger institutions and better tion for this situation is thought to lie in the
outcomes: specifically, broader access, higher relative weakness of African practices, policies,
service quality, and more financially efficient and agencies (that is, institutions) that guide
service. This chapter looks at the different and oversee African infrastructure sectors and
institutional models applied, the approaches firms, public or private.
to strengthen infrastructure-relevant institu- The statistical analysis for this chapter
tions, and the effect of the various approaches suggests that institutions make a difference.
on performance. It reveals strong links between institutional
The standard infrastructure reform and pol- reforms and enhanced governance in the coun-
icy prescription package of the 1990s—market try, sector, and enterprise—and improvements
restructuring, private involvement up to and in the quantity and quality of infrastructure
including privatization, establishing indepen- services (with sectoral variation). Given the
dent regulators, and enhancing competition— link between institutional development and
yielded a fair number of positive results in performance improvements, and the high costs
Africa. This conclusion deserves stress: ben- of inaction, strengthening sectoral institutions
eficial outcomes following the application of and country and sectoral governance is a very
these reforms have often been unacknowledged worthwhile investment.
or at least underappreciated. Nevertheless, this Most African countries have undertaken
set of reforms has proved more difficult to preliminary institutional reforms, mainly the
apply in Africa than in other regions. One finds broader sectoral policy and legal measures,
in Africa numerous failures to implement, or many of which can be accomplished by the
fully implement, the policy package; renegotia- stroke of a pen. What has lagged are regulatory
tions or cancellations of contracts with private and governance reforms; they have taken much
105
106 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
more time to bear fruit. For instance, effective almost all African countries have embarked
regulation requires building organizations that on institutional reforms, on average they have
challenge established vested interests. Gover- adopted no more than 50 percent of good insti-
nance improvements, particularly in state- tutional practices. The variation in performance
owned enterprises (SOEs), require aligning across countries is roughly two to one, with the
internal and external incentives, which again most advanced countries (Kenya) scoring about
require broader reforms of the external envi- 70 percent and those furthest behind scoring
ronment for infrastructure service providers. 30 percent (Benin).
At the country level, progress in one infra-
structure sector is no guarantee of progress in
Institutional Reforms: A Glass another. That is, institutional development in
Half Full infrastructure sectors is uneven both among and
within countries. Countries that perform fairly
Africa’s institutional framework for infrastruc- well in one aspect of infrastructure do not neces-
ture is no more than halfway along the path to sarily do so in another. This finding suggests that
best practice. The components of the institu- sector-specific constraints may be as important
tional performance indicators developed for this as country-specific constraints. It also points to
study capture a wide range of characteristics of the potential for greater cross-fertilization of
the institutional environment (box 4.1). A coun- experiences across sectors within a country.
try’s aggregate score on this index suggests the In addition, the quality of the institutional
extent of institutional reforms. Overall, although framework differs across country groupings.
BOX 4.1
(continued)
Building Sound Institutions 107
BOX 4.1
(continued)
Reform Internal Governance
Legislation Ownership and Shareholder Quality
Existence of de jure reform Concentration of ownership
Implementation of reform Corporatization/limited liability
Restructuring Rate of return and dividend policy
Unbundling/separation of business lines Managerial and Board Autonomy
State-owned enterprise corporatization Autonomy in hiring/firing/wages/production/sales
Existence of regulatory body Size of board
Selection of board members
Policy Oversight
Presence of independent directors
Oversight of regulation monitoring outside the ministry
Dispute arbitration outside the ministry Accounting, Disclosure, and Performance Monitoring
Tariff approval outside the ministry Publication of annual reports
Investment plan outside the ministry International financial reporting standards/external audits/
Technical standard outside the ministry independent audit
Audit publication
Private Sector Involvement
Remuneration of noncommercial activity
Private de jure/de facto
Performance contracts/with incentives
Private sector management/investment/ownership
Penalties for poor performance
Absence of distressed/renegotiation/renationalization
Monitoring/third-party monitoring
Regulation External Governance
Autonomy Labor Market Discipline
Formal autonomy on hiring/firing Restriction on dismissing employees
Financial autonomy (partial/full) Wages, compared to private sector
Managerial autonomy (partial/full) Benefits, compared to private sector
Multisectoral agency/commissioners
Capital Market Discipline
Transparency No exemption from taxation
Publication of decisions via report/Internet/public hearing Access to debt, compared to private sector
Accountability No state guarantees
Existence of appeal Public listing
Independence of appeal (partial/full) Outsourcing
Tools Billing and collection
Existence of tariff methodology/tariff indexation Meter reading
Existence of regulatory review; length of regulatory review Human resources information technology
Jointly, the three sets of indicators (reform, regulatory, Separately, each indicator serves as a basis for measuring the
and governance) added together summarize the overall (aggregate and disaggregate) effect of progress in reforms
level and type of institutional reforms in any given country. and enterprise performance.
Reflecting countries’ broader characteristics, cantly further ahead with water reform, perhaps
the extent of institutional reforms differs reflecting the strong role of donors in this sector.
across these groups (figure 4.1). For example, For telecommunications reform, the resource-
middle-income countries are significantly fur- rich low-income countries have higher scores.
ther ahead with power sector reform, whereas A correlation exists between the quality
aid-dependent low-income countries are signifi- of infrastructure institutions and the overall
108 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 4.1 Institutional Progress across Countries, the overall level of governance and control of
by Income Group, Aid Dependence, and corruption,1 as well as the quality of public
Resource Richness
percentage score on institutional scorecard
administration. A polity executive constraint
indicator also measures the extent of checks
a. Income group and balances within a government (Center for
100 Systemic Peace 2006).
90
80 Nevertheless, some countries do well on
70 infrastructure despite broader governance
60
50
limitations, and vice versa. For Kenya and
40 Niger, lower scores on the country governance
30 indicators have not impeded the achieve-
20
10
ment of good scores across all utilities’ insti-
0 tutional reforms. For Zambia, low scores in
electricity water telecommunications
budget execution and financial management
low income middle income
have not prevented the country from earn-
ing reasonably good scores in infrastructure
b. Aid dependence
100 institutional reforms. By contrast, Benin and
90 Lesotho display high country governance and
80
70 polity executive constraint scores and have
60 decent budgetary and financial management
50
40 standards, but neither has a high overall insti-
30 tutional score in the utility sectors.
20
10 Institutional development in the utilities
0 sector is well ahead of that in the transport
electricity water telecommunications
sector (figure 4.2). Unsurprisingly, institutional
aid dependent not aid dependent
development is furthest ahead in telecommu-
nications, where technological change and
c. Resource richness
100 competition have driven momentous change,
90 bringing the overall average reform score to
80
70 just under 50 percent. Electricity and water are
60 not that far behind, with institutional reform
50 scores just over 40 percent. Although institu-
40
30 tional actions in power and water lag those in
20 telecommunications for implementing reform
10
0
agendas, they score somewhat higher in the
electricity water telecommunications quality of regulation. Also, the governance
not resource rich resource rich framework for the main service providers is
significantly better than that for fixed-line tele-
Source: Vagliasindi and Nellis 2009. communication incumbents. By contrast, insti-
Note: See Vagliasindi 2008c for the definition of the institutional
indicators. tutional scores for ports and railways are only
about half those for the utilities. These sectors
have made significant progress with reform but
quality of institutions in the country, though lag in developing the regulatory framework.
this correlation is much stronger for electricity Across countries and sectors, the greatest
than for water (table 4.1). A key question is the progress has been in sector reform. Average scores
extent to which a country can make progress in exceed 60 percent for reform legislation and 50
reforming infrastructure institutions if its wider percent for sector restructuring, policy oversight,
governance framework is deficient (Levy 2007). and private sector participation (figure 4.3,
Numerous indicators have been developed in panel a). Telecommunications, the most
recent years (for example, by Kaufman, Kraay, advanced, scores about 80 percent of the best-
and Mastruzzi 2008), attempting to capture practice index across all areas of sector reform.
Building Sound Institutions 109
Table 4.1 Correlation between Institutional Scores for Infrastructure and Measures of Broader Country
Governance
Infrastructure Polity executive Budget and financial Public Overall Control of
sector constraint management administration governance corruption
Electricity 0.34 0.29 0.53 0.49 0.46
Water 0.08 0.33 0.3 0.18 0.08
Sources: Vagliasindi and Nellis 2009; Center for Systemic Peace 2006 for polity executive constraint scores; IDA 2008 for Country Policy
and Institutional Assessment score; Kaufmann, Kraay, and Mastruzzi 2008 for governance and control of corruption.
The equivalent score for electricity is about from government interference while remaining
60 percent, and for water about 50 percent. accountable to society. These aspects of regula-
Transport scores about 50 percent on private tion have proved more challenging, with scores
sector participation, but this development has remaining relatively low.
not been accompanied by the broader legal and Governance lags behind other areas of
structural reforms seen in the utilities sectors. institutional development, and the limited
Interference from government continues to progress shows up mainly in internal mana-
undermine regulatory independence in many gerial practices. Whereas the relevance of
countries. Infrastructure regulation in Africa sectoral and regulatory reforms has generally
is still in its early days. Typically, new laws and been well recognized, the governance regime
regulatory bodies have been introduced for has received less attention from policy makers
telecommunications and electricity, whereas and analysts. Almost all Sub-Saharan countries
few countries have created water or transport ranked significantly and consistently lower on
regulators. The quality of regulation can be this dimension of institutional development
measured along several dimensions (figure 4.3, than on the others (figure 4.3, panel c). Most
panel b). On the technical side, regulation needs countries are doing better on internal gover-
to be founded on solid methodological tools, nance than on external governance. Internal
and the resulting decisions need to be commu- governance relates to structures within the
nicated to the public in a transparent manner. service provision entity, such as the extent
African regulators score the highest on these to which its structure approximates standard
dimensions, even if (in absolute terms) they still corporate forms; the qualifications and auton-
have some way to go. On the political side, reg- omy of its senior management and board of
ulation requires a certain degree of autonomy directors; the nature, quality, and timeliness
110 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 4.3 Institutional Progress on Reforms, Regulation, and Governance of the information it submits to its overseers;
percentage score on institutional scorecard and the adoption of accounting and disclosure
a. Sectoral reforms
standards. External governance, by contrast,
100 refers to external market disciplines: being
90 subject to private rather than public sector
80 accounting and auditing systems, contract-
70 ing out noncore activities to private provid-
60 ers, and being obliged to raise debt or equity
50 funds on private capital markets, domestic or
40 international.
30 Only Kenya and South Africa have raised
20 much from external capital markets. Kenya
10 corporatized its power distribution utility
0 and more recently its generation firm, and
n
ht
e
to
ag
tio
rin
ec
rsi
er
a
ctu
s
isl
av
ve
te
tru
leg
iva
s
re
lic
pr
po
ity
ols
e
m
nc
ag
bil
to
no
re
er
ta
pa
to
av
un
au
ns
co
tra
ac
ity
et
et
ge
om
cin
sh
k
il
ra
ar
ar
ab
r
e
ne
av
nt
tso
t
ow
au
or
al
ou
ou
pit
lab
c
ac
ar
ca
bo
through management contracts, leases, and five percent of contracts in water have been
concessions. Only in telecommunications has canceled, as have 15 percent in electricity
divestiture been widely applied. The amount (table 4.2). Note that these cancellations do
of private participation, of all sorts, varies by not include contracts that have undergone
sector, with the most in telecommunications renegotiation because of the complaints of one
and the least in water (figure 4.5). or both of the parties involved; nor do they
As noted previously, private participation account for cases where anticipated renewals
in Africa has had some problems. Twenty- of leases or, especially, management contracts
have not occurred, leading to a resumption of
state management. In all infrastructure sectors,
Figure 4.4 Implementation of Private Participation contract negotiation, monitoring, and enforce-
across Sectors ment have proved more time-consuming and
100 difficult than expected.
90
Despite these difficulties, the survey
80
undertaken for this chapter reveals significant
gains from private participation in some sec-
percentage of countries
70
tors and for certain aspects of performance.
60
A higher degree of private sector involvement
50
is associated with higher labor productiv-
40
ity (connections per employee), though the
30
link is statistically significant only in the case
20 of electricity and ports, and higher but not
10 statistically significant cost-recovery ratios.
0 In telecommunications, the countries with
tio m-
ity
er
rts
ay
po
ica co
ns
ilw
w
ct
un tele
ra
ele
allowed by law actually implemented segments of the market. More extensive private
involvement in ports is associated with above-
Source: Vagliasindi and Nellis 2009. average technical efficiency.
70
Concession contract 4 50
60
BTO/BOO 1 0
50
Divestiture 1 0
40
Total 28 25
30
Electricity
20
Management or lease contract 17 24
10
Concession contract 16 31
0
Independent power project 34 6
tio m-
ty
rts
te
ay
ici
po
ica co
wa
ns
ilw
ctr
un tele
ra
ele
Divestiture 7 —
m
Total 74 15
management investment
Source: Vagliasindi and Nellis 2009.
Note: BOO = build-operate-own; BTO = build-transfer-operate.
Source: Vagliasindi and Nellis 2009. — Not applicable.
112 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Earlier empirical studies also find some Privatization of fixed-line incumbents has
evidence of positive outcomes, albeit based affected access and productivity, and qual-
on case study evidence and by no means in all ity of services somewhat, though the change
instances. For example, a recent study of seven is not statistically significant (figure 4.6). The
infrastructure privatizations in Africa assessed growth in the number of subscribers has been
three factors: (a) efficiency gains and losses, low and in almost all countries negative, except
(b) nature and competence of the transaction, Nigeria, the only country where competition
and (c) who won and who lost (and by how has also been introduced in the fixed-line seg-
much) in society because of the transaction ment of the market. Productivity is also low,
(BIDE 2006). Three of the seven (Côte d’Ivoire compared with international benchmarks
electricity, Senegal Airlines, and Senegal water) (lines per employee).
were assessed as “unqualified success stories” Still, several private participation transac-
in efficiency terms, according to a variety of tions in the fixed-line segment of these mar-
financial and service quality measures. Three kets (the remaining natural monopoly) have
others (Mozambique water and Uganda water run into problems. In the last few years, stra-
and telecommunications) were assessed as tegic investors from developed countries have
producing “some positive changes but less largely withdrawn from African telecommuni-
than what most expect from privatization.” cation privatizations. Only three such sales have
Only Senegal electricity was classified as having occurred since 2001, and in those sales, no tra-
“no significant effect” (BIDE 2006: 2). None of ditional strategic partner obtained a controlling
the seven was assessed as negatively affecting stake. Recent telecommunication divestitures
efficiency. have either been public offerings (South Africa
Such studies found a close correlation and Sudan), sales to developing-country inves-
between the competence in negotiating the tors (ZTE of China in Niger and Maroc Telecom
transaction and the efficiency gains by the in Burkina Faso), or sales to domestic investors
new private operator. The distributional effect (Malawi and Nigeria). In several instances, gov-
could be assessed fully in two cases (Côte ernments have repurchased shares in incum-
d’Ivoire electricity and Senegal water) and bent operators. That happened in Ghana and
partially in the other five. Even with limited is planned in Rwanda. In Tanzania in 2005,
information, the study found a correlation the government repurchased shares previously
between institutional capacity and wider dis- privatized in its fixed-line operator, TTCL. In
tribution of benefits. That is, the cases having May 2007, the government placed TTCL under
the better institutional arrangements in con- a three-year management contract with a Cana-
ducting the transaction show better outcomes dian firm, SaskTel. As of early 2009, the govern-
for a broader range of stakeholders than do ment was considering canceling the contract,
the cases where the transaction process was claiming that SaskTel had failed in its commit-
rated lower. ment to raise nonguaranteed debt financing to
Using the analysis of the sectoral chapters rehabilitate and expand the network.
in this study, one can further assess the extent Access has spread quickly since 1998 because
and effect of private participation in African of the rapid rise of mobile telephony, largely
infrastructure, from the most extensive and from the combination of private participation
successful involvement to the least. and increasingly intense competition. Private
investment, the bulk of it greenfield, in cellular
Telecommunications phone technology has allowed new providers to
Private participation in telecommunications enter previously monopolistic markets, result-
has taken place in the majority of Sub-Saharan ing in greater access and declining, though still
countries. In 15 countries, at least partial priva- comparatively high, consumer prices. Analysts
tization of the state-owned fixed-line telecom- generally argue that the dramatic increases
munication incumbent has occurred. Licensing in African access to and coverage of telecom-
new private mobile operators for greenfield munication services owe more to the entry of
networks has been even more widespread. new mobile operators, thereby strengthening
Building Sound Institutions 113
competition, than to the improved information Figure 4.6 Links between Private Sector Participation and Performance Indicators
and incentives of private managers and owners. in Telecommunications
Most new mobile operators are controlled by
one of five multinational firms operating in total telephone subscribers
(per 100 people)
the region: France Telecom, MTC (Kuwait),
MTN (South Africa), Millicom (Luxembourg),
mobile subscribers (thousand
or Vodacom (South Africa). of subscribers per employee)
The salutary effects of competition are appar-
ent. A strong link exists between liberalization in main telephone line
this sector (and others) and better outcomes in (100 lines per employee)
access and productivity (figure 4.7). Countries
with lower market concentration in the mobile number of faults
(per main line per year)
segment of the market, measured by the Her-
findahl-Hirschman Index,2 have much higher 0 20 40 60 80 100
penetration rates and productivity in the same
segment of the market, as well as in the fixed- privatized state-owned enterprises
Ports
By 2006, 20 port concessions were operating in Figure 4.7 Links between Market Concentration and Performance Indicators in
Africa, with 6 more in process. Evaluations of Telecommunications
these concessions indicate that delays, costs, and
thefts were reduced and that port infrastructure total telephone subscribers
started to improve. Cargo-handling rates and (per 100 people)
the use of better handling systems in African
“concessioned” ports are significantly higher mobile subscribers (thousand
than in state-managed ports (figure 4.8). of subscribers per employee)
Figure 4.8 Links between Port Concessions and Performance Indicators Concessions have not resolved the key
issue of mobilizing finance. The drawback
annual (weighted average) of concessions is that they rarely produce the
growth rate of container
trade 1995–2000
anticipated (and in many cases, contracted)
investments for network rehabilitation and
expansion. The reason is that service revenues
average moves per
hour (unit: 100 moves)*
are too low to support investment finance,
partly due to low traffic volumes and intense
intermodal competition and partly due to the
container gantries as
failure by governments to compensate conces-
handling system sionaires for running loss-making obligatory
(percent)** passenger services. Thus, chronic underinvest-
0 10 20 30 40 50 60 70 80 90 ment and dilapidated infrastructure remain as
major railway problems. The limited invest-
concessioned others
ment capital forthcoming has been financed
by international financial institutions and
Source: Vagliasindi and Nellis 2009.
Note: *Performance differential is statistically significant at the 10 percent level; **significant at passed through government to the private rail-
the 5 percent level.
way operators.
identify and ensure that the terms of the agree- Figure 4.9 Links between Rail Concessions and Performance Indicators
ments are “fair and balanced” to all parties who
will be directly and indirectly affected by these
locomotive availability
transactions (Besant-Jones, Tenenbaum, and (% of time)***
Tallapragada 2008). The Nigerian energy regu-
lator has tested the methodology. cars (million ton-
Management contracts are the second kilometer per car)
for improvements in quality of supply and ser- 0 20 40 60 80 100 120 140 160 180 200
vice, including expanded access, so that con- management contract other
sumers experience tangible benefits. Effective
contractual oversight is needed to track per- Source: Vagliasindi and Nellis 2009.
Note: ***Performance differential is statistically significant at the 1 percent level.
formance, to fairly assess and award incentive
payments or penalties, and to reduce informa-
tion asymmetry. Finally, postcontract manage-
ment succession issues need to be addressed example, dating from the 1990s). The two in
early (Ghanadan and Eberhard 2007). Gabon and Senegal have been judged as suc-
cesses in producing service improvements,
Water network expansions, and financial stability
In water, leases have been applied rather widely, (BIDE 2006). The long-term leases in Côte
with management contracts as the second most d’Ivoire and Niger have been producing posi-
common form of private participation. Con- tive financial and operational results, despite
cessions have also been used in several African difficult external conditions (Marin 2008).
countries (in Chad, Gabon, and Senegal, for Only one divestiture has occurred, entailing
116 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
the sale of 51 percent of equity, which was in which are medium term in nature, broader
the water company in Cape Verde in 1999. improvements in operational efficiency had
Private participation in water has generated been observed. Access had also improved,
much hostility and opposition. Because water although funding continued to come from the
is a commodity essential for life, many feel that public sector.
its distribution should be free or at a very low
price. Despite considerable evidence of persis- What Have We Learned?
tently poor performance by publicly owned and Any evaluation of the African experience with
operated African water companies, many people private sector participation must be nuanced
believe that the service cannot and should not by the wide variation in sector experiences
be turned over to private delivery. To support (table 4.3), and both successes and failures
their position, they point to several notable pri- considered in tandem. Public and official
vate participation problem cases in water, such attention have focused more on the failures
as Tanzania (see box 4.3) and Uganda. and contentious cases, particularly in water,
A recent detailed empirical review of the but also in power and transport. However, the
main experiences with private sector partici- lesson from African private participation is
pation in the African water sector concluded not that the approach should be discarded, but
that there had been beneficial effects on enter- that it should be applied selectively and care-
prise performance in a number of cases but fully to those areas of infrastructure where it
that the extent of these impacts depended on has a proven potential to contribute.
the contractual forms that were used (Marin Expectations should also be kept realistic.
2008). In the case of management contracts, Experience has shown that there are only a few
which are short term in nature, there had been niches where, either through raising investment
benefits in terms of revenue collection and ser- finance or improving operational efficiency, the
vice continuity but little effect on other aspects private sector can contribute significantly to
of performance. In the case of lease contracts, investment finance—namely, ICT (particularly
mobile networks), power generation, port con-
tainer terminals, and a handful of high-traffic
BOX 4.3 road segments (table 4.3). While the overall
volume of private finance for infrastructure
investment is limited, it is nonetheless sub-
Lessons from the DAWASA Lease stantial, having (at least during the mid-2000s)
Contract (Tanzania) exceeded the volume of ODA for these sectors
The failed lease contract of DAWASA, the water authority for Dar es (recall table 2.1). But even in areas of infra-
Salaam, Tanzania, is instructive. The contract was signed in August 2002. structure that have not proved attractive for
It was supposed to run for 10 years, but the government canceled it in private finance—such as most roads, railways,
May 2005, after only 21 months of operation. The government claimed power and water distribution systems—private
the private provider had failed to meet water production and collection management can still make a significant contri-
targets, pay the lease and other fees, meet service quality and quantity bution to improving operational performance,
commitments, and pay the penalties assessed for noncompliance. For
and thereby help to recover the very substantial
its part, the private provider claimed that its bid and business plan were
funds that are currently being lost to various
based on inaccurate, out-of-date, or partial information provided by the
government. Arbitration in April–May 2005 failed; the government ter- kinds of inefficiency (recall table 2.8).
minated the contract, and the service returned to public management. Of course, one cannot deny the problems
Critics note that the private provider’s lawsuit against the government in African private participation, or that many
for breach of contract was rejected (in 2008) by a British court. They of them can be attributed, in all sectors, to
argue that this case shows how difficult it is to ensure that private provi- institutional deficiencies. Poor sectoral plan-
sion of water will fulfill either the anticipated financial objectives or the ning, vague or absent sector policies, and
distributional objectives. long-standing weak financial and operational
Sources: BIDE 2006; Marin 2008.
performance in the utilities helped create the
crises of demand and insecurity that led to
rushed decisions. In state-owned infrastructure
Building Sound Institutions 117
firms, poor management, inadequate or non- 2008–09 will further deplete investor appetites
existent record keeping (at both the firm and for comparatively high-risk ventures in emerg-
higher levels), and lax monitoring created ing markets, heighten the reluctance of African
organizational and informational chaos that officials to embark on innovative schemes, and
reduced the interest of potential investors add weight to the notion of the primacy of the
and severely complicated the due diligence public sector. Therefore, the existing level of
processes of those who bid. Nontransparent state ownership is likely to persist, and may
contract negotiation proceedings, substan- indeed increase, in the near to medium term.
dard procurement practices, and inadequate This likelihood requires renewed attention to
contract monitoring and enforcement mecha- a long-standing but recently neglected issue:
nisms have been among the factors contribut- improving the financial and operational per-
ing to higher prices, poorer-than-anticipated formance of state-owned firms (Nellis 2005;
outcomes, renegotiated or canceled contracts, Gómez-Ibáñez 2007).
and governance issues. These institutional
deficiencies must be dealt with if private par-
ticipation is to fulfill its potential. How Can State-Owned Enterprise
Despite the inroads of private participation Performance Be Improved?
over the past two decades, Africa remains the
region with the highest state ownership of its Africa has the highest percentage of state-
infrastructure utilities. The financial crisis of owned infrastructure utilities of any developing
118 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
region. Given the mixed record; the absence of Only limited success in achieving full cor-
investor appetite; and the antipathy of African poratization, including establishing limited
officials, nongovernmental organizations, and liabilities and introducing rate of return and
many observers of development toward private dividend policies, has been recorded in Africa.
participation—and the resulting increasing The telecommunication sector alone can be
reluctance of the donor community to push reported as a success story, with electricity
for privatization in infrastructure—state own- and water lagging, proving that even coun-
ership is likely to be the norm for some time. tries with high scores compare poorly with
The track record so far in governance other regions.
reforms is not encouraging and varies sub- More limited corporate governance reforms
stantially across sectors and countries. One have been started more evenly across all sec-
must start with the recognition that a few tors and are becoming a dominant feature in
African state-owned infrastructure firms have electricity and water. These changes include
sustained good performance in the absence of the introduction of boards of directors (even if
private participation. Botswana and Uganda the size tends to be either too large or too small
show that fully state-owned African utilities compared with international standards), selec-
can deliver high-quality performance (see tion of board members according to a com-
box 4.4 and box 4.5, respectively). petitive process rather than direct appoint-
The cost of inaction implies high hidden ments by line ministries, and the introduction
costs. The estimated hidden cost of inefficiency of independent directors (figure 4.11).
coming from mispricing, unaccounted losses, Performance contracts with incentives and
and collection inefficiency is on average equal independent external audits have become
to 0.6 percent of GDP in the water sector and dominant features of the governance reform pro-
1.9 percent of GDP in the power sector. The cess for both electricity and water (figure 4.12).
inefficiency of SOEs can also be measured Independent audits have also been good for
by excessive employment. In the telecommu- efficiency in both cases.
nication sector, the hidden cost of excessive Of governance reforms that appear to be the
employment is on average equal to 0.1 percent most important drivers of higher performance,
of GDP. two appear especially promising: performance
contracts with incentives and independent
external audits (table 4.4). Uganda has had
good experience with a performance contract
BOX 4.4
in its water company, providing the utility with
incentives for good performance and produc-
Lessons from Successful SOE Reforms in ing greater accountability (see box 4.5). The
introduction of independent audits has also
Botswana Power Corporation
positively affected efficiency for both electric-
The state-owned and -operated Botswana Power Corporation ity and water utilities.
has long provided reliable, high-quality service. Over the years, What can such cases teach? First, recast and
Botswana Power has expanded its network in both urban and rural reapply the performance contract approach
areas, covered its costs, posed no burden on the government bud-
to SOE reform. Initial attempts to improve
get, minimized system losses (10 percent), and earned a decent
African SOEs using this device were mini-
return on assets. Although the availability of cheap imported power
from South Africa (now severely threatened) is part of the expla-
mally effective, but recent efforts have had a
nation for good performance, analysts give five institutional factors stronger and much more positive effect. The
equal weight in explaining this success: (a) a strong, stable economy, more recent performance contracts applied
(b) cost-reflective tariffs, (c) lack of government interference in mana- with some success in Uganda (and, reportedly,
gerial decisions, (d) good internal governance, and (e) competent, in Kenya) should be studied and modified for
well-motivated staff and management. broader application to African utilities across
sectors.
Source: PPA 2005.
Second, renew efforts to strengthen the
financial and operational monitoring of SOEs.
Building Sound Institutions 119
BOX 4.5
Performance Agreement for the National Water and Sewerage Corporation (Uganda)
Between 1998 and 2004, the National Water and Sewer- also increased the head office’s commitment to provide
age Corporation (NWSC) system operated under two man- financial and material resources to enable different areas to
agement contracts with private providers. By the end of the implement rehabilitation and investment programs.
second contract, neither party had an interest in continuing. In 2002, automatic tariff indexation was introduced.
After 2004, public managers, operating under performance In addition, the Stretch-Out Program increased staff com-
contracts, were responsible for the service. A review of per- mitment by improving internal communication and setting
formance during the entire period concluded that the targets tougher performance targets and corresponding incentives.
set for the private management contracts were fulfilled but A one-minute management system was introduced to further
that the public management team furnished similarly good enhance individual staff members’ accountability for targets.
performance. The main stages in the enterprise reform pro- The government introduced a three-year performance
cess are described below. contract in 2000. The NWSC’s debt service obligations were
From February 1999 onward, the management of the suspended in return for a commitment to operational and
NWSC in Uganda has sequentially implemented a number financial improvements and an increase in coverage.
of reform programs. First, local officials, called area service In 2003, a second performance contract continued the
providers (areas), negotiated with central authorities a set of suspension of debt service and specified that NWSC’s debt
tightly defined performance targets. Second, area manag- would be restructured to a sustainable level. A review com-
ers were given control over running the process. Third, they mittee monitored implementation of the agreement. The
were held strictly accountable for specific results. main incentives of the agreements are bonuses for manag-
A number of measures including the 100-day program ers and staff, if performance targets are achieved.
and the service and revenue enhancement programs resulted
in better specification of targets for the areas. The programs Sources: Baietti, Kingdom, and van Ginneken 2006; Vagliasindi 2008a.
Figure 4.11 Prevalence of Good Governance Practices among State-Owned Enterprises for Infrastructure
90
80 80
70
% of utilities
% of utilities
60 60
50
40 40
30
20 20
10
0 0
telecommunications electricity water telecommunications electricity water
Some of the structures implied in the Organi- function through an independent agency ver-
sation for Economic Co-operation and Devel- sus a decentralized structure) have not yet
opment’s Principles of Corporate Governance been sufficiently “tested” in practice and may
for SOEs (favoring a centralized ownership not suit all developing countries. A centralized
120 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 4.12 Prevalence of Performance Contracts in their economic performance, including the
Electricity and Water detailed structure of subsidies and intersec-
100 toral arrears (Vagliasindi 2008b).
80
Do Independent Regulators Make
% of utilities Sense?
60
Table 4.5 Links between Regulation and Performance Indicators for Telecommunications, Electricity, and Water
Technical losses Connections per employee Access
Sector Regulation No regulation Regulation No regulation Regulation No regulation
Telecommunications 0.2 0.3 0.38*** 0.03 0.2*** 0.1
Electricity 23.3 25.3 155.3 117.3 22.3 11.9
Water 35.2 34.8 6.8a 8.3a 36.1 35.9
Source: Vagliasindi and Nellis 2009.
Note: ***Performance differential is statistically significant at the 1 percent level.
a. The sign of the links between the variables is not as expected.
Table 4.6 Links between Type of Regulator and Performance Indicators for Water
Technical losses (% production) Connections per employee Access (% households)
Regulation by Regulation by Regulation by Regulation by Regulation by Regulation by
Sector contract agency contract agency contract agency
Water 39.4 31.5 0.19 0.05 16.9 32.1
Source: Vagliasindi and Nellis 2009.
Note: None of these performance differentials was found to be statistically significant.
BOX 4.6
Leigland, James, and Gylfi Palsson. 2007. “Port Lessons from Private Participation in Infrastruc-
Reform in Nigeria.” Gridlines, Note 17. ture in Sub-Saharan Africa—Case Study of Bot-
Public-Private Infrastructure Advisory swana Power Corporation.” Report submitted to
Facility, World Bank, Washington, DC. World Bank, Washington, DC.
Levy, Brian. 2007. Governance Reforms, Poverty Svendsen, Mark, Mandy Ewing, and Siwa
Reduction and Economic Management. Msangi. 2008. “Watermarks: Indicators of
Washington, DC: World Bank. Irrigation Sector Performance in Sub-Saharan
Marin, Philippe. 2008. Public-Private Partnerships Africa.” Background Paper 4, Africa Infrastruc-
for Urban Water Utilities: A Review of Experiences ture Sector Diagnostic, World Bank,
in Developing Countries. Washington, DC: World Washington, DC.
Bank. Vagliasindi, Maria. 2008a. “The Effectiveness of
Minges, Michael, Cecilia Briceño-Garmendia, Boards of Directors of State Owned Enterprises in
Mark Williams, Mavis Ampah, Daniel Camos, Developing Countries.” Policy Research Working
and Maria Shkratan. 2008. “Information and Paper 4579, World Bank, Washington, DC.
Communications Technology in Sub-Saharan ———. 2008b. “Governance Arrangement for
Africa: A Sector Review.” Background Paper 10, State-Owned Enterprises.” Policy Research Work-
Africa Infrastructure Sector Diagnostic, World ing Paper 4542, World Bank, Washington, DC.
Bank, Washington, DC. ———. 2008c. “Institutional Infrastructure Indi-
Mundy, Michael, and Andrew Penfold. 2008. cators: An Application to Reforms, Regulation
“Beyond the Bottlenecks: Ports in Sub-Saharan and Governance in Sub-Saharan Africa.” World
Africa.” Background Paper 8, Africa Infrastruc- Bank, Washington, DC.
ture Sector Diagnostic, World Bank,
Vagliasindi, Maria, and John Nellis. 2009. “Evalu-
Washington, DC.
ating Africa’s Experience with Institutional
Nellis, John. 2005. “The Evolution of Enterprise Reform for the Infrastructure Sectors.” Working
Reform in Africa: From State-Owned Enter- Paper 23, Africa Infrastructure Sector Diagnos-
prises to Private Participation Infrastructure— tic, World Bank, Washington, DC.
and Back?” Research Paper 117, Fondazione Eni
World Bank. 2005. “Second Port Modernization
Enrico Mattei, Milan.
and Railway Restructuring Projects, Project
PPA (Power Planning Associates Ltd.). 2005. Performance Assessment Report.” World Bank,
“Towards Growth and Poverty Reduction: Washington, DC.
Chapter 5
Facilitating Urbanization
A
frica is urbanizing fast, a change that is to link national economies with regional and
predictable and beneficial. Economic global markets.
geography indicates that prosperity and African cities are growing fast, but because
density go together because higher productivity of insufficient infrastructure and poor institu-
requires agglomeration economies, larger tions, most new settlements are informal and
markets, and better connectivity. Concentra- not covered by basic services. This situation
tion and urbanization trigger prosperity in has severe consequences for health, incomes,
urban areas as much as in rural areas, and and market integration. A combination of
well-functioning cities facilitate trade and the institutional reform, land policy and planning,
transformation of rural production and non- housing policies, and basic services is required
farm activities. The debate over rural or urban for urban expansion that is more equitable and
development should thus be replaced by the inclusive in nature.
understanding that rural and urban develop- Many necessary investments are beyond the
ment are mutually dependent and that eco- limited fiscal and financial base of African cities.
nomic integration of rural and urban areas is Decentralization has increased the responsibili-
the only way to produce growth and inclusive ties of cities, but not their powers and incen-
development. tives to raise (and retain) revenues. Cities need
Populated places in Africa need infrastruc- access to predictable streams of revenue and
ture to enhance the competitiveness of their the flexibility to raise additional resources, to
businesses and the productivity of their work- safeguard service provision to their constituen-
ers. Energy, roads, water, and information and cies. They also need to improve their technical
communication technologies (ICTs) give Afri- and managerial capacity to deal with priorities
can economies the capacity to develop. Long- in investment and operation and maintenance,
run growth requires an efficient system of to guide the inevitable expansion, to attract
urban centers that includes small, medium, and private partners, and to understand their sur-
larger cities that produce industrial goods and rounding neighborhoods to develop synergies.
high-value services, along with well-functioning Africa’s large agricultural sector and rural
transportation networks (roads, rails, and ports) economy remain central for overall economic
125
126 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
growth and poverty reduction. Better infra- services. Addressing bottlenecks in city per-
structure is crucial for raising agricultural formance is an effective entry point into this
productivity and facilitating access to markets “virtuous” circle (Kessides 2006).
for agricultural products. The Asian experi- Proximity to cities (neighborhood effect)
ence suggests that successful economic growth is critical for enabling the shift from subsis-
demands higher agricultural productivity, tence to commercial agriculture, for increasing
which raises incomes and the demand for rural incomes, and for making living standards
nonagricultural products, lowers food prices, converge. Areas within two hours’ travel time
and frees up labor for (mainly urban) indus- of cities of at least 100,000 people seem to
trial and service employment (World Bank have diversified into nonagricultural activities
2008).1 (Dorosh and others 2008). Rural areas located
The policy challenge is to harness market between two and eight hours’ travel time from
forces that encourage concentration and pro- such cities account for more than 62 percent of
mote convergence in living standards between the agricultural supply and generate a surplus
villages, towns, and cities. Policy makers will sold to urban areas. In areas farther than eight
be more effective if they look at development hours from these cities, agriculture is largely
strategies for broad economic areas that inte- for subsistence, and less than 15 percent of
grate towns and cities with their surrounding the land’s agricultural potential is realized
rural hinterland. This chapter discusses and (table 5.1).2 Similarly, farmers closer to cities
estimates the infrastructure needs of rural and tend to use more and higher-quality fertilizers
urban areas in the context of rapid urbaniza- and pesticides and better equipment, result-
tion and its challenges for infrastructure, insti- ing in clear improvements in productivity. So
tutions, and targeted interventions. the growth of urban markets is a key factor in
raising the income of the rural population in
the hinterland.
Viewing Cities as Engines of
Growth
Strengthening Urban-Rural Links
The debate on growth strategies has often
looked at urban and rural areas as compet- An integrated approach to development rec-
ing for primacy in the national agenda and ognizes and facilitates the links between urban
in investment allocations, but now is the time and rural areas. Urban centers consume rural
to frame the debate differently. Cities exist products and offer inputs for rural produc-
because of the economic and social advan- tion; rural areas serve as markets for goods and
tages of closeness. Urban centers contribute services produced in urban areas. Migration
to national economic growth by increasing produces social and economic links between
individual, business, and industry produc- urban and rural areas. Migrants often remain
tivity through agglomeration economies; by connected to their families, which they support
increasing household welfare through social through remittances. In addition, rural people
mobility and human development; and by often receive health services and education from
promoting positive institutional change. Cit- nearby towns and cities. Institutional and fiscal
ies also drive rural development, serving as links are often present as well. In most cases,
primary markets for rural production and fiscal redistribution takes place—typically
generating income that flows back to rural from cities to rural areas, given the cities’ larger
areas. Links between urban and rural areas tax base.
constitute a virtuous circle, where access Rural-urban links are constrained by inad-
to urban markets and services for nonfarm equate transport networks, poor electricity and
production stimulates agricultural produc- water provision, and limited coverage of ICT.
tivity and rural incomes, which in turn gen- Weak institutions add to the constraints. For
erate demand and labor for more goods and example, in Ethiopia, the city of Dessie enjoys
Facilitating Urbanization 127
a strategic location and is a main distribution In several fragile states, civil war has con-
center for manufacturing products to the sur- tributed to urban expansion as people from the
rounding regions (World Bank 2007). However, affected regions seek refuge in cities. One-third
the lack of a developed agroprocessing industry of Africa’s urban population is concentrated
limits the market opportunities for higher-value in the region’s 36 megacities with more than
agricultural production and the benefits for the 1 million inhabitants. Much of the remain-
surrounding agricultural areas. der is spread across 232 intermediate cities
of between 100,000 and 1 million inhabitants
Urbanization in Africa and in periurban areas. The largest cities are
Africa’s population remains predominantly growing fastest, suggesting that Africa’s urban
rural. About 66 percent of the inhabitants live population will become more concentrated.
in rural areas, with significant variation across As is typical in the early phases of urbaniza-
countries (table 5.2).3 In African middle-income tion, urban household incomes in Africa are
countries, half of the inhabitants live in rural much higher than rural incomes, almost twice as
areas, whereas in the landlocked low-income high. The 2009 World Development Report notes
countries, they account for about 70 percent. that an economy’s transformation is seldom
The vast majority of Africa’s rural population geographically balanced (World Bank 2009).
(or half of the overall population) lives in the Productivity tends to increase where people and
rural hinterland within six hours’ travel time economic activities concentrate to take advan-
of cities having at least 50,000 inhabitants, tage of agglomeration economies. The initial
whereas about 16 percent of Africa’s population growth spurt is typically associated with a diver-
lives in isolated areas more than six hours’ travel gence in living standards between leading regions
time from the same cities.
The continent is urbanizing rapidly, how-
Table 5.1 Link between Agricultural Productivity and Distance to Urban Centers
ever, and will become predominantly urban by
2020. The share of urban population rose from Percentage of Percentage of Per capita
Percentage of total total crop production
15 percent in 1960 to 35 percent in 2006 and Travel time total area population production ($ per capita)
will reach nearly 60 percent by 2020. Urban Less than 1.7 hours 10.0 41.4 23.6 57.00
growth is presently estimated at 3.9 percent a
1.7–7.6 hours 50.0 46.0 62.5 135.80
year. Rural migration accounts for one-quarter
More than 7.6 hours 40.0 12.5 13.9 110.70
of that growth, with the remainder attributable
to urban demographic growth and adminis- Total 100.0 100.0 100.0 n.a.
(mostly urban) and lagging regions (mostly than 20 percent of the continent’s GDP, despite
rural), but as incomes increase, the divergence is accounting for more than 60 percent of the
followed by convergence.4 Essential household population (table 5.4; Kessides 2006). Recent
consumption converges soonest, access to basic work in Tanzania confirms these values: the
public services next, and wages and income later. country’s urban areas are home to 23 percent
Convergence occurs because of the mobility of of the population yet account for 51 percent of
people and resources across regions and declin- the GDP (Maal 2008). One could infer that the
ing economic distances among regions. average productivity in urban areas is at least
African countries stand at the beginning three times that of rural areas (Farvacque-
of this process. The difference between urban Vitkovic, Glasser, and others 2008).
and rural incomes explains the lower urban
poverty rates (35 percent) vis-à-vis rural pov- Infrastructure in Urban and Rural Areas
erty rates (52 percent) (table 5.3). In absolute For population centers to realize their full eco-
terms, the African rural poor are almost three nomic potential, the provision of infrastructure
times as numerous as the urban poor. This pic- and public services must be efficient. Basic ser-
ture holds for every country regardless of its vices for households in both urban and rural
geography. Similar differences are observed in areas can guarantee sustainable urbanization
access to services. and social equity, enhance living conditions,
The incomes in urban areas reflect the and prevent disproportionate flows of under-
higher productivity possible thanks to agglom- served rural people to the city. Investment in
eration economies—the gain in efficiency infrastructure can improve productivity in the
from having many businesses and workers in modern sector and connectivity with and across
proximity. In countries where traditional non- locations. Deficiencies in infrastructure and
mechanized agriculture still dominates the services, which limit the potential for agglom-
rural sector, the difference between urban and eration economies, hinder African economies
rural productivity can be large. Assuming that and may explain the underperformance of
agricultural output originates in rural areas businesses in Africa relative to other continents.
and industry and services originate in urban One-third of African businesses report a worri-
areas, the African rural sector contributes less some lack of electricity, and 15 percent identify
Table 5.3 Economic Differentials between Rural and Urban Populations, by Country Type
Monthly household budget Poverty rate
National Rural Urban National Rural Urban
Country type ($/month) ($/month) ($/month) (percent) (percent) (percent)
Sub-Saharan Africa 144 106 195 48 35 52
Low-income countries, landlocked 86 75 139 49 32 53
Low-income countries, coastal 145 115 209 47 38 51
Middle-income countries 535 256 691 n.a. n.a. n.a.
Source: Authors’ compilation based on household surveys reported in Banerjee and others 2008.
Note: n.a. = not applicable.
transportation as a major constraint. Poor- Every year since 1990, an additional 0.9 per-
quality roads and other transportation infra- cent of the urban population has gained access
structure endanger connectivity between rural to improved water and 1.7 percent to improved
and urban areas, between products and mar- sanitation, whereas the corresponding rural
kets, and between workers and labor markets. figures stand, respectively, at only 0.3 and 0.4.
The difference in the coverage of basic Electricity service has been expanded to an
infrastructure services is huge across urban additional 3 percent of urban residents but to
and rural areas. For the spectrum of household only an additional 0.8 percent of rural residents.
services, urban coverage rates are between 5 Even so, rampant demographic pressure in
and 10 times those in rural areas (figure 5.1). urban areas has caused coverage rates to decline
The absolute differences are largest for power for all urban services (particularly improved
and smallest for ICT services. Electricity and water), whereas coverage of all rural services
improved water supply (such as piped connec- has increased (particularly power and ICT). As
tion or standpost) extend to a majority of the a result, the gap between urban and rural cover-
urban population, but to less than one-fifth of age rates has narrowed slightly but at the cost
the rural. An even smaller share in rural areas of leaving urban dwellers and businesses with-
uses septic tanks or improved latrines; and access out infrastructure for domestic and industrial
to ICT services remains negligible. In almost purposes (figure 5.2). This finding shows that
half of the countries, energy coverage barely urban service providers have struggled to keep
reaches 50 percent of the urban population and pace with accelerating urbanization.
5 percent of the rural. In addition, fewer than Africa’s sparse road density often leaves
40 percent of African urban households enjoy rural areas isolated from urban markets. Only
a private water connection, a septic tank, or an one-third of Africans living in rural areas are
improved latrine, a share that falls to 5 percent within 2 kilometers of an all-season road.
in rural areas. The paved-road density is 134 kilometers per
Growth in urban and rural coverage of 1,000 square kilometers of arable land, and the
network infrastructure tends to be positively unpaved, 490 kilometers. Moreover, the qual-
correlated. Countries with faster expansion of ity of the rural network is perceptibly lower
urban coverage of water and electricity also than that of the main network, with almost
tend to have faster expansion of rural cover- half in poor condition (figure 5.3). The lack
age, suggesting that an urban network eases of adequate urban transport is an obstacle for
expansion toward rural areas. It may also sug- businesses and for labor mobility.
gest that urban customers cross-subsidize rural The spatial footprint of infrastructure net-
water networks and electrification. works is larger than the coverage rates would
suggest. In the rural hinterlands, where the bulk
of Africa’s rural population lives, 40–50 percent
Figure 5.1 Access to Infrastructure by Location of people live within range of an infrastructure
network. Even in isolated rural areas, the share is
100
as high as 15 percent. This information suggests
90
80
that hookup rates to infrastructure networks are
71
lower in rural areas. In some cases, that likely
% of population
70 63
60 reflects much lower rural purchasing power. In
50 42 others, technical limitations might prevent rural
40
inhabitants from connecting to infrastructure
30 22
20
20
14
networks, even when close to one.
12
10 2 3
7 Infrastructure investment (particularly in
0 rural areas) continues to focus on sector-specific
landline cell phone improved improved electricity
sanitation water interventions rather than spatially synchroniz-
rural urban ing and concentrating the provision of different
infrastructure services in larger “bundles.” Avail-
Source: Banerjee and others 2008. able evidence suggests that bundling of services
130 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
80 80
% of population
% of population
60 60
40 40
20 20
0 0
1990–95 1996–2000 2001–05 1990–95 1996–2000 2001–05
c. Electricity d. Landline
100 100
80 80
% of population
% of population
60 60
40 40
20 20
0 0
1990–95 1996–2000 2001–05 1990–95 1996–2000 2001–05
rural urban
Figure 5.3 Quality Differentials between Main and costs because the interaction across services
Rural Road Networks compounds positive effects, such as time savings
100 or increased connectivity. With barriers to pro-
90 ductivity reduced faster than when services are
80 taken individually, poor households have more
70
60
chances to access economic opportunities.
percent
Table 5.5 Households with Access to One or More Modern Infrastructure Services
percentage
Quintile
Number of services National Rural Urban Poorest Second Third Fourth Fifth
Any one modern
infrastructure service 33 15 76 4 17 23 44 78
Any two modern
infrastructure services 17 4 47 0 2 7 19 56
Any three modern
infrastructure services 9 1 28 0 0 3 11 32
Any four modern
infrastructure services 4 0 12 0 0 1 4 16
Source: Banerjee and others 2008.
Note: Household coverage rates are population weighted for the latest available year. Modern infrastructure services include piped water,
flush toilet, power, and landline telephone.
will look somewhat different from those of Figure 5.4 Affordability of a Basic Package of Household Infrastructure
the past: 34 percent of the total will need to 18
00
00
00
00
00
38
13
individual sectors also.
34
87
27
99
61
45
18
,0
,0
,0
,0
,0
9,
4,
3,
2,
2,
1,
1,
30
25
20
15
10
population density (no. people/sq. km)
Infrastructure Financing
Source: Authors’ compilation.
Note: Affordable represents those population densities in which the capital cost per hectare of a basic
The jurisdiction responsible for infrastructure package of infrastructure services represents no more than one year of income for the inhabitants of
that hectare. Not affordable represents those population densities in which the capital cost per hectare
financing and provision differs hugely across of a basic package of infrastructure services exceeds, by several multiples, the annual income for the
inhabitants of that hectare.
sectors and countries. Although ICT and power
are usually national responsibilities, respon-
sibility for the water supply in urban areas is
widely decentralized (figure 5.7). Nevertheless, Figure 5.5 Spatial Split of Historic Infrastructure Investments
in many countries, markedly the francophone
100
countries, operation is entrusted to utilities
that remain national. Where municipal utili- 80
ties do exist, municipal governments own only
percent
60
a few in whole or in part. Responsibility for
transport infrastructure is divided between 40
Figure 5.6 Spatial Split of Future Infrastructure Investment Needs taxing authority, even though the potential
100
revenue base is large. For large cities, most of
the revenues collected are transferred to the
80 national treasury.
The use of local government budgets var-
60 ies widely across countries. In Ghana, revenues
percent
Organisation for Economic Co-operation and land market and the monopoly by traditional
Development member states, such as London landowners lead to shortages of urban land
and New York. Built-in areas are growing faster supply and rising prices. Lack of land titles
than urban populations in 7 of 10 African hurts business development and the estab-
cities, suggesting falling densities (Angel, lishment of new firms. With no access to land
Sheppard, and Civco 2005).8 As a result, the and located in underserved and peripheral
already high costs for infrastructure in urban areas, the poor suffer from poor connectivity
Africa will increase with further sprawl, hin- and low access to labor markets. The resis-
dering the affordability of basic services and tance of landowners and the lack of registries
adding to the environmental and carbon also hinder cities from raising revenues on
footprints. urban land (box 5.1).
In many African countries, land institu- African cities face shortages of housing and
tions are still incipient, reflecting the political shelter. In most countries, the real estate mar-
economy and colonial legacy. In Africa, land- ket and government agencies supply at most
ownership is made difficult by the extreme one-fourth of the annual demand for hous-
centralization of procedures, the costs of ing; the remaining three-fourths turn to the
titling, and the rapid depletion of central gov-
ernments’ land reserves. In this context, spon-
taneous settlements have developed. Limited Table 5.7 Overview of Local Infrastructure Financing Mechanisms
land supply and high prices affect location Mechanism Urban Rural
decisions and exclude low-income house- Direct central provision ICT, power, sometimes water ICT, irrigation, occasionally power
holds from the official land market. Many gov- and water
BOX 5.1
housing, but the effect of these incentives has
been minimal (Farvacque-Vitkovic, Glasser, and
others 2008). Construction costs are very high,
Land Issues in Tanzania especially for landlocked countries.9 Cement,
iron, and other materials are imported, making
Tanzania is at the early stage of its urbanization, and the population
is growing in most urban areas. Since the 1960s, demand for urban
housing prices unaffordable. Most new houses
land with services has significantly and systematically exceeded what are captured by medium-income households.
the government has supplied. Official demand for plots averages According to estimates, of five applications,
75,000, but the supply is below 6,000 a year. Most applicants lose only one housing unit is allocated to a needy
hope and turn to informal markets to obtain land for their develop- household.10
ment needs.
Serious difficulties also exist in accessing land for investment pur- Urban Growth and Informal
poses. The Tanzania Investment Centre has so far registered 4,210 Settlements
investment projects, 80 percent of which require access to land par- A major problem for Africa’s growing cities is
cels. However, the Tanzania Investment Centre estimates that only
the rapid spread of informal settlements. Lack
one-quarter of those registered can acquire land through the exist-
of affordable serviced plots and zoning poli-
ing formal system. Between 2004 and 2007, 440 applications for
land allocation were received, but only 13 applicants received their
cies have often excluded the poor from being
titles, evidence that the formal land delivery system in urban areas is integrated with urban development, leaving
not working. In 1998, informal settlements in Dar es Salaam covered them in underserviced shelters (slums) both
48 percent of the built-up area, and recent estimates place the share in and on the outskirts of major cities. The
as high as 70–80 percent. UN-Habitat standard definition of people
The scarcity of formal plots stems from the underperformance of living in slums is overcrowding, low access
the public system, the lack of institutional or financial resources, and to water and sanitation, lack of secure ten-
the competition of the formal system with informal land develop- ure, and poor housing quality. Based on this
ment, the latter being much cheaper and easier. Institutional con- definition, as much as 70 percent of Africa’s
straints also exist. The 1999 Land Act concentrated most powers of
urban population resides in slums (UN-Habitat
land management in the central government, depriving local govern-
2003), and from 1990 to 2001, the slum pop-
ment authorities of the institutional structures to deal with their land-
related responsibilities. Goals for the land are not explicit, and laws
ulation grew at 4.4 percent a year, faster than
are not applied consistently. The division of responsibilities among the urban population. If this trend persists, an
actors is unclear, and there is no coordination. additional 218 million Africans will be living
in slums by 2020, and almost one-third of the
Sources: Muzzini and Lindeboom 2008; Raich and Sarzin forthcoming.
world’s slum population will be in Africa.
Great heterogeneity exists across cities
in the living and income conditions of slum
residents. Not all people living in these settle-
ments are income poor, although significant
informal market and their own construction. overlap occurs, and living standards differ
In Burkina Faso and Mali, more than 45 per- across countries. In many cases, informal
cent of the population in the capital city lives settlements are scattered across cities, side by
in informal settlements (Farvacque-Vitkovic side with wealthier residences. In Tanzania,
and others 2007). In Ghana, the market pro- urban dwellers in the periurban areas do well,
duces only 20 percent of the annual demand enjoying the freedom of informality, even
for shelter (Farvacque-Vitkovic, Raghunath, without land titles or finished walls. Indeed,
and others 2008). In Accra, people’s increased informal periurban areas are sometimes the
purchasing power because of remittances has most dynamic, precisely because overzealous
led to higher real estate prices, and zoning regulation does not affect them. A strong “city
and legal restrictions artificially limit the land effect” also exists. Thanks to leadership, land
available for housing. security, ownership, and civic participation,
Many governments have tried to help the the inhabitants of Dakar’s slums (Senegal)
poor by subsidizing construction, providing tax have living standards far superior to Nairobi’s
incentives to developers, or producing public (Kenya) even though the latter have higher
Facilitating Urbanization 137
incomes and education levels (Gulyani, Taluk- the creditworthiness to raise their own debt
dar, and Jack 2008). finance, and the few examples (South Africa
The main problem of slums and informal- and Zimbabwe) have illustrated the weakness
ity is exclusion from basic amenities. Nai- of the system and the need for impartial credit
robi’s slum residents pay up to 11 times more ratings.
for water sold by private vendors than those For larger cities, the economic base is larger,
who have access to piped water (Farvacque- and they have greater autonomy to raise their
Vitkovic, Glasser, and others 2008). In Africa own taxes. However, tax receipts are often sent
as a whole, the price of piped water is $0.50 to the central governments, and political fac-
per cubic meter, whereas water purchased tors hinder the use of property taxes. Although
from private vendors in mobile carts is $4.75 African cities generate 80 percent of the coun-
per cubic meter. Inadequate access to basic try’s tax revenues, they end up with less than
services also has implications for health and 20 percent of the resources.11 On the other
human development. Moreover, spatial mis- hand, larger urban areas are likely to have
matches and distance constrain accessibility to fewer spending responsibilities because they
education and livelihood opportunities. are often covered by national service providers
(such as power and water utilities), thus reliev-
ing these municipalities of some budgetary
Policy Issues and Implications expenses they would otherwise incur.
In cities such as Dar es Salaam, Tanzania,
Infrastructure Financing and Nairobi, Kenya, improvement in land
Cities should be spending more and spending management institutions could open the door
more wisely. Although trunk infrastructure to increased property revenues for munici-
and services with substantial spillover effects palities, more land use and sales revenues for
are clearly the responsibility of the central gov- the government, and complementary private
ernment, cities are responsible for solid waste, finance. For example, Cairo has held sev-
sewerage, drainage, and lighting. In many eral auctions of land for conversion, adding
cases, they are called on to help with shelter as 10 percent to the city budget. It has also
well. The purported benefits of decentraliza- swapped land permits for private infrastruc-
tion are not realized because decentralization ture in public land. For large Chinese cities,
policies have given cities more responsibili- land leases are the usual method to mobilize
ties (notably in social sectors) but not more resources, as well as in Mumbai and Bangalore
resources. Without independent or predict- in India. In 11 African countries, street address-
able sources of revenues, African cities can ing (or addressage) has been established in the
rarely plan or decide on the best way to allo- major municipalities, increasing municipal tax
cate their resources. billings by about 50 percent, with 90 percent
In principle, cities have greater potential collection rates. In Burkina Faso, Mauritania,
to raise local revenues. First, the larger urban and Togo, street addressing has helped in
economy provides a significant local tax base, inventorying the local tax base and implement-
although its predominantly informal nature ing residential taxation (Farvacque-Vitkovic
prevents the authorities from capturing taxes. and others 2007; Farvacque-Vitkovic, Glasser,
Second, high-value urban properties constitute and others 2008; Kessides 2006). In Benin,
a major potential tax base, although the lack decentralized management fostered increases
of clear property titles prevents it from being in city revenue of 82 percent in Cotonou,
realized. Third, the higher purchasing power 131 percent in Parakou, and 148 percent in
and tighter agglomeration of urban house- Porto Novo, with better collection rates as
holds make recovery of a significant propor- well (Kessides 2006).
tion of infrastructure financing requirements In rural areas, transfers complemented by
with user fees easier. centralized funds dominate. Many countries
Long-term debt is always an alternative (at have tried to fund rural investment by intro-
least in theory), but few African cities have ducing centralized funding mechanisms to
138 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
channel earmarked central government funds businesses and households to locate in the city
and donor resources to rural infrastructure. and pushing prices up, thereby leading to a
These mechanisms include rural water funds relocation of activities and residents to non-
(90 percent of countries), rural electrification regulated places. Land acquisition delays are
funds (76 percent), and rural telecommunica- very long in Ethiopia and Zambia. In Mozam-
tion funds (29 percent). For power, rural funds bique, businesses pay on average $18,000 in
bring faster expansions in the electrified rural processing fees for land, and in Nigeria, they
population: an annual increase of 0.72 percent must register land to use it as collateral, a pro-
is seen in countries with such funds, compared cess that can take up to two years and cost
to an annual contraction of −0.05 percent else- 15 percent of the land value (Kessides 2006).
where. For rural water and rural ICT funds, no Land institutions can improve information,
significant differential was seen in the rate of strengthen property rights, record market trans-
service expansion. actions, and steadily move toward more open
Central funds can also support the mainte- land markets. With an endowment of 63 square
nance of rural infrastructure. To deal with poor miles of land, the Tema Development Corpo-
maintenance, many countries are allocating ration in Ghana is planning and laying out the
part of their national road fund revenues to the Tema area, constructing roads and sewerage sys-
maintenance of the rural network (60 percent tems, preparing and executing housing projects,
of countries). This decision may be a good and managing rental units. Permits for housing
strategy: countries that allocate at least $0.015 construction are submitted to the authority,
per liter of their fuel levies to the rural road which charges a permit fee based on the value
networks have a significantly higher share of the property to be developed (Farvacque-
of their rural roads in good condition than Vitkovic and McAuslan 1993).
those that allocate less, 36 percent versus Urban planning should guide urban expan-
21 percent. sion and the associated infrastructure needs.
Because of their top-down approach and weak
Land Policies and Urban and implementation, urban planning and master
Territorial Planning plans have lost their meaning in many Afri-
Adequate land policies and markets are the can cities. Urban dynamics are seldom cor-
backbone of an efficient urban transition. Land rectly foreseen, and in most cases, the political
management institutions include a compre- economy has the last word in determining the
hensive land registry, credible mechanisms for location of infrastructure or major develop-
contract enforcement and conflict resolution, ments. To be efficient and useful, planning
flexible zoning laws, and versatile regulation should be flexible, participatory, and indica-
of subdivisions that help rather than hinder tive (10–15 years). Urban reference maps
the conversion of land for different uses. Prop- should lay out major roads and city services,
erty rights embodied in land titles are essential the areas for urban expansion, and the reserves
for converting assets into usable wealth. The for amenities.
transformation of the agricultural sector from Planning should check sprawl, enhance
communal land rights to individual property densification, prevent development in pre-
rights is important for the rural-urban trans- carious environmental zones, and enhance the
formation, but it may take a long time. delivery of affordable serviced land and infra-
Land use and building regulations become structure. Ideally, planning should be rooted in
more important as urbanization advances. participatory strategies and linked to local and
Governments regulate land markets to ensure central budgets. Without realistic projections
separation of land between different uses for resource availability, urban plans often fall
and to ensure the integration of public and into discredit. Dakar (Senegal), Lagos (Nigeria),
private uses of land, such as providing space and Maputo (Mozambique) recently prepared
for transport infrastructure in densely popu- city development strategies as frameworks to
lated areas. Land regulations can be overzeal- encourage participation from the community
ous, however, distorting the incentives for in discussing challenges and opportunities.
Facilitating Urbanization 139
Territorial planning is critical in rural areas basic infrastructure services, and shelters are
to promote a more integrated vision of devel- upgraded into better and more durable con-
opment and to enhance growth opportunities. structions. Investment in transportation and
Rural development requires the coordinated social programs is also used to strengthen
provision of infrastructure services to support links between slum areas and the rest of the
agricultural production and off-farm activities, city and to facilitate social integration. The
such as irrigation infrastructure, rural roads, and Accra District Rehabilitation Project in Ghana
associated transport services, as well as storage is an example of successful upgrading, as are
and distribution infrastructure for agricultural several national upgrading programs in Ethio-
products. In rural areas, the limited administra- pia, Kenya, and Uganda.
tive capacity hinders an integrated vision. Coor- The second, more controversial, approach
dination may be further complicated because focuses on resettling slum dwellers, by moving
some services, such as irrigation, may be a cen- them either to existing neighborhoods or to
tral government competency, whereas others new, less crowded, and safer locations. In either
(such as roads) are local. Chongqing, China’s case, slum dwellers are compensated for reset-
experience with implementing a territorial tlement and disruptions to their livelihoods.
development plan on a regional scale provides Much can be done, starting with the provi-
an example that may be of interest to Africa. sion of basic services and infrastructure com-
Bundling infrastructure services can sub- bined with effective land policy. Legislation that
stantially increase the return of infrastructure boosts land prices and excludes the poor will
investments. Bundling infrastructure not only need to be revised. Basic packages of services
secures larger welfare gains to households, (street lighting, paving, drainage, roads) should
urban and rural alike, it also maximizes the be extended to the broadest number of people at
economic and social effect of infrastructure the lowest possible cost. In Kenya, Mozambique,
service provision in rural areas by granting and Nigeria, major improvements are possible
better access to economic opportunities and for a low of $150 per capita, compared with
reducing the gap between poor and nonpoor. $1,800 or more for finished solutions. Urban
Therefore, investment policies, especially when transportation can bring the urban poor into
rural infrastructure is concerned, should more the large labor markets. A consistent and exten-
attentively seek complementarities across sec- sive policy of land titling would provide clarity
tors. This requires that institutional coordina- and predictability to the land market, develop
tion and planning and financial capacity are people’s ownership, and promote private invest-
deployed as needed. Although bundling is an ment. In many countries, however, land titling
opportunity to realize larger returns on invest- will continue to be difficult (more difficult than
ments, alone it is not enough to drive economic extending basic services) because of the political
and social development in rural areas. A more economy and weak administrative capacity.
far-reaching vision of rural development that
maximizes coordination and complementa-
rities across sectors beyond the infrastructure Six Principles for Efficient
field—for which, however, bundling consti- Urbanization
tutes an important tool—is needed. This is the
scope of territorial development. Based on the foregoing discussion, there are six
key principles for achieving efficient urbaniza-
Informal Settlements tion. First, adopt a solid analytical framework to
Preventing the formation of slums and upgrad- help define priorities and sequencing. In places
ing existing ones are major concerns to policy that are mostly rural, governments should be
makers; one of two possible approaches nor- neutral and establish the foundations for effi-
mally prevails. The first approach focuses on cient urbanization (World Bank 2009). Good
improving the living standards of slum dwell- land policies and universal provision of basic
ers in their existing locations. They are given services are central. Where urbanization has
land tenure; slums areas are equipped with accelerated, the priority should be investments
140 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
in connectivity to ensure that the benefits of broadly planning for the extension of urban
rising economic density are widely shared. In settlements, taking into account transporta-
highly urban places, targeted interventions may tion, connectivity, and environmental factors.
be needed to deal with slums and exclusion. Sixth, improve the fiscal soundness of cities:
Second, recognize that the political econ- (a) improve transparency and predictability of
omy influences the urban transition. African transfers; (b) strengthen and simplify local
cities are not very powerful. Unlike cities in taxation, changing the focus of property tax
East Asia or Latin America, African cities have from ownership to occupancy; (c) take advan-
little autonomy and depend on the central gov- tage of cost recovery from revenue-producing
ernment for resources, infrastructure projects, services, such as markets and bus stations—
and even land development. Chinese mayors they can make up 70 percent of medium-size
are appointed by the party, but their political city revenues; and (d) use municipal contracts
careers depend on how well they develop their (between central and local governments) and
cities. Especially if they are elected, U.S. and adressage to help local governments manage
European mayors may see their cities as steps their resources.
to higher political positions, including the
presidency. Mayors in Africa have limited free-
dom of action. Many of the difficulties Afri- Notes
can cities have in collecting property taxes are The authors of this chapter are Elvira Morella,
related to the political influence of the major Maria Emilia Freire, and Paul Dorosh, who
landowners, who oppose such taxes. drew on background material and contribu-
Third, be pragmatic. While the long-term tions from Alvaro Federico Barra, Catherine
Farvacque-Vitkovic, Matthew Glasser, Sumila
goal is to have well-defined property rights
Gulyani, Darby Jack, Austin Kilroy, Barjor
and land titling, in the short run, cities may Mheta, Stephen Mink, Siobhan Murray, Madhu
need to “finesse” land titling and use occu- Raghunath, Uri Raich, Raj Salooja, Zmarak
pancy as a basis for land registration and Shalizi, and Debabrata Talukdar.
taxation. Resource-constrained governments 1. Note that the early stages of this transition need
should invest in minimum packages of water- not involve movements of activities or people;
sanitation-energy in informal underserviced rural households increasingly earn incomes
quarters at the citywide level and resist the from rural nonagricultural activities (in agri-
idea of transforming slums into perfect neigh- cultural processing, construction, commerce,
borhoods. For about $1,200 per capita, many and private services).
African slum dwellers can be provided with 2. This correlation between agricultural produc-
tion and proximity (as measured by travel time)
basic services, compared with $18,000 spent in
to urban markets holds even after taking agro-
more comprehensive and sophisticated proj- ecology into account.
ects in Latin America (Farvacque-Vitkovic, 3. No internationally accepted standard exists for
Glasser, and others 2008). identifying urban areas, and each country tends
Fourth, focus on cities and areas important to use its own definition. This situation hinders
for the economy. The priorities should be to any effort to make sensible comparisons across
improve the institutional framework (espe- countries. In this chapter, urban areas, from sec-
cially on land markets), to provide techni- ondary cities to megacities, are identified using
cal and financial resources for planning and a subset of the GRUMP (Global Rural-Urban
developing infrastructure and basic services, Mapping Project) urban extents layer (CIESIN
to harness agglomeration economies, and to 2004). The GRUMP urban extents were joined to
a data set of city populations compiled by Hend-
deal with congestion.
erson (2002) and of urban extents classified by
Fifth, improve land policies so that mar- population size. To complete the urban-rural
kets are more flexible and can respond to the gradient, nonurban areas were classified by dis-
increase in demand. That requires compiling tance or travel time to the nearest city. The com-
inventories of government land and of for- bination of urban extents and city populations
mal and informal developers, gathering prices allowed creation of a density-based typology of
and costs for land plots and construction, and cities. However, given the limitations associated
Facilitating Urbanization 141
with these input data, the calculation of density Banerjee, Sudeshna, Quentin Wodon, Amadou
is approximate at best. Even so, the density-based Diallo, Taras Pushak, Hellal Uddin, Clarence
characterization of “urban” areas allows com- Tsimpo, and Vivien Foster. 2008. “Access, Afford-
parisons across regions and reflects the relation ability, and Alternatives: Modern Infrastructure
between density and agglomeration economies. Services in Africa.” Background Paper 2, Africa
Infrastructure Country Diagnostic, World Bank,
4. World Development Report 2009 expressed geo-
Washington, DC.
graphic transformation as the development of
the leading and lagging regions. Although both Bertaud, Alain. 2003. “Order without Design.”
regions may consist of both urban and rural http://www.alain-bertaud.com.
areas and hence agriculture and nonagriculture, Briceño-Garmendia, Cecilia, Karlis Smits, and
in South Asia, lagging regions are predomi- Vivien Foster. 2008. “Financing Public Infra-
nantly rural and agriculture remains the main structure in Sub-Saharan Africa: Patterns, Issues,
source of livelihood. and Options.” Background Paper 15, Africa
Infrastructure Country Diagnostic, World Bank,
5. As mentioned, the most important hindrance
Washington, DC.
to economic production is the supply of energy.
In many countries, it can account for half of the CIESIN (Center for International Earth Science
value of the final output. Information Network). 2004. Global Rural-Urban
Mapping Project (GRUMP): Urban Extents.
6. These funds include annual public investments,
Palisades, NY: CIESIN, Columbia University.
annualized official development assistance, annu-
alized emerging financiers that do not belong to Chong, Alberto, Jesko Hentschel, and Jaime
the Organisation for Economic Co-operation Saavedra. 2007. “Bundling of Basic Public Ser-
vices and Household Welfare in Developing
and Development, and annualized private par-
Countries: An Empirical Exploration for the
ticipation in infrastructure. Case of Peru.” Oxford Development Studies 35
7. These figures include energy and roads; no data (3): 329–46.
are available on the share of investment on ICT
Dorosh, Paul, Hyoung-Gun Wang, Liang You,
and water that serves industrial production.
and Emily Schmidt. 2008. “Crop Production
8. Population figures are derived from national and Road Connectivity in Sub-Saharan Africa:
censuses. A Spatial Analysis.” Working Paper 19, Africa
9. The price of construction is estimated at $222 per Infrastructure Country Diagnostic, World Bank,
square meter, so a 75-square-meter home would Washington, DC.
cost about $17,000 (excluding land) (AGETIPE Eberhard, Anton, Vivien Foster, Cecelia Briceño-
2005). Average incomes are $850 a year, produc- Garmendia, Fatimata Ouedraogo, Daniel
ing a ratio of housing price to income of over Camos, and Maria Shkaratan. 2008. “Under-
2,000 to 1, one of the world’s highest. powered: The State of the Power Sector in Sub-
10. Efforts to develop mortgage systems for low- Saharan Africa.” Background Paper 6, Africa
income households in Africa have encountered Infrastructure Country Diagnostic, World Bank,
several problems: lack of credit history, lack of Washington, DC.
regular income, shallowness of the financial Farvacque-Vitkovic, Catherine, Alicia Casalis,
market, lack of long-term funding, lack of land Mahine Diop, and Christian Eghoff. 2007.
and house registries, high lending rates, and “Development of the Cities of Mali: Challenges
high credit risks. and Priorities.” Africa Region Working Paper
11. Too often, the central government appropriates 104a, World Bank, Washington, DC.
the city-generated taxes and distributes them to Farvacque-Vitkovic, Catherine, Matthew Glasser,
sectors that are not necessarily the most pro- Barjor Mehta, Madhu Raghunath, Austin Kilroy,
ductive and certainly not the ones that will feed Alvaro Federico Barra, and Raj Salooja. 2008.
into the urban economy. “Africa’s Urbanization for Development: Under-
standing Africa’s Urban Challenges and Oppor-
tunities.” World Bank, Washington, DC.
References Farvacque-Vitkovic, Catherine, and Patrick
AGETIPE (Agence d’exécution des travaux McAuslan. 1993. “Politiques Foncière des Villes en
d’intérêt public pour l’emploi). 2005. Contract Développement.” World Bank, Washington, DC.
documents. www.agetipe.org. Farvacque-Vitkovic, Catherine, Madhu Raghunath,
Angel, Schlomo, Stephen C. Sheppard, and Daniel Christian Eghoff, and Charles Boakye. 2008.
L. Civco. 2005. The Dynamics of Global Urban “Development of the Cities of Ghana: Chal-
Expansion. Washington, DC: Transport and lenges, Priorities and Pools.” Africa Region Work-
Urban Development Department, World Bank. ing Paper 110, World Bank, Washington, DC.
142 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Gulyani, Sumila, Debabrata Talukdar, and Darby Empirical Base for Policy Dialogue. Washington,
Jack. 2008. “A Tale of Three Cities: Understand- DC: World Bank.
ing Differences in Provision of Modern Raich, Uri, and Zara Sarzin. Forthcoming.
Services.” Working Paper 10, Africa Infrastruc- Financing the Urban Expansion in Tanzania.
ture Country Diagnostic, World Bank, Washington, DC: World Bank.
Washington, DC.
Torero, Maximo, and Javier Escobal. 2005. “Meas-
Gwilliam, Ken, Vivien Foster, Rodrigo Archondo-
uring the Impact of Asset Complementarities:
Callao, Cecilia Briceño-Garmendia, Alberto
The Case of Rural Peru.” Cuadernos de Economia
Nogales, and Kavita Sethi. 2008. “The Bur-
42 (May): 137–64.
den of Maintenance: Roads in Sub-Saharan
Africa.” Background Paper 14, Africa Infra- UN-Habitat (United Nations Human Settle-
structure Country Diagnostic, World Bank, ment Programme). 1998. “Global Urban
Washington, DC. Observatory.” http://ww2.unhabitat.org/
programmes/guo.
Henderson, J. Vernon. 2002. “World Cities Data.”
http://www.econ.brown.edu/faculty/henderson/ ———. 2003. “Global Urban Observatory.” http://
worldcities.html. ww2.unhabitat.org/programmes/guo.
Kessides, Christine. 2006. The Urban Transition in World Bank. 2007. “The Challenge of Urbaniza-
Sub-Saharan Africa: Implications for Economic tion in Ethiopia: Implications for Growth and
Growth and Poverty Reduction. Washington, DC: Poverty Alleviation.” Africa Region, Water and
The Cities Alliance. Urban Development Unit 2, Washington, DC.
Maal, Simen Jansen. 2008. “Measuring the Con- ———. 2008. World Development Report 2008:
tribution of Urban Centers to the National Agriculture for Development. Washington, DC:
Economy of Tanzania.” Report to World Bank, World Bank.
Dar es Salaam, Tanzania. ———. 2009. World Development Report 2009:
Muzzini, Elisa, and Wietze Lindeboom. 2008. Reshaping Economic Geography. Washington,
The Urban Transition in Tanzania: Building the DC: World Bank.
Chapter 6
W
ith many small, isolated economies, management and development of cross-border
Africa’s economic geography is public goods. Road and rail corridors linking
particularly challenging. Regional landlocked countries to the sea are an example
integration is likely the only way to overcome of such a regional public good, as are regional
these handicaps and participate in the global airport and seaport hubs. The same can be said
economy. Integrating physical infrastruc- of Africa’s 63 international river basins.
ture is both a precursor to and an enabler for Reaping these benefits, however, poses insti-
deeper economic integration, thereby allowing tutional challenges:
countries to gain scale economies and harness
regional public goods. For successful regional • Building a political consensus. The politi-
integration, countries must start small; build on cal obstacles can trump the economic case.
successes; think globally, linking Africa to more Regional infrastructure involves a high
external markets; and compensate the least for- level of trust between countries, not least
tunate, recognizing that benefits are not always because of the implied dependence on
evenly distributed. neighbors for key resources such as water
The benefits of regional integration are visible and energy.
across all aspects of infrastructure networks. • Establishing effective regional institutions.
For information and communication technol- Regional institutions have to facilitate
ogy (ICT) and power, regional infrastructure agreements and compensation. Africa has
provides scale economies that substantially an extensive architecture of regional politi-
reduce the costs of production. Thus, con- cal and technical bodies, but these face
tinental fiber-optic submarine cables could problems because of overlapping member-
reduce Internet and international call charges ships, limited technical capacity, and lim-
by one-half. Similarly, regional power pools ited enforcement powers.
that allow countries to share the most cost- • Setting priorities for regional investments.
effective energy resources can reduce electric- Given the daunting investment agenda,
ity costs by $2 billion a year. For transport and better sequencing and priority setting of
water, regional collaboration allows optimal regional projects has been elusive. Political,
143
144 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
economic, and spatial approaches to prior- the world requires large up-front investments
ity setting have all been widely discussed. in undersea cables or satellite communication.
For private engagement in ICT or energy, the
• Developing regional regulatory frameworks.
Physical integration of infrastructure net- opportunity to serve a larger regional mar-
works will be effective only with harmonized ket makes the extensive up-front investments
regulatory frameworks and administrative more attractive. Airports and seaports must be
procedures to allow the free flow of services organized as regional hubs to reach the scale
across national borders. necessary to attract airline and shipping ser-
vices from beyond the continent.
• Facilitating project preparation and cross-
Coordinated management and invest-
border finance. The complexity of regional
ment allow countries to reap the best from
infrastructure projects makes them costly
multicountry infrastructure systems. Some
and time-consuming to prepare. This is
regional infrastructure investments, such as
particularly true when projects are large
many types of transport investments, pro-
in relation to the size of the host economy
vide public goods, or they facilitate access
and essentially depend on financing from
to a common pool resource, as with water
downstream beneficiaries.
resource management for irrigation and other
uses. Both public goods and common pool
resources require strong coordination. Because
Why Regional Integration Matters the quality of a transport network depends on
its weakest link, broad participation is crucial,
Regional approaches can address the infrastruc- even where benefits are unequally distributed.
ture backlog in Africa and propel economic Water can be exhausted, and upstream users are
growth, overcoming the region’s difficult in a stronger position than downstream users.
geography (Limão and Venables 1999). Sub- Collective agreements, effective monitoring,
Saharan Africa has 48 countries, most with and conflict resolution mechanisms can ensure
small populations—more than 20 countries a fair distribution of costs and benefits.
have a population of less than 5 million. Econ- The goal of all regional infrastructure
omies are also very small—20 countries have efforts is to facilitate the spatial organization
a GDP of less than $5 billion. The small scale of economic activity as a catalyst for faster
means governments have difficulty funding the growth. Lessons from the new economic geog-
large fixed costs associated with infrastructure raphy, for which Paul Krugman received the
development. In addition, 15 African countries Nobel Prize in Economics in 2008, explain this
are landlocked, depending on their neighbors concept. Natural resource exports will remain
for access to global markets. important, but they provide few job oppor-
Most infrastructure investments share char- tunities and their benefits are seldom widely
acteristics of public goods, and all benefit to shared. Dutch disease, greater macroeconomic
varying degrees from scale economies. Infra- volatility, and weak governance have slowed
structure sharing addresses the problems of growth in some resource-rich African coun-
small scale and adverse location. Joint provision tries (Collier 2007). Fast employment growth
increases the scale of infrastructure construc- and sustained welfare improvements in devel-
tion, operation, and maintenance. It reduces oping countries require a move toward mod-
costs, pools scarce technical and managerial ern, mostly export-oriented manufacturing
capacity, and creates a larger market. Econo- activity. The shift in trade that allowed East
mies of scale are particularly important in the Asia’s rapid growth can also benefit Africa. In
ICT and power sectors. Big hydropower projects the world’s fastest-growing regions, the largest
that would not be economically viable for a sin- increase in trade has been within industries, for
gle country make sense when neighbors share parts and components produced in one loca-
their benefits. Although new ICT systems— tion and assembled in another. Manufacturing
especially mobile telephones—allow provision is more about specialized “tasks” than finished
at the local retail level, connecting Africa to products (Collier and Venables 2007).
Deepening Regional Integration 145
is limited to a few million. Enabling Africa to Africa Submarine Cable (SAT-3), which passes
develop regional manufacturing clusters that along the Gulf of Guinea and down to South
can compete globally requires lowering barri- Africa. The entire east coast of Africa lacks
ers to both productive interaction and (at least access to an underwater cable. Intraregional
temporarily) preferential access to world mar- connectivity by fiber-optic cables is also limited.
kets with liberalized rules of origin. Regional Most countries rely on satellites for interna-
integration is essential, and regional infrastruc- tional telecommunications, including Internet
ture sharing must be a key priority. National access, leading to prices for dial-up and broad-
infrastructure programs such as those in India band Internet at least twice as high as in other
or China (for example, the Golden Quadrilat- regions. Transmission capacities are low, and
eral highway program) will involve agreement costs are high. Countries without access to an
by numerous countries in Africa. However, the underwater cable have only 3 bits of bandwidth
payoffs to greater coordination and integration per capita, whereas those with access have 24
of infrastructure will be large. bits. The average cost of an international fixed-
line telephone call within Sub-Saharan Africa
is $1.23 a minute, almost twice the cost of a call
Opportunities for Regional to the United States ($0.73 a minute). Intrare-
Cooperation across Infrastructure gional call traffic is barely 113 million minutes,
Sectors compared with intercontinental call traffic of
250 million minutes.1
Africa’s regional infrastructure networks have Several projects are already under way
major gaps that increase the costs of doing busi- to complete the loop of underwater cables
ness and prevent the realization of scale econo- around Africa, with an estimated value of
mies. Supply of infrastructure as a public good $1.8 billion (table 6.1). Most are commercially
and management of common pool resources sponsored and privately financed, such as the
have been deficient. More efficient regional inte- Eastern African Submarine Cable System proj-
gration of infrastructure is needed in all sectors: ect to link South Africa and the Horn of Africa
ICT, transport, power, and water (figure 6.1). (box 6.1). The cost of completing the nascent
fiber-optic network interconnecting the capital
ICT—Slashing Costs of International cities of Sub-Saharan Africa and the main sub-
Voice and Internet Connectivity marine cables is modest at $316 million, based
As in other parts of the world, mobile tele- on a cost of about $27,000 per kilometer.
phones have greatly improved telecommuni- The most immediate direct benefits of
cations in Africa. However, the benefits have enhanced connectivity are reduced prices and
been limited to local and domestic communi- better service for international voice and Internet
cations. The region’s national telecommunica- connectivity. Prices for most services in countries
tion networks are poorly integrated with each with underwater cable access are half those
other and the rest of the world. Optical-fiber in countries without access (table 6.2). Such
technology provides the least expensive and large price reductions could boost demand
highest-capacity transmission of telephone, for these services and, ultimately, economic
Internet, and other data traffic. Submarine and productivity. However, too often, access to
land-based communication cables combine underwater cables remains with the incum-
high speed and large capacity. Although initial bent operator, which (without adequate regu-
investments are high, marginal transmission latory controls) charges monopoly prices that
costs are very low. prevent consumers from reaping the full cost
Access to the global network of submarine advantage of this technology. Countries with
cables is low in Africa, especially for landlocked multiple international gateways see some com-
countries that depend on neighbors for access. petitive pressure, which keeps service prices
Both coordination and massive investments significantly lower than in countries where
are required. The region’s main international an underwater cable is the only international
underwater cable is the South Atlantic 3/West gateway (table 6.2).
Deepening Regional Integration 147
TUNISIA TUNISIA
MOROCCO MOROCCO
ALGERIA ALGERIA
MAURITANIA
MAURITANIA MALI
• MALI NIGER ERITREA ERITREA
CAPE VERDE CAPE VERDE NIGER
CHAD
SUDAN
SENEGAL • • DJIBOUTI
SENEGAL
DJIBOUTI
• GAMBIA CHAD
GUINEA- • • SUDAN
•
BISSAU • • • • GUINEA-
NIGERIA BISSAU
ETHIOPIA
GUINEA • • •
SOMALIA GUINEA NIGERIA
SOMALIA
SIERRA LEONE
•
ETHIOPIA SIERRA LEONE
• • CENTRAL AFRICAN CENTRAL AFRICAN
LIBERIA •
• • REPUBLIC LIBERIA
BENIN
REPUBLIC
CÔTE D’IVOIRE BENIN
• • TOGO
GHANA TOGO CAMEROON CÔTE D’IVOIRE GHANA
BURKINA FASO CAMEROON
• BURKINA FASO KENYA
•
EQUATORIAL GUINEA
•
EQUATORIAL GUINEA GABON GABON
SAO TOME •
• KENYA UGANDA SAO TOME UGANDA
AND PRINCIPE
AND PRINCIPE CONGO, DEM REP • CONGO, DEM REP
SEYCHELLES RWANDA
• TANZANIA
RWANDA
BURUNDI
SEYCHELLES
REP. OF. CONGO BURUNDI
• REP. OF. CONGO
TANZANIA
•
ANGOLA COMOROS COMOROS
ANGOLA
• MALAWI MAYOTTE
ZAMBIA ZAMBIA
• MALAWI
SAINT HELENA
•
•
MOZAMBIQUE MOZAMBIQUE
• NAMIBIA REUNION
NAMIBIA
• ZIMBABWE
Broadband network BOTSWANA
•
• • ZIMBABWE
BOTSWANA
MADAGASCAR
SOUTH MADAGASCAR
SWAZILAND
existing or under construction • SWAZILAND AFRICA
SOUTH international river basins LESOTHO
missing AFRICA
LESOTHO
• TUNISIA TUNISIA
MOROCCO • MOROCCO
•
ALGERIA ALGERIA
LIBYA LIBYA EGYPT
EGYPT
• •
COMOROS COMOROS
ANGOLA
ANGOLA • ZAMBIA • MALAWI
ZAMBIA MALAWI
• •
• •
• •
MOZAMBIQUE MOZAMBIQUE
• NAMIBIA •
NAMIBIA ZIMBABWE ZIMBABWE
• •
Beyond direct benefits, enhanced con- specialized production that exploits econo-
nectivity supports closer regional economic mies of scale—to move from low-technology,
networks and integration with international standardized manufacturing to internationally
markets (Leamer and Storper 2001). Good competitive production and to ensure access
communication is a precondition for the to global markets for manufactured outputs
emergence of buyer-supplier networks for and business services.
148 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
BOX 6.1
Not So EASSy
The Eastern African Submarine Cable System (EASSy) is an has meant that prices remain high and the benefits for cus-
underground fiber-optic cable that runs from South Africa tomers limited.
to Sudan, allowing all countries along the route to connect Designed to ensure effective competition and regula-
to the global submarine cable system. tion, EASSy is owned by a consortium, which also includes
Developed and owned by a consortium of about a special-purpose vehicle owned by a group of smaller
25 telecommunication operators, mostly from eastern operators from the region. Development finance support
and southern Africa, the cable is expected to cost $230 for EASSy is provided as loans to this special-purpose vehi-
million. About one-third of the funding will come from cle, which can sell capacity in any market in the region to
nonconcessional debt finance provided by five develop- licensed operators on an open-access, nondiscriminatory
ment finance institutions (the International Finance Cor- basis, thus competing with other consortium members.
poration, the European Investment Bank, the African As traffic volumes increase, the special-purpose vehicle is
Development Bank, the French Development Agency, and required to pass on cost reductions to customers.
the German Development Bank), with the remainder from Reaching consensus about these access arrangements has
commercial equity. been difficult, leading to project delays. Meanwhile, Kenya
The project took about three years to develop. The first moved forward with its own underwater cable, The East Africa
stage involved discussions and negotiations with stake- Marine System (TEAMS), with links to the United Arab Emirates.
holders to determine the project’s structure. The second Technically much simpler, that project enjoys significant private
stage focused on the technical and financial details of backing. Unless the system can be integrated in a regional net-
implementation. The third stage, laying the cable, began work, however, costs will be higher and benefits less broadly
in 2008. shared than with a regional effort. A third, privately funded
Policy makers and development finance institutions have effort, South Africa–East Africa–South Asia–Fiber Optic Cable
focused on not repeating the experience of the SAT-3 cable, (SEACOM), is planned to connect South Africa and several East
which runs along the west coast. That project was also African countries to global networks by mid-2009.
financed, built, and managed by a consortium of operators,
but each member of the consortium has exclusive control over Source: Based on interviews with staff from the World Bank’s ICT Policy
access to the cable in its own country. Lack of competition Department, 2008.
raise the billions of dollars needed to develop Regional trade also puts Africa on a cleaner
them. For example, 60 percent of the region’s development path in terms of carbon emissions.
hydropower potential is in the Democratic Regional power trade would increase the share
Republic of Congo and Ethiopia. Because of hydropower in the continent’s generation
21 of 48 Sub-Saharan African countries have portfolio from 36 percent to 48 percent, dis-
total generation capacity of less than 200 placing 20,000 megawatts of thermal power in
megawatts—efficient scale of production the process and saving 70 million tons a year
below minimum—they pay a heavy penalty: of carbon emissions (8 percent of Sub-Saharan
costs reach $0.25 per kilowatt-hour, compared Africa’s anticipated emissions through 2015).
with $0.13 per kilowatt-hour in the region’s Applying the Clean Development Mechanism at
larger power systems. $15 per ton of carbon would reduce the region’s
The desire to pool energy resources and lever- carbon emissions another 4 percent. Closely
age scale economies in power sector develop- integrating power grids will also help balance
ment led to formation of regional power pools loads when other renewable energy resources,
in southern and western Africa during the mid- such as concentrated solar and geothermal
1990s, and more recently in eastern and central energy, are deployed on a large scale.
Africa. However, trade has yet to take off. Cross- The 10 largest potential power-importing
border power trade accounts for only 16 percent countries could reduce their long-run marginal
of the region’s power consumption, more than cost of power by $0.02–$0.07 per kilowatt-hour
90 percent of it within the Southern African (figure 6.2). Those that stand to gain most tend
Power Pool, and much of that between South to be smaller countries or those heavily reliant
Africa and its immediate neighbors. Without on thermal power, such as Angola, Burundi,
physical or regulatory impediments, about 40 Chad, Guinea-Bissau, Liberia, Niger, and
percent of eastern and southern Africa’s power Senegal. However, reaping the full benefits of
consumption would be traded across national power trade would require a political willingness
borders (Rosnes and Vennemo 2008). to depend heavily on power imports. As many as
If pursued to its full economic potential, 16 African countries would be better-off (from
regional trade would shave about $0.01 per a purely economic standpoint) importing more
kilowatt-hour off the marginal cost of power than 50 percent of their power needs.
generation in each of the power pools. The The future of power trade depends on the
resulting overall potential savings of regional health of the power sector in a handful of key
power trade would amount to about $2 billion exporting countries endowed with exception-
a year in the costs of power system develop- ally large and low-cost hydropower resources.
ment and operation (equivalent to about 5 per- In descending order of export potential,
cent of total power system costs). The savings these include the Democratic Republic of
come largely from substituting hydropower Congo, Ethiopia, Guinea, Sudan, Camer-
for thermal power, even if higher immediate oon, and Mozambique (table 6.3). The first
investments are required. three account for 74 percent of potential
150 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 6.2 Savings from Power Trade for Major investments needed to realize these export
Potential Power-Importing Countries volumes is daunting. Each of the top three
would need to invest more than $700 million
Guinea-Bissau
a year for the next decade to develop the gen-
Liberia eration capacity for export, or more than 8
Niger percent of their GDP. Thus, supporting such
Angola investments would be difficult without exten-
Chad sive cross-border financing arrangements that
Burundi
allow importing beneficiaries to make up-
front capital contributions.
Senegal
To make trade possible, countries would
Mali need to develop some 22,000 megawatts of
Congo, Dem. Rep. of interconnectors to allow the free flow of power
Equatorial Guinea across national borders, at a cost of more than
Mozambique $500 million a year over the next decade. The
returns to the interconnectors can be as high
Sierra Leone
as 120 percent for the Southern African Power
Lesotho
Pool, and more typically 20–30 percent for
Namibia the other power pools. For countries with the
South Africa most to gain from power imports, investments
Gabon in cross-border transmission have exception-
Kenya
ally high rates of return, typically paying for
themselves in less than a year.
0 1 2 3 4 5 6 7 8
US cents per kilowatt–hour
Transport Infrastructure—Facilitating
Source: Derived from Rosnes and Vennemo 2008. Internal and External Trade
Transport infrastructure is critical for linking
Africa to the global economy and promoting
Table 6.3 Profile of Top-Six Potential Power-Exporting Countries economic integration within the continent.
Potential However, the infrastructure demands are quite
Net revenues Required investment
exports different in each case.
(terrawatt $ millions Percentage $ millions Percentage
Country hours per year) per year of GDP per year of GDP
External Trade. For external trade, the conti-
Congo, Dem.
Rep. of 51.9 519 6.1 749 8.8 nent’s economic geography makes transport
Ethiopia 26.3 263 2.0 1,003 7.5 connections with the world something of a
Guinea 17.4 174 5.2 786 23.7
public good. Major corridors to the sea con-
nect the continent’s 15 landlocked countries to
Sudan 13.1 131 0.3 1,032 2.7
the major seaports, through a combination of
Cameroon 6.8 68 0.4 267 1.5
road and rail infrastructure. The main ports
Mozambique 5.9 59 0.8 216 2.8 include Douala (Cameroon) for central Africa;
Source: Derived from Rosnes and Vennemo 2008. Durban (South Africa) and Maputo (Mozam-
Notes: Net revenue is calculated as the estimated volume of exports multiplied by an illustrative profit
margin of $0.01 per kilowatt-hour exported. Required investment represents the investment needed bique) for southern Africa; Dar es Salaam
for the country to be able to exploit its full economic power-exporting potential.
(Tanzania) and Mombasa (Kenya) for eastern
Africa; and Abidjan (Côte d’Ivoire), Cotonou
(Benin), and Dakar (Senegal) for western
power exports, which could become big busi- Africa (table 6.4). About $200 billion a year in
ness for them. Based on a purely illustrative imports and exports moves along these cor-
profit margin of $0.01 per kilowatt-hour on ridors, barely 10,000 kilometers long. About
power sales, the net export revenues for these 70 percent is in good or fair condition, with
top three exporters could be 2–6 percent of donors channeling more resources to improve
their respective GDP. However, the size of the infrastructure along the routes.
Deepening Regional Integration 151
However, regulatory and administrative Table 6.4 Average Delivery Time for Containers from Vessel to Consignee
hurdles continue to inflate costs and pro- Distance Transit time
long delays for freight movements along Gateway Destination (kilometers) (days)
these strategic arteries (table 6.4). Despite Mombasa, Kenya Kampala, Uganda 1,100 20
the reasonably good physical condition of the Mombasa, Kenya Kigali, Rwanda 1,700 27
roads, the implicit velocity of freight move- Dar es Salaam, Tanzania Bujumbura, Burundi 1,800 21
ment is no more than 10 kilometers per hour Abidjan, Côte d’Ivoire Ouagadougou, Burkino Faso 1,200 7
(roughly the speed of a horse and buggy). The
Abidjan, Côte d’Ivoire Bamako, Mali 1,200 7
cause is the extensive delays of 10–30 hours
Dakar, Senegal Bamako, Mali 1,200 10
at border crossings and ports. Transit times
between major cities are very high by inter- Cotonou, Benin Niamey, Niger 1,000 11
national standards (table 6.4). Member coun- Douala, Cameroon Ndjamena, Chad 1,900 38
tries have organized corridor associations to Lagos, Nigeria Kano, Nigeria 1,100 21
address nonphysical barriers to transit, estab- Source: Arvis 2005, quoting an international logistics company.
lishing one-stop integrated border posts and
improving ports and custom administration.
Tariffs for road freight can be several times Table 6.5 Key Transport Corridors for International Trade, Sub-Saharan Africa
higher than in other parts of the developing Trade density Implicit Freight
world (table 6.5), which is attributed not to Road in good ($ millions velocitya tariff
Length condition per road (kilometers ($ per ton-
higher road transport costs in Africa, but to Corridor (kilometers) (percent) kilometer) per hour) kilometer)
exceptionally high profit margins in the truck-
Western 2,050 72 8.2 6.0 0.08
ing industry (Teravaninthorn and Raballand
Central 3,280 49 4.2 6.1 0.13
2008). These margins in turn reflect carteliza-
tion and restrictive regulatory frameworks, such Eastern 2,845 82 5.7 8.1 0.07
as market entry barriers; technical regulations; Southern 5,000 100 27.9 11.6 0.05
and the tour de role system that allocates freight Source: Teravaninthorn and Raballand 2008.
a. Includes time spent stationary at ports, border crossings, and other stops.
business based on queuing, particularly in cen-
tral and western Africa. This system favors large
fleets with mostly older trucks in poor condi- its own airport facility and jet fleet. Instead,
tion. Moreover, it fosters corruption, because a regional hubs serving multiple countries are
transport operator can increase its volume of needed, with fleets of smaller commuter jets
cargo only by bribing the freight bureaus, the to move passengers along spokes and into
government entities that allocate freight among hubs. Liberalization begun under the Yamous-
transport operators. soukro Decision in 1999 should allow carriers
Upgrading the remainder of the corridors serving key routes to consolidate and a better
to the sea to good condition is estimated to package of intraregional services to emerge.
cost about $1.5 billion, while the annual cost However, implementation has been lagging,
of maintenance is close to $1.0 billion. Simu- especially in harmonizing competition rules
lations suggest that corridor rehabilitation and removing nonphysical barriers such as
would yield an internal rate of return of 20–60 landing rights and tariffs. In eastern and
percent in eastern Africa’s northern corridor. southern Africa, this consolidation of carriers
However, low traffic, poor truck use, and an and hubs has already occurred, with Ethio-
aging fleet in central and western Africa would pian Airlines (Addis Ababa), Kenya Airways
undermine the economic viability of corridor (Nairobi), and South African Airways (Johan-
upgrades. Investments in these regions would nesburg) emerging as the main ones. Yet, in
likely become attractive only when more fun- central and western Africa, hubs are conspicu-
damental regulatory and institutional reforms ous by their absence (figure 6.3). The collapse
improve trucking productivity. of key regional carriers, notably Air Afrique,
For air transport, the size of the market is partly caused this gap. Particularly striking
simply too small to support a proliferation is the failure of Lagos to emerge as a hub for
of national carriers with each centered on western Africa.
152 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 6.3 Uneven Distribution of Airport Hubs across Africa: Traffic Flows between Emirates; Jeddah, Saudi Arabia; and Salalah,
Top-60 Intraregional City Pairs Oman). For southern Africa, the government
of South Africa has decided to develop a sizable
hub port in Richards Bay, which will likely cap-
ture a significant portion of the seaborne trade
between Asia and subequatorial Africa. On
the west coast, despite the growth of Tangiers,
Morocco, room may still exist for one or two
regional hubs. Dakar, with its recent container-
terminal concession and port expansion plans,
is a strong candidate. Apapa (Lagos), though
more centrally positioned, is already strug-
gling with its local market and facing heavy
congestion.
allows them to address early any negative benefit especially from regional power trade
effects of water management development on that reduces the costs of energy supply. As long
agriculture and fisheries. as regional integration provides a substan-
tial economic dividend to some of the par-
ticipating countries, designing compensation
Meeting the Challenges of mechanisms that benefit all of them should be
Regional Integration of possible. The concept of benefit sharing was
Infrastructure in Africa pioneered through international river basin
treaties, such as that for the Senegal River, and
The benefits of regional infrastructure develop- could be applied to regional infrastructure more
ment are clear, but reaping those benefits poses broadly.
political, institutional, economic, and financial A key prerequisite for any regional initiative
challenges that are far from trivial. The starting is building political consensus both nationally
point is building political consensus among and across borders. Although methods will
neighboring states that may have diverging vary from country to country, some broad
national agendas or even recent histories of principles apply.
conflict. Thereafter, effective regional institu-
tions are needed to take forward a collabora- Get High-Level Buy-In. Africa needs more
tive cross-border infrastructure development high-level advocacy and leadership to promote
program and to ensure an equitable distribu- regional integration for infrastructure develop-
tion of benefits. Given the vast needs and lim- ment and beyond. Regional integration issues
ited resources, some form of priority setting is remain only a small part of parliamentary
necessary to guide efforts on the regional inte- debate in most countries. Between infrequent
gration agenda. Even when priorities are clear, regional meetings of heads of state, a sense of
however, funding and implementing extensive inertia and lack of follow-up frequently exist.
project preparation studies and arranging Governments and international institutions
cross-border finance for complex multibillion- must provide leadership. The African Union
dollar projects is far from straightforward. has the mandate to coordinate the regional
Furthermore, once the regional infrastructure integration program of Africa as spelled out
is in place, its efficacy will ultimately depend in the Abuja Treaty (1991), which created the
on harmonizing the associated regulatory and African Economic Community, with regional
administrative procedures. economic communities as building blocks.
The African Union identified infrastructure
Building a Political Consensus and regional integration as major components
Regional infrastructure is only one aspect of of economic growth and poverty reduction in
broader regional integration. In contrast to Africa. The main vehicle is the New Partner-
economic or political integration, however, ship for Africa’s Development (NEPAD), which
cooperation in infrastructure provision is eas- has not always received the requisite support
ier to achieve, because benefits are more clearly from political leaders to build a consensus
defined, and countries need to cede less sover- around financially and economically viable
eignty. Regional infrastructure cooperation is projects. The NEPAD Heads of State Imple-
therefore an effective initial step on the path to mentation Committee, set up to help remove
broader integration. political blocks to projects, has not been effec-
Some countries have more to gain from tive and now meets less regularly than origi-
regional integration than others. Landlocked nally intended. A strong commitment from
countries depend particularly on effective road regional leaders is essential to move projects
and rail corridors to the sea, as well as on intra- forward. When the West African Gas Pipeline
continental fiber-optic backbones that link ran into political differences, for example, it
them to submarine cables. Coastal countries was the shuttle diplomacy of Nigeria’s Presi-
depend particularly on sound management dent Olusegun Obasanjo that kept the project
of water resources upstream. Small countries on track.
Deepening Regional Integration 155
Build Trust. Trust is important for regional agreement, contributing to broader political
integration—especially when some countries and economic cooperation and stability.
stand to benefit more than others. Coun-
tries that do not trust each other may fail to Think Regionally, Even When Developing
reach a cooperative solution. For example, National Policies. Regional interdependence
implementation of the road-railway bridge is a fact of life in all parts of Africa. It is criti-
project between Kinshasa and Brazzaville cal for not only landlocked countries but also
and extension of the Kinshasa-Ilebo rail- larger and coastal countries, which deal with
way are intended to accelerate trade between regional trade, labor migration, and expand-
the Democratic Republic of Congo and the ing markets. National policy makers should
Republic of Congo. Trust between the two therefore consider the regional consequences of
countries will be a key factor in the decision national policy making. Donors can encourage
to proceed with this project. Starting small, this approach. For instance, developing an assis-
with relatively well-defined projects, is one tance strategy for Burkina Faso without consid-
way to build that trust. Frequent interaction ering its place in the region in relation to Côte
among policy makers at all levels of govern- d’Ivoire, Ghana, and Mali makes little sense.
ment builds relationships that help overcome
inevitable disagreements. Supranational orga- Establishing Effective Regional
nizations can be honest brokers for sharing Institutions
gains and resolving disputes. No shortage exists of regional institutions in
Africa, but few are effective. The institutional
Invest in Credible Information. Trust is easier architecture that supports African integration
to build when the facts are available to all. Good comprises more than 30 executive continental
evidence must be gathered and made acces- bodies, regional economic communities with
sible to decision makers to gauge the full costs many overlapping memberships, sectoral tech-
and benefits of regional infrastructure invest- nical bodies, and national planning bodies. The
ments, many of which involve large allocations result is a high degree of complexity, unclear
of funds and ceding of some sovereignty. The functional responsibilities for strategy and
regional economic communities must commu- project development, and uncertain financing
nicate the potential benefits to all stakeholders strategies. This lack of clarity has slowed prog-
to help build consensus. Countries are unlikely ress on coherent regional strategies, realistic
to bear the full cost of public goods if the ben- programs for integration priorities (such as
efits are not clear. Because regional integration regional infrastructure and trade integration),
can create winners and losers, assessing the and technical plans for specific projects.
likely benefits and costs will help. Becoming more effective is easier for agen-
cies with a narrow set of tasks and responsibili-
Focus on Sharing Benefits, Not on Sharing ties than for those with a broader design. The
Resources. Regional projects often fail because African Union Commission has struggled with
of the perception of unequal access to a natu- its mandate because of a lack of human and
ral or infrastructure resource. However, what financial resources and unclear responsibili-
matters is how the economic benefits from ties. The regional economic communities have
resources or infrastructure are shared. This limited capabilities and resources and, above
philosophy is best illustrated in the manage- all, weak authority to enforce decisions. A dis-
ment of transboundary water resources, where connect often exists between what is written in
benefits include flood protection, hydropower, treaties and what happens on the ground. Insti-
irrigation, fisheries, leisure, tourism, and peace tutions will be more effective if a greater willing-
and security. One country may benefit most ness exists to cede some sovereignty in return
from hydropower, while another requires steady for greater economic benefits. Greater use of
access to water for irrigation. Successful benefit qualified majority rules in some areas of policy
sharing includes the Lesotho Highlands Water making would lead to more streamlined deci-
Project and the Incomati Basin water-sharing sion making. This approach has been debated
156 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
for some time in many regional economic com- implementation by high-level government
munities, without resolution. Adequate financ- officials.
ing is also a problem because member states Five actions to improve the effectiveness of
often fail to pay assessed contributions in full, institutions can aid regional cooperation in
if at all. Regional economic communities have infrastructure provision. First, the roles and
multiple functions, with infrastructure not nec- responsibilities of regional bodies concerned
essarily prominent (ICA 2008). As a result, they with regional integration must be clarified.
often fail to attract and retain professional staff Second, increased legal authority is required
with the experience to identify and promote for regional entities to improve and acceler-
complex regional infrastructure projects. ate decision-making processes. Third, the key
In 2006, the regional poverty reduction regional bodies must boost their professional
strategy for West Africa by the Economic capacity. Fourth, national planning agencies
Community of West African States and the must improve their ability to strengthen links
West African Economic and Monetary Union between regional strategies and national devel-
was a significant milestone. Other regions also opment plans. Fifth, delivery mechanisms for
have completed strategic planning exercises: priority programs (for example, regional infra-
for example, the Regional Indicative Strate- structure) should be strengthened to underpin
gic Plan (developed by the Southern African confidence in integration by delivering tangi-
Development Community) and the East Afri- ble results.
can Community Master Plan. However, links Africa’s efforts to strengthen regional
between these regional strategic plans and integration have focused on the fifth action.
country programs remain limited. Improved However, national priorities have limited the
links are critical for coordinated implementa- support for regional programs overall. Poor
tion of regional programs, which is essential reflection of regional priorities in national
to leverage outcomes at the country level. For plans has slowed priority programs, sapping
example, the Common Market for Eastern and government willingness to cede sovereignty to
Southern Africa, the East African Community, other regional initiatives and creating a vicious
and the Southern African Development Com- circle. Progress requires rebalancing efforts
munity have been successfully coordinating among the five institutional challenges.
programs through a tripartite task force.
Regional special-purpose entities or sectoral Setting Priorities for Regional
technical bodies have been more effective. A Infrastructure
power pool, for example, has a clear mandate, With a very large backlog of infrastructure
sufficient autonomy to execute its responsi- investments as well as limited fiscal space and
bilities, a dedicated funding mechanism, and borrowing ability, countries must set effective
career opportunities that attract and retain priorities for infrastructure investment in
high-caliber staff. It also receives substantial Africa. Projects should be well justified to com-
capacity building. The members of a power pete with investments in other sectors such as
pool are national electricity utilities, which health or education. The long life of infrastruc-
similarly have clear functions and roles within ture means that bad decisions are locked in for
their national contexts and are less subject to decades. An unwise investment can saddle gov-
immediate political pressures than less techni- ernments with an ineffective project that will
cal public agencies. also require costly maintenance. How should
The need to increase capacity and streamline priorities be set? Suitable criteria include pre-
decision making extends to national agencies. dicted economic returns, spatial targeting, and
For complex regional infrastructure projects, scope for private participation.
several line ministries are often involved in each
country. This practice complicates consensus Economic Returns. Projects with the highest
building and assigning of clear responsibilities. returns may not always be new infrastruc-
A further problem is the frequent lack of ture. Strategic investments that improve the
follow-up on regional commitments to national performance of infrastructure systems, such
Deepening Regional Integration 157
Africa-EU Energy Partnership, which aims to operators relying on private financing, a firm
agree on an Electricity Master Plan for Africa, planning horizon is even more critical than for
and the 2009 African Union Summit, which the public sector (ICA 2009).
agreed on a short-list of flagship projects for Support for regional projects by the Infra-
priority support and investment. In addi- structure Consortium for Africa grew sharply
tion, many regional economic communities from about $430 million in 2005 to $2.8 billion
and other technical regional institutions have in 2007 (ICA 2007).3 Although the bilateral share
10-year investment plans that present a large has increased over time, multilateral members
investment menu for external financiers. accounted for 60 percent of total commitments
A requirement for any priority setting is in 2007 (World Bank 2008). Multilateral institu-
transparency in decision making and agree- tions have been developing specific mechanisms
ment on selection criteria. Decision making for dealing with regional projects.
must be well documented and motivated, The World Bank has four criteria for
using sufficiently detailed data and a clear regional projects to qualify for concessional
explanation of assumptions, all publicly acces- funding from the International Development
sible. The investments in better information at Association (IDA): (a) at least three countries
the country and regional levels will be small must participate, though they can enter the
relative to the public and private funds at stake, project at different stages; (b) countries and
but the benefits of better decision making will the relevant regional entity must demonstrate
be large. a strong commitment; (c) economic and social
benefits must spill over country boundaries;
Facilitating Project Preparation and and (d) projects must provide a platform for
Cross-Border Finance policy harmonization among countries and be
Project design is complex. The appraisal phase priorities within a well-developed and broadly
establishes social, economic, financial, techni- supported regional strategy. A recent evalua-
cal, administrative, and environmental feasibil- tion of World Bank regional integration proj-
ity (Leigland and Roberts 2007). For regional ects concluded that regional programs have
projects, coordination among national agencies been effective, if still on a relatively small scale
with different procedures, capacity, and admin- (World Bank 2007).
istrative constraints adds to the complexity. The African Development Bank adopted
Thus, the project preparation costs for regional similar principles in 2008, though it requires
projects tend to be higher, and the process can that only two countries participate. To help
take longer than that for national projects. with country ownership, both institutions
Preparation costs are typically about 5 per- use a one-third, two-thirds principle, whereby
cent of total financing, or approximately double participants are expected to use one IDA or
the cost of preparing national projects. These African Development Fund (ADF) credit from
outlays occur up front when the success of the their country allocation, supplemented by two
project and the likelihood of a sufficient return credits from regionally dedicated resources.
from the investment are still uncertain. Regional Currently, 17.5 percent of ADF and 15 percent
institutions and donors have tried to address of IDA resources in Africa are dedicated to
these issues, setting up more than 20 project regional programs.
preparation facilities, many of which explicitly For the EU-Africa Infrastructure Trust
support regional activities (ICA 2006). How- Fund, eligible projects must have African
ever, the resources do not match the regional ownership and long-term project sustainabil-
needs. African countries need to commit more ity. They must also be cross-border projects or
funds and more people with the proper techni- national projects with a regional effect on two
cal, legal, and financial skills for infrastructure or more countries. Regional projects funded
planning and project implementation. Timely by the Development Bank of Southern Africa
execution of project preparation activities and a must either involve a minimum of two coun-
sufficient pipeline of new projects also encour- tries or be located in a single country with ben-
age participation of the private sector. For efits to the region.
Deepening Regional Integration 159
Some challenges remain. Although recipi- In the energy sector, African borders limit
ents of funds from the ADF and the IDA can market size through political and regula-
leverage their country allocations by partici- tory barriers to electricity trade and through
pating in regional projects, those receiving a physical barriers. Regional power infrastruc-
small allocation may be reluctant to use a large ture requires harmonized power pricing and
percentage on one regional project for which third-party access regulations, effective cross-
the benefits are unclear. How such concessional border trading contracts, and reliable and
resources are allocated and whether enough of creditworthy national utilities. In much of
the overall envelope is dedicated to regional Sub-Saharan Africa, bilateral arrangements
projects remain issues of debate. In addition, between vertically integrated utilities guide
limited financing instruments exist for middle- cross-border electricity exchanges, although,
income countries, which is an issue in North increasingly, regional power pools are liberal-
Africa (for connectivity with countries south of izing the electricity markets.
the Sahara) and in southern Africa (for projects Worldwide experience in developing power
that might involve Botswana or South Africa). pools has led to a consensus on three key
Regional organizations may not always building blocks for success: a common legal
qualify for grants or concessional finance from and regulatory framework, a durable frame-
donor institutions because of their suprana- work for systems planning and operation, and
tional character, limiting the availability of an equitable commercial framework for energy
resources for capacity building. Furthermore, exchanges (USAID 2008).
some projects with significant regional spill- The four power pools in Sub-Saharan Africa
overs may not involve three or more countries are at different stages of development, but as
and thus not qualify for regional financing, countries move from bilateral to multilateral
such as the Ethiopia-Sudan interconnector or power exchanges, a commercially acceptable
any national power-generation project that framework is essential. In 2006, the West African
may have export potential.
Power Pool was granted special status to rein- Collier, Paul, and Anthony Venables. 2007.
force its autonomy, and the 2007 ratification “Rethinking Trade Preferences: How Africa Can
Diversify Its Exports.” World Economy 30 (8):
of an overarching legal framework (ECOWAS
1326–45.
[Economic Community of West African States]
ICA (Infrastructure Consortium for Africa). 2006.
Energy Protocol) will help promote security for
Infrastructure Project Preparation Facilities in
investors and enshrine the principle of “open Africa: User Guide for Africa. Tunis: ICA. http://
access” to national transmission grids across the www.icafrica.org/en/documentation.
region. In 2008, the ECOWAS Regional Elec- ———. 2007. Annual Report 2007. Tunis: ICA.
tricity Regulatory Authority was established
———. 2008. “Mapping of Donor and Govern-
to regulate cross-border electricity exchanges ment Capacity-Building Support to African
between member states. RECs and Other Regional Bodies.” Report of
Economic Consulting Associates to the Infra-
Notes structure Consortium for Africa, Tunis.
The authors of this chapter are Souleymane ———. 2009. Attracting Investors to African Public-
Coulibaly, Andrew Roberts, Vivien Foster, and Private Partnerships: A Project Preparation Guide.
Uwe Deichmann, who drew on background Washington, DC: World Bank. http://www
material and contributions from Alvaro Feder- .icafrica.org/fileadmin/documents/guides/
Attracting-investors-to-African-PPP.pdf.
ico Barra, Pinki Chauduri, Siobhan Murray, and
Alex Rugamba. Leamer, Edward E., and Michael Storper. 2001.
1. Both figures exclude South Africa. “The Economic Geography of the Internet Age.”
2. See http://www.afrol.com/News/maw008 Journal of International Business Studies 32 (4):
641–65.
_growth_triangle.htm.
3. Infrastructure Consortium of Africa members Leigland, James, and Andrew Roberts. 2007. “The
are the Group of Eight countries, World Bank African Project Preparation Gap: Africans
Group, African Development Bank, European Address a Critical Limiting Factor in Infrastruc-
ture Investment.” Gridlines, Note 18 (March),
Community, European Investment Bank, and
Public-Private Infrastructure Advisory Facility,
Development Bank of Southern Africa. World Bank, Washington, DC.
Limão, Nuno, and Anthony Venables. 1999. “Infra-
References structure, Geographical Disadvantage, and
AfDB (African Development Bank) and UNECA Transportation Costs.” Policy Research Working
(United Nations Economic Commission for Paper 2257, World Bank, Washington, DC.
Africa). 2000. Review of the Implementation
Mayer, Rebecca, Ken Figueredo, Mike Jensen, Tim
Status of the Trans-African Highways and the
Kelly, Richard Green, and Alvaro Federico Barra.
Missing Links. Volume 2. Tunis and Addis Ababa:
2008. “Connecting the Continent: Costing the
African Development Bank and United Nations
Needs for Investment in ICT Infrastructure in
Economic Commission for Africa.
Africa.” Background Paper 3, Africa Infrastruc-
Arvis, Jean François. 2005. “Transit and the Special ture Country Diagnostic, World Bank,
Case of Land-Locked Countries.” In Customs Washington, DC.
Modernization Handbook, ed. Luc De Wulf and
Minges, Michael, Cecila Briceño-Garmendia, Mark
José B. Sokol, 243–64. Washington, DC: World
Williams, Mavis Ampah, Daniel Camos, and
Bank.
Maria Shkaratan. 2008. “Information and Com-
Bofinger, Heinrich C. 2009. “Air Transport Sector munications Technology in Sub-Saharan Africa:
Review.” Background Paper 16, Africa Infra- A Sector Review.” Background Paper 10, Africa
structure Country Diagnostic, World Bank, Infrastructure Sector Diagnostic, World Bank,
Washington, DC. Washington, DC.
Buys, Piet, Uwe Deichmann, and David Wheeler. Rosnes, Orvika, and Haakon Vennemo. 2008.
2006. “Road Network Upgrading and Overland “Powering Up: Costing Power Infrastructure
Trade Expansion in Sub-Saharan Africa.” Policy Investment Needs in Sub-Saharan Africa.” Back-
Research Working Paper 4097, World Bank, ground Paper 5, Africa Infrastructure Country
Washington, DC. Diagnostic, World Bank, Washington, DC.
Collier, Paul. 2007. “Poverty Reduction in Africa.” Teravaninthorn, Supee, and Gaël Raballand. 2008.
Proceedings of the National Academy of Sciences Transport Prices and Costs in Africa: A Review of
104 (43): 16763–68. the Main International Corridors. Directions in
Deepening Regional Integration 161
Development Series. Washington, DC: World Development Framework.” White Paper, USAID,
Bank. Washington, DC.
UNESCO (United Nations Educational, Scientific World Bank. 2007. The Development Potential of
and Cultural Organization). 2003. Conflict Pre- Regional Programs: An Evaluation of World Bank
vention and Cooperation in International Water Support of Multicountry Operations. Washington,
Resources. Paris: UNESCO. DC: World Bank, Independent Evaluation Group.
———. 2006. The 2nd United Nations World Water ———. 2008. “Africa Regional Integration Assis-
Report: Water, a Shared Responsibility. Paris: tance Strategy.” World Bank, Washington, DC.
UNESCO. ———. 2009. World Development Report 2009:
USAID (U.S. Agency for International Develop- Reshaping Economic Geography. Washington,
ment). 2008. “Sub-Saharan Africa’s Power Pools: DC: World Bank.
Par t 2
Sectoral Snapshots
Chapter 7
I
nformation and communication technolo- to address the specific challenges facing the
gies (ICTs) have been a remarkable success ICT sector in Africa:
in Africa. Sector reform, particularly in
the mobile segment of the market, has trans- • Complete the reform agenda by estab-
formed the availability, quality, and cost of lishing full competition throughout the
connectivity across the continent. In less than sector.
10 years, mobile networks have covered 91 • Revise the licensing framework to accom-
percent of the urban population, and cover- modate rapid technological change and
age in rural areas is growing. However, these emerging competition.
high overall levels of coverage hide significant
variation between countries, and particularly
• Reform the state-owned enterprises
(SOEs) that hinder sector growth and
in the proportion of their populations that development.
have access to services. Some countries have
been much more successful in providing basic • Ensure low-cost international access infra-
structure by preventing monopoly control
voice services than others, and some segments
over bottleneck facilities.
of the market, such as fixed-line telephone
service and broadband Internet, have been less • Promote the development of high-bandwidth
successful than the mobile segment. Fixed-line backbone infrastructure (the networks that
penetration rates remain low and are falling carry communications traffic between fixed
in most countries, while broadband Internet points in a network).
is expensive and available to only a small pro- • Stimulate innovation in the use of wireless
portion of the population. technologies by reforming the way the radio
Although large parts of the ICT sector have spectrum is allocated and managed.
been transformed, much remains to be done. • Promote universal access to ensure that ICT
Policy makers need to take the following steps availability is as extensive as possible.
165
166 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
The African ICT Revolution Figure 7.2 Global System for Mobile Communications
Coverage in Africa, 1998 to Third Quarter of 2006
% of population
ICT sector has had a strong, positive effect on 60 62
Africa’s GDP.
40 43
43
This rapid growth in access to ICT in Africa Figure 7.3 Price of One-Minute, Peak-Rate Call to the
has happened despite the relatively high price United States, 2006
of services. In 2007, a representative basket of Chad 4.9
prepaid mobile services cost $12.58 a month, Cape Verde 2.1
six times the $2 that it cost in Bangladesh, India, Zambia 1.8
and Pakistan. Prices in Africa are declining but Ethiopia 1.1
not as fast as in other world regions. In 2000, Malawi 1.0
each mobile subscriber paid about $39 a month Namibia 0.9
Madagascar 0.8
in African countries and in Bangladesh, India,
Tanzania 0.7
and Pakistan. By 2005, that figure had fallen to Niger 0.7
$7 in Bangladesh, India, and Pakistan but only Lesotho 0.6
to $20 in Africa. If prices were to fall to the lev- Rwanda 0.5
els seen in South Asia, access to ICT in Africa Uganda 0.5
could be significantly higher. Sudan 0.4
The average price of international calls in Congo, Dem. Rep. of 0.4
Mozambique 0.4
Sub-Saharan Africa has fallen significantly
Burkina Faso 0.4
since 2000, but prices for calls to countries Benin 0.4
outside the region remain much lower than Senegal 0.3
for calls within the region. The average peak Nigeria 0.3
price of a one-minute call from Africa to the Côte d’Ivoire 0.3
United States is $0.45, compared with $1.23 for Kenya 0.2
an international call within Africa. These aver- Cameroon 0.2
South Africa 0.2
ages mask significant variation among coun-
Ghana 0.2
tries (figure 7.3). Price variation is much lower South Asia 0.3
for calls within Africa. 0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0
US$ per minute
Access to the Internet
Source: Minges and others 2008.
Unlike the expanded access to basic voice ser- Note: Peak rate includes taxes.
vices, rates of access to the Internet are low
and growing only slowly in Africa. High prices
and limited availability are key reasons, com- Figure 7.4 Price Basket for Internet Access, 2005
pounded by poor fixed-line access networks, 70 1.0
limited access to the broadband radio spec- 60
0.9
trum, poor domestic backbone networks, and 0.8
50
US$ per month
0.7
limited use of computers (figure 7.4).
% of GDP
40 0.6
0.5
30 0.4
0.3
ICT Sector Developments 20
0.2
10
0.1
Growth in the ICT sector in Africa has taken 0 0
South Asia high income East Asia & Pacific Sub-Saharan Africa
place primarily in mobile phones through
global systems mobile (GSM) networks. The Internet access price (US$) Internet access price (% of GDP)
standard for wireless voice services, Code- increasing investment in ICT services results
Division Multiple Access, is also making slow in higher long-run rates of economic growth
but steady progress in Africa. Operators in the (Roeller and Waverman 2001); according
region have also pioneered innovative services to estimates, the ICT revolution in Africa is
using the mobile telephone networks, such as responsible for about 1 percentage point of
mobile banking and remittance payments. the improvement in Africa’s per capita eco-
The fixed-line market in Sub-Saharan Africa nomic growth rate between the mid-1990s and
continues to be dominated by incumbents— the mid-2000s (Calderón 2008). This positive
operators that were either formerly or are effect will continue as investment in the sector
currently owned by the state—and their perfor- continues and as the use of ICT raises produc-
mance remains relatively poor (figure 7.5). The tivity in all types of businesses.
productivity of these operators is low, and most Large-scale private investment, reaching
of them have higher levels of personnel than a cumulative value of about $20 billion, has
international benchmarks, as measured by the driven the expansion of access to ICT. Between
number of lines per employee. In July 2008, the 1992 and 2005, the vast majority of the 82 pri-
Sub-Saharan operators ranged from 20 to 346 vate sector transactions in the ICT sector were
lines per employee, whereas the figure was 427 for new operations in mobile communica-
in Latin America and the Caribbean region, and tions (World Bank 2009). SOE privatizations
700 in Organisation for Economic Co-operation and license fees generated a further $3.3 bil-
and Development member countries. lion of revenues for the state. This investment
The low productivity of African incumbent continues today, as new deals in the region are
telecommunication companies creates hidden announced regularly. The current financial
costs for the economy, through suboptimal crisis has adversely affected investment rates,
allocations of resources to the sector and low however, limiting operators’ access to finance.
consumption of telecommunication services. Overall ICT employment has grown as the
The cost of this excess labor can be on the mobile sector directly and indirectly added
order of 0.4 percent of GDP, and even higher in jobs in African countries. Multiplier effects and
some cases (figure 7.6). For Cameroon, Ghana, new lines of business (mobile airtime agents
and Namibia, among other countries, the level and m-transactions) also add to employment
of this inefficiency exceeds the cost of meeting growth and income generation. In East Africa,
universal access targets. the mobile industry directly and indirectly pro-
vides employment for close to 500,000 people
Economic Impact of the ICT Industry (GSMA 2007).
The ICT sector has positively affected eco- The new ICT infrastructure and related
nomic growth in Africa. Research shows that reforms have increased government revenues
through one-time fees for licenses and ongoing
payments through licenses and taxes. The rev-
enue generated by the ICT industry in African
Figure 7.5 Net Change in Fixed-Line Market,
2001–05 countries ranges from 1.7 percent to 8.2 percent
of GDP, with an average of 4.0 percent. The tax
600 and license revenues generated by the industry
net change in no. of subscribers
500
have also had a significant positive fiscal effect
400
(figure 7.7).
(thousands)
300
200
100
0 Institutional Reforms in the
–100 ICT Sector
–200
2001 2002 2003 2004 2005
South Africa Nigeria others
Market liberalization has been the most
important cause of the ICT sector’s growth
Source: Minges and others 2008. in Africa. Regulatory reforms and the
Information and Communication Technologies: A Boost for Growth 169
privatization of SOEs have complemented Figure 7.6 Costs of Overstaffing for Fixed-Line Incumbents in Selected Countries
these market reforms. 1.0
% of GDP
0.6
Africa and the emergence of competition, par-
ticularly among mobile operators, have been 0.4
the main drivers of the ICT sector’s perfor-
0.2
mance (figure 7.8).
Countries with more competitive markets 0
pia
nia
nin
a
cover, on average, 64 percent of their popu-
u
ny
ibi
ric
an
a
oo
biq
Ch
hio
za
Be
Ke
Af
m
Gh
er
n
Na
am
m
Et
h
Ta
lation with mobile networks, compared with
Ca
ut
oz
So
M
57 percent for the less competitive markets. investment needs
overstaffing costs
Among the low-income African countries,
Source: Minges and others 2008.
those with more competitive markets have
31 percent higher mobile penetration rates,
6 percent lower mobile prices, and 39 percent
lower international call prices (as measured by
the price of a call to the United States). Figure 7.7 Value Added and Excise Taxes on Mobile
The benefits of market liberalization Telephone Services, 2006
increase as competition intensifies. In gen-
eral, the annual increase in penetration rises Uganda
as more firms enter the market. Relatively Tanzania
little growth occurs in market penetration in Kenya
the initial change from monopoly to duopoly, Niger
but when a country issues its fourth mobile Cameroon
license, penetration rates increase, on aver- Senegal
age, by almost 3 percentage points every year. Madagascar
A country’s average income also affects the Côte d’Ivoire
performance of the telecommunication sector. Chad
In poorer countries, increased competition is Burkina Faso
felt most strongly when a market reaches four Benin
operators, whereas for middle-income coun- Zambia
tries, the effect is strongest when a third opera- Malawi
tor is introduced. Mozambique
Some countries that have established a Rwanda
legal framework for a liberalized market have Namibia
nonetheless failed to establish effective com- Ghana
petition. Few countries have legislation with Ethiopia
outright prohibitions on competition in tele- Cape Verde
communications, but many have restrictions South Africa
on competition arising from exclusivity clauses Congo, Dem. Rep. of
granted in licenses to existing operators. In Sudan
12 countries where data were available, a gap of Nigeria
at least two years elapsed between ending the Lesotho
legal restrictions on competition and granting 0 5 10 15 20 25 30
new licenses. Twelve Sub-Saharan countries % surcharge on mobile service
have competition in the fixed-line and inter- value added tax excise tax
national markets, but only a few of them have Source: Minges and others 2008.
more than two operators in these segments. Note: In Kenya, Tanzania, and Uganda, the excise taxes shown are
applicable to mobile calls. Rwanda is planning to implement an
Even in the mobile segment, barely half of the excise tax on mobile airtime.
170 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 7.8 Status of Mobile Competition, 1993–2006 Figure 7.9 Tariff Rebalancing in African Countries
with a Liberalized Telecommunication Sector,
24 1993–2006
20
number of countries
250
16
0 50
1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006
no network monopoly duopoly three or more operators 0
2000 2001 2002 2003 2004 2005
fixed basket mobile basket
Source: Minges and others 2008.
three-minute peak rate call to U.S.
the fixed-line business has been complemented to benefit from lower prices and better quality
by sales to developing-country investors, par- of service.
ticularly from the Middle East and South Asia.
The performance of these privatizations and Domestic Backbone Infrastructure
partnerships has been mixed. In some cases, Backbone network infrastructure to carry
such as Uganda, the privatization of the state- communications traffic between fixed points in
owned fixed-line incumbent was part of a the networks is limited, thus constraining the
successful overall reform of the sector. In oth- development of broadband Internet. Mobile
ers, private investors have withdrawn, resulting operators do not require high-capacity back-
in the renationalization of ICT assets. Ghana bone networks to carry voice traffic and have
and Rwanda resold the businesses after the typically developed their own using wireless
first privatization failed, indicating a sustained technologies. Broadband Internet backbone
commitment to reform. networks need much greater capacity, how-
Despite the notable successes, the gov- ever, typically using fiber-optic cables. The
ernments of many African countries retain limited extent of these networks is a constraint
ownership of at least one telecommunication on the development of the broadband market
operator, which distorts the market and creates in Africa.
inefficiencies. Thus, the region has some dis- Considerable variation exists across the
tance to go before it has a fully privately owned region in how markets for domestic backbone
and competitive telecommunication market. infrastructure operate. In many countries,
both implicit and explicit constraints limit
Access to International Connectivity development of this type of infrastructure. For
One of the main drivers of the high cost of example, mobile operators may be required to
Internet and of international voice calls is use the incumbent’s backbone network, or
the price of international connectivity, deter- they might be allowed to build their own but
mined by physical access to submarine fiber- not to sell backbone network services to other
optic cables and the level of competition in the operators on a wholesale basis. These types of
international market. Countries with access to regulations limit the development of back-
submarine cables have lower international call bone networks and hinder the development of
prices than those without access. Nevertheless, broadband.
countries that have competitive access have Countries that have fully liberalized the
significantly lower prices than those retaining market for backbone networks have seen
a gateway monopoly (table 7.1). rapid growth in infrastructure competition. In
Access to high-capacity submarine fiber- Nigeria, at least four of the major operators are
optic infrastructure is therefore a necessary developing high-capacity fiber-optic cable net-
but insufficient condition for low-price inter- works capable of supporting high-bandwidth
national voice services. Countries also need to services, and a similar pattern is emerging in
ensure that the international facilities segment Kenya. These networks are concentrating on
of the market is competitive if customers are major urban areas and on interurban links
Table 7.1 Prices for Access to International Voice and Internet Connectivity
Price for a Price for a Price for 20 hours Price for ADSL
call within call to the of dial-up broadband
Share of Sub-Saharan Africa United States Internet access Internet access
Access level countries (%) ($ per minute) ($ per minute) ($ per month) ($ per month)
No access to submarine cable 67 1.34 0.86 67.95 282.97
Access to submarine cable 32 0.57 0.48 47.28 110.71
Monopoly on international gateway 16 0.70 0.72 37.36 119.88
Competitive international gateways 16 0.48 0.23 36.62 98.49
Source: Minges and others 2008.
Note: ADSL = asymmetric digital subscriber line.
172 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
where the majority of customers are. If high- Nevertheless, the size of the coverage gap var-
capacity backbone networks are to extend ies immensely across countries (figure 7.10),
beyond these areas, some form of public sup- and in a handful of cases (the Democratic
port will likely be needed, preferably in part- Republic of Congo, Madagascar, and Zambia)
nership with the private sector. can exceed 15 percent of the population.
These analytical results are robust; the size
of the coverage gap increases only from 4.4
Completing the Remaining percent to 5.9 percent of the population if the
Investment Agenda amount spent on telecommunication services
drops from 4 percent of GDP per capita (the
baseline assumption) to 3 percent. Similarly,
Voice Services
even if costs were three times greater than in the
The cost of completing mobile network cover-
base case, the coverage gap would increase from
age for voice in Africa is relatively modest. By
4.4 percent of the population to 12.6 percent.
adopting a spatial approach to modeling the
cost of providing access to mobile phone net-
works, reliable estimates have been developed of Internet Services
the capital and operating expenditures required Despite the anticipated positive economic
for completing the rollout of GSM voice signal effect that widespread use of broadband would
throughout Africa. Potential revenues are esti- have on African economies, mass-market
mated based on population density and income broadband Internet at speeds seen in other
distribution. Potential costs are estimated based parts of the world is unlikely to be commer-
on terrain characteristics and cell size and the cially viable in Africa for the near future. The
resulting number of additional base stations broadband Internet available in most African
needed to complete national GSM coverage. countries is typically limited to major urban
These raw base station numbers drive estimates areas and to Internet cafés, businesses, and
of capital and operating expenditures. high-income residential customers. Network
Reaching all the unserved population coverage is limited, prices are high, and speeds
would require investments of $0.8 billion a are lower than in other regions of the world.
year over 10 years. Currently, 43.7 percent of This limited current level of service could be
the population lives in areas not covered by expanded to national coverage using wireless
wireless voice networks. If the right competi- network infrastructure with the same techni-
tive environment is established, the private cal and economic advantages as GSM voice
sector could fill most of this gap, reaching 39 networks (lower operating and security costs
percent of the population—the vast majority than wired networks and the potential to use
of the unserved—with a voice signal. Only $0.3 prepaid billing systems). The investment to
billion per year of public investment would be cover the entire population using limited-
needed to reach the remaining 4.7 percent of performance wireless broadband technology has
the population in the coverage gap (table 7.2). been estimated at approximately $0.9 billion.
Table 7.2 Investments Needed to Close Gaps in Voice and Broadband Coverage in Sub-Saharan Africa
Voice coverage Broadband coverage
Total Efficient Coverage Total Efficient Coverage
Indicator investment market gap gap investment market gap gap
Average annual
investment ($ billions) 0.8 0.5 0.3 0.9 0.7 0.2
Percentage of
population affected 43.7 39.0 4.7 100.0 89.0 11.0
Source: Mayer and others 2008.
Note: Efficient market gap is the portion of total investment need that the private sector could meet under commercial terms if all regulatory
barriers to entry were dismantled to allow the market to function efficiently. Coverage gap is the portion of the total investment need that
the private sector could not meet even under efficient market conditions. This gap would require public subsidy because the service lacks
commercial viability.
Information and Communication Technologies: A Boost for Growth 173
Figure 7.10 Voice Coverage Gaps in 24 Sub-Saharan Figure 7.11 Broadband Coverage Gaps in 24 African
Countries Countries
Zambia Mozambique
Niger Zambia
Mozambique Niger
Chad Namibia
Namibia Chad
Lesotho Malawi
Tanzania Lesotho
Ethiopia Tanzania
Cameroon Cameroon
Cape Verde Ethiopia
Burkina Faso Cape Verde
Malawi Burkina Faso
Sudan Senegal
Kenya Kenya
Senegal Sudan
Côte d’Ivoire Côte d’Ivoire
Benin Benin
Ghana Uganda
Uganda Ghana
Rwanda Rwanda
Nigeria South Africa
South Africa Nigeria
0 20 40 60 80 100 0 20 40 60 80 100
% of population % of population
existing coverage efficient market gap efficient market gap coverage gap
coverage gap
Source: Mayer and others 2008.
Source: Mayer and others 2008. Note: Efficient market gap is the portion of the unserved market
Note: Efficient market gap is the portion of the unserved market that the private sector could serve under commercial terms if all
that the private sector could serve under commercial terms if all regulatory barriers to entry were dismantled to allow the market
regulatory barriers to entry were dismantled to allow the market to function efficiently. Coverage gap is the portion of the unserved
to function efficiently. Coverage gap is the portion of the unserved market that the private sector could not serve even under efficient
market that the private sector could not serve even under efficient market conditions. This gap would require public subsidy because
market conditions. This gap would require public subsidy because the service lacks commercial viability.
the service lacks commercial viability.
analysis in telecommunication regulation. They Internet Protocol (VoIP), limited mobility, and
are also adapting their regulatory approach to Internet protocol television. Many licensing
reflect the evolving marketplace: for example, regimes restrict either the VoIP technology or
by relaxing controls on tariffs as competition its derived services. Direct consumer access to
becomes more effective at controlling them. VoIP allows voice calls over Internet connections
African countries will continue to see ben- instead of the public switched telephone net-
efits as competition intensifies and access to works. Such services offer much lower prices for
ICT increases. As prices fall, even as far as the long-distance and international calls; however,
$0.01–$0.03 per minute range currently seen in restrictions are common because the widespread
South Asia, mobile phone services will become use of VoIP could undermine the main sources
affordable to much of the African population, of voice revenues for incumbent operators.
bringing with it positive economic and social Licensing constraints on the mobility of
benefits. Governments will also benefit from specific wireless telecommunication operators
the expansion of telephone services. First, lower are common in Africa. Operators with limited-
prices will fuel uptake and access to services, mobility licenses can provide wireless telecom-
directly reducing the costs involved in deliver- munication services while allowing customers
ing universal service. Second, greater competi- to move around within a limited area. No
tion will expose the hidden costs of the incum- technological reason exists why these networks
bent state-owned operators, which represent cannot offer full mobility; the restrictions
a burden on government finances and a more are often imposed to protect existing mobile
general effect on the economy. The expansion licensees. As competition in the full-mobility
of the ICT sector resulting from market liber- market increases, these restrictions will seem
alization will increase the tax and license-fee increasingly anachronistic.
revenues earned by governments, and ICT ser- Finally, the use of Internet networks to
vices themselves will become a more effective provide television services is increasing in
platform for delivering public services. Africa as the number of broadband subscrib-
ers increases. This raises many challenges for
Revising the Licensing Framework regulatory systems that have traditionally dealt
The traditional model of licensing is becoming with communication and broadcasting media
obsolete. In the first wave of market liberaliza- through separate institutions and through sep-
tion, licenses were linked to market segments arate legal and regulatory frameworks. These
and technologies. GSM licenses granted the separations are creating obstacles to invest-
right to provide mobile communications ment and competition as convergence blurs
in specific spectrum bands using a specific the boundaries between the technologies.
technology, and data licenses were granted to The initial response of policy makers to
operate in specific value added markets. Two these trends has been to move toward unified
factors are making this traditional approach to licenses that remove technological distinctions
licensing obsolete. First, the growth of compet- and allow operators to provide a full range of
itive ICT markets in Africa has demonstrated services to customers. The design and imple-
that multiple players can compete successfully, mentation of a unified licensing system can be
even in small markets. Managing liberaliza- complex, however, magnified by the need to
tion through technology and service-specific adjust a wide range of existing rights and obli-
licenses has therefore proved to be ineffective gations, annual fees, and acquisition costs. This
as a policy tool. Second, technological conver- adjustment can be done in a transparent way
gence allows networks to deliver multiple ICT through public consultation, but the migration
services, thereby reducing costs and promoting process has to be managed carefully to avoid
service innovation. The traditional approach undue destabilization of the market.
to licensing often prevents operators from tak- In the medium term, licenses will have to
ing advantage of this convergence. become simpler and less restrictive to facili-
The negative effect of current licensing tate the development of new services at lower
frameworks is especially evident in Voice over prices. Ultimately, the challenge for African
176 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
countries will be to migrate from the current Providing SOEs with monopoly control over
licensing regime to one in which controls on specific segments of the market to make them
market entry and the services delivered by more attractive to potential buyers will ulti-
market parties would be largely abandoned. mately be unsuccessful, for it will distort the
The countries of the European Union have market and constrain its development.
taken this approach, moving from a system
of individual licenses to a general authoriza- Ensuring Low-Cost Access to
tion regime. International Infrastructure
Creating the conditions for widespread broad-
Reforming State-Owned Enterprises band access is a complex policy issue facing
Reform of state-owned, fixed-line incumbents the ICT sector in Africa. The markets in the
remains a major policy challenge for govern- region are so different from those in other
ments in the region. The last decade has seen parts of the world that governments have no
the fixed-line incumbent operators eclipsed. obvious models to draw on. Some lessons are
Compared with the mobile operators, they beginning to emerge, however. One, in par-
now play a minor role in telecommunication ticular, is the importance of access to high-
service in most African countries. Incum- capacity, low-cost bandwidth via submarine
bent operators can be a disruptive force in fiber-optic cable infrastructure.
the economy through misallocation of public The private sector has demonstrated its
resources, use of incumbents as social buf- capacity to develop, finance, and operate such
fers, and the regulatory uncertainty created by cables in Africa. The challenge for govern-
their presence in the market. In some cases, ments is to minimize the obstacles to this type
preferential treatment of these operators— of investment by readily issuing cable opera-
exclusivity agreements (for example, in con- tors permits and licenses. The development of
trol of international gateways and backbone infrastructure, on its own, will not guarantee
capacity), banning of innovative services such better services for customers. The experience
as VoIP, and distortion of prices—inhibits of the South Atlantic 3/West Africa Subma-
innovation and investment and amplifies the rine Cable (SAT-3) cable on the west coast
economic burden of SOEs on national econo- of Africa shows that physical access to a cable
mies. This issue has emerged again as some is necessary but not sufficient for low-cost
Sub-Saharan governments finance the devel- connectivity. A consortium of private opera-
opment of fiber-optic backbone networks tors with little direct regulation controls access
through their SOEs. to the SAT-3 cable. Because these operators
At a minimum, SOEs should be brought are protected from competition on the cable,
fully within the regulatory and licensing frame- customers have not received the full potential
work so that they are treated in the same way benefit of the facility. The challenge for gov-
as private operators. This move will stimulate ernments seeking to improve access to inter-
competition and efficiency in resource alloca- national infrastructure is to avoid creating
tion. Encouraging greater private participation infrastructure bottlenecks and to encourage
in SOEs to transform and grow the businesses competition between submarine cables and
may also be appropriate. Given the state of landing stations. Where they cannot do this,
many incumbent operators’ networks, that may regulators should ensure access to the facilities
require some form of financial and manage- on equitable terms.
ment incentives to attract partners and inves- Landlocked countries face a special chal-
tors. The challenge for governments will be to lenge in ensuring that their operators have
ensure that this transition is achieved without access to submarine fiber-optic infrastruc-
distorting the market. It can be done by allocat- ture. If the private sector does not provide
ing mobile and other wireless spectrum to these competitive infrastructure in the intervening
operators, offering management control, and countries, the government may have a role to
minimizing network coverage commitments. play through public-private partnerships.
Information and Communication Technologies: A Boost for Growth 177
Recent developments in wireless technology these countries, a service target of 100 per-
have allowed multiple users of the same radio cent coverage may be economically feasible.
spectrum bands to operate without undue In countries where the gap left by the market
interference. Allowing anyone to share the is larger, a more modest target will likely be
radio spectrum, with little or no registration necessary.
and usually without a fee, reduces the cost of When a universal service target is set, the
entry into the market and therefore encour- major challenge is to establish a mechanism
ages innovation in technology and service to achieve it. The majority of countries across
delivery. the region currently apply a universal service
Changing how the radio spectrum is man- levy on private operators, using the funds for
aged requires political will. The establishment specific ICT projects. This approach has had
of a property rights scheme can arbitrarily cre- very limited success, particularly when con-
ate windfall gains and losses for current and trasted with the commercially driven network
future users. Some users of the radio spectrum expansion into rural areas. Universal service
may be difficult to incorporate into a pure funds often suffer from bureaucratic obstacles
market-based system. For example, requiring and political interference in expenditure, and
users in the military or emergency services to frequently they are not spent on the sector at
participate in spectrum markets may be par- all. Universal service policy in Africa there-
ticularly challenging (although not impos- fore requires new thinking. The challenge is
sible) and would certainly have budgetary to meet the government’s policy objectives of
implications for those agencies. universal service at minimum cost to taxpay-
Reforming the allocation and management ers while harnessing the beneficial effects of
of the radio spectrum would change the role competition. An alternative to the traditional
of government. Its primary role in spectrum fund-based approach is to provide direct
management would no longer be to make incentives for operators to deliver services in
technical and licensing decisions. Rather, the rural areas. For example, governments could
government’s role would be to design, oper- offer operators a reduction in license-fee pay-
ate, and regulate the market in the radio spec- ments in exchange for providing services in
trum. Such a change would require changes specified areas, or they could establish pay-or-
in the legal framework governing the radio play schemes in which operators can choose
spectrum and the capacity of the regulatory between building networks in specified areas
institutions involved. and contributing to a universal service fund,
which is then used to subsidize operators that
Promoting Universal Access do provide services in unprofitable areas. The
As more people in Africa gain access to ICT major advantage of these approaches is the
services, those who remain outside the range reduction in transfers between operators and
of networks are at a disadvantage. Several gov- the government, thereby lessening bureau-
ernments in Africa have attempted to extend cratic delays or the diversion of funds.
access to ICT beyond the perceived limits of the Once a coverage target is defined, gov-
market. The quickest and most effective way ernments may also wish to address the
of getting infrastructure to poor rural users issue of access for low-income groups.
is through competition. Malawi and Uganda Call-by-call resale of services has signifi-
have set up effectively competitive mobile mar- cantly reduced the costs of accessing the
kets that already cover over 80 percent of the network, and these systems are widespread
population and are continuing to expand. in Africa (for example, the VillagePhone
For the majority of countries in Africa, only program of cell phone company MTN). Uni-
competition will result in mobile networks versal service targets could potentially include
that cover the whole population. With a few subsidies for prepayment directed at specific
exceptions, such as the Democratic Republic target groups of the population. However, they
of Congo, the additional cost to make voice would have to be carefully designed to avoid
network coverage universal is modest. In mistargeting and leakage.
Information and Communication Technologies: A Boost for Growth 179
Power: Catching Up
A
frica’s chronic power problems have importing countries that could potentially
escalated in recent years into a crisis meet more than half their power demand
affecting 30 countries, taking a heavy through trade.
toll on economic growth and productivity. The The operational inefficiencies of power utili-
region has inadequate generation capacity, lim- ties cost $3.3 billion a year, deterring investments
ited electrification, low power consumption, in electrification and new capacity, while under-
unreliable services, and high costs. It also faces pricing of power translates into losses of at least
a power sector financing gap of approximately $2.2 billion a year. Full cost-recovery tariffs would
$23 billion a year. It spends only about one- already be affordable in countries with efficient
quarter of what it needs to spend on power, large-scale hydropower- or coal-based systems,
much of which is on operating expenditures to but not in those relying on small-scale oil-based
run the continent’s high-cost power systems, plants. If regional power trade comes into play,
thus leaving little for the huge investments generation costs will fall, and full cost-recovery
needed to provide a long-term solution. tariffs could be affordable in much of Africa.
Further development of the regional power The key policy challenges are to strengthen
trade would allow Africa to harness larger- sector planning capabilities, too often over-
scale, more cost-effective energy sources, looked in today’s hybrid markets. A serious
thereby reducing energy system costs by recommitment to reforming state-owned
$2 billion a year and saving 70 million tons of enterprises (SOEs) should emphasize improve-
carbon emissions annually. Economic returns ments in corporate governance more than
to investments in cross-border transmission purely technical fixes. Improving cost recov-
are particularly high, but reaping the prom- ery is essential for sustaining investments in
ise of regional trade depends on a handful of electrification and regional power generation
major exporting countries’ raising the large projects. Closing the huge financing gap will
volumes of finance needed to develop gen- require improving the creditworthiness of
eration capacity for export. It would also utilities and sustaining the recent upswing in
require political will in a large number of external finance to the sector.
181
182 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Africa’s Chronic Power Problems scale of most national power systems and the
widespread reliance on expensive oil-based
Africa’s generation capacity, stagnant since the generation make the average total historic
1980s, is woefully inadequate today. The entire cost of producing power in Africa exception-
installed generation capacity of the 48 Sub- ally high: $0.18 per kilowatt-hour with an
Saharan countries is 68 gigawatts, no more than average effective tariff of $0.14 per kilowatt-
Spain’s, and without South Africa, the total falls hour.2 Compare that with tariffs of $0.04 per
to 28 gigawatts (EIA 2006). As much as one- kilowatt-hour in South Asia and $0.07 in East
quarter of that capacity is unavailable because Asia. Rising oil prices, lower availability of
of aging plants and poor maintenance. hydropower, and greater reliance on emer-
The growth in generation capacity has been gency leases have put further upward pressure
barely half that in other developing regions. In on costs and prices.
1980, Sub-Saharan Africa was at approximately Power consumption is tiny and falling.
the same level as South Asia in generation capac- Given limited power generation and low
ity per million people, but it has since fallen access, per capita electricity consumption in
far behind. Sub-Saharan African countries lag Sub-Saharan Africa (excluding South Africa)
even compared with others in the same income averages only 124 kilowatt-hours a year,
bracket (Yepes, Pierce, and Foster 2008). barely 1 percent of the consumption typical
Only about one-fifth of the Sub-Saharan in high-income countries. Even if that power
population has access to electricity, compared were entirely allocated to household light-
with about one-half in South Asia and more ing, it would hardly be enough to power
than four-fifths in Latin America. Since 1990, one lightbulb per person for six hours a day.
East Asia, Latin America, and the Middle East Sub-Saharan Africa is the only region in the
have all added at least 20 percentage points to world where per capita consumption is falling
their electrification rates, but access rates in (World Bank 2005).
Sub-Saharan Africa are relatively stagnant, as Power shortages have made service even
population growth and household formation less reliable. More than 30 African countries
outstrip new connections. now experience power shortages and regular
At current trends, less than 40 percent of interruptions in service (figure 8.1). From
African countries will reach universal access to 2001 to 2005, half of the countries in Sub-
electricity by 2050 (Banerjee and others 2008). Saharan Africa achieved solid GDP growth
Overall, household access to electricity in urban rates in excess of 4.5 percent. Their demand
areas is 71 percent, compared with only 12 per- for power grew at a similar pace, yet generation
cent in rural areas. Moreover, access rates in the capacity expanded only 1.2 percent annually.
upper half of the income distribution exceed South Africa shows what happens when gen-
50 percent, whereas they are less than 20 percent eration capacity fails to keep up with demand
in the bottom half. Given that rural areas account (box 8.1). In some countries, supply shocks
for about two-thirds of the population, extend- exacerbated the situation. Causes of the sup-
ing access presents a major challenge. Only ply shocks include droughts in East Africa;
15 percent of the rural population lives within oil price inflation, which made it difficult for
10 kilometers of a substation (or within 5 kilo- many West African countries to afford diesel
meters of the medium-voltage line) and could imports; and conflicts that destroyed the power
thus be added to the electricity grid at relatively infrastructure in some fragile states.
low cost. As much as 41 percent of the rural Inadequate power supplies take a heavy
population lives in areas considered isolated or toll on the private sector. Many African enter-
remote from the grid1 and is reachable in the prises experience frequent outages: in Sen-
medium term only by off-grid technologies such egal 25 days a year, in Tanzania 63 days, and
as solar photovoltaic panels, which typically cost in Burundi 144 days. Frequent power outages
$0.50–$0.75 per kilowatt-hour (ESMAP 2007). mean big losses in forgone sales and damaged
The cost of producing power in Africa equipment—6 percent of turnover on average
is exceptionally high and rising. The small for formal enterprises, and as much as 16 percent
Power: Catching Up 183
MAURITANIA
CAPE VERDE MALI
NIGER
SE CHAD SUDAN ERITREA
NE
GA
GAMBIA L BURKINA DJIBOUTI
GUINEA FASO BENIN
GUINEA-BISSAU
NIGERIA
GHANA
D'IVOIRE
CÔTE
ON
AFRICAN REPUBLIC
RO
LIBERIA
IA
ME
AL
TOGO
DEMOCRATIC
CA
M
SO
GO
EQUATORIAL GUINEA REPUBLIC OF UGANDA
S
LE
REP
BURUNDI HEL
TANZANIA YC
SE
MALAWI
COMOROS
Main cause or trigger MAYOTTE
ANGOLA
natural causes (droughts) ZAMBIA
E
IQU
oil price shock MB
AR
ZA
ZIMBABWE
ASC
MO
DAG
high growth, low investment/structural issues BOTSWANA
MA
SWAZILAND
SOUTH
AFRICA
LESOTHO
BOX 8.1
of turnover for informal enterprises unable to the private provider. At least an estimated
to provide their own backstop generation 750 megawatts of emergency generation is
(Foster and Steinbuks 2008). Therefore, many currently operating in Sub-Saharan Africa,
enterprises invest in backup generators. In representing for some countries a large pro-
many countries, backup generators repre- portion of their national installed capacity.
sent a significant proportion of total installed Because of the preponderance of small diesel
power capacity: 50 percent in the Democratic units, the costs have typically been $0.20–$0.30
Republic of Congo, Equatorial Guinea, and per kilowatt-hour, and for some countries, the
Mauritania, and 17 percent in West Africa as price tag can be 4 percent of GDP (table 8.1).
a whole. The cost of backup generation can
easily run to $0.40 per kilowatt-hour or several
times higher than the utility’s costs of generat- Figure 8.2 Economic Cost of Outages in Selected
ing power (Foster and Steinbuks 2008). Countries
The economic costs of power outages are
substantial. The immediate economic cost of Malawi
power shortages can be gauged by looking at South Africa
the cost of running backup generators and Uganda
forgoing production during power shortages.
Tanzania
These costs typically range between 1 and 4
Kenya
percent of GDP (figure 8.2). Over time, the lack
of a reliable power supply is also a drag on eco- Senegal
in the nonfragile low-income countries, this (IPPs). In recent years, 34 IPP contracts
share increases to 2.9 percent of GDP. Of the in Africa have involved investments of
external capital flows, finance from coun- $2.4 billion for the construction of 3,000
tries not belonging to the Organisation for megawatts of new power generation capacity.
Economic Co-operation and Development Those projects have provided much-needed
(OECD) is the most significant, accounting for generation capacity. An independent assess-
$1.1 billion a year, primarily from the Export- ment concluded that they have also been
Import Bank of China. Official development relatively costly because of technology choices,
assistance follows at $0.7 billion a year and procurement problems, and currency devalu-
then private capital flows of $0.5 billion a ations (calling for adjustments in dollar- or
year (table 8.3). euro-denominated off-take agreements)
Most of the private sector finance recorded (Gratwick and Eberhard 2008).
relates to independent power producers The existing resource envelope would
go significantly further if the sector oper-
ated more efficiently. Addressing the operat-
Table 8.3 Financing Flows to the Power Sector ing inefficiencies of the power utilities could
$ billions annually reduce the funding gap by $3.3 billion a year,
Operation improving cost recovery would bring an addi-
and
maintenance Capital spending tional $2.2 billion a year, and $0.3 billion a year
could be recouped by improving execution of
Public Public Non-OECD Total
Country type sector sector ODA financiers PPI Total spending the capital budget.
Sub-Saharan Even if all these inefficiencies could be
Africa 7.00 2.40 0.70 1.10 0.50 4.60 11.60 eliminated, a sizable power sector financing
Middle-income gap of $23 billion a year would remain (table
countries 2.66 0.80 0.03 0 0.01 0.80 3.50 8.4). Three-quarters of this financing gap is
Low-income a shortfall in capital expenditure, while the
fragile countries 0.60 0 0.04 0.20 0.01 0.30 0.80 remaining quarter is a shortfall in operation
Low-income and maintenance spending. The largest por-
nonfragile
countries 2.00 0.40 0.60 0.10 0.20 1.30 3.20 tion of the gap—nearly $11 billion per year—
Resource-rich
corresponds to the middle-income countries.
countries 1.60 1.20 0.10 0.70 0.30 2.30 3.90 However, the largest financing burden relates
Source: Briceño-Garmendia, Smits, and Foster 2008. to the low-income fragile states, where the
Note: Operation and maintenance includes other current expenditures. ODA = official development financing gap amounts to roughly 7 percent of
assistance; OECD = Organisation for Economic Co-operation and Development; PPI = private
participation in infrastructure. Totals may not add exactly because of rounding errors. their GDP.
Figure 8.3 Economic Burden Associated with Power 6 percent in countries with higher efficiency
Utility Inefficiencies in Selected Countries (figure 8.4).
Institutional reform measures hold the key
South Africa
to improving utility performance. Countries
Zambia
that have advanced the institutional reform
Benin agenda for the power sector show substantially
Chad lower hidden costs than those that have not,
Kenya as do countries with more developed power
Mozambique regulatory frameworks and better governance
Ethiopia of their state-owned utilities (figure 8.5). Mea-
Lesotho
sures that seem to have a substantial effect on
reducing hidden costs are private participation
Rwanda
in the power distribution sector and (among
Madagascar
state-owned utilities) performance contracts
Burkina Faso that incorporate clear incentives. The case of
Tanzania Kenya Power and Lighting Company is par-
Senegal ticularly striking (box 8.2).
Nigeria Labor redundancy is another source of
Cameroon utility inefficiencies. Power utilities in Africa
Malawi
have overemployment of 88 percent relative
to a developing country benchmark of 413
Cape Verde
connections per employee. Overemployment
Niger
by utilities results in labor overspending
Uganda in the range of 0.07 percent to 0.6 percent
Ghana of GDP.
Congo, Dem. Rep. of The application of management contracts
0 1 2 3 4 5 has been more complex than originally sup-
% of GDP posed. More than 20 African countries have
unaccounted losses undercollection experimented with private sector participation
in power distribution, split evenly between
Source: Briceño-Garmendia, Smits, and Foster 2008. concessions and management contracts. Man-
Note: Power utility inefficiencies include undercollection of rev-
enues and unaccounted-for distribution losses. agement contracts have attracted interest
1.4 14
1.2 12
1.0 10
0.8 8
0.6 6
0.4 4
0.2 2
0 0
high efficiency low efficiency high efficiency low efficiency
because they are a simpler way of addressing Figure 8.5 Effect of Reform Measures on Hidden Costs
inefficiencies, but their application has proved
complex and contentious, and they have not high reform
always proved sustainable. Of 17 African man-
agement contracts, 4 were canceled before the high regulation
originally designated expiry date, and at least 5
more were not renewed after their initial term, high governance
reverting to state operation. Only 3 manage-
ment contracts remain in place. management contract or
concession
Problems with management contracts have
included unrealistic expectations and limited performance contracts
with incentives
ability to address broader sector challenges.
First, many management contracts were under- 0 50 100 150 200 250 300 350 400
average hidden cost of inefficiencies as a percentage of utility revenue
taken with donor involvement. Donors saw the
no yes
contracts as an initial step on the road to more
extensive sector reform that would be extended Source: Eberhard and others 2008.
BOX 8.2
long enough to allow parallel policy and Most utilities score better on internal gover-
institutional changes to be enacted and to take nance criteria, such as board structure and
root. In contrast, many African governments accountability, than on external governance
saw them as costly reform measures needed to criteria, such as outsourcing and labor and
secure donor finance and had no intention of capital market disciplines.
taking the process any further. Second, although The acute need to improve the management
management contracts can produce financial of utilities and the frameworks they operate
and efficiency gains, they cannot overcome under has long been acknowledged. Over the
broader policy and institutional weaknesses. years, substantial sums have been spent on
Moreover, the efficiency gains do not always institutional reforms: training management,
provide tangible improvements for customers, improving internal accounting and external
even though they impose substantial adjust- auditing, strengthening boards of directors,
ment costs on management, making political providing financial and operational informa-
support for these measures hard to build. tion, building reporting systems, creating and
Most African power utilities remain state reinforcing supervisory and regulatory agen-
owned and operated. On average, Africa’s cies, and much more. Some enduring suc-
state-owned power utilities embody only cesses have been registered (box 8.3; further
40 percent of good governance practices for discussion of institutional issues can be found
such enterprises (Vagliasindi and Nellis 2009). in chapter 4 of this volume).
BOX 8.3
The Challenge of Cost Recovery where large strategic customers have purchased
power at heavily discounted rates of just a few
Underpricing power costs the sector at least cents per kilowatt-hour. These arrangements
$2.2 billion a year in forgone revenues (0.9 per- were initially justified as locking in base-load
cent of GDP on average). Underpricing power demand to support very large power proj-
is widespread across Africa. In the worst cases ects that exceeded the country’s immedi-
(Malawi, Tanzania, and Zambia), underpricing ate demands, but they are now questionable
can result in utilities’ capturing less than half because competing demands have grown to
of the revenues they need and creating an eco- absorb this capacity.
nomic burden in excess of 2 percent of GDP Power prices have risen substantially in
(figure 8.6). recent years, but they have nonetheless failed to
These figures probably understate the keep pace with escalating costs. Because of ris-
underpricing because of the difficulty of cap- ing oil prices, lower availability of hydropower,
turing subsidies to large industrial and mining and greater reliance on emergency leases, the
customers, which are usually contained in costs of power production in Africa rose sub-
bilateral contracts and not reflected in the stantially in the early to mid-2000s (figure 8.7,
general tariff structure. Key examples include panel a). In response, several countries have
the aluminum-smelting sector in Cameroon increased power tariffs, so that the average
and Ghana and the mining sector in Zambia, revenue of power utilities almost doubled over
the same period (figure 8.7, panel b). Even so,
because of historic pricing shortfalls, overall
Figure 8.6 Underpricing of Power in Selected
average revenues by the end of this period had
Countries barely caught up with average operating costs
at the beginning of the period.
Uganda Most countries are achieving no more
South Africa than operating-cost recovery. The correlation
Ghana between average revenues and average oper-
Congo, Dem. Rep. of
ating costs across Sub-Saharan countries is as
high as 90 percent, indicating that operating-
Burkina Faso
cost recovery is the driving principle behind
Benin
power pricing in most cases. Cameroon,
Mozambique
Cape Verde, Chad, Malawi, Niger, Rwanda,
Kenya Senegal, and Tanzania (countries under the
Nigeria 45-degree line in figure 8.8, panel a) fail to
Cape Verde meet even operating-cost recovery, and sev-
Chad eral of them face particularly high operating
Lesotho costs (figure 8.8).
Madagascar
The longer-term cost-recovery situation is
somewhat more hopeful. Comparing existing
Senegal
average revenues and average operating costs
Ethiopia
misrepresents long-term cost recovery for two
Rwanda reasons. First, because of major inefficiencies
Cameroon in revenue collection, the average revenue col-
Niger lected per unit of electricity sold is substantially
Tanzania lower than the average effective tariff charged
Zambia today. Second, because of the major inefficien-
Malawi
cies in generation technology and the potential
for regional trade, for more than two-thirds of
0 1 2 3 4
the countries the average incremental cost of
% of GDP
power looking forward is lower than the aver-
Source: Briceño-Garmendia, Smits, and Foster 2008. age historical cost of power production looking
192 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 8.7 Electricity Costs and Revenues by Type of Power System, 2001–05
US$ per kilowatt-hour
0.25 0.25
0.20 0.20
0.15 0.15
0.10 0.10
0.05 0.05
0 0
2001 2002 2003 2004 2005 2001 2002 2003 2004 2005
0.6
average revenue
0.3
0.5
0.4 0.2
0.3
0.2 0.1
0.1
0 0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0 0.1 0.2 0.3 0.4
average operating cost average incremental cost
backward and including both historic operat- In most cases, the state or donors have almost
ing and capital costs. entirely subsidized the historic capital costs of
A truer picture of long-term cost recovery power development. Although the residential
comes from comparing today’s average effec- sector accounts for 95 percent of power utility
tive tariff with the average incremental cost customers in Africa, it contributes only around
looking forward (figure 8.8). At least in some 50 percent of sales revenue. Thus, the pricing of
countries, even the current tariff would be power to commercial and industrial consumers
adequate for cost recovery, if only all revenues is just as important for cost recovery. Neither
could be collected and the power system could commercial nor residential customers are close
move toward a more efficient production to paying full cost-recovery prices.
structure. In other countries, however, signifi- Subsidies to residential consumers are
cant tariff adjustments would still be needed in highly regressive. Across the bottom half of
the long term. the income distribution, barely 10 percent of
Power: Catching Up 193
households have access to electricity (Wodon available (table 8.5). Such modest bills would
2008). Indeed, three-quarters of the house- be affordable to all but the poorest 25 percent
holds with electricity come from the top of the population. In eastern and western
two quintiles of the income distribution. Africa, the subsistence monthly bill would
Because poorer households are almost entirely fall in the $7.00–$9.00 range. Although this
excluded, they cannot benefit from subsidies amount would likely be affordable for exist-
embedded in electricity prices. In many cases, ing customers, it would represent a problem
targeting performance is further exacerbated as power access is expanded to lower-income
by poor tariff design, with the widespread use populations. When a more efficient power sys-
of increasing block tariffs that provide large tem develops, full cost-recovery tariffs would
lifeline blocks of highly subsidized power to be affordable for the vast majority, except per-
all consumers. haps in West Africa.
With subsistence consumption of 50 If regional trade is pursued, the average costs
kilowatt-hours a month, the cost of a monthly of power production could be expected to fall
utility bill priced to recover full historic costs toward $0.07 in central and southern Africa,
of production would be as much as $24.30 in $0.12 in eastern Africa, and $0.18 in western
central Africa, which is manifestly unafford- Africa. Assuming, again, subsistence consump-
able for the vast majority of the population tion of 50 kilowatt-hours a month, a monthly
(table 8.5). Elsewhere in Africa, a subsistence utility bill under full cost-recovery pricing would
monthly bill priced at full historic cost would be about $4 a month in central and southern
range between $7.00 and $10.70 and would be Africa, $6 a month in eastern Africa, and $9 a
affordable to the relatively affluent sections month in western Africa. Based on an affordabil-
of the population that already enjoy access to ity threshold of 3 percent of household income,
power, but not to the poorer segments of the full cost-recovery tariffs would prove affordable
population that remain unconnected. Indeed, for the vast majority of the population of low-
affordability of cost-recovery power bills for income countries in central, eastern, and south-
existing customers is today really only a prob- ern Africa (see figure 8.9). In West Africa, about
lem in low-income countries reliant on small- half the population of the low-income countries
scale, oil-based generation. would face affordability problems. A number of
Looking into the future, pricing at the lower West African countries—notably Côte d’Ivoire,
long-run marginal cost of power would reduce
the subsistence monthly bill to the $3.00–$4.00
range in central and southern Africa where Figure 8.9 Affordability of Subsistence Consumption of Power at Cost-Recovery
abundant low-cost hydropower would become Pricing
than 5% of their monthly budgets
% of households spending more
1.2
1
Table 8.5 Cost and Affordability of Monthly Power
Bills at Cost-Recovery Prices: Past and Future 0.8
$ per month
0.6
Long-run
Location Historic cost marginal cost 0.4
Ghana, Nigeria, and Senegal—already have are simply too small to support any meaning-
power coverage of around 50 percent and would ful competition. The new reality is thus one of
face affordability issues as coverage broadens. At “hybrid markets,” with the state-owned utility
any of these levels, power tariffs do not repre- remaining intact and occupying a dominant
sent a significant affordability issue in the mid- market position. At the same time, because
dle-income countries. (For a fuller discussion of many governments and utilities lack suffi-
the social issues associated with utility pricing in cient investment resources, the private sector
Africa, see chapter 3 in this volume.) participates, typically as IPPs. Africa’s hybrid
electricity markets pose new challenges in
policy, regulation, planning, and procure-
ment. The widespread power shortages across
Policy Challenges
the continent and the increasing reliance on
emergency power indicate the seriousness of
The depth and extent of Africa’s power cri-
those challenges.
sis and its associated costs demand renewed
Too often, the planning function has fallen
efforts to tackle the policy and institutional
between the cracks. Traditionally, planning and
challenges needed to improve performance
procurement of new power infrastructure were
and financing. The key challenges can be char-
the province of the state-owned utility. With
acterized as follows:
the advent of power sector reforms and the
• Strengthening sector planning IPPs, those functions were often moved to the
ministry of energy or electricity. A simultane-
• Recommitting to the reform of SOEs
ous transfer of skills did not always occur, how-
• Increasing cost recovery
ever, resulting in plans that were not adequately
• Accelerating electrification informed by the complexities on the ground: a
• Expanding regional trade in power new hybrid market of private and public actors.
• Closing the financing gap. In many cases, planning has collapsed. Where
still present, planning tends to take the form of
These interdependent challenges must be outdated, rigid master plans. The lack of strate-
dealt with simultaneously. Efforts to boost gic policy and planning for the electricity sector
generation through regional power trade will at the central government level is a critical weak-
stumble if the utilities, which will continue ness. Interventions have been piecemeal rather
to be central actors, remain inefficient and than integrated; many countries have focused
insolvent. Expanding electricity distribution on generation without investing in efficient
systems without addressing the shortages in transmission and delivery of power.
generation and improving transmission capac- This situation has led to very costly delays
ity would clearly be futile. In addition, focusing in commissioning new plants. In the absence of
exclusively on utility reform would be fruitless strong political leadership, good information,
without a start on substantial, long-gestation and the requisite planning capability, incum-
investments in both generation and access to bent state-owned utilities often undermine the
improve the quality of service and make the entry of IPPs by arguing that they can supply
utilities viable. In short, these strategic priori- power more cheaply or quickly than private
ties must progress together. alternatives, even if they lack the resources to
do so. Poor understanding of the hybrid market
Strengthening Sector Planning deprives policy makers of clear and transpar-
Most African power markets present an insti- ent criteria for allocating new plants between
tutional “hybrid,” with public and private the incumbent, state-owned utility and the
actors operating in parallel. The 1990s reform IPPs. New plants are rarely ordered on a timely
prescription of unbundling and privatization, basis, thereby opening power gaps that prompt
leading to wholesale and retail competition, recourse to temporary power and discour-
did not prove very relevant to Africa, not least age investors. When procurement is (finally)
because most of the region’s power systems undertaken, the authorities may not take the
Power: Catching Up 195
trouble to conduct international competitive focused too heavily on technical issues to the
bidding. This outcome is unfortunate, because exclusion of governance and accountability.
a rigorous bidding process lends credibility Future SOE reforms seem justified as long
and transparency to procurement and results as they focus on these deeper institutional
in more competitively priced power. issues.
Restoring and strengthening planning The starting point for SOE reform should
capabilities are imperative. Hybrid power be to reform corporate governance. Key mea-
markets will not disappear from the African sures include greater decision-making auton-
landscape any time soon. To make the best of omy for the board of directors, more objective
them, African governments and their develop- selection criteria for senior managers, and rig-
ment partners must strive to develop a robust orous disclosure of conflicts of interest, as well
institutional foundation for the single-buyer as more transparent and merit-based recruit-
model, with clear criteria for power purchase ment processes.
(off-take) agreements and dispatches of power Parallel efforts are needed to strengthen
under those agreements. Governments must financial and operational monitoring of SOEs
restore a strong sector planning capability at by their supervisory agencies, whether they are
the line ministry level, establish clear policies line ministries or ministries of finance. Trans-
and criteria for allocating new plant oppor- parency and accountability of SOEs depend on
tunities between the state-owned utilities and solid financial management, procurement, and
IPPs, and commit to competitive and timely management information systems. Today, basic
bidding processes. A well-articulated plan for operational and financial data on firm perfor-
the sector will allow governments to move mance are not produced, reported, or acted
beyond the “firefighting” that has reduced their on. Without information or, perhaps worse,
ability to anticipate exogenous shocks, such as without action based on whatever informa-
drought or high oil prices. tion is produced, better outcomes cannot be
Development partners need to tread care- expected. Key measures include auditing and
fully in the hybrid marketplace. They can help publishing financial accounts and using com-
by providing advice on transparent contract- prehensive cost-based accounting systems that
ing frameworks and processes and by lend- allow functional unbundling of costs and a
ing expertise to governments and utilities as clearer sense of cost centers.
the latter seek to reach financial closure with In principle, regulation can be an important
project sponsors and private investors. Lend- part of this process, but in practice, it proves
ing to public utilities needs to be handled care- challenging to develop. Electricity regulators
fully; if done without adequate attention to have been set up across Africa, precisely to insu-
the peculiarities of the hybrid market, it may late utilities from political interference while
have the unintended effect of deepening the closely monitoring enterprise performance.
contradictions inherent in those markets and Some critics argue that regulatory agencies
even crowding out private investment. What is have simply created additional risks because of
needed above all is to strengthen public insti- their unpredictable decisions, resulting from
tutions to enable them to engage effectively excessive discretion and overly broad objec-
with the private sector. tives. Moreover, regulatory autonomy remains
elusive; in some countries, turnover among
Recommitting to the Reform of commissioners has been high, while the gap
State-Owned Enterprises between law (or rule) and practice is often
Renewed efforts on SOE reform should favor wide. The challenge of establishing new
governance over technical fixes. State-owned public institutions in developing countries is
utilities are still prevalent across Africa, and often underestimated. Independent regula-
their performance is generally poorer than in tion requires a strong political commitment
other regions. Fortunately, improving the gov- and competent institutions and people. Where
ernance of SOEs can improve performance. some or all are lacking, it seems wise to consider
Past efforts at improving utility management complementary or transitional options that
196 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
reduce discretion in regulatory decision making local management teams, each of which offers
through more explicit rules and procedures, or advantages. Nonetheless, clarity about what
outsourcing the regulatory functions to advisory they can and cannot achieve, particularly given
regulators and expert panels (Eberhard 2007). their short time horizons, is important. At
When this foundation is in place, con- best, a management contract can improve per-
tracting mechanisms can be used to improve formance on a handful of manageable aspects
performance. These mechanisms could be of efficiency, such as revenue collection and
performance contracts in the public sector labor productivity. It cannot solve deficien-
or management contracts with the private cies in the broader institutional framework,
sector. which ideally should be addressed earlier. Nor
Public sector performance contracts need can a management contract raise investment
to incorporate strong performance incen- finance or significantly affect service quality
tives. Initial attempts to improve African if substantial investments or long gestations
SOEs through performance contracts with are required.
the line ministry or other supervisory agency Utilities that have the institutional basics
were minimally effective. Recent efforts in the in place would likely benefit from technical
water sector (in Uganda, for example) have assistance (box 8.4). In particular, operational
had a stronger and much more positive effect. efficiency programs are needed to reduce the
The key feature of these contracts is to incor- high rates of technical, nontechnical (electric-
porate incentives for good managerial (and ity theft), and collection losses. Such programs
staff) performance and, more rarely, sanctions can include capacity building and technical
for failure to reach targets. This approach to assistance to improve management, business
more comprehensive performance contracts practices, and planning. The priorities are
deserves further consideration. improved load management (to better match
Creating effective performance incentives supply with priority customer needs), theft
within a public sector context can be quite reduction initiatives, and increased revenue
challenging. Management contracts with collection (through enhanced metering and
the private sector are thus a relevant option. better-run customer service units). Capital
They can be applied with either expatriate or spending can also be reduced by using low-cost
BOX 8.4
technology standards, as in Guinea and Mali. West Africa, where the costs of power will
Innovations have included adjusting technical remain relatively high even with regional
design standards to meet the reduced require- trade.) A case thus exists for these countries
ments of low-load systems, maximizing the to start moving their tariffs toward longer-
use of material provided by local communities term cost-recovery levels, accepting that the
(such as locally sourced wooden poles), and sector will continue to register financial defi-
recruiting employees and supervisors from the cits in the short term.
local community. Cost recovery is particularly important
Finally, institutional change is a long-term for emergency power leases, to avoid divert-
matter, but well worth the wait. Victories on ing budgetary resources from long-term
this front will be small and slow in coming. investments. Numerous African countries
Donors may prefer the large and the quick, have responded to the power crisis by leasing
but they must recognize that positive changes emergency power generation. This solution is
in this field lie at the heart of African power rapid and effective but simultaneously costly
sector reform. and temporary. Charges typically amount to
$0.20–$0.30 per kilowatt-hour, without con-
Increasing Cost Recovery sidering transmission and distribution costs
The financial viability of incumbent utilities or associated losses. Given that the cost of
is a key foundation of a healthy power sector. backup generation for the private sector is
Financially viable utilities are more effective approximately $0.40 per kilowatt-hour, and
operationally, because they are able to finance that the value of lost load may well be higher
timely maintenance activities. They are also than that, the private sector should be will-
more creditworthy and thus may begin to ing to pay the full cost of this emergency
secure their own access to domestic or inter- power. Nevertheless, when emergency power
national capital markets. Achieving this goal is provided without any adjustment to power
demands power tariffs that are high enough to tariffs, the resulting fiscal drain can be very
cover operating costs and to contribute as much large, diverting scarce resources from the
as possible to covering capital costs as well. investments needed to provide a longer-term
Cost recovery already looks feasible in coun- solution to the power problem. To avoid this
tries with relatively low-cost domestic power fiscal drain, utilities must price emergency
sources. In the continent’s larger countries, power at cost-recovery levels for nonresiden-
and in those that rely on hydropower and coal- tial customers.
based generation, cost-recovery tariffs already Power subsidies will still be needed, but
appear affordable for the majority of the popu- they should be well targeted and focus initially
lation, and certainly for the affluent minority on expanding access. Existing power subsidies
that enjoys access to power. A case thus exists are captured largely by higher-income groups
for these countries seriously to consider mov- and do little to broaden access to electricity.
ing closer to full cost recovery. Redesigning power subsidies would free scarce
For countries with high-cost domestic fiscal resources that could be redirected to
power, cost recovery may become feasible subsidize the expansion of power networks
in the medium term as regional trade devel- to serve lower-income rural and periurban
ops. In the continent’s smaller countries, and communities, or for other poverty-alleviation
those reliant primarily on oil-based genera- programs. In some of Africa’s poorest coun-
tion, cost-recovery tariffs are largely unaf- tries, even low-cost power will remain unaf-
fordable. As regional trade develops and fordable for a significant minority of the
access to more cost-effective sources of power population, so well-targeted subsidies would
generation open up, however, the total cost be needed as part of the strategy for reaching
of power production will fall, making cost universal access. What is clearly untenable,
recovery a much more reasonable goal in the however, is the situation where power subsi-
medium term. (The possible exception is dies that benefit only a privileged minority of
198 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
the population create a significant fiscal drag responsible for extending the grid, have been
on the economy. more successful than those that followed
decentralized approaches, where a rural elec-
Accelerating Electrification trification agency attempted to recruit multiple
From a social and political perspective, expand- utilities or private companies into the electrifi-
ing access is imperative. Yet financing expan- cation campaign. Therefore, expecting special-
sion to lower-income households will further ized agencies to solve the rural electrification
strain the financial viability of the power sec- challenge on their own may be unrealistic.
tor. Tackling this dilemma will require sig- They may be most productive in promoting
nificantly higher concessional financing from minigrids and off-grid options as extensions of
development partners for access programs, the national utility’s efforts to extend the grid,
as well as improved financial and operational as in Mali (box 8.5).
performance from utilities. Rural electrification may need to follow
Given the scale of investments needed, a urban electrification. In an African context,
systematic approach to planning and financing one can legitimately ask how far rural electri-
new investments is critical. The current ad hoc fication can progress when the urban electrifi-
project-by-project approach in development cation process is still far from complete. Across
partner financing has led to fragmented plan- countries, a strong correlation exists between
ning, volatile and uncertain financial flows, urban and rural electrification rates, as well as
and duplicated efforts. Engagement across a systematic lag between the two. Countries
the sector in multiyear programs of access with seriously underdeveloped generation
rollout, supported by multiple development capacity and tiny urban customer bases are
partners as part of a coherent national strat- not well placed to tackle rural electrification,
egy, will channel resources in a more sustained either technically because of power shortages
and cost-effective way to the distribution or financially because of the lack of a basis for
subsector. Coordinated action by develop- cross-subsidization.
ment partners will also reduce the unit costs Finding ways of spreading the benefits of
of increasing access, by achieving economies of electrification more widely is also impor-
scale in implementation. tant. Because universal household electrifica-
Completing the urban electrification pro- tion is still decades away in many countries,
cess requires careful attention to the social sectorwide programmatic approaches need
issues raised. Chapter 3 of this volume found to ensure that the benefits of electrification
that approximately half of the nonelectrified touch the poorest households that may be too
urban population lives in proximity to the far from the grid or just unable to pay for a
grid. Densification is thus a key challenge. grid connection. Street lighting may be one
Demand-side barriers, including high connec- way of doing that in urban areas. In rural
tion charges and insecure tenure, need to be areas, solar-powered electrification of clinics
addressed as part of this process. Expansion and schools that provide essential public ser-
into periurban slums will need to face power vices to low-income communities is another
theft, for which technical fixes are available way of allowing them to participate in the
(see box 8.4). benefits of electrification. Another is appro-
For rural electrification, emerging evidence priate technology, such as low-cost portable
favors more centralized approaches (Mostert solar lanterns that are much more accessible
2008). Countries with dedicated rural electrifi- and affordable to the rural public. The Light-
cation funds have higher rates of electrification ing Africa initiative is supporting the develop-
than those without. Of greatest interest, how- ment of the market for such products.
ever, are the differences among the countries
with funds. Case studies indicate that the coun- Expanding Regional Trade
tries that have taken a centralized approach A strategic priority is to tackle head-on the
to electrification, with the national utility generation capacity deficit through major
Power: Catching Up 199
BOX 8.5
significant thermal generation projects in Figure 8.10 External Financing Commitments for the African Power Sector,
Nigeria and Sudan. 1994–2007
Private finance was also buoyant until 2007, 6,000
but significantly lower than official finance.
Private commitments to Africa’s power sector 5,000
averaged about $1 billion a year in 2005–07,
putting it in third place behind non-OECD 4,000
US$ millions
finance and traditional official development
3,000
assistance. The bulk of private resources has
gone into 3,000 megawatts of independent 2,000
power projects. Although it will not be enough
to close the financing gap, private finance is 1,000
very much needed. Successful private invest-
ments in energy projects in Africa are still rare, 0
1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007
however, and increased private investment will
official development assistance private participation in infrastructure
not materialize simply because of large financ-
non-OECD
ing gaps. The lessons from past failures must
be addressed because private investment will Source: Briceño-Garmendia, Smits, and Foster 2008.
flow only where rewards demonstrably out- Note: OECD = Organisation for Economic Co-operation and Development.
weigh risks. Some early but encouraging signs
indicate that scaling up generation capacity
through large private sector–led projects is Notes
starting to gather momentum. A prominent The authors of this chapter are Anton Eberhard,
example is the privately owned 250-megawatt Vivien Foster, Cecilia Briceño-Garmendia, and
Bujagali hydropower plant in Uganda, sup- Maria Shkaratan, who drew on background
ported by World Bank Group guarantees and material and contributions from Daniel Camos-
funded by a private consortium. Ambitious i-Daurella, Gabriel Goddard, Jaakko Hellaranta,
regional projects undoubtedly present techni- Rob Mills, Fatimata Ouedraogo, Timo Reiss,
cal, financing, and political risks and will con- Orvika Rosnes, Jevgenijs Steinbuks, Prasad
tinue to be underpinned by substantial public V. S. N. Tallapragada, Maria Vagliasindi, Tjaarda
sector and donor contributions. Storm P. Van Leeuwen, and Haakon Vennemo.
1. Isolated areas are more than 50 kilometers from
Shorter-term measures on energy effi-
a substation and are either in the power plant
ciency can ease the transition. Most of the buffer (within 10 kilometers for capacity below
measures described here are medium term 10 megawatts, 20 kilometers for capacity below
and cannot be implemented overnight. Many 100 megawatts, and 50 kilometers for capacity
Sub-Saharan countries will continue to face below 100 megawatts) or within 10 kilometers
a very tight demand-supply balance in the of a lit urban area or lit pixel. Remote areas
coming years. Therefore, shorter-term mea- are more than 50 kilometers from a substation
sures to soften the economic and social effects and are not in the power plant buffer or within
of power scarcity must complement longer- 10 kilometers of a lit urban area or lit pixel.
term efforts at addressing the underlying 2. These costs are calculated at the consumption
structural causes of the power supply crisis. level of 100 kilowatt-hours a month.
Recent experiences from countries such as
Brazil show that well-designed demand-side References
management measures (for example, a quota Banerjee, Sudeshna, Quentin Wodon, Amadou
system with price signals, combined with a Diallo, Taras Pushak, Hellal Uddin, Clarence
public energy-efficiency campaign) can go Tsimpo, and Vivien Foster. 2008. “Access, Afford-
ability and Alternatives: Modern Infrastructure
a long way toward trimming peak demand, Services in Africa.” Background Paper 2, Africa
substantially reducing power rationing at Infrastructure Country Diagnostic, World Bank,
fairly low economic and social cost. Washington, DC.
202 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Briceño-Garmendia, Cecilia. 2008. “Quasi-Fiscal Gratwick, Katharine Nawaal, and Anton Eberhard.
Costs: A Never Ending Concern.” Internal note, 2008. “An Analysis of Independent Power
Africa Infrastructure Country Diagnostic, World Projects in Africa: Understanding Development
Bank, Washington, DC. and Investment Outcomes.” Development Policy
Briceño-Garmendia, Cecilia, Karlis Smits, and Review 26 (3): 309–38.
Vivien Foster. 2008. “Financing Public Infra- Molefhi, Benjamin O. C., and L. J. Grobler. 2006.
structure in Sub-Saharan Africa: Patterns, Issues, “Demand-Side Management: A Challenge and
and Options.” Background Paper 15, Africa Opportunity for Botswana Electric Energy
Infrastructure Country Diagnostic, World Bank, Sector.” North-West University, Potchefstroom,
Washington, DC. South Africa.
Calderón, César. 2008. “Infrastructure and Mostert, Wolfgang. 2008. Review of Experience with
Growth in Africa.” Working Paper 3, Africa Rural Electrification Agencies: Lesson for Africa.
Infrastructure Country Diagnostic, World Eschborn, Germany: EU Energy Initiative Part-
Bank, Washington, DC. nership Dialogue Facility.
Eberhard, Anton. 2007. “Matching Regulatory PPA (Power Planning Associates). 2005. “Towards
Design to Country Circumstances: The Potential Growth and Poverty Reduction: Lessons from
of Hybrid and Transitional Models.” Gridlines, Private Participation in Infrastructure in Sub-
Note 23 (May). Public-Private Infrastructure Saharan Africa—Case Study of Botswana Power
Advisory Facility, World Bank, Washington, DC. Corporation.” Summary report, World Bank,
Eberhard, Anton, Vivien Foster, Cecilia Briceño- Washington, DC.
Garmendia, Fatimata Ouedraogo, Daniel Rosnes, Orvika, and Haakon Vennemo. 2008.
Camos, and Maria Shkaratan. 2008. “Under- “Powering Up: Costing Power Infrastructure
powered: The State of the Power Sector in Investment Needs in Southern and Eastern
Sub-Saharan Africa.” Background Paper 6, Africa.” Background Paper 5, Africa Infrastruc-
Africa Infrastructure Country Diagnostic, ture Sector Diagnostic, World Bank,
World Bank, Washington, DC. Washington, DC.
EIA (Energy Information Administration). 2006. Vagliasindi, Maria, and John Nellis. 2009. “Evalu-
International Energy Annual. Washington, DC: ating Africa’s Experience with Institutional
U.S. Department of Energy. http://www.eia.doe Reform for the Infrastructure Sectors.” Working
.gov/emeu/international/contents.html. Paper 23, Africa Infrastructure Country Diag-
ESMAP (Energy Sector Management Assistance nostic, World Bank, Washington, DC.
Program). 2007. “Technical and Economic Wodon, Quentin, ed. 2008. “Electricity Tariffs
Assessment of Off-Grid, Mini-Grid and and the Poor: Case Studies from Sub-Saharan
Grid Electrification Technologies.” ESMAP Africa.” Working Paper 11, Africa Infrastructure
Technical Paper 121/07, World Bank, Country Diagnostic, World Bank, Washington,
Washington, DC. DC.
Foster, Vivien, William Butterfield, Chuan Chen, World Bank. 2005. African Development Indicators.
and Nataliya Pushak. 2008. Building Bridges: Washington, DC: World Bank.
China’s Growing Role as Infrastructure Finan-
cier for Africa. Trends and Policy Options no. 5. ———. 2007. Enterprise Surveys database. http://
Washington, DC: Public-Private Infrastructure www.enterprisesurveys.org/.
Advisory Facility, World Bank. Yepes, Tito, Justin Pierce, and Vivien Foster.
Foster, Vivien, and Jevgenijs Steinbuks. 2008. 2008. “Making Sense of Sub-Saharan Africa’s
“Paying the Price for Unreliable Power Supplies: Infrastructure Endowment: A Benchmarking
In-House Generation of Electricity by Firms in Approach.” Working Paper 1, Africa Infrastruc-
Africa.” Policy Research Working Paper 4913, ture Country Diagnostic, World Bank,
World Bank, Washington, DC. Washington, DC.
Chapter 9
S
o much variety exists among the trans- Increasing the scope for competition among
port modes and their infrastructure, in carriers is a challenge confronting most trans-
both policy and technical matters, that port modes. In most parts of the world, com-
the following four chapters deal with them petition has involved extending the role of the
separately. Those modes interact and intercon- private sector. Franchises and concessions per-
nect in complex ways, however, raising com- mit this extension without loss of government
mon issues. influence over activities regarded as strategic.
Many movements of passengers and freight Bus transport and trucking are already pre-
involve more than one mode of transport, with dominantly private. Many African railways are
impediments at the point of interchange, often now concessioned, and the role of the private
caused by corrupt customs administration, or sector is increasing in ports and air transport.
restrictions on entry to the transport market. In However, the regulation of these markets often
freight, Sub-Saharan Africa suffers particularly remains obstructive rather than constructive,
from such blockages, which delay shipments, with tour de role traffic allocation and dispatch-
increase costs, and hinder the development ing1 reducing efficiency and increasing costs.
of logistics systems so critical in sophisticated A common reason for restricting competi-
global markets. Because Africa has so many tion is the belief that the government can ensure
landlocked countries, this problem is multina- provision of socially desirable services only
tional and must be confronted on a regional or when they are supplied by a public enterprise or
a corridor basis. strongly protected private enterprises. However,
One of the main sources of restrictive entry this belief is largely fallacious, particularly if the
is the perceived need to protect national carri- protected suppliers have no incentive to be effi-
ers or the existing carriers. Although this pro- cient or no segment of the market is profitable
tection has largely withered away in interna- enough to support unremunerative “social” ser-
tional shipping and inter-African air transport, vices. Publicly owned airlines, railways, shipping
it still prevails in road freight. companies, and bus companies have all fallen
203
204 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
short. Thus, much work remains, both nationally can increase their profits through predatory
and regionally, in developing regulatory regimes practices, as in Zambia’s treatment of copper
that reconcile public and private interests. exports by the Democratic Republic of Congo
(see chapter 11 in this volume). In East Africa,
joint concessioning of railways is part of a
Integrating Multimodal Transport donor–funded corridor, where border-crossing
arrangements have been reformed. Some coun-
Inland transport costs often seriously inhibit tries are now trying to develop coordinated
foreign trade, which passes through a seaport corridor systems, as in the Ghana Gateway and
or airport and then one or more land modes. Maputo corridors.
For example, inland transport accounts for an Other transport modes may also be involved.
estimated 40 percent of the total cost from the Inland waterway transport historically carried
point of origin to the port of destination for primary products from landlocked countries but
West African cocoa and coffee exports. is now in decline. The three major lakes in East
The transport chains are no stronger than and Central Africa—Victoria, Tanganyika, and
their weakest links, usually the interchanges. Malawi—used to be important in transit and
The weaknesses are partly physical, with no intraregional trade. Particularly on Lake Victo-
physical connection between the modes and ria, the lake services were part of the railway sys-
no infrastructure for transshipment; partly tems linking the railheads at the inland ports of
institutional, with responsibility for the inter- Bell (Uganda), Kisumu (Kenya), and Mwanza
changes not falling clearly to one modal agency (Tanzania). The Kenya and Uganda lake opera-
or the other; and partly operational, with the tions were concessioned together with the rail-
government collecting taxes and duties or staff ways in those countries, whereas in Tanzania, the
collecting bribes, slowing movements, and lake services were separated from the railways.
pushing up costs. Only one service now operates on Lake Victo-
The port-rail link is the first major weakness. ria, and some of the railway track leading to the
Although rail transport is advantageous for ports is in poor repair, especially in Kenya.
long-distance, time-insensitive commodities, it The story is similar in West and Central
depends heavily on internationally traded traf- Africa, where the Congo basin has a navigable
fic. To hold this traffic, it must be linked effi- network of 12,000 kilometers, covering nearly
ciently to good port connections, but conflicts 4 million square kilometers in nine countries.
between rail and port jurisdictions over rail seg- In principle, this waterway could be a very
ments in port areas often inhibit this link. Except valuable resource in a multimodal transport
in South Africa, inland transport and facilities network serving the region, particularly given
are poorly aligned with port development. The low associated transport costs of $0.05 per ton-
stripping and stuffing of containers in port kilometer versus $0.15 per ton-kilometer for
areas also increases congestion in many ports. It road or rail freight in Central Africa, albeit at
is no accident that some of the most successful significantly lower speeds. In practice, however,
rail lines in Africa operate in national corridors the river corridor suffers from an outdated and
where specialized rail and port facilities are ver- insufficient infrastructure, inadequate chan-
tically integrated (for example, Spoornet coal nel markings and maintenance, feeble regula-
and ore lines and Gabonese manganese ore). tion, and numerous nonphysical barriers to
Good links between complementary rail movement. Thus, it is ever more marginal in
systems are also essential. Some railway orga- transport. Recognizing this untapped poten-
nizations already contribute to this. The bina- tial, in October 2005, the Executive Secretary
tional railways in Côte d’Ivoire–Burkina Faso of the Economic and Monetary Community of
and Senegal-Mali provide good examples, as Central Africa encouraged the governments of
does the operation of contiguous railways by Cameroon, the Central African Republic, the
the same concessionaire (Central East African Democratic Republic of Congo, and the Repub-
Railways in Mozambique and Malawi). How- lic of Congo to establish the Commission Inter-
ever, they also create local monopolies that nationale du Bassin Congo-Oubangui-Sangha
Transport: More Than the Sum of Its Parts 205
to improve the physical and regulatory arrange- developed in Africa. Whatever the main mode
ments for inland navigation. A consultancy of transport, the most serious impediments are
study is examining the current arrangements administrative. For example, in road transport,
in the four participating countries and identi- regulation and market structures of the road
fying steps to begin an effective redevelopment freight industry, rather than the quality of the
of inland navigation (CICOS 2007). road infrastructure, are the binding constraints
on international movements (Teravaininthorn
and Raballand 2008). African governments must
Developing Logistics Systems understand the importance of the qualitative
aspects of logistic performance and act to remove
The transport problems of doing business in administrative obstacles to that performance.
Africa are not just a matter of poor infrastructure Corruption, through bribery of a range of
or high transport costs. Modern logistics systems officials having a combination of discretion-
also emphasize the efficiency of customs and ary power and a quasi-monopoly position in
other border agencies, the ease and affordability the logistics chain, is a critical problem. For
of arranging international shipments, the com- example, corruption can increase the total
petence of the local logistics sector, the ability to shipping costs, including costs of overland
track and trace shipments, and their timeliness. transport, port clearance, and ocean ship-
These various aspects of logistic performance ping, of a standard 20-foot container trav-
are encapsulated in a logistic performance eling between South Africa’s economic hub
index (figure 9.1). Except for South Africa, Sub- and eastern Africa or the Far East by up to
Saharan African countries perform poorly not 14 percent and the total port costs by up
only on infrastructure quality, but also on all the to 130 percent. Contrary to common belief,
main aspects of logistics competence. no robust evidence links corruption with low
Africa is thus viewed as logistics-unfriendly; wage levels or lack of job rotation. Rather, it
third-party logistics systems, so important in is highly correlated with the extent to which
increasing production and distribution effi- rules, regulations, and the organizational fea-
ciency in the industrial countries, are poorly tures of bureaucracies give public officials the
Figure 9.1 Africa Registers Low Scores on the Logistics Performance Index, 2007
1 <= LPI <= 2.29 2.29 <= LPI <= 2.53 2.53 <= LPI <= 3.14 3.14 <= LPI <= 5 no data
bargaining rights to extort bribe payments Most of the critical trade facilitation imped-
from shippers. In a comparison between the iments stem from the procedure and services
ports of Maputo (Mozambique) and Durban in the transit country, which often sees no
(South Africa), bribery of customs officials advantage for its own economy in addressing
accounted for 80 percent of total bribery the needs of transit traffic. This view may, in
in Maputo, but only 10 percent in Durban fact, be mistaken if transit traffic contributes
(Sequeira and Macchi 2009). The explanation to scale economies in the transit country’s
was that, in Maputo, a low level of automa- own port and maritime or air service sectors,
tion existed and both monitoring and sanc- and particularly if the goods transported com-
tioning were weak, whereas in Durban, the plement, rather than compete with, its own
opposite was the case. In contrast, bribery of products. Even in those circumstances, how-
port officials was lower in the privately con- ever, the vested interests of the beneficiaries
cessioned Maputo port, where a higher level of bribes and providers of unnecessary border
of automation, monitoring, and sanctioning service may be difficult to overcome. Further-
exists than in the publicly operated Durban more, governments of transit countries with
port, where automation is low and monitoring potential benefits may give the wrong signals
and sanctioning are weak. (see box 9.1).
These problems with transit countries
are increasingly recognized. In August 2003,
Developing Transit Corridors for attendees at an intergovernmental conference
Landlocked Countries at Almaty, Kazakhstan, agreed to a Program of
Action calling for joint efforts by landlocked
The weakness of logistics systems impinges and transit countries to revise their regulatory
most on the landlocked countries. In southern frameworks and improve transit-related infra-
Africa, operating costs are not excessively structure and procedures with the assistance of
higher than in Europe ($0.08 per ton kilo- other countries and microfinance institutions.
meter in 2008), but queues at borders and Some critical elements of a reform program
restrictions on access to the market mean that can be identified. Reform of trucking regulation
rates on international routes are higher by is essential and is discussed further in chapter
10–30 percent. Being landlocked adds four 10 of this volume. This must be supplemented
days to exports and nine days to imports by a reengineering of the transit system. In the
for land distribution of equivalent distances longer term, something similar to the European
within a seaport country. However, this differ- Transports Internationaux Routiers (TIR),
ence is usually less than the sea transit time. which provides a reliable system based on a
Much more problematic is the high variability single manifest (the TIR carnet) and a chain
of transit time, especially for imports. The of guarantees, could speed freight through
unreliability of import and export chains international borders, though the immediate
stems largely from inadequate transit proce- introduction of a system is doubtful. More
dures, overregulation, multiple controls, and immediately, one-stop border posts, in which
poor border service (World Bank 2008). In the two border countries share the informa-
addition, bribes and unnecessary charges can tion provided, could eliminate duplication and
add 50 percent to the transport costs between a speed transit. Customs must be automated,
port and a landlocked country. Because of these limiting direct contact between officials and
impediments, landlocked countries have a agents; monitoring and sanctioning must be
much smaller involvement in world trade than strengthened; and port management must be
the coastal countries, and they have remained privatized to reduce corruption. Inland clear-
predominantly dependent on exports of raw ance centers or dry ports, which deal with traf-
materials. Although increasing prices for raw fic under bond, can circumvent the traditional
materials over the past five years have helped impediments of the seaports. Development of
landlocked countries in Africa, they remain airfreight can also overcome the constraints for
vulnerable to a change in that trend. some traffic.
Transport: More Than the Sum of Its Parts 207
BOX 9.1
Some progress has been made in this direc- the planned Madarail bonded container ter-
tion. Trucking regulation is more liberal in minal near Antananarivo, Madagascar.
southern Africa than in other parts of Sub- Much remains to be done to fill gaps in the
Saharan Africa. Transit system reform is being implementation of the Almaty program. Gen-
introduced to facilitate transit traffic through erally accepted corridor-monitoring indica-
Cameroon to Chad and the Central African tors are required to check progress. Trucking
Republic (see box 9.2). markets and the associated freight-forwarding
One-stop border post initiatives apply sector reforms need to be extended. Agreement
to the border between Kenya and Uganda at between neighboring countries on transit sys-
Malaba; between Zambia and Zimbabwe at tem reforms must be encouraged by donor
Chirundu; between Zimbabwe and Mozam- contributions and facilitated by the develop-
bique at Forbes/Machipanda; along the ment of best practice standards.
Trans-Kalahari Corridor; and in West Africa
on selective borders of Burkina Faso, Ghana,
Mali, and Togo. Some landlocked countries Increasing Competition
already have bonded warehouses at the ports
in West Africa, and rail concessionaires are Competition between suppliers of an individ-
developing facilities to speed transits, such as ual transport mode improves service quality
the Sitarail intermodal terminal proposal in and reduces costs, as shown in chapter 10 on
Ouagadougou, Burkina Faso; the ZRS com- road transport and chapter 13 on air transport
pany customs bond at the Victoria Falls border in this volume. Competition between modes
crossing between Zambia and Zimbabwe; and may also result in the replacement of one by
208 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
BOX 9.2
the other. For example, road improvement in rail service is expected to cover its full costs,
Mauritania effectively eliminated domestic air including track. The viability of privately
transport, and all African railways have had concessioned railways can be hurt by under-
difficulty retaining passenger traffic where charging for road maintenance and not enforc-
a road alternative exists. However, various ing axle load limits on road traffic.
modes may be complementing each other or National transport strategies that put
competing with each other at the same time. infrastructure charging for road and rail on a
Reconciling these tensions is very difficult. common basis would probably reduce budget
Where modes compete, efficient alloca- burdens and improve road conditions. The net
tion of traffic between them would normally effect on modal split is more difficult to predict
depend on their relative prices reflecting their because of security, reliability, and other non-
relative costs; however, that can be difficult to cost items considered in determining choices.
interpret. Where traffic is heavily imbalanced For example, a comparison of road and rail
(as in international trade) and cost structures tariffs on five concessioned railways showed
differ between modes, both commercial pres- road tariffs exceeding rail tariffs by 44–213 per-
sure and economic efficiency may call for cent. More competition between road and rail,
widely varying price–marginal cost ratios, and abetted by increasing road costs, might reduce
predatory pricing is difficult to define. trucker profit margins rather than increase the
Attention to the fairness of competitive mode share or pricing power of rail. Given
conditions is important. Roads are publicly the cartels and high profit rates in road haulage,
provided at costs to users that do not even even greater benefits could likely be obtained
cover maintenance, whereas privately provided by promoting competition in road haulage.
Transport: More Than the Sum of Its Parts 209
passenger fares. Where fares are commercially competition and usually results in low vehicle
unviable, however, they are likely to under- use and high unit costs.
mine the supply of the public services. Even 2. A negative concession is one offered for a loss-
where the controls are enforced only on pub- making service, in which the competition is
conducted in terms of the minimum public
licly provided modes (notably rail or conven-
subsidy required by the operator to take on the
tional large bus services), they typically shift
loss-making service. An affermage contract is
business to an informal sector with higher similar to a management contract, but the pri-
fares or lower service quality, failing to assist vate operator takes responsibility for all opera-
poorer citizens. tion and maintenance functions, both technical
Governments could pay compensation for and commercial.
the imposed social obligations. However, only
in rare cases, such as the support of the public References
bus company in Addis Ababa, are public service CICOS (Commission Internationale du Bassin
obligation payments adequate and timely. More Congo-Oubangui-Sangha). 2007. Plan d’action
generally, the failure to achieve this timeliness strategique pour la promotion de la navigation dans
le bassin Congo-Oubangui-Sangha. Hamburg:
postpones, but does not avoid, the collapse HPC Hamburg Port Consulting GmbH.
of public bus service. The same appears to be
Department of Transport. 2005. “National Freight
happening with the provisions for subsidizing Logistics Strategy.” Department of Transport,
rail passengers in some rail concessions. Where Republic of South Africa.
subsidized rail services have economically Raballand, Gaël, and Patricia Macchi. 2008. The
viable road alternatives, the subsidy (if paid) Critical Importance of Strengthened Regional
drains resources from the economy. More com- Integration for Growth and Development in
prehensive and well-thought-out strategies for Mozambique. Washington, DC: World Bank.
fare controls and subsidies, probably by pro- Sequeira, Sandra, and Patricia Macchi. 2009. “The
curing subsidized services through competitive Importance of Soft Transport Infrastructure:
tendering, can ensure the most effective use of Customs Officials in Maputo versus the Port
Operators in Durban.” Afrique Contemporaine
transport infrastructure. 230 (2).
Teravaininthorn, Supee, and Gaël Raballand. 2008.
Notes Transport Prices and Costs in Africa: A Review
The author of this chapter is Kenneth Gwilliam, of the Main International Corridors. Directions
who drew on background material and con- in Development Series. Washington, DC:
tributions from Jean Francois Arvis, Rodrigo World Bank.
Archondo-Callao, Jose L. Guasch, Alberto World Bank. 2007. Connecting to Compete: Trade
Nogales, Gael Raballand, Sandra Sequeira, and Logistics in the Global Economy—the Logistics
Kavita Sethi. Performance Index and Its Indicators. Washington,
1. Tour de role dispatching is a system in which DC: World Bank.
service suppliers are allocated traffic strictly in ———. 2008. Improving Trade and Transport for
turn. It is widely applied in the public transport Landlocked Developing Countries. Washington,
and trucking markets. This system limits direct DC: World Bank.
Chapter 10
A
frican governments have been address- and financial flows for the interurban roads,
ing the low density and poor condition other challenges have surfaced that will require
of their road networks. Institutional different types of solutions.
reform since the mid-1990s has progressed well, First, the reforms to the interurban road
with a remarkable consensus on the content. network have affected rural roads much
Most countries have second-generation road less. Even though agriculture is viewed as
funds supported by fuel levies, and many oth- an engine of growth, only one-third of rural
ers have autonomous road agencies. Specialist inhabitants live within 2 kilometers of an
maintenance management agencies have been all-season road. Doubling this percentage
established, and new forms of contract-based would be very costly, absorbing more than
maintenance are being introduced. Although 1 percent of GDP a year for a decade. The
important funding gaps remain, results are dis- rural environment presents particular insti-
cernible. On average, 80 percent of the main tutional challenges for road maintenance.
road network is in good or fair condition, and Second, surface transportation is about more
the current value of the national road net- than good roads. Africa continues to be handi-
works is at least 70 percent of their potential. capped by very high road freight tariffs, driven
The limited time series available also suggests primarily by high profit margins rather than
that a number of countries have improved road high costs (or defective roads). In Central and
conditions over time. West Africa particularly, trucking industry car-
Despite this progress, the reform agenda tels and restrictive tour de role traffic allocation
is incomplete. In many cases, fuel levies have and dispatching practices are responsible for
been set too low to be effective, and road funds low vehicle mileage and poor fleet quality. The
and agencies do not always meet all good- most urgent reforms are to liberalize trucking
practice design criteria. Modern contracting while mitigating associated social effects. With-
and contract management methods are far out such measures, further improvements in
from universal. Furthermore, while policy mak- road quality will only translate into higher profit
ers’ attention has focused on the institutions margins for the trucking industry.
211
212 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Third, Africa’s rapidly growing cities face the road conditions lag somewhat behind those
major mobility problems. Urban road den- of other developing regions, but not so much
sity is low by developing-country standards. for the main trunk road network as for other
Moreover, following the demise of large buses roads (Gwilliam and others 2008).
in many cities, myriad informal minibus With accelerating urbanization, Africa needs
operators largely dominate urban transport to develop intraurban roads, but networks in
services. Services are costly, and availability is 14 African cities were found to be substan-
inadequate. Few countries capture sufficient dard (Kumar and Barrett 2007). Road den-
financial resources to develop and main- sity (paved-road density in particular) lags far
tain the urban road network. Overlapping behind that in other developing cities. Capacity
national, metropolitan, and municipal juris- is generally limited. The majority of roads have
dictions present serious institutional chal- one lane in each direction, and where roads
lenges. Furthermore, the cross-sectoral links are wider, pedestrians and parked vehicles
between urban transport and land use plan- often take up one lane. Intersections are close
ning are unexploited. together and are ill designed for turning. Ser-
vice lanes are absent, pavement is deteriorating,
and street lighting is minimal. Because traffic
Road Infrastructure—Lagging management is limited, accidents are frequent,
Other Regions Somewhat with pedestrians accounting for two-thirds
of fatalities.
The region’s trunk road network comprises For rural roads, beyond the classified ter-
strategic trading corridors linking deep-sea tiary network, which is typically the responsi-
ports to economic hinterlands. These corri- bility of local government, a vast unclassified
dors, which carry about $200 billion of trade network of tracks providing service to rural
a year, include no more than 10,000 kilome- areas is usually the responsibility of local com-
ters of road. The concept of an intraregional munities. Nevertheless, African rural commu-
trunk network, or Trans-African Highway, nities have by far the lowest accessibility to an
remains a distant reality because of missing all-season road in the developing world. Evi-
links and poor maintenance on key segments. dence indicates that physical isolation prevents
Between 60,000 and 100,000 kilometers of large areas of the continent from reaching their
road are required to provide such intraconti- agricultural potential. With low population
nental connectivity. density, achieving good overall rural accessi-
Africa’s national road density is substantially bility would imply at least doubling the length
lower than that in other developing regions: of the classified network for most countries
only 204 kilometers of road per 1,000 square (Starkey and others 2002).
kilometers of land area, with only one-quarter Traffic volumes remain low and heav-
paved, compared with a world average of ily concentrated on the main road network
944 kilometers per 1,000 square kilometers, (table 10.1). In most countries, at least
with more than half paved. That density is less 90 percent of reported traffic on the classified
than 30 percent of the next-lowest region, South network is carried on the main networks.1
Asia. However, Sub-Saharan African road den- Except in Nigeria and South Africa, the traffic
sity in relation to population is slightly higher on the main road network in Africa averages
than South Asia’s and only slightly lower than only about 500 vehicles a day.
the Middle East’s and North Africa’s. Rural networks typically carry less than
Relative to GDP, however, Sub-Saharan 10 percent of the classified network’s traffic;
Africa has a large road network. In Madagascar, however, in Ethiopia, Malawi, and Nigeria,
Malawi, Mozambique, and Niger, the asset they carry more than 20 percent. Except in
value of the road network exceeds 30 percent Nigeria, the absolute volumes of traffic on the
of GDP, an indication of the consequently large rural network are very low, averaging about
economic burden of maintenance. As a result, 30 vehicles a day.
Roads: Broadening the Agenda 213
Tanzania
independent auditing Namibia
Kenya
Rwanda
road user charges Malawi
Madagascar
Ghana
revenue allocation rules
Ethiopia
Chad
Zambia
separation of functions
Niger
Mozambique
Cameroon
clear legal basis
Lesotho
Côte d’Ivoire
Burkina Faso
0 20 40 60 80 100 0 20 40 60 80 100
% of countries % of score
Source: Gwilliam and others 2008, based on data collected by World Bank 2007.
to neighboring countries so that more than responsible for maintaining a level of service
40,000 kilometers of Latin American roads are linked to the condition of the road surface. Under
now being maintained under such contracts. a performance-based maintenance contract, the
In Africa, Ethiopia, Ghana, and Zambia have condition of the road improves steadily, whereas
begun to use them. under the traditional approach, the road condi-
Cost savings from performance-based tion improves for a short period following the
maintenance contracts on paved roads have work and then starts deteriorating quickly until
ranged between 10 and 40 percent in industrial- new maintenance is carried out.
ized countries and between 10 and 20 percent A parallel institutional development, par-
for several developing countries (Stankevich, ticularly relevant to rural roads, involves
Qureshi, and Queiroz 2005). Even where such delegating project management to special-
cost savings were not achieved, the benefits have ist agencies. In many countries, these agen-
been substantial. In Chad, the only example in cies, agences d’exécution des travaux d’intérêt
francophone Africa, the cost per kilometer of public (AGETIPs; public works implementing
routine maintenance under a performance-based agency), now manage private consultants and
maintenance contract for a set of gravel roads contractors on behalf of the public authority
was significantly higher ($5,000) than under and perform all the necessary functions for
a traditional maintenance contract ($1,500) contract preparation, implementation, and
but with the benefit that the contractor was supervision (box 10.1).
Roads: Broadening the Agenda 215
BOX 10.1
4.5 35
4.0
30
3.5
25
3.0
2.0 15
1.5
10
1.0
5
0.5
0 0
n
in
ire
Rw a
da
Ni r
ria
Ta a
ia
Ug l
da
ag d
Za r
bia
am ia
ue
wi
pe a
Na e
Le a
o
ga
ge
ca
rd
ny
an
Ca fric
ibi
oo
M Cha
th
an
p
Cô Ben
ala
biq
an
an
Ivo
ge
as
ne
M hio
Ni
Ve
so
Ke
m
Gh
er
nz
A
M
Se
d'
m
Et
h
Ca
ad
ut
te
oz
So
as % GDP US$ per capita
are the main culprits. Middle-income coun- Figure 10.3 Rehabilitation and Maintenance
tries and those with road funds and fuel levies Spending Relative to Norms
fare best in executing their capital budgets.
High capital spending may be justified by South Africa
wide price spreads across bids, looks to be the substantially better road conditions than do
main culprit (Van Zyl, Coetzer, and Lombard those that lack either. Moreover, both the road
2008). Substantial delays in project implemen- fund design and the level of the fuel levy appear
tation also add to the costs. to significantly affect the quality of the main
These higher costs double the importance road network, although again the effect on rural
of ensuring that engineering standards are the road quality is much less pronounced. The main
most cost-effective possible. Network analy- exception, South Africa, has very good primary
sis reveals that on average, about 30 percent roads without a conventional second-generation
of main road networks are overengineered road fund, but it does have a very effective road
relative to observed traffic volumes, and only agency and consistent government commitment
10 percent are underengineered. Particularly to finance road maintenance adequately.
where rapidly expanding the extent of the Surprisingly little variation exists in the
passable road network is desirable, as in the road network’s asset value as a percentage of its
Democratic Republic of Congo, adopting potential value if it all were in good condition.
lower standards may be sensible, upgrading All countries realize at least 70 percent of this
only when traffic growth or local bottlenecks potential, suggesting that maintenance efforts
make it necessary. About 15 percent of rural are fairly well concentrated on preserving the
networks are underengineered. Greater efforts high-value paved-road network. Moreover, the
are needed to adapt road design standards to limited time series available suggests that most
local conditions and materials to avoid exces- countries have improved the quality of their
sive costs in road construction, particularly for roads in recent years.
low-volume sealed roads. Good governance is thus critical for safe-
guarding road quality through good budget
finance and a professionally competent public
sector implementation agency. Countries with
Road Conditions—Reflecting road funds and high fuel levies are substan-
Quality of Sector Governance tially more successful at raising finance that
translates into higher road maintenance expen-
About half of the main network is in good con- ditures. Countries with road funds and quasi-
dition, and an additional one-third is in fair independent road agencies show substantially
condition—whereas only one-quarter of the higher quality on main road networks.
rural network is in good condition and a further
one-quarter in fair condition (figure 10.4).
The large variation in road quality reflects
several interacting factors. First is affordabil- Infrastructure Spending
ity. GDP per capita is most strongly correlated Needs—an Average of
with the percentage of the main road network 1.5 Percent of GDP a Year
in good condition, suggesting that richer
countries tend to spend more on maintenance. A modest set of connectivity objectives might
However, no such clear relationship exists for include the following: (a) connecting large cities
rural roads. Second are some fundamental top- and international frontiers with a good-quality,
ographic and climatic influences. Mountain- two-lane paved road; (b) connecting intermedi-
ous and wet countries generally have poorer ate cities and the provincial capital with a good-
conditions, in both main and rural networks, quality, one-lane paved road; (c) increasing to
associated with climate rather than traffic. 100 percent the proportion of the rural popula-
Even controlling for income and climate, tion living within 2 kilometers of an all-season
however, substantial variation still exists in road in the agricultural areas comprising the 80
observed road quality across countries, vary- percent highest production value within each
ing with the quality of their road sector insti- country; and (d) putting the urban population
tutions and financing framework. Countries within 500 meters of a road supporting all-sea-
with both road funds and road agencies show son bus access.
218 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 10.4 Distribution of Road Network Length across Condition Classes, by Country
Kenya Kenya
Namibia Malawi
Ethiopia Chad
Malawi Namibia
Ghana Cameroon
Zambia Ethiopia
Mozambique Lesotho
Niger Niger
Cameroon Tanzania
Tanzania Benin
Chad Senegal
Senegal Mozambique
Lesotho Zambia
Benin Nigeria
Rwanda Madagascar
Uganda Uganda
0 25 50 75 100 0 25 50 75 100
% road network length % road network length
good fair poor
On the basis of this package, spending Existing spending on the sector amounts to
needs for the road sector amount to $9.6 $6.9 billion a year, significantly less than what
billion a year, skewed toward capital expen- is needed (table 10.3). Spending in the fragile
diture (table 10.2). Because of their mature states is particularly low, barely one-tenth of
road network, the middle-income countries what is required. The public sector finances two-
account for little more than 10 percent of thirds of road sector spending and more than
this total. Except in middle-income coun- one-half of road sector investment. In the low-
tries, about two-thirds of spending require- income countries—whether fragile or not—
ments relate to capital expenditure, with about half of road sector expenditure is donor
the remainder attributable to operation financed. The contribution of the private sector
and maintenance. Thus, overall, the region to road finance in Africa is almost negligible.
needs to spend 1.5 percent of GDP on roads, Financiers from outside the Organisation for
of which 0.6 percent of GDP is needed for Economic Co-operation and Development are
road maintenance. However, the burden for not making a major contribution to this sector.
low-income fragile states is very high—in Implementation of efficiency-oriented
excess of 7 percent of GDP. reforms could raise a total of $3.8 billion a year,
Roads: Broadening the Agenda 219
largely eliminating the funding gap, except in “hard” infrastructure. However, what ulti-
fragile states. The greatest scope for efficiency mately matters from an economic perspective
gains lies in practicing sound preventive main- is the extent to which roads support efficient,
tenance, which in the medium term would reliable, and safe transport services for various
substantially reduce the investments needed kinds of freight, as well as for urban and rural
to clear the rehabilitation backlog, saving an populations. Although the private sector typi-
estimated $1.9 billion per year. Low ratios of cally provides these services, the government
capital budget execution are also holding back has a critical role to play as regulator and
public investment in roads. Addressing this facilitator of service provision. These “soft”
issue would capture a further $1.3 billion annu- transport issues deserve more attention.
ally. Finally, some countries face difficulties in
collecting revenues owed to their road funds; Road Freight
solving this problem would capture another The regulation and market structures of the
$0.6 billion a year of resources. road freight industry, rather than the quality
of the road infrastructure, are the binding con-
straints on performance in the international
Transport Services—the Forgotten corridors (Teravaninthorn and Raballand 2008).
Problem Although the associated road infrastructure is
generally in good condition, the administrative
Road sector interventions have traditionally bottlenecks at borders and ports keep the effec-
focused on constructing and improving the tive velocity of transit along these routes very
220 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
low (typically less than 10 kilometers an hour). Out-of-village travel by motorized transport,
Even for national traffic, the exceptionally high from villages to market towns and from towns
road freight tariffs in parts of Africa—reaching to cities, is less common. Supply is typically
$0.13 per tonne-kilometer in Central Africa— fragmented and informal, and rural com-
are attributable more to high profit margins munities are often captive markets for local
(60–160 percent) than to any inherent cost dis- monopolists.
advantage (table 10.4).
Marked performance differences occur Urban Road Transport
across subregions. Performance is worst in Buses are the common mode of public tran-
Central and West Africa and best in southern sit in most cities (Kumar and Barrett 2007).
Africa, with East Africa in between. The dif- Except in a handful of cases, however, mini-
ference can be explained by industry cartels in buses are much more prevalent than large
Central and West Africa, together with the tour buses. About twice as many trips are taken by
de role regulatory framework. That framework minibuses and shared taxis than by large
is based on market sharing and centralized buses. The use of motorcycles for commercial
allocation of freight that limits vehicle mile- transport has also grown very rapidly in recent
age and undermines incentives for improving years, mainly because of the poor state of the
fleet quality. In southern Africa, by contrast, roads and the inability of bus companies to
a much larger share of freight traffic is allo- meet growing demand. Small-scale suburban
cated through competitive bilateral contracts rail networks exist in a few cities, but nowhere
between clients and shippers. do they account for more than 2 percent of
the market.
Rural Road Transport Supply is inadequate and tariffs are high.
Most rural transport takes place near vil- Most African cities have 30–60 public trans-
lages. Trips generally involve short distances port vehicle seats per 1,000 residents, but only
and small loads carried on paths, typically 6 large bus seats per 1,000 residents (compared
for marketing, collecting water and firewood, with 30–40 in middle-income countries). Low
and tending crops and animals. Most trips fleet capacity is exacerbated by poor use of the
are walking trips. Nonmotorized transport, limited vehicle fleet, with vehicles achieving
such as a bicycle, is often unaffordable, and fewer than 200 kilometers a day. The quality
where mechanical transport is available to the of public transport is consequently poor, with
household, it tends to be appropriated by a long walking and waiting times typically dou-
male household head. Changing this practice bling the in-vehicle time. Extreme overcrowd-
may be a demanding cultural task; in some ing is also common, particularly on large
parts of Africa, women’s use of bicycles may buses. The average cost of a one-way trip, at
be considered unseemly. Nevertheless, when about $0.30, is high in relation to household
enough women begin to use them and the budgets. Regulations that keep fares for large
benefits to the whole household become buses below those of minibuses and inappro-
apparent, the practice may rapidly gain social priate cost benchmarks have contributed to
acceptance, as it has in parts of Burkina Faso. the demise of large buses.
Table 10.4 Overview of Key Road Freight Parameters on Main International Corridors
Trade density Implicit velocity Freight tariff
Roads in good ($ millions per (kilometers ($ millions per Profit
Corridor condition (%) kilometer) per hour) tonne-kilometer) margins (%)
Western 72 8.2 6.0 0.08 80
Central 49 4.2 6.1 0.13 70–160
Eastern 82 5.7 8.1 0.07 70–90
Southern 100 27.9 11.6 0.05 20–60
Source: Teravaninthorn and Raballand 2008.
Roads: Broadening the Agenda 221
will continue to formulate transport policy be achievable simply through spot improve-
and to regulate and oversee the new road insti- ments at vulnerable points (such as creeks and
tutions. If they do not perform these functions riverbeds) without the need to upgrade the
effectively, the new institutions are bound to surface of the road along its entire length.
suffer. For example, road funds will not get the Whatever the chosen objective for rural
revenue increases they need, and road agen- road development, financing will likely remain
cies will be unable to deliver realistic network a challenge. Local governments mobilize only
improvement programs. modest revenues of their own, with market and
business taxes as the main sources. Intergov-
Increasing Rural Accessibility ernmental transfers are thus the main source
Providing full road accessibility to 100 percent of domestic funding for local government
of Africa’s rural inhabitants would entail a vast spending in many countries. This situation
expansion in the all-season road network, vir- poses three main problems. First, throughout
tually tripling its length. For many countries, most of Sub-Saharan Africa, less than 5 percent
this goal is unlikely to be affordable in the of aggregate public revenue is generally made
medium term, highlighting the need to care- available to local governments managing rural
fully select and prioritize rural road invest- networks. Second, general budgets rarely allocate
ments. One way of doing so is to strategically adequate funds for maintaining main roads,
align rural road investments with agricultural much less rural roads. Third, capital and recur-
development programs at the national level, rent allocations to local governments are usually
to prioritize those rural roads likely to have not fungible, and the allocation for recurrent
the largest effect on agricultural productivity expenditures may barely cover the salaries of
and market access. Recent analysis of Central the rural road unit. Moreover, the budget cycle
Africa suggests that the most attractive rural dictates such transfers, so that central-to-local
road investments may lie in areas that are at transfers are unlikely to be adequate and timely
some distance from major urban markets but for maintaining local government roads.
still within reasonable reach, because in these Adequate, steady funding for local govern-
cases, rural road investments may be the criti- ment maintenance is more likely to be forth-
cal intervention needed to open up market coming from a dedicated road fund, as long as
accessibility (Briceño-Garmendia and others some formal commitment exists in the road
2009a, 2009b). By contrast, road investments fund law to ensure that it accepts responsi-
in very isolated rural areas may be less attrac- bility and provides for local roads. Although
tive because they would not make enough of 60 percent of road fund revenues is typically
a difference in overall travel times to provide allocated to the main interurban road network,
adequate market access. countries have, to varying degrees, channeled
Given the vast scale of Africa’s potential rural portions to the maintenance of rural road net-
road network, the issue of keeping down unit works. This approach appears to be effective.
costs becomes particularly critical. Policy mak- Countries that allocate at least $0.015 of their
ers face a stark trade-off between the standard fuel duties to rural roads have 36 percent of
to which rural roads are built and the length their rural networks in good condition, com-
of the rural network that can be developed for pared with 21 percent for those that do not.
any given budget envelope. This choice raises Building full capacity for all management
questions about what kind of rural roads farm- functions in each local government and com-
ers really need. In many rural communities, munity is unrealistic and inefficient. Individ-
volumes of production may be well below the ual local government networks are small, and
threshold needed to justify the use of a truck the management contract for an individual
to collect produce, and simpler roads targeted local government may be too modest to attract
more at ensuring accessibility for two-wheeled competent consulting firms. In Madagascar,
vehicles or animal-drawn carts may be more the average network for a local government is
suitable (Raballand and others 2009). In other 140 kilometers; in Cameroon and Nigeria, it is
settings, ensuring all-season accessibility may 180 kilometers; and in Tanzania and Zambia,
Roads: Broadening the Agenda 223
it is 280 kilometers. All fall far short of the Community infrastructure, including mainly
500–2,000 kilometers needed to justify unclassified roads and paths for which no level
employing an engineer in a local unit. Joint of the formal government accepts responsi-
service committees of local authorities can bility, faces particular problems. Community
achieve economies of scale in procurement for contributions in cash and kind are suitable
the authorities they represent, but they usually primarily for community roads and paths, but
require substantial technical assistance from in-kind contributions may be inefficient, and
central ministries or from the regional offices other sources of money are necessary. Cost
of a main road authority. In countries with an sharing between local communities and other
autonomous road authority responsible for government or external agencies can help raise
main roads, local governments can contract the volume of resources mobilized and thereby
with the road authority to manage the roads increase the proportion of the network that
on their behalf or to assist with planning and receives regular maintenance. Well-structured
procurement. donor financing through rural road projects or
Some countries centralize the technical through social and community or rural infra-
responsibility for rural roads. Relying on a road structure funds can support investment in com-
ministry or another central ministry to man- munity infrastructure. Cost sharing may also
age rural roads has the advantage of a formal be effective in maintaining community roads.
channel for technical support. However, the Many local authorities in Africa have more
disadvantage is that the ministry often operates roads to maintain than they can afford, so cost
completely independently of the local govern- sharing with communities has merit.
ment structure and thus is poorly connected Lack of technical know-how often impedes
to local needs and developments. In principle, community management. Communities in
a central coordinating unit for local govern- Sub-Saharan Africa need technical advice (on
ment roads should perform as well as a central road design and standards, appropriate materi-
government rural roads department. In prac- als, work planning) and managerial advice (on
tice, however, coordinating units for local gov- financial accounting, contract management,
ernment roads are not always as strong as they procurement) to perform the responsibilities
need to be, as is illustrated by the experience of that come with ownership.
Tanzania and Zambia in the late 1990s. For rural transport services, the main issues
Delegation of planning, procurement, and are increasing service quantity and keeping
management has already been improved in services affordable. The priority in transport
many countries through the establishment of services in rural areas must be to maintain
AGETIPs, and the national institutions have basic year-round vehicle access for the types of
benefited from their association in AFRICA- vehicles likely to be operating. The quantity of
TIP and from donor assistance. Considerable access is even more important than the quality.
room for improvement exists in the work Better rural telecommunications can provide
of the AGETIPs, however, particularly with the means of more effectively matching vehi-
respect to their technical capacity, the quality cles to loads. Given the monopoly power of
of preparatory studies and contract supervi- service providers, communities can organize
sion, and delays in project implementation to increase their bargaining power through a
(Diou, Henry, and Demy 2007). collective lobby. Operating subsidies are usu-
The options are not mutually exclusive. ally infeasible, but providing credit for vehicle
For example, a joint services committee can purchases, possibly through piggybacking on
use private consultants, hired through a con- agricultural credit programs, is an option.
tract management agency. The best option for
managing local roads depends on many local Developing Urban Transport Services
factors, including the size of the authorities, Urban public transport requires simultane-
the nature of the network, and the compe- ous and integrated attention to planning
tence of the private sector or higher public urban structures, building and maintain-
authority units. ing infrastructure, and organizing transport
224 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
services. In practice, these three primary resources to reestablish a large-bus sector while
functions are seldom housed in the same the private sector remains wary of investing in
institution, and even where they remain with large vehicles. The short-term options are thus
the central government, several ministries to rely on self-regulation of the sector, which
are usually involved. Only a handful of has usually failed, or to devise competitive
African cities have agencies with metropolitan structures, either “for” or “in” the market to
responsibilities and overarching functions, build private confidence in a managed private
and even those agencies lack the executive market. Attempts can also be made to stimu-
powers to implement their vision and must late increased vehicle size, as in South Africa.
work through other units of government. This sort of strategy has achieved limited
The institutional arrangements for urban success in some central Asian countries.
roads are frequently complex. Legislation In the very largest cities, exclusive road-
pertaining to roads is usually separate from based track systems, such as bus rapid transit
that governing transport services, and several or (more costly) light rail, may have a role. Such
national and local bodies often share juris- developments are being considered in Dar es
diction. In Conakry, Guinea, several institu- Salaam in Tanzania and Lagos in Nigeria, but
tions are responsible for segments of the road they are still experimental.
network. In Accra, Ghana, responsibility for
urban transport has been devolved from cen- Liberalizing Road Freight Transport
tral to local government—at least in principle. Freight tariffs in much of Africa are unneces-
However, local governments have neither the sarily high because of restrictive regulation
resources nor the technical know-how to carry and weak competition among truckers. The
out their assigned functions, so the ministry most damaging aspect of trucking organi-
of transportation (through the department of zation is the combination of self-regulation
urban roads) is effectively responsible for road with national protection. Both militate for the
maintenance and development. interests of the incumbent national operators
In the passenger transportation market, at the expense of their customers. They cre-
the self-regulation of operators associations ate scope for corruption while leaving socially
and cartels has ossified fragmentation of costly problems (vehicle overloading) relatively
informal operations and wasteful institution- untouched. Both areas offer alternatives.
alized procedures. Supply is inadequate and Self-regulation is a means of maintaining
expensive. Two main options exist to remedy on-the-road discipline in an excessively frag-
this situation: small and medium-size vehi- mented market. It fills a vacuum created by the
cles and, in the largest cities, light rail. absence of effective public regulation. By its
In principle, traditional, disciplined large- nature, it concentrates on ensuring an equita-
vehicle services could be reintroduced, but ble distribution of traffic between members of
such attempts—in Dakar (Senegal), Accra, the association, typically through the wasteful
and other cities—have failed. The association operating procedure of tour de role dispatching.
of large vehicles with traditional large public The alternative is a combination of freedom of
monopoly companies made them vulnerable entry and market pricing, with independent
to political intervention and to the failures enforcement of rules on quality and operating
in cost control that destroyed them in the behavior, as in efficient road-haulage markets
first place. Moreover, although subsidies may in Europe and the United States.
be required to sustain service levels and fare National protection appears to secure a
aspirations, open-ended subsidies of a public “fair” share of traffic for the haulers of each
operator will almost certainly pass the benefit of the national partners in a transit market. It
of subsidies to managers and employees rather operates through quotas that reduce the use of
than to passengers. vehicles and thus increase the costs. It is often
For some time, the small and medium-size supported by enforcing regulated rates, which
vehicles must be part of the structure, not deny the shipper the opportunity to shop
least because few governments have the fiscal around for a better deal. The alternative is to
Roads: Broadening the Agenda 225
combine free entry to the market with rigorous Dealing with Escalating Unit Costs
enforcement of national safety and operational The recent escalating costs of roadwork can
behavior rules in all countries. Regulatory sys- be attributed to rising input costs against a
tems would combine strict quality control backdrop of growing demand for contracting,
with liberal approaches to pricing and market which appears to have been exploited in an
entry. Moving in this direction would include environment of generally low competition for
developing internationally agreed strategies to contracts. No one solution exists. Inflation in
improve the range of elements on the main input costs lies beyond the control of policy
transit corridors and to strengthen enforce- makers, but they can take other measures.
ment on overloading. A key issue is to ensure effective compe-
Breaking the regulatory status quo in Cen- tition for contracts. Road agencies should
tral and West Africa is difficult because of a actively market contracts to obtain a set of
coalition of interest groups opposing change. good bidders. If at any stage in the bidding
Truckers have strong leverage on high-level a competitive choice set of bidders does not
authorities because they have enough monop- surface (say, at prequalification), something
oly power to block trade. Governance issues is seriously wrong, and the agency should
also intrude because some high-level authori- consider postponing the process until it has
ties own or indirectly control trucks or truck- identified and corrected the underlying issues.
ing companies and thus benefit from the status Continuing the bid without a proper choice
quo and current market-sharing schemes. set in the hope of achieving an acceptable bid
Deregulating the trucking industry in Cen- price is an unnecessary gamble.
tral and West Africa is thus more of a politi- A better understanding of the underlying
cal and social challenge than a technical one. cost trends and their links to contract pric-
The main concern is the potential reduction ing is also critical. Although cost inflation
in the number of trucks to match demand in lies beyond the control of sector authorities,
road transport. That reduction could lead to they can increase the accuracy of the design
a drop in trucking employment and profits, cost estimates, improve the allowance for
because some companies (or owner-operators) cost fluctuations, and monitor cost increases
would disappear and others would shrink, and through the procurement period. To this end,
these social effects would need to be mitigated. agencies need to understand the cost struc-
Some chance exists that the coalition of inter- ture underlying road contracts more clearly
est groups opposing change in the transport and to track international price trends for key
market in most countries in Central and West inputs over time.
Africa might not resist reforms if compensation The capacity of project-executing agencies
schemes pay, at least partly, for the social costs. also needs to be strengthened to support the
The southern African international trans- timely implementation of contracts. Delays
port market is a good model for the rest of often result from deficiencies in the planning
the continent because it combines liberalizing and procurement of sector agencies, making
entry with enforcing quality and load-control this a third area for attention.
rules applicable to all operators. Operations Whatever the improvements to road agency
to and from South Africa are governed by procurement processes, the unit costs of road
bilateral agreements that provide for sharing infrastructure are likely to remain on an
information on traffic development and define upward path, straining already limited sector
the types of permits that can be issued. This budgets. Beyond measures to improve procure-
system restricts the carriage of bilateral trade ment, considering how to design roads to keep
to operators from the two countries concerned costs down is important. Overengineering of
and prohibits cabotage.2 It does not establish roads—beyond the surface type needed for the
quotas, however, and it allows rates to be deter- anticipated traffic volume—is an issue in parts
mined by the market to enable direct contract- of Africa and represents a waste of resources
ing between shippers and transporters and that should be avoided. Careful economic
giving incentives to efficient operators. analysis of road investments can avoid the
226 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Public-Private Infrastructure Advisory Facility, Stankevich, Natalya, Navaid Qureshi, and Cesar
and AFRICATIP. http://www.africatip.net/fr/ Queiroz. 2005. “Performance-Based Contract-
publications/downloads/2008-11-11%2006: ing for Preservation and Improvement of Road
28:29/Rapport_AGETIP_MOD_vfinale.pdf. Assets.” Transport Note TN-27, World Bank,
Gwilliam, Ken, Vivien Foster, Rodrigo Archondo- Washington, DC.
Callao, Cecilia Briceño-Garmendia, Alberto Starkey, Paul, John Hine, Simon Ellis, and Anna
Nogales, and Kavita Sethi. 2008. “The Burden Terrell. 2002. “Improving Rural Mobility:
of Maintenance: Roads in Sub-Saharan Africa.” Options for Developing Motorized and Non-
Background Paper 14, Africa Infrastructure Motorized Transport in Rural Areas.” Technical
Country Diagnostic, World Bank, Washington, DC. Paper 525, World Bank, Washington, DC.
Harral, Clell, and Asif Faiz. 1988. Road Deteriora- Teravaninthorn, Supee, and Gael Raballand. 2008.
tion in Developing Countries. Washington, DC: “Transport Prices and Costs in Africa: A Review
World Bank. of the Main International Corridors.” Working
Kumar, Ajay, and Fanny Barrett. 2007. “Stuck in Paper 14, Africa Infrastructure Country Diag-
Traffic: Urban Transport in Africa.” Background nostic, World Bank, Washington, DC.
Paper 1, Africa Infrastructure Country Diagnos- Van Zyl, Willem, Lynette Coetzer, and Chris
tic, World Bank, Washington, DC. Lombard. 2008. “Unit Costs of Infrastructure
Raballand, Gaël, Somik Lall, Arnaud Desmarchelier, Projects in Sub-Saharan Africa.” Background
and Patricia Macchi. 2009. “Economic Paper 11, Africa Infrastructure Country Diag-
Geography and Aid Effectiveness in Transport nostic, World Bank, Washington, DC.
in Sub-Saharan Africa.” Report, Transport World Bank. 2007. “Road Maintenance Initiative
Department, Africa Region, World Bank, Matrix.” Sub-Saharan Africa Transport Program,
Washington, DC. World Bank, Washington, DC.
Chapter 11
A
frican railroads have changed greatly stock. Moreover, various conflicts and wars
in the past 30 years. Back in the 1980s, have rendered several rail sections unusable.
many railway systems carried a large As a result, some networks have closed and
share of their country’s traffic because road many others are in relatively poor condition,
transport was poor or faced restrictive regu- with investment backlogs stretching back over
lations, and rail customers were established many years.
businesses locked into rail either through Few railways are able to generate signifi-
physical connections or (if they were para- cant funds for investment. Other than for
statals) through policies requiring them to use purely mineral lines, investment has usually
a fellow parastatal. Since then, most national come from bilateral and multilateral donors.
economies and national railways have been Almost all remaining passenger services fail
liberalized. Coupled with the general improve- to cover their costs, and freight service tariffs
ment in road infrastructure, liberalization has are constrained by road competition. More-
led to strong intermodal competition. Today, over, as long as the railways are government
few railways outside South Africa, other than operated, bureaucratic constraints and lack
dedicated mineral lines, are essential to the of commercial incentives will prevent them
functioning of the economy. from competing successfully. Since 1993, sev-
Rail networks in Africa are disconnected, eral governments in Africa have responded by
and many are in poor condition. Although concessioning their systems, often accompa-
an extensive system based in southern Africa nied by a rehabilitation program funded by
reaches as far as the Democratic Republic of international financial institutions.
Congo and East Africa, most other railways For the most part, concessions have
are disconnected lines reaching inland from improved operational performance. Although
the ports, serving small markets by mod- results have been mixed, many concessionaires
ern railway standards. Most were built rela- have increased traffic volumes and have gener-
tively lightly, and few, other than Spoornet ally performed more efficiently, and there has
in South Africa, have invested in rehabilitat- been little evidence of monopolistic behavior.
ing and renewing infrastructure and rolling Relations with governments have often been
229
230 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
uneasy, however, especially concerning ade- followed a similar pattern in almost all Afri-
quate compensation for loss-making passen- can countries. Typically, isolated lines headed
ger service obligations, and many governments inland from a port to reach a trading center
clearly had unrealistic expectations about the or a mine, and over time, a few branch lines
private sector’s ability to improve operations were built. Many of the lines were state owned,
and generate investment. but some were constructed as concessions or,
Concessionaires appear willing to spend their in the case of some mineral developments,
own funds only on day-to-day maintenance, not as part of a mining company’s operation.
on infrastructure. Financing asset renewal and Although continental rail master plans have
upgrading remains an open question for most existed for over a century, most of the Afri-
of the African rail network. Without infrastruc- can network remains disconnected, operating
ture investments, the competition from road within a single country or linking a port and
networks will thwart railway survival except to its immediate regional hinterland. The only
carry large-scale mineral traffic. Although con- significant international network is centered
cessioning has generally improved service and in South Africa and stretches north to Zim-
reduced the financial burden on governments, babwe, Zambia, and the Democratic Republic
it does not appear to be a full solution to financ- of Congo (figure 11.1). Trade between African
ing the investment needs of African railways. countries (other than to and from South Africa)
has always been minimal, largely because of the
similarity in the products exported, which sug-
Africa’s Rail Networks gests that interregional links would be lightly
used even if they existed.
At the end of 2008, 47 railways were operating in
32 countries in Africa. Railway development has Low Rail and Traffic Density
African railway networks’ spatial density, a
Figure 11.1 Map of African Rail Networks metric that compares track mileage with the
size of a country, is low (UIC 2008).1 The
highest measurement of spatial density is 16
in South Africa, but most other countries fall
in the range of 1 to 6, and 13 countries have no
operating railway at all. Too much should not
be read into this indicator, however; network
PORT SUDAN density is strongly affected by the pattern of
population. Australia, Canada, China, and the
DAKAR
DJIBOUTI Russian Federation, all with vast undeveloped
CONAKRY and sparsely populated areas, have densities
LAGOS of between 5 and 7, whereas most European
LOME PORT
ABIDJAN TEMA HARCOURT DOUALA countries range from 20 to 100.
A complementary indicator is the net-
POINTE NOIRE MOMBASA work density per million inhabitants, which is
LUANDA
DAR ES SALAAM highest in Gabon (520) and Botswana (480),
followed by South Africa (460). Most other
African countries range from 30 to 50. Euro-
TOAMASINA
BEIRA
pean countries range from 200 to 1,000, and
WALVIS BAY Australia and Canada exceed 1,500. China is
MAPUTO much lower, at 50.
railway DURBAN These metrics alone cannot justify network
major port CAPE TOWN EAST LONDON expansion in Africa. To be an economical
PORT ELIZABETH investment, a new line needs a minimum level
of traffic, and the geographical distribution of
Source: Bullock 2009. potential customers within a country and the
Railways: Looking for Traffic 231
passenger-km (billions)
70 10
60
percent
8
50
40 6
30 4
20
2
10
0 0
route-km passenger-km net ton-km southern Africa Central Africa East Africa West Africa
southern Africa Central Africa East Africa West Africa nonurban suburban
35 120
route-km (thousands)
30
net ton-km (billions)
100
25
80
20
60
15
40
10
5 20
0 0
southern Africa Central Africa East Africa West Africa southern Africa Central Africa East Africa West Africa
general mineral general mineral
level of usage that can be expected are more Traffic density on African railways is gen-
important than these national averages. erally low.2 The highest average network
South Africa has the most important net- traffic density outside Spoornet is in Gabon
work (figure 11.2). Specialized mineral lines in (2.7 million traffic units), with Cameroon and
western and southern Africa carry over half of Swaziland having the only other railways over
the region’s freight, most of it on the Spoor- 1 million; many railways average fewer than
net coal and ore export lines. Southern Africa 300,000 units (figure 11.3). By comparison, the
dominates general rail freight, handling over average traffic density of the Maghreb systems
80 percent of the freight traffic on the non- (Algeria, Morocco, and Tunisia) is nearly 2 mil-
mineral lines. Southern Africa also dominates lion units, and the Arab Republic of Egypt, with
the passenger business, with over 70 percent of its heavy passenger traffic, exceeds 8 million.
passenger traffic, largely because of its heavy Most European systems average 2 million to
commuter passenger business in cities. Some 5 million, with densities under 1 million found
other African cities also operate commuter ser- only in Albania and Montenegro. With such
vices, but with the exception of Dakar, Senegal, light usage, many networks struggle to gener-
they mostly provide one or two trains at peak ate enough funds just to maintain, much less
hours along a short line. renew, their infrastructure.
232 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 11.3 Average Railway Network Traffic Density, 2001–05 low-speed, small-scale, undercapitalized net-
works ill suited to modern requirements. Many
Madarail, Madagascar
structures and some of the track work are now
CEAR, Malawi over 100 years old. Many sections of track have
CDN, Mozambique deteriorated almost beyond repair. Although
CCFB, Mozambique this situation can be tolerated on low-volume
RVRC-URC, Uganda feeder lines, and indeed may be the only way
concessioned
Transrail, Mali some can be viably operated, it is a major
RSZ, Zambia
handicap when competing against the modern
Sitarail, Burkina Faso and Côte d’Ivoire
roads being constructed in major corridors.
Most rail systems have considerable sections
Camrail, Cameroon
of track in need of repair or replacement. Some
TRC, Tanzania
have major sections that are not in operation
RVRC-KRC, Kenya and will require rehabilitation before operations
SETRAG, Gabon can resume. Even where service exists, poor track
condition forces speed restrictions, resulting in
FCE, Madagascar lower railway competitiveness and rolling-stock
NRC, Nigeria productivity.
CFMK, Dem. Rep. of Congo In some countries, parts of the network are
OCBN, Benin
not operated because of war damage, natu-
ral disaster, or general neglect. Much of the
CDE, Ethiopia
Mozambican central and northern networks
GRC, Ghana
and railways in Angola, Côte d’Ivoire, Eritrea,
state owned
and central Africa uses the Cape gauge, which Traffic—Low and Growing Slowly
is also used in some anglophone countries far- Outside South Africa, the traffic volumes
ther north. The meter gauge is used in most of serviced by African railways are very small;
francophone Africa and much of East Africa. about half of the 26 railway operators sur-
A number of isolated standard-gauge lines are veyed carried traffic of less than 500,000
used primarily for mineral traffic, although traffic units annually, while only 5 of them
Nigeria is developing a new standard-gauge exceeded 1 million traffic units annually—a
network to serve its capital, Abuja. Narrow- volume comparable to a moderately busy
gauge lines have operated at various times, branch line on other railways (figure 11.4).
but most are now derelict. Apart from the By comparison, Spoornet in South Africa
network in East Africa and the one extending carries 1 million traffic units every three days
north from South Africa, few railways cross (Thompson 2007). In some cases, the light
international borders. Instead, they reach rail- traffic is caused by a lack of demand; in oth-
heads from which traffic can be carried farther ers, it is caused by shortages of rolling stock,
by road. particularly locomotives.
Despite the multiplicity of gauges, interop- Although the average haul on African net-
erability is not a major problem in Africa. Two works is relatively long with regard to their
gauges exist in the same location in only three size, it is not especially so vis-à-vis road trans-
places—two in Tanzania and one in Guinea. port. Some railways carry mostly end-to-end
However, mixed gauges will become a problem traffic; Tanzania Railways Corporation, Tazara
if some of the proposed connecting lines are (Tanzania-Zambia Railway Authority), and
constructed. Transrail (Dakar-Bamako Railway) all haul
In summary, most African railways are con- freight an average distance of 1,000 kilometers,
fronting major infrastructure problems pri- and some smaller railways, such as Uganda
marily associated with aging track: insufficient Railway or CEAR (Central East African Rail-
ballast, rail wear, deteriorating earthworks and ways), act as feeders to other systems, which
formation, decrepit structures, and rail signal- carry the traffic a few hundred kilometers
ing and telecommunications with obsolete farther. These systems have a good chance of
equipment and lack of spare parts. The cost of competing for general freight traffic, even as a
rehabilitating the networks is large compared road network improves, as long as satisfactory
with the existing traffic volumes and revenues. service levels can be achieved, but the shorter
The means by which rehabilitation can be systems that require transshipment to road at
done on a sustainable basis is the central ques- railheads will generally find they can compete
tion faced by most African railways. effectively only for bulk traffic.
Most systems operate only limited passen-
ger commuter services, if any, and the aver-
The African Rail Market age distance of passenger trips is the distance
between the capital of a country and major
Typically railways in Africa are small, carrying provincial centers. The only significant cross-
no more traffic than a moderately busy branch border flows are on the Sitarail (Côte d’Ivoire),
line in other parts of the world. African rail- Tazara, and Transrail networks.
ways carry far more freight than passengers, Since the mid-1990s, most African coun-
with freight averaging about 80 percent of tries experienced steady economic growth.
traffic between 1995 and 2005. Almost all rail- Average annual GDP grew 4 percent, with cor-
ways carry passenger traffic; only Swaziland responding increases in trade. Per capita GDP
and Uganda have freight-only railways. The grew by about 1.5 percent a year. Countries
passenger business is steadily shrinking, how- such as Mali, Mozambique, and Tanzania that
ever, and several of the railways still retaining avoided political upheaval grew as much as
a reasonable passenger business do so only 50 percent faster. Despite the generally favor-
because competing road networks are in poor able economic background, only four African
condition or do not exist. railways increased both their passenger and
234 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 11.4 Average Railway Traffic Volumes, 2001–05 have been concessioned, freight traffic has gen-
erally increased, whereas passenger traffic has
Madarail, Madagascar
generally stagnated or declined.
CEAR, Malawi The growth or decline of traffic on many
CDN, Mozambique systems over the last decade often had little to
CCFB, Mozambique do with changes in the underlying demand.
RVRC-URC, Uganda War or natural disaster has had a major effect
concessioned
Transrail, Mali in some cases; on other railways, the volume
RSZ, Zambia carried reflects the availability of rolling stock,
Sitarail, Burkina Faso and Côte d’Ivoire
particularly locomotives. Many railways are
short of locomotives. When this situation
Camrail, Cameroon
improves with new or secondhand locomo-
TRC, Tanzania
tives or through a locomotive rehabilitation
RVRC-KRC, Kenya
project, traffic will increase accordingly.
SETRAG, Gabon
0
0
40
60
80
00
20
40
60
80
00
1,
1,
1,
1,
1,
2,
traffic units (millions) more frequent. Buses have the lion’s share of
net tonne-km passenger-km the market, although they suffer from the same
problems as rail: unreliable departures, delays
Source: Bullock 2009. and breakdowns, and overcrowding.
Note: Traffic units are passenger-kilometers in the case of passenger traffic and net tonne-kilometers in
the case of freight traffic. BR = Botswana Railways; Camrail = Cameroon Railway Corporation; CCFB = The long-term prospects for nonurban
Companhia dos Caminhos de Ferro da Beira (Mozambique); CDE = Chemin de Fer Djibouto-Ethiopien; rail services are generally poor (Amos and
CDN = Corredor de Desenvolvimento do Norte (Mozambique); CEAR = Central East African Railways
Corporation (Malawi); CFCO = Chemin de Fer Congo-Océan (Republic of Congo); CFMK = Chemin de Bullock 2007). Rail services start competing
Fer Matadi-Kinshasa (Democratic Republic of Congo); CFM = Caminhos de Ferro do Mocambique;
FCE = Fianarantsoa Côte Est (Madagascar); GRC = Ghana Railways Corporation; NRC = Nigeria Railways
with roads at speeds higher than 70 kilometers
Corporation; OCBN = Organisation Commune Bénin-Niger; RSZ = Railway Systems of Zambia Ltd; RVRC- per hour. However, the cost of maintain-
KRC = Rift Valley Rail Corporation-Kenya Railways Corporation; RVRC-URC = Rift Valley Rail Corporation-
Uganda Railways Corporation; SETRAG = Société Transgabonnaise (Gabon); SNCC = Société Nationale ing track and signaling systems that would
des Chemins de Fer du Congo (Democratic Republic of Congo); SR = Swaziland Railways; SRC = Sudan enable these commercial speeds is signifi-
Railways Corporation; TAZARA = Tanzania-Zambia Railway; TRC = Tanzania Railways Corporation.
cantly more than the cost of maintaining the
30- to 40-kilometer-per-hour commercial
freight traffic over the period, two of which speed needed for a freight railway. In addi-
had been concessioned. One other railway saw tion, a very large capital investment would
an increase in average passenger traffic, and be required to construct new medium-speed
all others saw a reduction. Fifteen railways (for example, 200 kilometers per hour) inter-
increased their freight traffic. Where railways urban railways. Such investment is justified
Railways: Looking for Traffic 235
only on the basis of substantial demand (sev- differences in the commodity mix, with many
eral million passengers a year) and relatively requiring specialized cars, mean freight trains
high-income passengers who can afford to are rarely fully loaded in both directions. In
cover at least operational costs. Few, if any, some cases, this natural imbalance in traffic is
corridors in Africa could justify such invest- accentuated for rail because road vehicles deliv-
ments, at least for the medium term. ering imports tend to backload freight at mar-
Formal compensation schemes, such as pub- ginal cost, leaving rail to transport the remaining
lic service obligations, have been introduced in freight without a compensating return load.
a few cases to support passenger rail services, Average freight tariffs range from $0.03 to
but they rarely provide timely compensation $0.05 per net tonne-kilometer, similar to tariffs
for service operations. Payment may be delayed on other general freight railways in compara-
several years or may otherwise take the form ble countries. Tariffs are generally constrained
of a subsidy calculated to break even, limiting by competition, either from road or alternate
the ability of railways to increase their main- routes (particularly in the Great Lakes region,
tenance and negating any attempts to improve Malawi, West Africa, and Zambia) and are also
the financial performance of the freight ser- influenced by the traditional value-based tar-
vices. As a result, most long-distance passenger iff structures, the relative cost of carrying dif-
services in Africa are trapped in a cycle of mini- ferent commodities (as reflected in net tons
mal investment, deteriorating services, declin- per railcar round-trip), direction of travel,
ing patronage, and financial losses. and volume. Although most rail rates are well
The few instances in which local trains serve below comparable road rates, especially for
villages with no road connection pose a differ- containers, rail typically carries only 20–50
ent problem. These trains are used by traders percent of the traffic in a corridor, and some
bringing goods to and from regional centers, of the smaller state-owned railways have an
and although heavily loaded with passengers, even smaller share.
they nonetheless incur major losses. Although Line-haul tariffs are only part of the cost
such services can be funded through govern- equation for freight traffic. Much is often
ment subsidy, the long-term solution is to cre- made of the inherent lower cost of rail com-
ate feeder roads for motorized access, enabling pared to road. This is true where minerals
more cost-effective means of transporting must be transported from a rail-connected
goods and greatly improving accessibility to mine to a rail-connected port but is not so
such locations. clear for medium-distance general freight that
also must be transported by road to and from
Freight—Needs Improving railheads. Haulage between the railway and
Freight traffic on railways is mostly bulk and the ultimate origin and destination can be
semibulk commodities, principally to and from surprisingly expensive, often as much as the
ports. The actual commodities transported by equivalent of 200–300 kilometers of line-haul
rail reflect the economic structure of countries transport, negating any advantage rail may
served by the railway, with mining products have in pure line-haul tariffs. New sidings are
important in several countries and timber and sometimes constructed, but they need a certain
export crops important in West Africa. Imports amount of traffic to be economical. Traffic that
are mostly manufactures, such as cement and needs to be collected at a central depot before
petroleum products, and general freight. On being dispatched by rail is more vulnerable to
some systems, much of the general freight is road competition, and even bulk traffic is not
containerized (cash crops with high value are immune if distances are not too long. In many
increasingly traveling this way), particularly countries, collection and distribution chains
when the trip involves crossing an intermedi- are being streamlined, often eliminating up-
ate border before reaching the port. Unlike pas- country depots and distribution centers, and
senger services, significant imbalances between marketing channels have become more diver-
traffic in the two directions are common. Even sified. The railways have often been slow to
where tonnage is approximately balanced, the respond, steadily losing market share.
236 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Level of service is a key factor in the freight to merely a line-haul operation, can reduce the
business. For rail to play a significant role in price discount between rail and road, increas-
the general freight transport system, it must ing the contribution that freight can make to
improve its service (specifically, overall tran- the maintenance and renewal of infrastructure.
sit time, reliability, security, and service fre- This improvement is one of the major benefits a
quency) and ensure that it is addressing the concessionaire can offer a state-owned railway.
needs of customers. Too often, what rail has Moreover, because of the lack of inter-
offered as transport has been quite different connection services and cross-border service
from what the competing road hauler can contracts, rail freight suffers huge delays in
offer, and road carriers can charge a significant crossing national borders. For example, a
premium. In general, freight markets in Africa rail freight journey of 3,000 kilometers from
require reliable services (a commercial speed Kolwezi on the Democratic Republic of Congo
of 40 kilometers per hour is usually sufficient) border to the port of Durban in South Africa
rather than high-speed services, with (a) rail takes 38 days to complete, an effective speed
infrastructure and rolling stock maintained for of only 4 kilometers per hour. Only 9 of these
service, (b) operating discipline to ensure that days are spent traveling, with the remainder
schedules are maintained, and (c) commercial (a staggering 29 days) taken up primarily with
arrangements that ensure that customers fulfill loading and interchanging freight, as well as
their contractual responsibilities. some time for customs clearance. Each day of
Most railways can win bulk mineral traffic delay costs $200 per railcar. The main cause of
when it is offered, but general freight requires the problems in the rail sector is the absence
a reasonable level of service from rail if it is of reliable interconnection services when
to compete with road without offering a sig- trains cross borders. Locomotives from one
nificant price discount. By 2025, any remain- country are currently not allowed to travel on
ing monopolies for general freight will have another country’s network, mainly because of
run their course, and the only traffic on which the inability to provide breakdown assistance
African railways will have an undisputed grip to foreign operators. As a result, rail freight
will be minerals (although mining compa- crossing borders must wait to be picked up
nies are increasingly running even this traf- by a different locomotive. The delays are often
fic directly, either as third-party operators or extensive, partly because of the lack of reli-
on their own private networks). Experience in able, well-maintained locomotives. Delays also
many countries has shown that general freight reflect the lack of clear contractual incentives
transport requires operators to be flexible, to service traffic from a neighboring country’s
responsive, and adaptable. Fewer custom- network. Reducing such delays would there-
ers are fellow parastatals under order to use fore require totally rethinking the contractual
a state-owned railway, and few government- relationships and access rights linking the rail-
owned organizations, no matter how cor- ways along the corridor. It would also likely
poratized they may be, have the commercial require the establishment of a regional clear-
freedom to operate effectively in a fully com- inghouse to ensure transparency and fairness
petitive environment. in reciprocal track access rights.
Rail in Africa must become a transport busi-
ness in the broadest sense and must be able to
adapt to new markets. The predicaments of the How Much Investment Can Be
remaining government-owned railways, how- Justified?
ever, show that rail cannot compete effectively
while it is handicapped by the bureaucratic con- Providing an estimate of the investment needed
straints and lack of commercial incentives and by African railways is a daunting task (Carruthers,
accountability of a government organization. Krishnamani, and Murray 2009). In addition to
Achieving an acceptable level of service, com- building detailed inventories and assessments
bined with flexible pricing policies and a strat- of infrastructure and determining how much
egy of providing a transport service as opposed needs to be repaired or replaced, the question of
Railways: Looking for Traffic 237
how much investment is economically justified by using assumed average asset lives. Excluding
must be asked. Lines that have been superseded South Africa, the Sub-Saharan network carries
by road developments and those with low traffic about 15 billion net tonne-kilometers a year,
levels will rarely merit reconstruction and invest- excluding the mineral lines, and about 4 billion
ment, and funds should instead be directed to passenger-kilometers. That level of traffic will
those parts of the network with long-term value. require, on average, replacing 500 freight cars,
Although a government’s desire to reinstate 20 passenger cars, and about 20 locomotives
such links is understandable, doing so is often a year. As with infrastructure, much of that
extremely expensive. stock will be secondhand (from India or South
Investment has historically been used Africa), but the estimated cost will still average
for new construction and rolling stock, for about $80 million a year, equivalent to about
replacement of rolling stock, and sometimes $0.04 per net tonne-kilometer or passenger-
for rehabilitation and replacement of track. kilometer. The steady-state investment in the
Long-term maintenance neglect has caused a African network north of South Africa should
huge backlog investment of up to $3 billion thus be about $200 million a year (allowing
for Africa’s railways. In practice, this one-time $20 million for facilities, maintenance, equip-
expenditure needed to eliminate the rehabili- ment, and other costs).
tation backlog could be spread over a 10-year That amounts to a combined annual pro-
period at an annual rate of $300 million. gram of about $500 million for 10 years, after
After the network is restored to good con- which investment would drop to the steady-
dition, the annual bill would fall substantially state level of $200 million (Bullock 2009). The
to cover only what was needed for ongoing $500 million a year requirement refers to the
track rehabilitation and renewal. Excluding period during which the rehabilitation back-
South Africa, the Sub-Saharan network con- log is being cleared. These calculations are only
sists of about 44,000 kilometers of track, of broad order-of-magnitude estimates. However,
which about 34,000 kilometers is operational. the amount needed to overcome these prob-
The infrastructure on this network will have lems is large, equal to the annual revenues of
a life of at least 40–50 years, given the gener- some of the railways and well beyond their
ally low traffic volumes; the cost of periodic capacity to self-finance. The only option in
reconstruction (about $350,000 per kilome- most cases is to seek large concessional loans
ter) is thus equivalent to an annual cost of or grants from third parties.
about $8,000 per kilometer. Few lines with In addition to reinvestment in the cur-
an average density of fewer than 1 million rent network, investment in new projects is
net tons a year are likely to warrant this kind a possibility. For years, proposals have been
of major rehabilitation expenditure, because floated to create new routes for landlocked
traffic would need to earn $0.08 per net countries and to integrate the isolated net-
tonne-kilometer to fund the reconstruction, works. The most comprehensive proposal was
whereas typical rail freight tariffs are no more the 1976 master plan of the Union of African
than $0.05 per net tonne-kilometer. Lines with Railways for a pan-African rail network that
a density under 250,000 tons a year probably included 26,000 kilometers of new construc-
cannot support anything more than routine tion. Designed to create a grid to support
maintenance. Even if low-volume lines are intra-African trade development and regional
reconstructed using cheaper, secondhand economic integration, the plan was approved
materials, this level of expenditure is unlikely by the Organization of African Unity in 1979,
to be justified for more than 20,000 kilometers but few, if any, of the proposed links have
of the network. Overall, the ongoing annual gone beyond the drawing board. The Union
cost of track reconstruction would thus aver- of African Railways is now concentrating on
age approximately $100 million a year. a revised plan containing a subset of 10 cor-
Sustaining an adequate fleet of rolling stock ridors, some of which are already partially
will cost an additional $80 million a year. The constructed, and the proposal has generated a
cost of replacing rolling stock can be estimated number of regional studies and action plans.
238 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Several proposals for individual segments have countries suffering or recovering from civil
been made, and mining companies have pro- disruption (Angola, the Democratic Republic
posed a number of dedicated mineral lines. of Congo, and Zimbabwe), most countries are
Few of these projects will be financially or at various stages of reform. Of the 30 African
economically viable. The cost of new construc- countries with publicly owned railways, 14 have
tion of a single-track, nonelectrified railway on opted for a concession arrangement and 1 oper-
relatively flat terrain is at least $1.5 million per ates under a management contract (figure 11.5).
kilometer, increasing to about $5 million in Four others have begun the process.
more rugged country. In many cases, the pro-
posed new routes would compete with existing Concessions—Becoming the Norm
road and rail routes, which would constrain The introduction of concessions has required
the rates that typically could be charged to at substantial changes in the legal and regulatory
most $0.05 per net tonne-kilometer. In the case framework in many countries. In the franco-
of export mineral traffic, the potential rate is phone countries, concessions can generally be
generally constrained to about $0.02–$0.03 per done within the existing legal system, but most
net tonne-kilometer by the long-term delivered anglophone countries have had to amend their
market price. Because a serviceable two-lane railway acts. Arrangements have also been
road can generally be constructed for approxi- made for the economic and safety regulation
mately $1 million per kilometer, the additional of concessions, and new government bodies
rail investment would be economically justified have been established to own the assets leased
only if expected traffic was at least 2 million– to the concessionaires.
4 million tons a year. If the capital costs of the Those railways that have not been conces-
infrastructure do not have to be recovered, the sioned remain subject to significant political
lines can probably be operated successfully at and governmental influence. Arrangements
0.5 million–1.0 million ton. vary across countries, but the sectoral minis-
try (normally transport) exercises political and
administrative control, while the ministry of
Institutional Arrangements and finance exercises financial control. Board direc-
Performance tors are generally a combination of ministry
officials and internal senior management, who
Until the 1980s, almost all African railway are often appointed by the government. Over-
companies were publicly owned corporations, sight is nominally assigned to the parliament,
with varying degrees of financial and manage- but in practice such control may be limited to
ment autonomy. Attempts at commercializa- an audit of the company accounts in its annual
tion while retaining public ownership were report (often several years in arrears). Although
generally unsuccessful, and concessions were the governing regulatory frameworks nominally
introduced in the 1990s. Under concessional provide financial and management autonomy,
arrangements, the state remains the owner in practice this arrangement is considerably
of all or some of the existing assets, typically limited by the many opportunities for state
the infrastructure, and transfers the other intervention permitted under the legal and
assets (normally the rolling stock) and the regulatory frameworks at both the institutional
responsibility to operate and maintain the rail- and jurisdictional levels. This conflict between
way to a concessionaire. the control and decision functions, as well as
Most countries in Central, East, and West frequent reviews by political authorities of ini-
Africa have moved all or part of the way to con- tiatives taken by the government’s authorized
cessioning, often under the pressure of multi- representatives in the corporation, discourages
lateral and bilateral organizations that have management initiative and effectiveness.
until recently been the only source of large The first railways to be concessioned were
loans for asset rehabilitation and renewal. With in West Africa, beginning in 1995 with the
the exception of southern Africa (Botswana, Sitarail concession linking Burkina Faso and
Namibia, South Africa, and Swaziland) and Côte d’Ivoire and followed in the late 1990s by
Railways: Looking for Traffic 239
Cameroon, Gabon, and Malawi. The reform Figure 11.5 Private Participation in African Railways since 1990
momentum accelerated in the 2000s, but
Sizarail
implementation has often been a slow process, Transrail Dem. Rep. of Congo
Senegal/Mali 1995–97
typically taking three to five years, sometimes 2003 2008
much longer. RVRC
Most African networks leave little room Sitarail Kenya/Uganda
Côte d’Ivoire/Burkina 2006
for competition, and few governments have Faso
seriously considered the European model of 1995
full vertical separation. However, third-party CANAC/WACEM TRC
Togo Tanzania
operators run on government lines in Kenya 1995/2002 2007
Camrail
and Senegal, and a through freight service has Cameroon Transgabonaise CEAR
Gabon Malawi
operated for some years from South Africa to 1999
1999 1999
Tanzania. Concessions do not always include railway operated by state RSZ
Madarail
Madagascar
the entire network, with lightly used branch railway company Zambia 2003
private sector participation 2003
lines sometimes excluded. projects planned or under way
The initial duration of concessions varies part of rail network now BBR CDN
under private management Zimbabwe Ressano Garcia Mozambique
from 15 to 30 years, and the concessionaire 1997 Mozambique 2005
railway now under private canceled CCFB
is free to operate its activity as a business, with management Mozambique
freight tariffs generally determined by supply 2005
Norte–Mozambique) and is also a significant productivity is similarly low, with the source
shareholder in the adjacent CEAR conces- generally being low availability caused by a
sion. In Madagascar, the government holds lack of spare parts.
25 percent of Madarail, while governments Labor and asset productivity have improved
own 10–20 percent in Abidjan-Ouagadougou steadily in most concessions, typically doubling
Railway (Sitarail), Dakar-Bamako Railway because of workforce reductions either before
(Transrail), and Cameroon Railway Corpo- or at the time of concessioning, the scrapping
ration (Camrail). Local private participation of obsolete rolling stock, and increased traffic
in concessions has generally been relatively volumes (figure 11.7).
low and is often fraught with problems Safety is also an important aspect of opera-
during the bidding process. Employee share- tional performance. Rail travel is still safer
holding remains under 5 percent where it than road travel, but rail’s record in Africa is
exists at all. much worse than that of comparable railways
elsewhere, caused by obsolete track infrastruc-
Operational Performance— ture, poorly maintained rolling stock, and lack
Concessioning Helps of operational discipline. As with productivity,
Both labor productivity and asset productivity however, safety has generally improved follow-
(locomotive and railcar use) are low in most ing concessioning.
African networks, compared with railways
elsewhere, because of the poor condition of Financial Performance—Generally
the infrastructure and rolling stock, low traf- Unsustainable
fic levels, and government ownership. Under Most state-owned railways in Africa just about
concessions, however, these indicators have break even cashwise after receiving govern-
improved sharply, partly because of growth in ment support. Often, this balance occurs only
traffic but mostly from major reductions in because a significant amount of maintenance
the workforce. has been deferred; when the maintenance back-
Since about 1990, almost all railway com- log becomes too great, it is typically addressed
panies have streamlined their workforces. This by a loan that is treated as investment. The
measure has often been the prelude to conces- two companies that have been concessioned
sioning, but in some cases, it has also been a the longest (Camrail and Sitarail) make mod-
general policy to improve efficiency. Still, labor est operating profits. The performance of RSZ
productivity on most African systems is rela- (Railway Systems of Zambia) is unknown, and
tively low by world standards, with few railways the cases of Kenya and Tanzania are too early
achieving over 500,000 traffic units per staff a to judge.
year, compared with an average 3.3 million Passenger services generally do not con-
traffic units per staff a year for the South Afri- tribute significantly to the cost of maintaining
can operator Spoornet (figure 11.6). This low infrastructure or to covering corporate over-
productivity not only reflects the continuing head. In a few cases, they cover their marginal
use of labor-intensive methods with relatively costs (train crew, rolling-stock maintenance,
little outsourcing, but it is also the consequence fuel or traction electricity, and passenger-
of a decline in traffic without adjustments to handling costs). Passenger tariffs on many
staff levels. With low wages, the direct financial railways are essentially regulated, often within
impact is not always catastrophic, but having a a framework that includes only a subset of
large number of underemployed staff members total costs. However, many of the poorer per-
corrodes morale and is a strong disincentive forming systems in Africa would be unable to
for those who wish to improve efficiency. An cover above-rail working expenses on a sys-
important effect is that railways have difficulty temwide level even if they could set their own
recruiting and retaining technically competent tariffs.
staff or introducing the technology required to Freight services normally cover their avoid-
improve service levels, for which a better-paid able operating costs. Some also earn enough
and more skilled workforce is essential. Asset to cover infrastructure costs and even capital
Railways: Looking for Traffic 241
costs for rolling stock. Earnings are a function Figure 11.6 Labor Productivity on African Rail Systems
of the tariff rate and the average carload on
Madarail, Madagascar
the revenue side, and factors such as train size,
commercial speed, and rolling-stock use and CEAR, Malawi
concessioned
replacement of rolling stock, and very rarely Transrail, Mali
can it earn enough to finance infrastructure
RSZ, Zambia
renewal.
Sitarail, Burkina Faso and Côte d’Ivoire
Where railways have been concessioned,
Camrail, Cameroon
low-interest sovereign loans to concessionaires
have usually made a substantial contribution TRC, Tanzania
state owned
rolling stock transferred to the concessionaire CFCO, Rep. of Congo
more than compensates for the equity put into CFM, Mozambique
the concessions in most cases, the result is a SNCC, Dem. Rep. of Congo
significant transfer of the financial risks asso- SR, Swaziland
ciated with infrastructure investment from the BR, Botswana
private sector to the public sector. The busi-
TAZARA, Tanzania
ness fundamentals of many concessions are
SRC, Sudan
insufficient to support major investment on a
Transnamib, Namibia
commercial basis, and they are all too prone
to significant liquidity problems. Major asset 0 500 1,000 1,500 2,000 2,500
maintenance and reinvestment are thus likely traffic units/employee (thousands)
Figure 11.7 Rail Concession Labor Productivity have generally been unenthusiastic about run-
ning passenger services, which do not generate
labor productivity/(traffic units [thousands]/staff)
700
the same revenues as freight; this situation has
600 not been helped by delays and disputes about
date of the payment of government compensation for
500 concessioning unprofitable services. Further problems have
400
arisen over the level of concession fees, the
length of the concession, and arrangements for
300
Rivi-Rivi bridge
redundant staff. In some cases, these issues have
collapse led to renegotiation of the concession contract.
200
Despite these vicissitudes, the results to date
100
civil war in
are encouraging. Even if not all expectations
Côte d’Ivoire have been met, most of the concessioned rail-
0 from late-2003
–5 –4 –3 –2 –1 0 1 2 3 4 5 6 7 8 9
ways have improved their traffic levels and their
years since concessioning
productivity and are providing better service
Camrail CEAR RSZ Sitarail
to users, albeit after a solid injection of invest-
ment by donors and international financial
Source: Bullock 2009. institutions. Arguably, some of this improve-
Note: The overall traffic units carried by a railway are the sum of the passenger-kilometers and the net ment might have occurred anyway. In addition,
tonne-kilometers of freight carried. This simple standard measure is widely used as a means of aggre-
gating freight and passenger traffic. The relative weighting of passenger and freight is conventionally responsibility for the ongoing rehabilitation
taken as 1:1. Rivi-Rivi bridge refers to the Rivi-Rivi River bridge in Balaka, Malawi. Camrail = Cameroon
Railway Corporation; CEAR = Central East African Railways Corporation (Malawi); RSZ = Railway and maintenance of track is rapidly emerg-
Systems of Zambia Ltd; Sitarail = railway operator for Burkina Faso and Côte d’Ivoire. ing as a key issue between concessionaires and
governments. A key government objective in
many railway concessions is to obtain finance
The Verdict on Concessions—Generally (whether private or through international
Beneficial but Not the Full Answer financial institutions) to rehabilitate track infra-
Since 1992, there have been 16 rail concessions structure. For most private operators, however,
in Africa. Two of the 16 have been canceled, track rehabilitation, especially track renewal, is
1 has been badly affected by war, and 1 has a major expense that drains available funds, but
suffered from natural disasters and procedural it is also one that can be easily deferred.
delays. Six have operated for five years or more The greatest effect of concessionaires has
but only 2 of those without a significant dislo- been improved operations. Given the weak
cation of some sort.3 investment and regulatory climate in many
Except for the railways immediately adja- African countries, investment flows have been
cent to South Africa, those that have not been limited. Under concessioning, operations
concessioned have deteriorated continuously have been positive, and efficiency has clearly
since the mid-1990s. In a number of cases, improved. Labor productivity has increased
these declines will prove to be terminal. Many steadily in all the concessions in operation for
governments in Africa will consider conces- over five years, and similar figures will likely
sions only as a last-ditch solution, but in many come from recent concessions. Asset produc-
cases, the railways have been left to deteriorate tivity has also generally increased. Although
for too long, and rectifying the situation will concessionaires in Africa typically have a more
be a struggle. appropriate cost structure than their predeces-
The concessions have not been without sors, it is rarely the ideal cost structure. Operat-
their problems. In many cases, finding more ing costs on railways are a function of capital
than a few bidders has been difficult, and in invested, as well as operating efficiency, and
several cases, bidders’ financial resources have many African railways have been starved of
been insufficient to finance the major invest- capital, substantially increasing overall operat-
ments required. As a result, the state has had to ing costs.
guarantee investments; even then, mobilizing Allocative efficiency is difficult to measure
the financing has been slow. Concessionaires directly, but the evidence is generally positive.
Railways: Looking for Traffic 243
Improved productivity, an active search for the low-volume operators, the sensible choice
new traffic by concessionaires, and better inter- is to find secondhand equipment. Much of
nal business practices have all improved rail- the investment to date has been for mainte-
way cost and pricing structures and lifted the nance and renewal backlogs, without which
level of service, thus helping attract traffic to the railway often would not function, and can
the mode that can carry it most efficiently and be characterized as one-time investment to get
improve intermodal competition. the systems running. Even that investment has
Most concessionaires have fulfilled the pas- been slow, more than four years in Cameroon
senger service requirements in their concession and five years on the Nacala line—a long time to
agreements, even where it has been operation- wait when a business is barely breaking even.
ally difficult or where agreed public service Are concessions a long-term answer? Or
obligation payments have not been forthcom- are they merely quick fixes that are living off
ing. Many of these services were inherited, investment by third parties and will prove
and passenger service would often be more unsustainable in the long term? What more
economical with a road-based system. must be done to ensure a sustainable sector?
A recent review of four concessions found Many of the answers to those questions must
little evidence of monopolistic behavior by con- come from governments.
cessionaires (Pozzo di Borgo and others 2006;
World Bank 2006). This review examined freight
rates and whether services were being reduced Key Issues for Governments
so resources could be redeployed to favored
users, beyond changes in services that any com- Classic concession schemes4 in Africa are
mercialized railway undertakes in response to unlikely to be financially attractive to bidders
changing traffic patterns. Few concessions are other than those who can secure financial ben-
immune from road competition, except in the efits not directly linked to the railway opera-
few cases where roads still must be constructed tions.5 Consequently, unless the structure of
or where heavy mineral movements occur. No African rail concessions changes or the mar-
evidence exists that personal travel has been ket environment in which they operate alters
made more expensive for the poor. favorably, private operators will continue to
The greatest disappointment for govern- show limited interest in African railway con-
ments has been the lack of infrastructure cessions. Two key areas need to be addressed:
funding from sources other than international the financing of passenger services and major
financial institutions. Concession agreements track renewals and rehabilitation, both requir-
clearly put the responsibility of financing track ing substantial public funding in most conces-
maintenance and renewal on private opera- sions. If this funding is provided, governments
tors. Likewise, rolling-stock financing has been will also need to strengthen their regulatory
left to concessionaires under their contracts. capacity to ensure that the conditions are met
However, most concessionaires initially rely and that the effect on the rail sector in general,
on loans from international institutions, with and concessionaires in particular, is properly
below-market borrowing costs, lengthy loan considered when policies in other sectors of
terms, and grace periods to finance infrastruc- the economy are developed.
ture. (The exceptions are the Beitbridge Railway
[Zimbabwe to South Africa], which relies on Passenger Services
take-or-pay clauses that guarantee minimum If governments want the concessionaire to
revenues; the Nacala Railway in Mozambique, operate passenger services, they should make
which is being funded at semicommercial rates; clear compensation arrangements that can be
and Zambia and the Rift Valley Railways [cover- monitored. Few passenger train services will
ing Kenya and Uganda], where the investment likely cover even their above-rail costs. Their
program is modest and is funded directly by financial contribution to infrastructure costs
the concessionaire.) Loans have been provided is minimal, and few services would justify
for rolling stock in some cases, but for many of investment in rolling stock, whether hauled by
244 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
locomotive or self-propelled. If these services cash, on a small network when the expenditure
are to operate for more than the initial years may not occur for 5 or 10 years, a concessionaire
of a concession, governments need to develop is unlikely to reserve funds annually and hold
a simple compensation scheme with timely them in reserve that long. Furthermore, raising
payments. Any scheme should enable the con- debt financing for rail repair will generally be
cessionaire to keep all the revenue, which will possible only through a general corporate loan,
encourage maximum operation, and should which is almost impossible for a small stand-
include a public contribution, possibly per alone railway.
carriage-kilometer, toward the cost of run- Profits to the concessionaire need to be
ning unprofitable passenger services. The boosted, or supplementary funding sources
scheme should be easily audited and should be need to be developed, or both. Today, Afri-
reviewed periodically, perhaps every five years. can railway concessions offer two models for
If such schemes are not introduced, pas- financing infrastructure. In the first, govern-
senger services will be a constant source of ments finance initial track rehabilitation and
conflict between the government and the renewal costs, generally by securing loans from
operator. Moreover, the issue will divert the international financial institutions. These loans
focus of the concessionaire from the freight are then made to private operators and tend to
services, where improvement is far more cover only the initial five-year investment plan
important economically for the country. in the hope that they will propel each conces-
sionaire’s traffic to a level that will then enable
Capacity or Willingness of Private it to self-finance future track investments. This
Operators to Finance Track Renewal approach is commonly used for railways with a
Few, if any, concessions are generating signifi- high ratio of initial track investment compared
cant profits for their operators and certainly not with revenues and that are thus unlikely to be
enough to fund long-term renewals. Although able to mobilize sufficient private financing. In
most concessionaires pay fees into general gov- the second model, governments do not finance
ernment revenue, none can afford to do so and initial track renewal but commit to compen-
accrue funds for future renewals at the same sating concessionaires for their investment by
time. Whether a purely privately financed rail the end of the concession (for example, Kenya,
concession model is sustainable in much of Uganda, and Zambia railways). In such cases,
Africa remains doubtful. Track structures have the initial amount to be invested is relatively
(or should have) lives of several decades, given small in relation to expected revenues, and pri-
the traffic volumes typically carried on an Afri- vate operators are assumed to be able to secure
can railway. On a small system, track renewal private financing on the merits of their busi-
is needed somewhere on the network only ness case. Under both models, governments
about every 20 years. It is almost always pos- usually agree to purchase at the end of their
sible to defer renewals for several years, albeit concessions the nonamortized portion of any
at the cost of deteriorating track conditions infrastructure investment concessionaires have
and reduced operating speeds. For any conces- financed. However, the ability of many govern-
sionaire who is uncertain about the future, the ments to make such a payment is uncertain,
safest decision is to do as little track renewal which often affects infrastructure investment
as possible. in the later stages of a concession, although a
Even if they do want to renew track, private partial risk guarantee can strengthen the gov-
operators will often struggle to generate suffi- ernment’s reimbursement commitment.
cient cash flow for it. Few concessions are strong Three conditions must be met to secure
financially. If a government makes the level of privately financed track investment: (a) gov-
the concession fee or rolling-stock purchase ernments ensure that the concession (and thus
price the ultimate measure of a successful deal, it the proposed track investment) is financially
will limit the successful bidder’s ability to renew sound, (b) the nonamortized value of the
infrastructure. Even if an operator has sufficient assets owed to the concessionaire at the end
Railways: Looking for Traffic 245
Pozzo di Borgo, Pierre, Alain Labeau, Raphael UIC (International Union of Railways)-Statistics
Eskinazi, Julien Dehornoy, Alan Parte, and Centre. 2008. Railway Time-Series Data 2007.
Marouane Ameziane. 2006. “Review of Selected Paris: Railway Technical Publications.
Railway Concessions in Sub-Saharan Africa.”
World Bank. 2006. “Sub-Saharan Africa: Review
World Bank, Washington, DC.
of Selected Railway Concessions.” Africa
Thompson, Louis. 2007. “Spoornet and Transnet Transport Sector Report 36491, World Bank,
Sectoral Reference Paper.” World Bank, Washington, DC.
Washington, DC.
Chapter 12
A
frican shipping has been largely dereg- Kenya) are not major hubs on the main inter-
ulated. However, many African coun- national itineraries, and they appear unlikely
tries are trapped in a vicious circle of to become so. Several ports suffer from low
high tariffs discouraging traffic and further capacity, particularly in terminal storage,
increasing costs. Poor inland links and waste- maintenance, and dredging capability. Overall,
ful and costly port administration accentuate however, total use of African port capacity is
this problematic situation. The lack of an inte- estimated at 80 percent and likely to remain at
grated land distribution system, particularly this level in the near future.
for transit, impedes container traffic. Many ports are poorly equipped and inef-
Since the mid-1990s, both general cargo and ficiently operated. Container handling rates
containerized cargo passing through African fall well below international norms. Port
public ports have trebled. Southern Africa has charges for both containers and general cargo
had the fastest growth in general cargo traffic are substantially higher than in other regions.
and West Africa in container traffic, albeit from Security standards are still extremely variable,
a low base. Dry bulk traffic (coal, grain, and and few ports are prepared for the dramatic
some chemicals) and liquid bulk traffic (mostly changes in trade and shipping patterns now
oil) have also been growing rapidly. By inter- occurring.
national standards, however, these traffic cat- The main requirements are organizational.
egories are unbalanced, increasing the costs for Many capacity constraints could be overcome
African trade. Export volumes greatly exceed simply by making the existing ports more effi-
import volumes for dry and liquid bulks, while cient. Regional port planning is required to
imports dominate exports for general cargo counter the costs of fragmentation. Port pric-
and container trades. ing and regulatory policies need to be more
Many ports handle the traffic, few of them commercial and to better respond to the inter-
large by world standards. The main transship- national shipping market. Comprehensive
ment points for regional traffic (Abidjan, Côte policies are required for modal integration
d’Ivoire; Dar es Salaam, Tanzania; Djibouti, and administrative simplification, and modern
Djibouti; Durban, South Africa; and Mombasa, port management structures are essential.
249
250 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
The landlord port system has been more accounts for less than 1 percent of total world
successful than the service port in Africa (as container traffic and for little more than
elsewhere) and is best suited to introduce the 2 percent of all African traffic. East Africa has
private sector. Within a landlord port structure, a heavy concentration in Mombasa (6 percent
attracting container line operators and major of the Sub-Saharan African total, according to
international terminal operators could produce the United Nations Conference on Trade and
efficiency improvements. Ghana and Nigeria Development) while West Africa has five ports
have moved toward the landlord port, and sev- handling more than 350,000 TEUs each.
eral francophone countries operate a hybrid The lack of an integrated land distribution
model. However, development is slow, and system, particularly for transit traffic, impedes
the involvement of the efficient private global container traffic. Handling of dry and liquid
operators is very low. bulk exports is making the most progress, with
many port facilities privately owned and inte-
grated in a comprehensive logistic system.
The African Shipping Market From 1995 to 2005, general cargo has grown
at an average annual rate of 6.6 percent and
Africa’s maritime traffic has been growing rap- at a rate as high as 15.7 percent in southern
idly across all cargo types, although container Africa (table 12.2), rates higher than in the rest
traffic is highly imbalanced and faces major of the world because of later containerization.
challenges because of the lack of efficient General cargo has traditionally been the major
transportation links back to the hinterland. type of cargo moved to landlocked countries.
Shipping markets are small and thin, contrib- Little congestion occurs in the ports, but han-
uting to relatively high costs. dling is inefficient, and the transfer of some
of this traffic to containers is contingent on
Maritime Traffic—On the Rise but Out inland distribution systems.
of Balance Dry bulk traffic is sometimes handled at
Except in South Africa, container transport in common-user general cargo facilities, but the
Sub-Saharan Africa is still at an early stage of major flows (grain from Mombasa, ferro-
development; however, it is growing rapidly chrome from Maputo, and coal from Richards
from a very low base, with an average annual Bay [South Africa]) pass through privately
growth rate of 7.2 percent and as high as owned and operated dry and liquid bulk termi-
13.8 percent in West Africa (table 12.1). Of the nals, for which the traffic volumes are generally
7.6 million 20-foot equivalent units (TEUs) not well reported. In 2007, the total throughput
handled by all Sub-Saharan African ports in of Richards Bay was 66 million tons, making it
2005, Durban handled nearly 2 million TEUs, the world’s ninth-largest bulk exporting port.
and the three main South African ports together Because major global interests control these
handled more than 3 million TEUs. West Africa businesses, the port and shipping arrangements
likely conform to the best international stan-
dards. Some dry bulk traffic is still handled on
general cargo quays, suggesting the possibility
Table 12.1 Traffic Trends for Container Trade, Sub-Saharan Ports, by Region, for further specialization, though that depends
1995–2005
on having a large enough basic flow.
Average Liquid bulk traffic is predominantly oil,
TEUs Overall annual
percentage percentage with 11 countries (dominated by Nigeria and
Region 1995 2005 change change Angola) supplying 12 percent of world demand
East Africa–Indian Ocean 505,100 1,394,956 176 5.8 and 19 percent of U.S. demand. In 2006, oil
Southern Africa 1,356,000 3,091,846 128 2.5 made up 85 percent of exports by value from
West Africa 673,400 3,126,901 364 13.8 West and Central Africa. It is a growing sector,
Total 2,534,500 7,613,703 200 7.2
with Asian countries (China, in particular) and
Source: Mundy and Penfold 2008.
the United States making significant invest-
Note: TEU = 20-foot equivalent unit. ments in Africa, including placement of the
Ports and Shipping: Landlords Needed 251
proper export platforms. Again, international Table 12.2 Traffic Trends for General Cargo, 1995–2005
standards are generally met. Average
For the most part, African countries are Thousand tons Overall annual
exporters of minerals (including oil) and agri- percentage percentage
Region 1995 2005 change change
cultural products, handled either by specialized
East Africa–Indian Ocean 13.84 38.42 177 5.9
or dedicated dry or liquid bulk terminals or by
general cargo facilities. For example, agricul- Southern Africa 2.73 14.52 431 15.7
tural products are often handled over the quay West Africa 19.57 51.68 164 5.1
at general cargo facilities by grabs or mobile Total 36.14 104.62 189 6.6
conveyors. Export volumes (loadings) greatly Source: Mundy and Penfold 2008.
exceed import volumes (unloadings) for dry
and liquid bulks, while imports dominate Figure 12.1 Balance of Sub-Saharan African Container
exports for general cargo. Trade, 2005
The dominance of imports is most pro-
nounced in the container trades, increasing the 3,500,000
LIBERIA
TOGO
COTE D’IVOIREGHANA
CENTRAL AFRICAN REPUBLIC
ETHIOPIA
to extend the contract for a much longer period
CAMEROON
with further equipment purchases planned. In
EQUATORIAL GUINEA
SAO TOME AND PRINCIPE
GABON
UGANDA KENYA
Mombasa, a contract is being awarded to deepen
CONGO CONGO, DEM REP
RWANDA
BURUNDI
the port, and a new container terminal to com-
SEYCHELLES
10–50 MOZAMBIQUE
mance, and hinterland connections. Meanwhile,
ZIMBABWE MADAGASCAR
REUNION
Djibouti may soon provide competition for
> 50 NAMIBIA
BOTSWANA
Salalah and Dubai, with DP World scheduled to
railway
SWAZILAND bring on stream a new container terminal facility
major road
SOUTH AFRICA
LESOTHO at Doraleh, targeted specifically at offering sig-
nificant transshipment capacity for East Africa
and the Indian Ocean. In addition, container
Source: Mundy and Penfold 2008. terminal facilities in Jeddah (Saudi Arabia) are
Ports and Shipping: Landlords Needed 253
being extended, and in 2008, DP World signed as its main hubs for West African container
a concession for the operation and further trade, relaying West African cargo moving
development of the Aden Container Terminal to and from Europe and Asia. Nevertheless,
in Yemen. Since 2008, a consultant has been the number of public-private partnerships is
finalizing the National Port Strategy project increasing, and competition between the pri-
for Tanzania. vate port operators in the area is fierce. Con-
In South Africa, Durban struggles to han- sequently, some of the big global operators
dle its own national traffic and experiences have become willing to look at medium-size
recurring berth congestion crises during the and even small terminal projects, which they
peak season. Shipping lines are threatening to previously snubbed (Harding, Pálsson, and
reintroduce a surcharge for berthing delays, as Raballand 2007).
existed between 2002 and 2005. Durban also
has problems of environment, security, hinter- Port Ownership and Management—
land connections, and space. Although plans Still Mainly a Public Service Model
exist to bring on stream major new capacity, Port planning and management are generally
such as the new Pier One scheme, demand is outdated, though seven sampled countries are
very strong, and over the short to medium developing new port master plans, several with
term, it may well outstrip the new capacity. a focus on institutional reform. Port regula-
The number of carriers seeking alternative tion is normally undertaken by a ministry of
locations for transshipment in the Indian transport, rather than by a quasi-independent
Ocean islands (notably Mauritius) reflects agency; thus, it tends to be highly politicized.
Durban’s problems. Although superstruc- With its independent regulator, South Africa is
ture and infrastructure usually are separated the exception.
in South Africa, the common ownership of The dominant port management model in
both within the publicly owned Transnet has Africa is still the public service port: the state
clearly failed to deliver the necessary improve- enterprise owns the port infrastructure and
ments required of a great world port ideally undertakes all port operations. This model is
located to act as a transshipment center for beginning to change. Some statutory incorpo-
southern Africa. Currently, South Africa is rated port agencies are being reestablished as
an end-of-the-line country, and unlike other limited liability commercial companies. Ghana
major or global hubs located on one of the and Nigeria have moved toward the landlord
very large east-west routes that make econo- port, where the state owns and operates major
mies of scale possible, its problems arguably port infrastructure but allows the private sector
lie at least partly in the organization and the to provide basic services. In addition, several
provision and management of equipment francophone countries have a hybrid model,
and handling space, as much as in basic quay called amodiation, in which the port author-
capacity. The significance of that distinction ity rents on-dock storage space to privately
is that the solution lies in institutional reform owned, licensed stevedoring companies hired
and the mobilization of private sector capa- by shipping lines for cargo handling.
bilities in port service management, as well as Since 2000, some major container terminals
in public sector investment. have been concessioned to the major interna-
On the West African coast, Abidjan has tional terminal operators (table 12.3). How-
enjoyed some success as a container trans- ever, involvement of the efficient private global
shipment center, but it has suffered because operators is still low; in 2007, the top 20 global
of the country’s internal strife and the specific terminal operators handled only 16 percent of
problems relating to ownership of operating throughput in Africa, compared with about
rights for the container terminal. The need 70 percent in other regions of the world. Con-
for an alternative to Abidjan is indicated by cessioning has proved controversial in some
the Maersk Line’s (and its affiliate Safmarine) cases, with the results contested in both Luanda
using the Spanish port of Algeciras and the and Dakar. No generally accepted “clean” model
new container terminal at Tangier (Morocco) exists, and influence and corruption remain.
254 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Port Performance—Room for Figure 12.3 Average Moves per Hour by Category
of Port
Improvement
Container handling falls below international 20
standards in most ports. Even when container 16
d
ea
rie
ne
ne
’g
nt
sio
sio
London (South Africa), Matadi (the Democratic
ga
ips
es
es
sh
nc
nc
Republic of Congo), and Walvis Bay (Namibia).
co
co
un
The low performance is partly explained by the
Source: Mundy and Penfold 2008.
lower number of containers handled per call
with smaller vessels. However, management is
even more important. Higher handling rates bulk port facilities, less documentation exists
are generally achieved in locations where pri- about their efficiency, but a recent study of
vate operators have been in residence for some South African ports showed the bulk ports
time; although the hybrid Mozambique model, performing well on international benchmarks
in which the government retains a major share, (Bell and Bichou 2007).
has not been so successful. The growing interest in Africa as a source of
Rates for general cargo handling are also energy products, agricultural products, timber,
lower in Africa at 7 to 25 tons per crane hour, and minerals might aid in creating the proper
compared with more than 30 tons in other maritime export capacity. The funding for this
world ports.2 Almost all handling is through new capacity is invariably provided as part of
public ports. a turnkey project, not under a traditional port
Specialized oil and coal terminals usually authority budget. Thus, financial impediments
do not fall under public port management. are not envisaged.
Traditionally, state-owned organizations, The quality of container handling inland is
private interests, or a combination have indicated by the cycle times of trucks dropping
developed the facilities, which fall outside off and picking up containers at the terminal
the mainstream of port operations and fol- and the average container dwell times in a ter-
low an integrated supply-chain logic. Because minal. The typical target for an efficient truck
of the private involvement in dry and liquid cycle is 1 hour. Average cycle times are estimated
Ports and Shipping: Landlords Needed 255
at 5–6 hours in East Africa, 4 hours in south- Table 12.4 Average Port Delays
ern Africa, and 10 hours or more in West Africa Range of truck Range of container
(table 12.4). Average container dwell times are Region cycle times dwell times
6 days in southern Africa, 12 days in East Africa, East Africa 3.5 hours to 1 day 5 to 28 days
and 15 days or more in West Africa, more than Southern Africa 2 to 12 hours 4 to 8 days
an accepted international standard of 7 days West Africa 6 hours to 1 + day 11 to 30 days
or less. The range of dwell times in southern Source: Mundy and Penfold 2008.
Africa (4–8 days) is a much tighter band than
in East and West Africa, thanks largely to better
organization and control of container storage Table 12.5 Typical Gateway Container and General Cargo Handling Charges in
at the terminal. World Markets
dollars
Like all ports in countries that are signatories
to the International Convention for the Safety Container handling General cargo over
Region from ship to gate the quay per ton
of Life at Sea, African ports have been required,
West Africa 100–320 8.00–15.00
since 2004, to comply with the International
Ship and Port Facility Security Code. Although East Africa 135–275 6.00–15.00
standards of security are still extremely vari- Southern Africa 110–243 11.00–15.00
able, the estimated costs of compliance for Southeast Asia 80 8.00
African ports—averaging about $2 per TEU Far East 144 8.00
and about $.04–$.05 per ton of general cargo— Middle East/South Asia 96 7.00
have not increased overall costs significantly
United Kingdom 100 8.50
and may have generated compensating bene-
Northern Europe 110 7.50
fits, including reduced losses through pilferage
and higher customs yields (Kruk and Donner Southern Europe 95 7.00
2008). The long-term concern is a sensible bal- Latin America 154 9.00
ance between security and costs. Australasia 130 9.00
Port charges for both containers and gen- Source: Mundy and Penfold 2008.
eral cargo are substantially higher in African
ports than in other regions (table 12.5). For
container handling, the charges applied in with finding appropriate responses to changes
Sub-Saharan Africa can be more than twice in international trade and shipping markets.
those typically applied for the same service in
other parts of the world, with at least 50 per- Responses to Changes in the
cent more as the norm. Normal charges for International Shipping Market
general cargo handling per ton offloaded from The problem is not just port capacity. East
a vessel in Sub-Saharan Africa are also about Asian ports use vessels in the 8,000–11,000
40 percent above world rates. TEU range, but most African ports cannot
efficiently handle container vessels above
2,000 TEUs. Moreover, an upper limit exists
Policy Issues and Implementation to optimal vessel capacity because of the low
Challenges total volume of freight to African ports. Serv-
ing multiple African ports directly with vessels
World trade and shipping patterns are chang- of 8,000 TEUs or more is therefore unlikely in
ing. Expanding containerization, in ever-larger the near future. Thus, a tendency will exist to
vessels, requires port facilities to handle large transship through a small number of African
vessels quickly and efficiently. The financial regional hubs with container transshipment
crisis of 2008–09 adds to the turmoil. By inter- facilities to distribute traffic along the coasts.
national standards, African public port capac- The direction of the trade may also be
ity is low, and its performance is poor, bringing changing as some lines consider liner services
higher costs and further losses in world trade from Asia by way of southern Africa to the east
shares. African governments are thus faced coast of Latin America and the Caribbean.
256 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
For this trade, however, average vessel size is Governments will need to choose how
unlikely to increase because ships are cascaded best to develop state-of-the-art ports, with
down from longer-distance trade. appropriate technologies and management
As the shipping industry has grown more skills. This determination will almost certainly
capital intensive, more technically demanding, involve the international private sector, partic-
and more subject to global regulatory change, ularly in the container terminal business. Stra-
the number of active African shipping lines has tegic port planning must set the roles of the
severely decreased. Liberalizing the shipping public and private sectors and identify the pro-
market has already brought down deep-sea cesses to attract and select private partners.
shipping costs; it should also facilitate the devel- Countries with congested city ports or with
opment of less-costly feeder services for con- draft limitations will need to consider whether
tainer shipping. As the major traders attract the to rehabilitate existing ports or develop new
global operators, they may also develop a niche ones. Developments in the deep-sea shipping
market in African feeder services, reestablishing markets may also trigger the need to change
African-owned shipping companies. For exam- location. For example, an east-west axis between
ple, the establishment of Togo-based Ecoma- Asia and Latin America would be economical
rine in the West African feeder market in 2003 only for vessels of 6,000 TEUs or larger. Any
was the first indigenous development since the such service would necessitate a port of call in
decline of West African national companies South Africa, unlikely to be satisfied by Durban.
in the 1990s.3 Where collusion or barriers to Cape Town is developing a new container termi-
market entry remain in the shipping market, nal, but it is too far from the industrial heart of
governments will need to assess the level and the country in Gauteng province to be a strong
distribution of benefits from the restraints and South African hub port. Richards Bay, which
compare them to the widely distributed benefits has deep water and a spacious environment,
of lower shipping costs in a deregulated market. might be better. It recently launched plans for
the staged development of a megaport, includ-
Strategic Port Planning ing a container terminal.
The expected changes have implications for
port planning. Africa can support only a few Port Pricing and Regulation
regional hubs and possibly one major hub (in Having the economy benefit from lower
South Africa). Competition already exists for transport costs typically requires regulating
the hub in East Africa (between Dar es Salaam port tariffs to obtain the most efficient sup-
and Mombasa) and will intensify as facilities ply and the lowest real costs, thereby prevent-
are upgraded in Djibouti, and regional collab- ing any monopoly, whether state owned or
oration—though desirable—seems unlikely. private, from exploiting its advantage in the
Simply investing in port capacity will not turn a market. In many countries, however, a single
port into a hub unless it has a strategic location, port is a natural monopoly, tempting gov-
adequate water depth, and the facilities and ernments to maintain direct ownership and
performance to ensure low handling costs. operation and, thus, to use the port as a “cash
A strong corridor for transit traffic also cow” to support other government activities.
helps. This requires facilitating traffic on the For example, in South Africa, all major ports
main trade corridors from the port to the are owned and operated by the National Ports
landlocked hinterland. A common problem is Authority (responsible for infrastructure) and
the failure to address international, intermodal South African Port Operations (responsible
transport holistically. Inland movement, par- for port operations), both part of the state-
ticularly across green borders, has been slow owned Transnet, a wider monopoly cover-
and expensive, thereby stifling trade. Although ing rail, pipelines, and ports. Transnet has
logistics chains are a commercial matter, they presided over an extensive but opaque struc-
require the facilitation of integrated port, cus- ture of cross subsidies, allowing the whole
toms, and inland transport arrangements—a operation to exist without government sub-
matter for government. sidy. Thus, South African ports suffer from
Ports and Shipping: Landlords Needed 257
underinvestment. This type of port strategy in terminal operations and with the benefits of
inevitably reduces the broader benefit to the an efficient hub-and-feeder structure in the
economy of having lower transport costs. deep-sea trade.
The need is thus to mobilize private capital
Port Management and management skills to improve efficiency
African ports do not necessarily lack basic and develop a logistics system. In Africa, as else-
quay capacity (though some ports appear to be where, the landlord port system has generally
straining their limits). However, they are ineffi- enjoyed greater success than the public service
cient in using the basic infrastructure. The lack port and is the best way to attract the private
of modern superstructure, particularly cranes, sector. Attracting major international container
inhibits fast vessel turnarounds and imposes lines and terminal operators can increase effi-
costs on customers. Continuing reliance on ciency. However, mobilization of that potential
the public service port structure accentuates still requires appropriate public sector actions
overmanning and forgoes the advantages of in the administration of customs and the regu-
the modern technologies and management lation of inland transport, as recent experience
practices that have revolutionized world mar- in Lagos demonstrates (see box 12.1).
kets for shipping and cargo handling (see table Private participation by the most efficient
12.4). The prevalence of state-owned service international port operators must be stimu-
ports is also associated with the low concentra- lated by a landlord port philosophy conducive
tion of global operators in African ports. Hav- to their participation and by transparent ten-
ing global terminal operators in this business dering. Such port reforms are likely to involve
would almost certainly improve matters: they retrenchment and compensation. Govern-
are well acquainted with the advantages of scale ments should develop advance strategies to
BOX 12.1
Private Participation and Port Efficiency: The Case of Apapa Container Terminal,
Lagos, Nigeria
Lagos port has long been notorious for inadequate facili- mid-April, to enable terminals to clear “alarming” backlogs.
ties and congestion. As part of a broader program of port The controller of the Nigeria Customs Service for Apapa
reform in early 2006, the Nigerian Ports Authority awarded blamed the low clearance volume on the need to physically
a concession to APM Terminals to manage, operate, and examine every container because of the high incidence of
develop the Apapa container terminal, increasing capacity concealment and false declaration by importers. However,
from 220,000 TEUs per year to 1.6 million TEUs. Within even cleared containers were not being collected. At the end
months of the award of that concession, delays for berth- of January, of the reported 9,741 containers in the port for
ing space dwindled significantly, and shipping lines reduced delivery to the importers, only 851 had been cleared by cus-
their congestion surcharge from $740 to $105 per TEU, toms, with all charges paid and documentation completed
saving the Nigerian economy $200 million a year. By early but not picked up by agents. The Nigerian Ports Authority
2009, new gantry cranes had been acquired to triple the consequently proposed introducing demurrage charges of
original capacity. $4 per TEU in a bid to force owners to move their contain-
However, that was not the end of the story. Although ers out of the ports. In their turn, however, the containers’
the port’s equipment is able to handle more than 500 con- agents blamed a lack of trucks, arguing that many had been
tainers per day for customs examinations, the majority of booked to empty containers. Although the moratorium on
the containers are returned to stacking by the end of each entry of new vessels was lifted in early March, some back-
day. By January 2009, the port was clogged by uncollected logs and delays remained and significant organizational and
containers, and at the end of February, the head of the Nige- regulatory problems still remain.
rian Ports Authority announced a temporary suspension of
ship entry with immediate effect, lasting until sometime in Source: Press reports assembled by C. Bert Kruk, World Bank.
258 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
create employment alternatives and to handle efficiency (and thereby reduce costs), but also
the administration and finance of the adjust- to respond to the growing importance of and
ment in the terms of the concession contracts. future obligation in supply chain security.
A
ir transport can stimulate regional eco- cases, the protection of small national carriers,
nomic development. In Africa, 120,000 as an instrument for this cross subsidy, adds to
people are employed directly in air the budget burden and hinders efficient service.
transport, and 20 percent of tourism jobs Generally, infrastructure capacity is not a
are associated with travelers by air. Air cargo serious problem. The number of airports is sta-
is also important in some export trades (such ble, and enough runways exist to handle traffic
as flowers from Kenya and fish from Tanzania). with better scheduling and modest investment
Overall, traffic has been growing at about in parallel taxiways and some terminal facili-
6 percent a year from 1997 to 2006. In south- ties. Aircraft fleets are being modernized, but
ern and eastern Africa, the market growth is air traffic control facilities need substantial
strongest, with three vigorous hubs and three improvement. Revenues from airports and air
major African carriers dominating interna- traffic are probably high enough to finance the
tional and domestic markets. In central and necessary investments, but the sector does not
western Africa, however, the market is stag- capture them. The problem is both political
nating, with the vacuum created by the con- and organizational.
flict in Côte d’Ivoire and the demise of several Two other serious challenges remain. First,
regional airlines still unfilled. liberalization of the international regime
Air transport in Sub-Saharan Africa is still within Africa must be completed, as commit-
expensive by international standards. Landing ted to in the Yamoussoukro Decision of 1999.
charges are high, partly caused by the absence of Many international agreements within Africa
the support from concessions revenue enjoyed have been liberalized, resulting in routes and
by many airports in the world. Because of rela- aircraft sizes that are better adapted to the
tively low volumes of traffic on many routes in market and some large, viable indigenous
Africa, airfares are also high, despite the efforts carriers that are expanding. However, the
of some governments to subsidize domestic fares domestic and intercontinental markets often
from protected intercontinental routes. In many remain protected, and many small, nonviable
259
260 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Banjul Accord Group of countries has experi- Figure 13.2 Top-60 International Routes within Sub-Saharan Africa, 2007
enced negative growth. Of 19 countries that lost
international connections since 2004, 16 are in
these areas of West and Central Africa. Most
worrisome are the Central African Republic
(only one flight a week in November 2007),
Chad, Eritrea, Mauritania, and the Seychelles.
Not only are they minimally connected, but
also their connectivity plunged between 2004
and 2007.
Figure 13.3 Regional Growth Zones in Seats Offered, All Travel Categories
been to increase concentration, both in the by the national flag carrier and is highly con-
total market and on individual routes. By early centrated. Of the 286 routes with service
2008, 15 airlines accounted for 59 percent of in 2007, only 54 had more than one service
all seat capacity offered in Africa. The top-20 provider—generally, the national flag carrier.
intercontinental routes each average 3.45 com- Standing out among the larger countries for
peting airlines. International service within allowing competition are South Africa, where
Sub-Saharan Africa is less competitive, with competition occurs only on the heaviest routes,
just 15 airlines providing more than 82 percent and Tanzania, where more than one service
of capacity, and the big 3 (Ethiopian Airlines, provider exists on all of its 17 domestic routes.
Kenya Airways, and South African Airways)
providing 57 percent. The Equipment Operated
Sixteen of the top 60 routes, and 66 of the The two most significant trends for Sub-
206 total routes, have only one carrier. Ethi- Saharan African air transport have been the
opian Airlines and, to a lesser extent, Kenya downsizing of aircraft toward the city-jet size
Airways have contributed to this concentra- (Boeing 737 or Airbus 319) and, contrary to
tion by developing new routes as sole carri- many accounts, the renewal of the fleet. The
ers. In most cases, domestic travel is serviced proportion of seat miles flown on older Western
Airports and Air Transport: The Sky’s the Limit 263
aircraft has increased to accommodate growth Figure 13.4 Airfares by Distance on African Routes, Including North Africa, 2008
quickly and inexpensively only in some domes- 3.00
tic markets, and even that proportion is only
4 percent of seat miles overall. A slight increase 2.50
has already occurred in the use of commuter
propeller aircraft on the low-volume routes 2.00
in Tanzania and in West and Central Africa,
US$/mile
and such aircraft may have an important role 1.50
in the recovery of the air transport market in
that subregion. 1.00
than Outside
0
Air travel within Africa is considerably more 0 1,000 2,000 3,000 4,000 5,000 6,000 7,000 8,000
expensive per mile flown than intercontinental distance (miles)
travel, especially on routes of less than 2,000 international intercontinental domestic
nautical miles (figure 13.4). This cost reflects between Sub-Saharan Africa and North Africa
the fact that domestic and international mar-
kets are less dense and less competitive than Source: Bofinger 2009.
On October 17, 1988, the ministers in liberalizing fares. Signatory states were obliged
charge of civil aviation in African states met in to ensure the fair opportunity to compete on
Yamoussoukro, Côte d’Ivoire, and expounded a nondiscriminatory basis. A monitoring body
a new African Air Transport Policy, later called would supervise and implement the decision,
the Yamoussoukro Declaration. Although the and an African air transport executing agency
ministers foresaw the gradual elimination would ensure fair competition. The decision
of traffic restrictions, the declaration was paid special attention to improving air trans-
aimed primarily at cooperation among Afri- port safety. However, even though the decision
can air carriers to better compete with non- is a pan-African agreement to which most
African carriers. Nevertheless, it did stimulate African states are bound, the parties decided
the United Nations Economic Commission that separate regional economic organizations
for Africa to initiate a further conference in should implement it.
Yamoussoukro, which resulted in the historic The monitoring body has met only a few
agreement on pan-African liberalization of times. Competition rules and arbitration
air services, the 1999 Yamoussoukro Decision procedures remain pending. Although an
(Schlumberger 2008). executing agency was finally created in 2007
The decision’s main objective was the grad- by assigning the responsibilities and duties to
ual liberalization of scheduled and nonsched- the African Civil Aviation Commission, a spe-
uled intra-African air services, abolishing limits cialized institution of the African Union, the
on the capacity and frequency of international commission has yet to prove its effectiveness.
air services within Africa, universally grant- In contrast, operational implementation has
ing traffic rights up to the fifth freedom,1 and been much more productive (table 13.2), with
greater freedom to negotiate bilateral agree- Johannesburg, and Nairobi) act as gateways to
ments in all of the subregions. the continent for intercontinental traffic and
In West Africa, the Economic Community as hubs for its distribution. They currently
of West African States was unable to take any handle 36 percent of all international traffic
significant steps toward liberalizing air services, in Africa. Lagos could perform a similar func-
but the smaller West African Economic and tion for West and Central Africa but is lag-
Monetary Union went even beyond the ging. Second, about 40 medium-size airports
Yamoussoukro principles with a regime that are connected to the hubs and primarily serve
includes cabotage rights. The Banjul Accord international and domestic traffic. Third, more
Group, also in West Africa, agreed to a multi- than 200 small and often nonviable airports
lateral air service agreement, fully compatible act as the distribution points for frequently
with the Yamoussoukro Decision. In Central declining domestic air traffic. Except for those
Africa, the Economic and Monetary Commu- in the Banjul Accord countries (Cape Verde,
nity of Central Africa has implemented all the the Gambia, Ghana, Guinea, Liberia, Nigeria,
necessary legislative and regulatory elements to and Sierra Leone), the number of airports with
comply with the provisions of the Yamoussoukro scheduled service dropped by 20–40 percent
Decision. In East and southern Africa, the Com- between 2001 and 2007.
mon Market for Eastern and Southern Africa Nearly all airports that had scheduled
achieved the most progress, but full application advertised service in November 2007 have at
of the decision principles remains suspended. least one paved major runway. Only a dozen
The East African Community has chosen the or so airstrips are unpaved, most of them in
effective strategy of directing the bilaterals to countries either in or just out of military con-
conform to the decision, but signing of the flict. An exception is Tanzania, which has five
agreement remains pending. The Southern airports with scheduled service and with alter-
African Development Community has pro- natively surfaced runways, though projects for
gressed the least. Apparently, the dominant posi- resurfacing are in progress.
tion of South Africa remains the main obstacle Airports in Africa do not often have “pure”
toward implementing the decision. Overall, runway capacity constraints. With a five-
about two-thirds of air transport service within minute separation between flights, a single-
Africa is now liberalized. runway airport could accommodate 144 flights
in 12 hours, or more than 1,000 flights a week.
With an average load of 120 passengers, more
African Air Transport than 17,000 passengers a day could be handled.
Infrastructure Even with a 20-minute separation, passenger
numbers would not exceed 4,300 a day. Very
The existing air transport infrastructure can few airports in Africa handle more passengers
fairly well accommodate Africa’s current air than this. Capacity constraints may appear,
traffic and foreseeable growth; only a hand- however, on taxiways, aprons, and jetways.
ful of cases warrant investments in taxiways or In many African airports, aircraft must taxi
terminal upgrades. Instead, the main invest- to the turning bay, turn around, and taxi back
ment need lies in air traffic control and surveil- toward the access to the apron, usually in the
lance equipment, which with few exceptions is center of the runway. This procedure is accept-
largely inadequate. able in most airports where enough time elapses
between departing and arriving aircraft to
Airports—a Declining Number in do so. Only a relatively few high-volume air-
Service, but Adequate Runway Capacity ports require parallel taxiways with multiple
At least 2,900 airports exist in Africa, of which turn-off ramps from the runway. Despite this
an estimated 261 in Sub-Saharan Africa generally adequate flight capacity, policy makers
received scheduled services in 2007. These in several countries are urging the construction
airports fall into three groups. First, three of entirely new international airports for which
major international airports (Addis Ababa, no economic justification exists.
266 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
The airside infrastructure in such major hubs capacity issues are already being addressed.
as Johannesburg and Nairobi meets high inter- For example, Nairobi’s passenger terminal is
national standards in runway length, instru- going through an extensive upgrade to allow
ment landing systems, and so on. For airports more than 9 million passengers a year. In other
with lower traffic volumes, however, significant cases, the declared capacity needs to be exam-
differences in the quality of the infrastructure ined. For example, Malawi’s airport in Lilongwe,
are apparent. Fewer than 50 percent of airport though clearly in need of some upgrades, does
runways are in excellent or very good condi- not appear on the ground to be as deficient as
tion, compared with 96 percent in North Africa. the figures suggest. Rescheduling to prevent
Although traffic to airports without paved run- the simultaneous arrival of too many flights
ways is low, the number of airports with poor can be of great help.
runways is relatively high in some countries
(table 13.3). Fortunately, 87 percent of the seats Airport Management—Limited
landed are on excellent or very good runways,
Privatization
and only about 4 percent of the traffic is to air-
Airports usually have some quasi-independent
ports with marginal or poor ratings.
operating agency, whether government owned
Instrument landing systems can be found at
or not. Even the company that owns South Afri-
nearly all airports with an estimated capacity
ca’s nine most important airports, including
of 1 million seats or more, but their presence
Johannesburg, is only partially privatized,
drops off rapidly below this traffic volume. In
with majority ownership still held by the state.
many smaller, older airports, nondirectional
Cameroon, Côte d’Ivoire, and Madagascar
beacon systems—now very outdated—still pre-
have concessioned their major airport groups,
vail. However, this circumstance does not nec-
although in Madagascar, the government has
essarily imply that new investment is needed in
a majority shareholding in the concessionaire.
ground-based navigation infrastructure; today’s
Kenya has concessioned the development of
satellite technology can easily replace many of
the cargo terminal at Nairobi’s international
the ground-based systems at a much lower cost.
airport. South Africa has divested some smaller
Nevertheless, in practice, few plans have been
airports completely. Even without full airport
made to replace obsolete technologies.
concessions, the range of airport service pro-
viders is wide. In Tanzania’s Dar es Salaam
Airport Terminals—Few Capacity International Airport, passenger services are
Constraints performed by Swissport, and at Nairobi’s Jomo
Some evidence exists of inadequate capacity Kenyatta International Airport, a broad range
of passenger terminals, though data are not of competition exists for landside services.
readily available. Many Sub-Saharan African Navigation and air traffic control still typically
terminals report traffic volumes at or above fall directly under governmental agencies, with
their declared capacity, and in some cases, some services subcontracted.
vintage, Russian-built aircraft are as safe as record. The U.S. Federal Aviation Adminis-
Western-built aircraft if properly maintained tration and the European Union also rate air
and operated. The problem is that many small safety in many African countries as poor.
carriers acquire one or more old aircraft on Because of the interaction between national
the nontransparent aircraft supply market and systems, air safety is a regional problem that
operate them without supervision by the civil needs to be addressed regionally. ASECNA
aviation authorities. Their pilots work long (Agence pour la Sécurité de la Navigation Aéri-
hours and regularly operate aircraft in a dan- enne en Afrique et à Madagascar), founded in
gerous environment, resulting in crashes. Even 1959, has 15 member states aiming to pool air
for U.S.-built aircraft, U.S. National Transpor- navigation services and other infrastructure. In
tation Board inquiries highlighted several cases addition to navigation infrastructure, the orga-
of pilot error, in which poor pilot training and nization manages eight airports. Further steps
assessment contributed to an accident. are under way. In East Africa, a centralized East
In general, the International Air Trans- African civil aviation authority has just been
port Association identifies poor regulatory formed with support from the U.S. Depart-
oversight as the top threat to safety in Africa, ment of Transportation’s Safe Skies for Africa
followed by inadequate safety management program. Though not yet fully implemented,
systems. Similarly, the International Civil Avia- the organization, now headquartered at the
tion Organization’s Universal Safety Oversight East African Community in Arusha, Tanzania,
Program shows that safety implementation in would supplement the existing civil aviation
Africa is very deficient. For example, West and authorities in the member countries by pro-
Central Africa, and East and southern Africa viding resources in a central pool available to
perform below the world average in all the community countries. Also, two projects for
critical elements of safety implementation— the Cooperative Development of Operational
in most cases, by a factor of two (figure 13.5). Safety and Continuing Airworthiness Program
These deficiencies are highly correlated with are being planned for the Southern African
accident rates, suggesting that institutional Development Community and the Economic
failings explain much of Africa’s poor accident and Monetary Community of Central Africa.
Thus, regional pooling of resources is address-
ing Africa’s shortcomings in oversight.
Figure 13.5 International Civil Aviation Organization Analysis of Safety
Implementation in Africa, 2004
Liberalization of Air Transport Markets
Formal implementation of the Yamoussoukro
70 Decision remains very slow, particularly in
60 southern Africa. Many national airline offi-
% of nonimplementation
50
cials have advised their governments that the
airlines are not ready for a free market and still
40
need protection. Many are concerned that an
30 open skies policy for intercontinental trans-
20 port, particularly to Europe, would drive Afri-
10
can carriers from the intercontinental market,
with adverse secondary effects for international
0
and domestic services in Africa.
at n
ion g
at ion
at nce
tio ty
ue f
er l
lat tin
at nica
iss o
nc fe
ion
lig at
s
ns
ty on
gu ra
fu sa
r if
s
ion
ion
ial
l
ne
leg avi
ob ific
pe al
e m ch
re pe
ob rvei
fe ti
ht &
al qu
sa solu
nc te
t
co
y
r
sig re
su
ce
ar
re
im
ed
&
ec
ov tru
pr
inu
sp
nt
A
en
ch
gu
CA
co
lic
te
critical element
wider liberalization. For example, liberalizing
the domestic market in South Africa in 1990
global East and southern Africa West and Central Africa
fueled passenger growth of 80 percent between
Source: ICAO 2004.
1994 and 2004, and eventually led to the
Note: CAA = Civil Aviation Authority. establishment of domestic low-cost carriers.
Airports and Air Transport: The Sky’s the Limit 269
Similarly, liberalizing the Nairobi-Johannesburg does the allocation of many airport revenues
route in 2000 increased flights from 4 a day to (such as those for overflight) to the state trea-
14 and increased passenger volumes by 69 per- sury, with only a small part left to the sector.
cent. Competition is important here. On the The sensible investment strategy would seem
Johannesburg-Lusaka route (for which South to be investment in existing infrastructure, not
African Airways had been the only carrier for new airports, and use of newer, cheaper tech-
over 10 years), designating the South African nologies for air traffic surveillance and naviga-
low-cost carrier, Kulula, as the Zambian carrier tion systems.
reduced fares by 33–38 percent and increased
passengers by 38 percent. Development and Maintenance of Skills
Recent modeling suggests that full liberal- The International Air Transport Association
ization in the Southern African Development has identified the lack of effective flight crew
Community would reduce airfares by 18–40 training and proficiency as a major source of
percent with a low-cost carrier entering the safety problems. Even where training is ade-
market and would increase traffic volume quate, highly trained flight crews in the poorer
by 20 percent (Genesis Analytics and others countries can usually command higher salaries
2006). Another half a million foreign tourists when working for a larger foreign airline. A
would arrive by air each year, spending more similar drain of skills affects regulatory staff.
than $500 million. The multiplier effect would Safety inspectors trained with donor funding
increase the Southern African Development abandon oversight almost immediately to earn
Community’s GDP by about $1.5 billion, or much more working for an airline.
0.5 percent growth. Another study on liber- Many issues are interconnected. Poor coun-
alizing routes on 20 city-pairs to and from tries trying to maintain a domestic flag carrier
Addis Ababa came to a similar conclusion do not pay enough to train and maintain good
(Abate 2007). flight crews. Because funds are usually insuffi-
cient to provide competitive salaries for safety
Financing of Infrastructure inspectors, oversight standards also fall. Main-
Two agencies are typically concerned with air taining staff is thus linked to the policies for
transport infrastructure. Air traffic control is flag carriers, market liberalization, and sector
usually the function of a national civil aviation finance. Only a coherent and fiscally affordable
authority, whereas airport infrastructure (both set of policies in these challenging areas can
airside and landside) is usually the responsi- produce a more sustainable air transport sec-
bility of a separate airports agency. Both have tor in Sub-Saharan Africa.
financing problems.
Civil aviation authorities survive on fees.
Where the land mass is large and the geographic Notes
location important, significant air navigation The authors of this chapter are Heinrich
charges (exceeding many of the other service Bofinger and Kenneth Gwilliam, who drew
charges relied on by authorities) can be gained on background material and contributions
from overflights. However, reallocating those from Michel Iches, Pierre Pozzo di Borgo, and
charges can be politically contentious. A truly Charles Schlumberger.
independent regulatory body could improve 1. The eight “freedoms of the air” are the focus of
services, but in many cases, the revenues end international regulation of air transport. The
up in the national treasury. first and second freedoms are technical free-
doms to overfly a foreign country or to land
The same situation applies to airports. In
for refueling. The third and fourth freedoms
a study by the International Civil Aviation are commercial freedoms to carry passengers
Organization using 2005 data (ICAO 2008), from a carrier’s home country to another or
the main access airports in Africa were con- vice versa. The fifth to seventh freedoms con-
sidered inherently profitable. However, the cern the rights to carry passengers between two
nonpayment of tariffs by weak flag carriers in foreign countries, either as an extension of a
small countries undermines airport finance, as flight from the home country (fifth), through
270 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
a stop in the home country (sixth), or without Bofinger, Heinrich C. 2009. “Air Transport: Chal-
ongoing service to the home base (seventh). lenges to Growth.” Background Paper 16, Africa
The eighth freedom, pure cabotage, is the right Infrastructure Country Diagnostic, World Bank,
to carry traffic between two points in a foreign Washington, DC.
country. Genesis Analytics, Andrew Myburgh, Fathima
2. The lack of radar installations should be discussed Sheik, Fatima Fiandeiro, and James Hodge.
as the lack of surveillance infrastructure, because 2006. Clear Skies over Southern Africa. Wood-
radar is now an obsolete technology. Newer, much mead, South Africa: ComMark Trust.
more accurate, and much less costly technologies ICAO (International Civil Aviation Organization).
are now being installed, as in the United States. 2004. “An Update of the ICAO Universal Safety
Similarly, navigation aids are being supplanted by Oversight Audit Programme.” Presentation to
technologies based on global positioning systems the ICAO TCB Seminar, Singapore, January
in modern aircraft. 12–14, pp. 9 and 11. http://www.icao.int/icao/
en/tcb/TCB-Singapore-2004/Attachements/
3. In a recent accident in Cameroon involving a
Presentations/Acrobat/ICAO%20USOAP%20
new Boeing 737, the aircraft could not immedi- &%20Follow-up%20Programme.pdf.
ately be located because the last known position
was the departure end of the runway. ———. 2008. “Financial Situation of Airports and
Air Navigation Service Providers 2005.” ICAO,
Montreal, Canada.
References Schlumberger, Charles E. 2008. “Air Transport
Abate, Megersa A. 2007. “The Economic Effects Policy Research: The Implementation of the
of Progressive Air Transport Liberalization Yamoussoukro Decision.” Working Paper 20,
in Africa.” School of Graduate Studies, Addis Africa Infrastructure Country Diagnostic, World
Ababa University, Ethiopia. Bank, Washington, DC.
Chapter 14
W
ater management is critical for Building water infrastructure will fuel
meeting Africa’s development chal- growth, reduce weather-induced risk, and
lenges. Though water is vital for alleviate water-related conflicts. Both large
agriculture, only 5 percent of Africa’s cultivated and small infrastructure projects need to be
land is irrigated. Hydropower is also largely part of a balanced water investment program
undeveloped in Africa; less than 10 percent of that provides reliable water supplies for human
its potential has been tapped. Water for people health and economic activities and that pro-
and animals is vital for health and livelihoods, tects natural water and environmental assets.
yet only 58 percent of Africans have access to Development of large multipurpose storage
safe drinking water. facilities (often combined with hydropower
African economies depend on a reliable generation) is necessary for mitigating the
and adequate supply of water, but high rain- economic effects of hydroclimatic variabil-
fall and hydrological variability result in fre- ity, for ensuring reliable water supply, and for
quent droughts and floods that stifle economic using available water. Small-scale approaches
growth. Moreover, water resources shared by to water management improve the ability of
countries pose complex political and manage- the rural poor to cope with water shocks by
ment challenges. increasing agricultural productivity and pro-
Achieving water security to support growth viding cost-effective water supply and drought
and to build climate resilience is at the heart of mitigation. Sound water management institu-
water resource management in Africa. Water tions are necessary to ensure sustained returns
security reflects a country’s ability to function on infrastructure investments and to optimize
productively in the face of water vulnerability. the use of the water by multiple users and
It is a precondition for sustaining and increas- across administrative and political borders.
ing investment returns and achieving dynamic The estimated annual capital cost of water
economic growth. A minimum capacity of resource infrastructure is approximately $10
infrastructure and institutions, backed by billion, of which almost 80 percent is for
robust water information systems, is needed development of large multipurpose hydro-
to ensure basic national water security. power storage, and about 10 percent each is
271
272 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
for development of large storage capacity for of about 3–4 degrees Celsius by the end of
urban water supply and investment in devel- the 21st century, compared with the period
oping small-scale infrastructure projects. As from 1980 to 1990 (IPCC 2007). The semiarid
a complement to these physical investments, margins of the Sahara and the central part of
Africa will need an additional $1.0 billion a southern Africa will be most affected, whereas
year to develop hydrological networks, meet equatorial latitudes and coastal areas will be
gaps in water information, and develop water least affected. The Intergovernmental Panel
management institutions. on Climate Change projects that the mean
annual precipitation and runoff will decrease
for northern and southern Africa and increase
Water Resources and Economic for eastern Africa by midcentury. Precipitation
intensity will also likely increase for the entire
Development: Challenges continent; the benefit of increased rainfall in
for Africa the wetter areas may be negated by the rainfall
being concentrated in more extreme weather
Africa faces difficult water legacies in the form events and thus less usable. In the drier areas,
of high hydrological variability and a mul- the spread between high and low runoff will
tiplicity of transboundary river basins. Both likely increase, substantially complicating the
challenges can be impediments to the conti- challenge of water resource management.
nent’s economic growth. The second challenge is Africa’s political and
geographic legacy in which several countries
Africa’s Water Legacies share the same river basins. Africa has more
Africa faces a complex challenge in water than 60 transboundary rivers, with many coun-
resource management because of two legacies. tries sharing the same basin. International river
The first is its natural legacy of high hydro- basins cover more than 60 percent of the con-
climatic variability. The amount of water in tinent, and virtually all the region’s rivers cross
Africa is comparable to that in other regions several borders: the Nile crosses 10, the Niger 9,
of the world; the continent has 9 percent of the Senegal 4, and the Zambezi 8 (figure 14.2).
the world’s water resources and 11 percent of Shared water resources present a manage-
its population. However, Africa’s water endow- ment challenge and require investment in
ment conceals the fact that rainfall across much transboundary water management capacity
of the continent is variable and unpredictable, and institutions, even if they also offer oppor-
both between and within years. Interannual tunities for joint action and cooperation.
rainfall variability in Africa, especially in east- Cross-border rivers have further implications
ern and southern Africa, is high. These regions for regional security and development, partic-
experience year-to-year variations exceeding ularly as Africa tries to develop and manage its
30 percent around the mean, a rate much water resources for economic development.
greater than the temperate climates in Europe These legacies, compounded by underde-
and North America (figure 14.1). High seasonal velopment of water infrastructure, present
variability compounds these effects, causing significant social, economic, and political risks.
droughts and floods. Runoff in Africa is extraor- The region’s weak capacity to buffer the effects
dinarily low, only half that in Asia, Australia, of hydrological variability creates uncertainty
Europe, and North America, despite having the and risk for economic activity. The expectation
same average precipitation. Low runoff coupled of variability and unpredictability in rainfall
with high rainfall variability explains the unpre- and runoff can encourage risk-averse behav-
dictable, and relatively low, seasonal and annual ior at all levels of the economy. It discourages
flows in many African rivers. investment in land, advanced technologies, or
Climate change is expected to increase agriculture. An unreliable water supply is also
this variability. The Intergovernmental Panel a significant disincentive for investments in
on Climate Change forecasts that Africa will industry and services. Growing demands for
experience a significant rise in temperature water generate competition over water use,
Water Resources: A Common Interest 273
Figure 14.1 Interannual Hydroclimatic Variability in Africa, by Selected Regions and Countries
60 130
40 110
20 90
0 70
–20 50
–40 30
–60 10
–80 –10
1880 1900 1920 1940 1960 1980 2000 1 2 3 4 5 6 7 8 9 10 11 12
b. East Africa b. Uganda
80 200
60 180
160
40
140
20 120
0 100
–20 80
60
–40
40
–60 20
–80 0
1880 1900 1920 1940 1960 1980 2000 1 2 3 4 5 6 7 8 9 10 11 12
c. Southeast Africa c. Malawi
80 350
mean monthly variability (mm)
annual rainfall anomaly (%)
60 300
40 250
20
200
0
150
–20
–40 100
–60 50
–80 0
1880 1900 1920 1940 1960 1980 2000 1 2 3 4 5 6 7 8 9 10 11 12
mean + standard deviation
mean
mean – standard deviation
posing social risks to poorer communities. those years (World Bank 2004). Mozambique’s
Weak international relations and institutions GDP growth is reduced by over 1 percentage
for managing the international river basins point annually because of water shocks (World
may cause regional tensions and possible con- Bank 2007). In Zambia, a study of how hydro-
flicts among riparian countries. logic variability affects the economy found that
rainfall variability will cost the country $4.3
Effects on Economic Development billion in lost GDP over 10 years, and it lowers
Hydrological variability causes significant eco- the country’s agricultural growth by 1 percent-
nomic loss and constrains growth (Grey and age point each year (World Bank 2008).
Sadoff 2006b). Africa lacks the capacity to buf- Underdevelopment of water resources leads
fer the shocks of frequent droughts and floods. to underuse of economic potential. Water
The abundance or shortage of rainfall affects resources in Sub-Saharan Africa compare well
national agricultural outputs. In Kenya, losses in absolute terms with other countries in the
from flooding caused by El Niño in 1997–98 world. The region has 9 percent of the world’s
and drought caused by La Niña in 1998–2000 water resources and about 6,000 cubic meters
ranged from 10 to 16 percent of GDP during of annual renewable water resources per capita
274 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 14.2 Africa’s International River Basins Weak institutional capacity in river basin
planning and management creates potential
conflicts and lost benefits. Conflicts are emerg-
ing over water allocation and use in different
parts of the region. Competing claims over
water have been asserted over many of the lakes
(for example, Lake Victoria and Lake Tana)
MAURITANIA
and river basins for economic and environ-
CAPE VERDE
MALI NIGER
CHAD ERITREA
mental uses. The growing demand for water
SENEGAL
GAMBIA
GUINEA-BISSAU BURKINA FASO
SUDAN
DJIBOUTI
from the major sectors of African economies,
GUINEA
SIERRA LEONE
BENIN
TOGO
NIGERIA
ETHIOPIA
SOMALIA especially agriculture, imposes a serious con-
CÔTE D’IVOIRE GHANA
LIBERIA CENTRAL AFRICAN REPUBLIC
CAMEROON
straint on the medium- and long-term growth
EQUATORIAL GUINEA
in water availability in some river basins. The
SAO TOME AND PRINCIPE
GABON
UGANDA
RWANDA
KENYA
expected growth in hydropower production
CONGO, DEM REP
CONGO BURUNDI
SEYCHELLES will likely require an increase in peak capac-
TANZANIA
ity that will add to the competition. However,
COMOROS
the water management institutions needed to
ANGOLA MALAWI MAYOTTE
SAINT HELENA
ZAMBIA address these conflicts in many African coun-
ZIMBABWE
MOZAMBIQUE
MADAGASCAR
tries are weak and fragmented. Agencies with
NAMIBIA
BOTSWANA
REUNION authority over a particular economic sector
often make uncoordinated decisions about
SWAZILAND
SOUTH AFRICA
water allocation and use, which lead to inef-
LESOTHO
international river basins ficiency and degradation of the resource. The
absence of water rights regimes and incentives
for efficient water allocation and conservation
Source: Oregon State University 2005. Reprinted with permission.
contributes to the problem.
Cooperative management of water resources
compared with Asia’s 4,000 cubic meters and in international river basins is necessary to
the Middle East and North Africa’s 1,500 cubic increase the basins’ yield of food, power, and
meters. However, Africa has the lowest water economic opportunities while strengthening
withdrawal per capita in the world (about environmental sustainability and mitigat-
170 cubic meters) because of hydrologic ing the effects of droughts and floods. The
variability, underdeveloped water infrastruc- cost of noncooperation is high, including
ture, and current water resource manage- the economic cost of negative environmental
ment. Asia and Europe use about three times impacts, suboptimal water resource develop-
the water per person; the water-scarce Middle ment, political tensions over shared resources,
East withdraws more than four times as much and the forgone benefits of joint water
and North America more than eight times as resource development (Sadoff, Whittington,
much (table 14.1). and Grey 2003).
The low level of water mobilization leads Lack of water infrastructure and inadequate
to underdevelopment of Africa’s economic water management mostly affect the poor.
potential: less than 5 percent of cultivated land Africa’s poverty is closely linked to its depen-
is irrigated, and less than 10 percent of hydro- dence on rain-fed subsistence farming. About
power potential is developed (figure 14.3). Irri- 28 percent of Africa’s working population is
gated land contributes only about 10 percent of engaged in agricultural production, ranging
the agricultural production in Africa, with only from 4 percent in South Africa to 47 percent
about 8.5 percent of cultivated land irrigated in Rwanda (You 2008). Because subsistence
(figure 14.4). Less than 58 percent of Africa’s agriculture is the dominant livelihood, rain-
population has access to drinking water, fall, droughts, and floods, combined with the
and 31 percent to sanitation services (WHO/ weak marketing network and difficult physical
UNICEF 2006). access to many areas, affect food security across
Water Resources: A Common Interest 275
the region. These factors, along with limited Figure 14.3 Africa’s Hydropower Potential
irrigation and underdevelopment of water
infrastructure, increase the rural economy’s
vulnerability to water shocks. Annually, some
220 million Africans are exposed to drought,
and more than 1.5 million were affected by
floods in 2007.
Degradation of water catchments under- MAURITANIA
CHAD ERITREA
DJIBOUTI
ment of Africa’s water catchments has led to SAO TOME AND PRINCIPE
GABON
UGANDA
RWANDA
KENYA
excessive soil erosion, increased costs of water REPUBLIC OF CONGO CONGO, DEM REP BURUNDI
SEYCHELLES
MADAGASCAR
missioning. The most important dams in the BOTSWANA
REUNION
NAMIBIA
LESOTHO
mentation. In Kenya, the rate of sediment out-
flow from the Athi-Galana-Sabaki River into
the Indian Ocean increased from about 50,000
Source: International Energy Agency 2006 (EI-Gazzar and others 2007, p.7).
tons a year in the 1950s to 8.4 million tons a
year by 1992 (World Bank 2004). In countries
with degraded water catchments, develop- Regular monitoring of the hydrological sys-
ment of water storage infrastructure needs to tem (such as meteorological stations, rain
be accompanied by improved protection of gauges, and river flows) is steadily declin-
watersheds to sustain the investment. ing, and most African countries have not
Reliable hydrological and water quality updated their assessments. Africa also lags
data are needed for effective water resource behind the rest of the world in the number
management and informed decision making. of meteorological stations where data can be
Hydrographic networks are outdated or in systematically collected for dissemination to
need of rehabilitation in many countries. users. According to the World Meteorological
276 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 14.4 Sub-Saharan Africa’s Irrigation Potential water vulnerability. It is a precondition for sus-
30° 20° 10° 0° 10° 20° 30° 40° 50° 60°
taining and increasing investment returns and
0 500 1,000 1,500 KILOMETERS
achieving dynamic economic growth.
Mediterranean
MOROCCO
TUNISIA Sea
0 500 1,000 MILES
The more vulnerable an economy is to
30° Canary Is.
(Sp)
30° water variability, the greater is the required
ALGERIA
FORMER
SPANISH
LIBYA ARAB
REP. OF
investment to achieve water security. If a
EGYPT
SAHARA
country cannot provide water security, it will
Re
d
20° 20°
not be resilient to water shocks, and it will not
Se
CAPE MAURITANIA
a
VERDE
SENEGAL
MALI NIGER
CHAD
ERITREA have reliable water supplies (figure 14.5, sce-
THE GAMBIA
GUINEA-BISSAU
BURKINA
FASO
SUDAN
DJIBOUTI nario 1; Subramanian, Yu, and Dankova 2008).
GUINEA
10° 10°
When an acceptable level of water security is
BENIN
NIGERIA
TOGO
GHANA
ETHIOPIA
SIERRA LEONE CÔTE
D’IVOIRE CENTRAL
LIBERIA AFRICAN REPUBLIC
SOMALIA achieved along with basic climate resilience
CAMEROON
EQUATORIAL GUINEA
SEYCHELLES
ture contributes to economic growth (figure
10° 10°
ANGOLA
COMOROS
Mayotte
14.5, scenario 2). Climate change will likely
MALAWI
Water-Managed Areas ZAMBIA
(Fr)
impose additional costs for achieving and sus-
as Percentage of Irrigation
Potentials ZIMBABWE
MOZAMBIQUE MADAGASCAR taining water security through its effect on the
20° 20°
76%–100%
NAMIBIA
BOTSWANA MAURITIUS spatial and temporal pattern of water demand
51%–75%
SWAZILAND
and availability, as well as by increasing hydro-
26%–50%
10%–25% SOUTH
LESOTHO
logical variability in certain areas.
30° 30°
–10% AFRICA
Water security is a dynamic state. Its defi-
INTERNATIONAL BOUNDARIES
This map was produced by the Map Design Unit of The World Bank.
The boundaries, colors, denominations and any other information
nition varies in different parts of the world,
Source: Aquastat 2005
shown on this map do not imply, on the part of The World Bank
Group, any judgment on the legal status of any territory, or any
endorsement or acceptance of such boundaries.
reflecting geographic, social, and political
30° 20° 10° 0° 10° 20° 30° 40° 50° 60° factors and the stage of economic develop-
ment. In Africa, hydrologic variability and
Source: ECA 2006.
extremes are at the heart of water vulnerability
because they weaken growth and retard devel-
opment. African countries must achieve water
Organization, Africa has only 1,150 observation
security to keep the risks of droughts, floods,
stations—one-eighth the recommended num-
and unreliable water supplies at a socially and
ber (UNFCCC 2006).
economically acceptable level. Several studies
have pointed to the strong correlation between
rainfall variability and national GDP in coun-
Addressing the Challenges tries as diverse as Ethiopia, Kenya, Lesotho,
Mozambique, and Zambia (World Bank 2004,
Achieving water security is a prerequisite for 2006, 2007, 2008, 2009). Emerging evidence
attracting investment and promoting eco-
nomic growth in Africa.
Figure 14.5 Water Security and Growth
Water Security, Investments, and GDP
Growth
Minimum water infrastructure and institu- Scenario 2: water
enhances growth...
tional capacity are required to ensure basic ... basic water security
is achieved
national water security (Grey and Sadoff
2006b). Water security is the ability of a coun- Scenario 1: water-
vulnerable economy...
try to function productively despite its inherent ...with climate change
Tunisia
Morocco
Algeria
Libya
A.R. of Egypt
Former
Spanish
Sahara
Mauritania Mali
Niger Chad Sudan
Eritrea
Senegal
Gambia
Guinea-Bissau Djibouti
Guinea Burkina Faso Nigeria
Somalia
Sierra Leone
Togo Central African Republic
Liberia Ethiopia
legend Tanzania
potential for poverty reduction
high
Angola
moderate
Zambia
low
Zimbabwe
country boundaries
Namibia Botswana Mozambique
rivers
waterbodies
Swaziland Madagascar
Lesotho
South Africa
Kilometers W E
0 250 510 1,000 1,500 2,045
S
However, at the same time, water is a limiting environmental sustainability and mitigating
factor for these rural livelihoods because of the effects of droughts and floods. Countries
significant seasonal and interannual variabil- around the Senegal and Niger Rivers have
ity and lack of water control. Such zones offer started to reap the benefits of a cooperative
the greatest opportunities for expanding food and joint approach to managing their shared
production through irrigation, rain-fed agri- waters (Andersen and others 2005; Yu 2008).
culture, rainfall harvesting, and conservation Countries surrounding Lake Victoria are gain-
of soil moisture. ing environmental benefits from their joint
Community-based management and devel- efforts to eliminate encroaching water hya-
opment of local watersheds and groundwater cinth and other weeds. The systemwide yield
resources in the poorest areas are crucial for of water in the Nile could likely be increased by
sustaining local livelihoods against the effects several percentage points a year if cooperation
of climate variability. These measures include led to water storage upstream and coordinated
constructing small hydraulic structures, intro- reservoir operation in the arid plains down-
ducing local-scale hydropower units, harvesting stream (Sadoff, Whittington, and Grey 2003).
water, developing smallholder irrigation, and The countries will also likely benefit from joint
installing flood protection measures. investments in water infrastructure, thereby
reducing infrastructure costs and maximizing
Building Climate Resilience returns. Unilateral and uncoordinated deci-
Managing rainfall variability is a critical task sions on water resources by riparian countries,
for African countries. Expanding water storage in contrast, may lead to increased political
capacity (as well as improving management tensions and conflicts, reducing development
of the existing hydraulic structures) would opportunities in the basins.
mitigate the effects of water shocks and build
climate resilience. Along with large storage
and hydraulic structures, small multiobjective Investing in Africa’s Water
water development is an important opportu- Security
nity for new investment. These investments
should take into account the potential effects Because most African countries have low
of climate change. Appropriately designed irri- stocks of hydraulic infrastructure, emphasiz-
gation investments would increase agricultural ing investments in infrastructure is appropri-
productivity and significantly mitigate the dan- ate for them. However, institution building
ger of rainfall variability. Investments to reduce and reform, improvements in water manage-
climate vulnerability should also include ment and operations, and strengthening of
hydrometeorological services and monitoring, water information systems must complement
catchment protection and management, and growth in infrastructure. Development of
risk assessment and mitigation. institutions is a lengthy and costly process, and
adequately sequenced and balanced, it should
Fostering Transboundary Cooperation be advanced in parallel with infrastructure
Optimizing economic productivity and envi- investments, paying particular attention to the
ronmental sustainability means managing riv- development of river basin organizations.
ers as hydrological units at the basin level. The
great hydrologic challenges that countries face Investing in Storage Infrastructure:
at the national level of water resource manage- Both Large and Small
ment (such as rainfall and runoff variability, Development of water infrastructure is a pre-
degrading water quality, and flood protection) requisite for water security in Africa and for
create opportunities to gain significant bene- meeting the targets of the eight Millennium
fits from cooperation over shared river basins. Development Goals. Artificial water storage of
Cooperative management of water resources adequate capacity is needed to ensure reliable
increases a basin’s yields of food, power, and water supply during droughts and to retain
other economic goods while strengthening excessive water during periods of flooding.
280 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Despite Africa’s vulnerability to frequent The direct and indirect long-term eco-
droughts and floods, storage capacity remains nomic benefits of investing in large storage
underdeveloped. Average per capita storage capacity are many, but they require consider-
capacity in Africa is about 200 cubic meters a able initial capital investments. The aggregate
year, much less than that of countries in other spending needs for African infrastructure
regions (figure 14.7). were reported in chapter 1 of this volume, but
Irrigation and hydropower have been the because water resource investments are typi-
main drivers for dam construction in Africa, cally buried within investment programs for
but a storage facility designed for a single other sectors, such as irrigation, power, and
purpose has limited capacity to serve other water supply, the specific water resource com-
economic and social needs, and its invest- ponents are explicitly highlighted here.
ments often have higher opportunity costs. Table 14.2 details the component of power
For example, the Cahora Bassa Dam on the sector investment needs explicitly attribut-
Zambezi River was constructed for the sole able to water storage in large dams. If regional
purpose of generating hydropower. The dev- power trade could be effectively harnessed,
astating 2000 floods in Mozambique showed some 50,000 megawatts of new hydropower
that the dam could play an important flood capacity would need to be built from 2006 to
mitigation role, but its operational rules do not 2015, but without expanding regional trade,
permit use of its storage capacity to mitigate only 33,000 megawatts of hydropower capac-
water shocks. To convert this dam into a multi- ity could be developed. The trade scenario
purpose reservoir would be too costly. Despite would translate into an annual average invest-
the need for power, other considerations, such ment requirement of $7.8 billion in large-
as flood control, salinity repulsion, irrigation scale water storage to support generation of
development, and environmental require- electricity over the 2006–15 period (Rosnes
ments, call for multipurpose storage develop- and Vennemo 2008, 2009).
ment. Multipurpose water projects generally Not only would expanded regional trade
result in optimal water development, maxi- lead to the development of more water storage,
mize economic returns on investments, and but it would also improve the cost-effectiveness
need to be implemented in the basinwide con- of water storage. The unit capital cost of
text. Two important principles for developing hydropower investments would fall from
large water infrastructure are equitable shar- $5.9 million to $5.4 million per megawatt
ing of the benefits with the people affected and (table 14.2) because cross-border collabora-
mitigation of possible negative environmental tion allows larger and more efficient storage
effects. Thus, stakeholder participation is nec- sites to be developed.
essary at all stages of decision making, project The key challenge is how to finance the
design, and implementation. multibillion-dollar large-scale water storage
projects needed to make this savings a real-
Figure 14.7 Water Reservoir Storage per Capita in ity. Often the countries with the best storage
Selected Countries, 2003
sites are those with the least financial capacity
7,000 to develop them. Regional collaboration offers
5,961
6,000 the possibility of cost-sharing arrangements
5,000 4,717 among countries for large water infrastructure,
cubic meters
ico
ina
il
ica
ali
an
iop
er
ex
Ch
Br
Af
str
ail
Am
h
Th
Au
Et
rth
So
No
Table 14.2 Capital Investment Needs in Large Multipurpose Hydropower Storage by 2015
Large storage-based hydropower Capital costs per unit Total capital costs
production (megawatts) ($ millions per megawatts) ($ billions per year)
Regional power No power trade Regional power No power trade Regional power No power trade
Region trade expansion expansion trade expansion expansion trade expansion expansion
Southern Africa 16,764 10,797 0.96 1.21 1.95 1.59
Eastern Africa 10,675 4,170 1.81 1.96 2.35 0.99
Western Africa 17,260 14,845 1.32 1.37 2.77 2.48
Central Africa 4,847 3,567 1.28 1.37 0.75 0.59
Total 49,546 33,379 5.37 5.91 7.82 5.65
Source: Authors’ calculations based on Rosnes and Vennemo 2008.
Note: These investment cost estimates are based on actual planned large storage-based (multipurpose) hydropower projects presented for each region in Rosnes and Vennemo
2008 and 2009. Totals do not add exactly because of averaging.
Governments have traditionally shouldered Studies in Botswana and Zimbabwe show that
the responsibility for financing large hydrau- 80–85 percent of all measurable rain can be
lic structures, including most multipurpose collected from outside catchment areas and
dams. Private sector participation in such stored (Dixit and Patil 1996).
investments is possible, however, for mul- The estimated cost for small reservoir facili-
tipurpose projects with revenue-creating ties is based on an assessment of small-scale
components (such as hydropower produc- irrigation development potential from 2006
tion, commercial irrigation, and urban water to 2015 (You 2008). This analysis used a spa-
supply functions) combined with nonrevenue tial allocation model on a 10-kilometer global
and public-good functions (such as flood grid, considering economic profitability, crop
control, fish breeding, recreation, small-scale pattern, prices, crop water productivity, water
irrigation, and so forth). Promoting public- balance and availability, and distance to mar-
private partnerships to finance multipurpose ket for each pixel. A macroscale hydrology
water resource projects is a key challenge of model used climatic data to calculate runoff
decision makers. for each pixel. The total small-scale irrigation
A continuum of options from large to area was estimated at about 10 million hect-
small infrastructure projects needs to be part ares, and the total water storage requirement
of a balanced investment approach. Small was estimated at about 35 billion cubic meters
water storage facilities can increase climate for the countries sampled. The total cost for
resilience and improve food security. Small- small irrigation storage facilities was based on
scale approaches to water management help the irrigation development area, required stor-
the rural poor by providing cost-effective age volume, and their average costs.
solutions to water supply and drought miti- A certain amount of storage is also needed
gation. They improve the ability of the rural to keep pace with growing demands for the
poor to address food insecurity by increas- urban water supply. The associated invest-
ing agricultural productivity. Small storage ment can be estimated based on demographic
options include (a) off-stream reservoirs, (b) growth, Millennium Development Goal tar-
on-farm ponds and networks of multipurpose gets, trends in water consumption, as well as
small reservoirs, (c) groundwater storage, and unaccounted for water and the availability of
(d) water storage through a root zone with a supply from various sources. The storage need
variety of water-harvesting techniques and is put at 5.4 billion cubic meters.
conservation of soil moisture. Structures to Total capital investment needs for the devel-
harvest rain require little space and are not opment of water resource infrastructure for
labor intensive, but they must be designed 2006–15 are estimated at approximately $10 bil-
in accordance with intended usage and local lion a year, included within estimates for power
circumstances to ensure cost-effectiveness. ($7.8 billion), water supply ($1.3 billion), and
282 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
support from multiple donors, enabling them Figure 14.8 Degree of Regional Water Cooperation
to successfully complete several years of insti-
Toward regional cooperation on shared waters
tutional development, confidence-building
Nile Basin Initiative Niger Basin Authority Senegal River Basin Authority
measures, and cooperative investment pro-
• 9 countries + 1 observer • 9 countries • 4 countries
gramming. The Niger Basin Authority, for • established 1999 • established 1963, revitalized 2001 • established 1972
example, undertook an institutional assessment • no formal treaty framework • formal treaty framework • formal treaty framework and
• confidence building and and formal river basin formal river basin organization
to prepare for reorienting its efforts in early investment programs in place organization • sustainable development
2000. Subsequently, it introduced an organi- • sustainable development plan in place
zational structure approved by the Niger Basin plan in place
Grey, David, and Claudia Sadoff. 2006a. “The World Bank. 2004. “Towards a Water-Secure
Global Water Challenge: Poverty Growth and Kenya: Kenya Water Resources Sector Memoran-
International Relations.” Paper presented at dum.” World Bank, Washington, DC.
Global Issues Seminar Series, World Bank, ———. 2006. “Ethiopia Country Water Resources
Washington, DC, January 25. Assistance Strategy.” Agriculture and Rural
———. 2006b. “Water for Growth and Develop- Development Department, World Bank,
ment: A Framework for Analysis.” Theme docu- Washington, DC.
ment of the 4th World Water Forum, Mexico ———.2007. “Mozambique Country Water
City, March. Resources Assistance Strategy.” Africa
Hulme, Mike, Ruth Doherty, Todd Ngara, Mark Region Water Resources Unit, World Bank,
New, and David Lister. 2001. “African Climate Washington, DC.
Change: 1900–2100.” Climate Research 17: ———. 2008. “Zambia Country Water Resources
145–68. Assistance Strategy.” Draft, Africa Region
Water Resources: A Common Interest 285
Water Resources Unit, World Bank, Africa Infrastructure Country Diagnostic, World
Washington, DC. Bank, Washington, DC.
———. 2009. “Lesotho Water Sector Improvement Yu, Winston. 2008. “Benefit Sharing in Interna-
Program (Second Phase): Metolong Dam and tional Rivers: Findings from the Senegal River
Water Supply Project.” Report 46272-LS, World Basin, the Columbia River Basin, and the
Bank, Washington, DC. Lesotho Highlands Water Project.” Working
You, Liang Zhi. 2008. “Irrigation Investment Needs Paper 1, Africa Region Water Resources Unit,
in Sub-Saharan Africa.” Background Paper 9, World Bank, Washington, DC.
Chapter
Chapter 15
A
large segment of Sub-Saharan Africa’s irrigation, involving an additional one-time
population lives in rural areas and investment of $17.8 billion. In general, economic
depends heavily on agriculture. Agricul- returns on small-scale schemes (on average
tural growth is clearly key to poverty reduction 26 percent) are substantially higher than those
and to achieving the Millennium Development on large-scale schemes (on average 17 percent).
Goal of halving poverty by 2015. Agricultural These results, however, are critically depen-
performance has significantly improved since dent on keeping investment costs down to
2000; nevertheless, agricultural productiv- best-practice levels of $3,000 per hectare for
ity remains the lowest in the world. Climate the water distribution component of large-
change and the global food price crisis provide scale irrigation and $2,000 per hectare for
additional challenges. A comprehensive effort small-scale irrigation, rather than the signifi-
is required to increase investment in agricul- cantly higher levels often observed at the proj-
tural intensification. Water for agriculture is a ect level in the recent past. Another key finding
critical ingredient of such programs. is that irrigation is in most cases viable only
Africa’s agricultural water remains com- for cash crops or high-value food crops (such
paratively underdeveloped, despite economi- as horticulture) that raise revenues in excess of
cally viable potential to expand irrigated areas. $2,000 per hectare; relatively few hectares are
Today, only 3.5 percent of Africa’s agricultural viable for irrigation of staple food crops.
land is equipped for irrigation, some 7 million What would be the effect on development
hectares concentrated in a handful of countries. of an aggressive expansion of Africa’s irrigated
However, further expansion of the irrigated agricultural area? If Africa does not increase its
area would be profitable. At least 1.4 million slow growth in irrigated area, the food supply on
hectares could be developed using existing or the continent will gradually diminish because
planned dams associated with hydropower of climate change, leading to a huge surge in
development, at a total additional one-time cereal imports and a significant increase in child
investment of $2.6 billion in distribution of malnutrition. In contrast, if Africa’s irrigated
agricultural water. In addition, at least 5.4 mil- area could be tripled by 2050, the food supply
lion hectares would be viable for small-scale would increase markedly, with a huge decline in
287
288 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
cereal imports. There would be 2 million fewer Despite these encouraging developments,
malnourished children than under the lower agricultural productivity is the lowest in the
irrigation scenario, or about the same level that world, with per capita output only 56 percent
would be expected in the absence of climate of the world average. Output has not kept pace
change. Thus, aggressive agricultural water with population increases, and growth has
development could reverse the adverse effects occurred largely through expansion of har-
of moderate global warming on food security. vested area (rather than through more intensive
Aggressive scaling up of investments in use of existing cropland): more than 80 percent
agricultural water raises issues associated with of output growth since 1980 has come from
the performance and sustainability of irri- expansion of cropped area, compared with less
gated agriculture. Considerations of economic than 20 percent for all other regions.
viability, farm-level profitability, and sustain- Climate change and the global food price
ability should guide future investment deci- crisis, developments likely to hit Africa dis-
sions. In particular, investments in agricultural proportionately, further challenge agricultural
water should be considered as part of a com- performance. Water sources will become more
prehensive package, including (a) empowered variable. Droughts and floods will stress agri-
farmer organizations; (b) sustainable, effi- cultural systems. The seas will inundate some
cient, and accountable agricultural support coastal food-producing areas, and food pro-
services; and (c) accessible, profitable markets. duction will fall in some places in the interior.
Furthermore, efforts to expand irrigation Nevertheless, substantial uncertainty remains
should be made in the context of national about where the effects will be greatest.
agricultural water development strategies that A comprehensive effort is required to
emphasize the importance of a more condu- advance agricultural productivity. Invest-
cive institutional environment and that form ments in more reliable access to agricultural
the foundation for sectoral programs that water1 are critical in support of that objective
combine investment in infrastructure with (World Bank 2008). More reliable access to
investment in institutional reforms. agricultural water increases the opportunity
to use productivity-enhancing inputs and
thus supports intensification and diversifica-
Agriculture and Poverty Reduction tion, as well as the scope for agricultural wage
employment. In addition, it reduces local food
Agricultural growth is a key to reducing poverty. prices, improving real net incomes. It can also
More than half of Sub-Saharan Africa’s popula- reduce poverty indirectly through increased
tion lives in rural areas, and agriculture accounts rural nonfarm and urban employment. More-
for a significant percentage of GDP. Of Africa’s over, investments in irrigation improve access
poor, 85 percent live in rural areas and depend to markets. A more reliable, year-round supply
largely on agriculture for their livelihoods. of products, a higher and more uniform qual-
GDP growth originating in agriculture is about ity of products, and the option to manipulate
four times more effective in raising incomes of harvest dates to capture higher seasonal prices
extremely poor people than GDP growth origi- put a high premium on irrigation.
nating outside the sector—and the potential African leaders have identified agricultural
multipliers from agricultural water investment water development as a key area for investment.
are even higher (World Bank 2008). The Comprehensive Africa Agricultural Devel-
Agricultural performance in the region has opment Program prepared under the New
improved significantly since 2000. Growth in Partnership for Africa’s Development in 2002
agricultural GDP in Sub-Saharan Africa has adopted land and water management as the
accelerated from 2.3 percent a year in the 1980s first of its four pillars for priority investment. It
to 3.8 percent a year from 2000 to 2005. Aver- proposed extending the area under sustainable
age incomes in Sub-Saharan Africa have been land management and reliable water control
rising in tandem with those in other regions systems to 20 million hectares (more than twice
since the mid-1990s. the area currently under water management in
Irrigation: Tapping Potential 289
Sub-Saharan Africa) by 2015 (NEPAD 2003). In irrigation in Sub-Saharan Africa varies consid-
response to this call for action, the Partnership erably by country but is generally very low, with
for Agricultural Water in Africa was recently only a few countries reaching the 20 percent
launched to scale up investments in agricultural mark (figure 15.1). In absolute terms, more than
water and to harmonize donor programs. 60 percent of the total area is concentrated in
just three countries—Madagascar, South Africa,
and Sudan—each with over a million hectares
Current State of Irrigation of irrigated area.
Donor investments in agricultural water
Irrigation carries significant potential to have declined sharply. From 1994 to 1996, the
increase agricultural productivity. Across Sub- total value of projects funded by all donors for
Saharan Africa, irrigated agriculture accounts irrigation and drainage was less than 10 percent
for about 25 percent of the value of agricultural of the levels 20 years earlier—just $127 million
output (table 15.1). This share is produced on from all sources (World Bank 2007). Significant
just 3.5 percent of the cultivated land, confirm- scope exists for scaling up investments. Average
ing the potential of irrigation to improve live- agricultural water withdrawals are 1.3 percent
lihoods in Sub-Saharan Africa and suggesting of renewable water resources; groundwater use
that more investment in irrigation would yield is less than 20 percent of renewable supplies,
substantial benefits. indicating significant scope for further surface
However, Sub-Saharan Africa’s agricultural water and groundwater development. Expan-
water remains underdeveloped. Of a cultivated sion of irrigated area has been slow in Sub-
area of 197 million hectares, only 7 million hect- Saharan Africa. Over the last 40 years, only
ares is equipped for irrigation, with a further 4 million hectares of new irrigation has been
2 million hectares under some other form of developed, by far the smallest expansion of any
water management. Overall, this area amounts region. Over the same period, China added
to only 23 percent of the 39 million hectares that 25 million hectares, and India 32 million.
is believed to be physically suitable (though not Annual growth in irrigation development in the
necessarily economically viable) for irrigation. region was 2.3 percent from 1973 to 2000 and
The share of cultivated area equipped for slowed further from 2000 to 2003, but it has
Figure 15.1 Percentage of Cultivated Area Equipped picked up in the last several years, particularly
for Irrigation, by Country for the Central African Republic, Kenya, Mau-
Madagascar
ritius, Nigeria, Senegal, and Zambia. An appro-
Swaziland
priate institutional framework is required to
Mauritius manage water for growth. Sub-Saharan Africa’s
São Tomé and Principe framework of irrigation-related institutions is
Somalia undeveloped. These institutions are important
Sudan for managing and sharing water over agroeco-
South Africa logically and hydrologically diverse areas and
Mauritania over transboundary basins. The establishment
Cape Verde
and sound functioning of water management
Zimbabwe
bodies will provide the institutional framework
Guinea
Mali
and will specify the location of investment
Senegal
planning and implementation responsibilities
Sierra Leone for irrigation infrastructure projects.
Seychelles
Guinea-Bissau
Tanzania Economic Investment Potential
Eritrea and Needs
Zambia
Mozambique
Some 39 million hectares of agricultural land
Ethiopia
in Africa is deemed physically suitable for irri-
Malawi
Angola
gation; however, physical suitability does not
Kenya
necessarily entail economic viability (box 15.1).
Niger Economic potential for irrigation is highly sen-
Burundi sitive to initial investment costs and requires
Côte d’Ivoire market access, complementary inputs, extension
Nigeria of credit, and a supportive enabling environ-
Namibia ment. Both large-scale irrigation schemes (dis-
Gabon tributing water collected in major dams) and
Lesotho
small-scale irrigation schemes (collecting water
Chad
locally at the farm level) are relevant for Africa.
Rwanda
Gambia, The
Profitable dam-based, large-scale irrigation
Burkina Faso
appears feasible on 1.35 million hectares (table
Ghana 15.2), if the attention is confined only to those
Congo, Rep. of projects that pass a threshold of 12 percent for
Cameroon internal rate of return. The associated one-time,
Botswana on-farm investment is $2.6 billion, which would
Benin nonetheless need to be spread out over at least a
Togo decade. The countries with the greatest potential
Liberia
for dam-associated large-scale investments are
Uganda
Ethiopia, Nigeria, Sudan, and Zimbabwe, all with
Congo, Dem. Rep. of
Comoros
more than 100,000 hectares of potential. These
Central African Republic
estimates are based on the assumption that dam
0 5 10 15 20 25 30 35
construction would be deemed profitable purely
percent from a hydropower perspective and that the only
Source: Svendsen, Ewing, and Msangi 2009.
costs that would need to be covered by related
Note: Djibouti, with 100 percent of its cultivated area equipped irrigation schemes would relate to water distri-
for irrigation, is not shown here for easier visibility of other coun-
tries. The percentage of area equipped for irrigation that is actu- bution infrastructure.
ally irrigated spans a wide range in Sub-Saharan Africa. Overall, The small-scale irrigation potential is much
the average use rate is 71 percent in the region, compared with
a similar but slightly lower 67 percent in Asia. greater because large, existing rain-fed areas
Irrigation: Tapping Potential 291
BOX 15.1
Table 15.2 Potential Investment Needs for Large-Scale, Dam-Based, and Complementary Small-Scale Irrigation in Sub-Saharan Africa
Large-scale irrigation Small-scale irrigation
Increase in Increase in
irrigated area Investment cost irrigated area Investment cost
Region (million hectares) ($ millions) Average IRR (%) (million hectares) ($ millions) Average IRR (%)
Sudano-Sahelian 0.26 508 14 1.26 4,391 33
Eastern 0.25 482 18 1.08 3,873 28
Gulf of Guinea 0.61 1,188 18 2.61 8,233 22
Central 0.00 4 12 0.30 881 29
Southern 0.23 458 16 0.19 413 13
Indian Ocean Islands 0.00 0 — 0.00 0 —
Total 1.35 2,640 17 5.44 17,790 26
Source: You 2009.
Note: See table 15.1 for definitions of regional groupings. The average value for IRR was weighted by the increase in irrigated area. Benin, Chad, and Madagascar have no
profitable large-scale irrigation. IRR = internal rate of return. — Not available.
a. Unlike cost estimates presented elsewhere in this report, these are one-time investment costs rather than annualized figures.
would nonetheless need to be spread out over under the Comprehensive Africa Agricultural
at least a decade (see table 15.2 and figure 15.2). Development Program of the New Partnership
In all regions except southern Africa, small-scale for Africa’s Development, African countries have
irrigation has a higher internal rate of return already committed themselves to raising alloca-
than does large-scale irrigation. By far the great- tions of national budgetary resources to agricul-
est potential is found in Nigeria, which accounts tural and rural development up to 10 percent
for more than 2.5 million (or almost half) of the of the total by 2015. This planned increase in
suitable hectares. Countries such as Cameroon, spending could go some way toward meeting
Chad, Ethiopia, Mali, Niger, South Africa, Sudan, the costs of an expanded irrigation program.
Tanzania, Togo, and Uganda each have at least Most of the hectares found to be viable for
100,000 hectares of potential. irrigation would be dedicated to higher-value
Full development of economic irrigation crops (table 15.3). In most cases, irrigation is
potential doubles the share of cultivated land found to be viable only when high-revenue-
under irrigation, raising the share of cultivated yielding crops are cultivated, be they traditional
land under irrigation from 3.5 percent to 7.0 per- cash crops (such as coffee) or higher-value food
cent. Annualized investment costs over 10 years crops (such as horticulture). More than half of
would constitute 1.8 percent of 2000 GDP and the viable hectares identified are associated
account for 88 percent of agricultural spending. with crops that can yield in excess of $2,000
For many countries, that would imply a substan- per hectare annually. Relatively few hectares
tial increase in agricultural spending. However, are found where irrigation investments can be
justified simply to grow staple food crops.
The investment cost estimates used here
Figure 15.2 Investment Potential for Dam-Based and Small-Scale Irrigation reflect best-practice experience, but actual
costs may often be higher. Studies suggest that
well-designed and well-implemented irrigation
projects in Africa can lead to costs of no more
than $2,000 per hectare for small-scale irrigation
schemes and $3,000 per hectare (the distribution
component) for large-scale irrigation schemes.
Therefore, they are the central parameters used in
this modeling exercise. Nevertheless, in practice,
irrigation projects in Africa may incur invest-
ment costs well in excess of $4,000 per hectare
(Inocencio and others 2005). Therefore, consid-
ering the sensitivity of these results to possible
changes in unit investment costs is important.
Estimates for both large- and (particu-
larly) small-scale irrigation potential are
highly sensitive to assumptions about invest-
ment costs (table 15.4). Results are sensi-
tive to the assumptions about investment
costs per hectare and other parameters (see
Profitable area in small scale
Profitable area in large scale
box 15.1). For large-scale irrigation, the
number of viable hectares would decline to
54 percent of the base case if investment costs
rose from $3,000 to $6,000 per hectare. This
situation might be the case, for example,
if irrigation were required to contribute to
water storage costs and not simply to water
Source: You 2009. distribution infrastructure. For small-scale
Note: Dark gray areas indicate positive internal rate of return for dam-based irrigation. Light gray areas
indicate profitable areas for small-scale irrigation. Countries left blank were not covered in the sample. irrigation, the story is much more dramatic.
Irrigation: Tapping Potential 293
Table 15.3 Share of Crops under Irrigation, Irrigation Investment Needs Assessment
Crops Average revenue ($/hectare/year) Percentage total viable hectares
Sugarcane, barley, soybeans, other pulses < 100 16
Bananas, beans, potatoes, sorghum, sugar beets,
ground nuts, cassava, maize, cotton lint 100–500 14
Coffee, rice, sweet potatoes, millet 500–1,000 16
Horticulture and other high-value crops > 2,000 54
Source: Derived from You 2009.
Raising the investment cost from the baseline the food economy looking ahead to 2020 and
case of $2,000 per hectare to $5,000 per hect- beyond.
are would all but eliminate the economic case A high but feasible 3.6 percent annual
for small-scale irrigation. This finding under- increase in investments in irrigation would
scores the fact that the economic viability of triple the irrigated harvested area to 22 mil-
much of Africa’s potentially irrigable land lion hectares in 2050. The irrigated yield would
depends critically on containing investment grow by 10 percent in parallel with a gradual
costs to best-practice levels. 10 percent increase in rain-fed crop productiv-
ity (by increasing effective rainfall by 5 percent
by 2020 and 10 percent by 2050). The effects
of these investments are discernible in three
Effect of Expanding Agricultural areas: food prices, food imports, and nutri-
Water Development tional outcomes.
Greater availability of food would help con-
What would be the development effect of an tain major projected increases in food prices
aggressive expansion of Africa’s irrigated agri- (table 15.5). Food prices in Africa are projected
cultural area? It has already been established to increase from the 2000 baseline level of $117
that an economic case exists for developing an per ton of cereals to $205 by 2050. The greater
additional 6.8 million hectares of irrigation food production brought about by irrigation
with respectable returns of at least 12 percent. scale-up would help limit this inflationary
This conclusion comes with two important pressure, keeping the price of a ton of cereals
caveats. First, investment costs would need to down to $177 by 2050.
be contained to the $2,000–$3,000 per hectare Irrigation scale-up would also reduce
range. Second, to ensure viability, most irri- Africa’s reliance on food imports (table 15.5).
gation development would need to focus on As of 2000, Africa was importing more than
higher-value crops. Assuming such conditions 23,000 tons of cereals annually. Africa’s food
were met, what would be the broader develop- trade deficit is projected to grow dramatically
ment effect? An illustrative modeling exercise to more than 98,000 tons in 2050, reflecting a
is used to see how the higher agricultural yields substantial rise in food demand from growing
resulting from irrigation scale-up would affect and increasingly urban populations combined
294 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Table 15.5 Food Price Changes for Various Indicators, 2020 and 2050
Rain-fed Irrigated
Average Net cereal Number of Caloric Rain-fed Irrigated production, production,
world price imports malnourished availability area, cereals area, cereals cereals cereals
for cereals (thousand children (kilocalories/ (thousand (thousand (thousand (thousand
Year ($/metric ton) tons) (thousands) person/day) hectares) hectares) metric tons) metric tons)
Projection of current investment levels with climate change
2000 117 23,638 32,669 2,277 74,303 3,783 75,283 6,829
2020 187 4,370 44,041 2,241 87,109 4,847 132,184 12,851
2050 205 75,417 33,756 2,761 92,908 6,294 203,680 26,011
Projection of current investment levels without climate change
2000 117 23,638 32,669 2,277 74,303 3,783 75,283 6,829
2020 179 6,398 43,646 2,263 86,908 4,858 132,125 12,891
2050 159 98,963 31,894 2,886 92,441 6,441 204,427 26,454
Increased investments with climate change
2000 117 23,638 32,669 2,277 74,303 3,783 75,283 6,829
2020 182 −7,331 42,507 2,235 85,793 7,666 138,904 18,625
2050 177 11,134 31,640 2,852 89,560 21,722 220,820 86,003
Source: Estimates provided by International Food Policy Research Institute, Washington, D.C., 2009.
1. Agricultural water and irrigation are used NEPAD (New Partnership for Africa’s Develop-
interchangeably, and both refer to the supply of ment). 2003. Comprehensive Africa Agricul-
additional water to augment rainwater (if any) tural Development Programme. http://www
for crops and livestock. In this report, agricul- .nepad-caadp.net/.
tural water and irrigation include drainage, Svendsen, Mark, Mandy Ewing, and Siwa Msangi.
where appropriate. 2008. “Watermarks: Indicators of Irrigation Sector
Performance in Sub-Saharan Africa.” Background
Paper 4, Africa Infrastructure Country Diagnos-
References tic, World Bank, Washington, DC.
Frenken, Karen, ed. 2005. Irrigation in Africa in
Figures: AQUASTAT Survey—2005. FAO Water World Bank. 2007. Investment in Agricultural Water
Report 29. Rome: Food and Agriculture for Poverty Reduction and Economic Growth in
Organization of the United Nations. Sub-Saharan Africa: Synthesis Report. Washing-
ton, DC: World Bank.
Inocencio, Arlene, Masao Kikuchi, Manabu
Tonosaki, Atsushi Maruyama, and Hilmy Sally. ———. 2008. World Development Report 2008:
2005. “Costs of Irrigation Projects: A Compari- Agriculture for Development. Washington, DC:
son of Sub-Saharan Africa and Other Developing World Bank.
Regions and Finding Options to Reduce Costs.” You, Liang Zhi. 2008. “Irrigation Investment Needs
Report of component study for Collaborative in Sub-Saharan Africa.” Background Paper 9,
Programme, International Water Management Africa Infrastructure Country Diagnostic, World
Institute, Pretoria, South Africa. Bank, Washington, DC.
Chapter
Chapter 16
S
ub-Saharan Africa as a whole is unlikely to forgoes at least $1.8 billion a year in revenues.
meet the Millennium Development Goal Typically, capital costs have been subsidized,
(MDG) for water supply. Coverage in but the subsidies are highly regressive. Full
urban areas has been declining as utilities have capital cost recovery should be affordable for
struggled to keep pace with population growth. half of the population, including the bulk of
In rural areas, more than 40 percent of the those that enjoy piped water access today, but
population continues to rely on surface water. would not be affordable to the remainder.
Overall, wells and boreholes are the fastest- Furthermore, the operational inefficiencies
growing sources of supply. of water utilities cost the region $0.9 billion
The price tag for reaching the MDG for a year and impede service expansion. Insti-
access to an improved water source is estimated tutional reforms of legal and regulatory
at $16.5 billion a year (roughly 2.6 percent of frameworks hold the key to improving per-
Africa’s GDP). For many countries, these costs formance. Private participation, particularly
look prohibitive. By emphasizing lower-cost lease contracts, has significantly affected
technologies, such as standposts and bore- utility performance, but state-owned utilities
holes, those countries could reduce the cost of will remain the central actors, and greater
meeting the MDG. However, standpost use is efforts are needed to improve their gover-
affected by institutional challenges that remain nance frameworks.
to be addressed. Even if all these inefficiencies could be
Spending on the water sector today is eliminated, the overall financing gap for the
$3.6 billion, one-fourth of what is required. water sector would still be $7.8 billion a year
However, some $2.7 billion available to the sector (1.2 percent of GDP).
is currently being wasted due to inefficiency. Looking ahead, the institutional reform
An important example of inefficiency is agenda remains as relevant as before, even
underpricing of services. Average water tar- if the focus has shifted toward a more plu-
iffs are about $0.67 per cubic meter, below the ralistic view of public and private sector
cost-recovery threshold of just over $1.00 per roles. The reform agenda also needs to move
cubic meter. By underpricing water, the sector beyond utilities to encompass line ministries
299
300 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
and the whole public expenditure framework Guinea, Mali, Mauritania, Senegal, Sudan,
that underpins, and too often hinders, sec- Uganda, and Zimbabwe. At the other end of
tor investment programs. Room also exists the spectrum, some of Africa’s most populous
for more cost recovery, so that scarce subsidy countries, such as the Democratic Republic of
resources are redirected to promote access Congo and Nigeria, are a long way from meet-
among the poorest. For the majority that ing the target. The proximity of 2015 and the
does not enjoy access to a piped-water con- daunting challenges highlight the importance
nection, greater thought needs to be given to of understanding the performance of the water
how standposts can become a more effective sector in the region, its achievements and
part of urban water supplies. The burgeoning shortcomings, and the factors most critical in
use of wells and boreholes for supply in urban expanding coverage.
areas demands urgent attention from policy Service options for water supply can be
makers both to improve their understanding organized on a hierarchical ladder accord-
and to develop suitable regulatory tools. In ing to the delivery method and quality of the
rural areas, the big challenge, in addition to associated service. At the top of the ladder is
continuing to expand access, is the high break- piped water, which is both potable and conve-
down rate from lack of maintenance, which nient. Standposts offer the same potability, but
threatens the sustainability of what has already through a less convenient channel and with
been achieved. some risk of the water becoming contaminated
during collection. Next come wells and bore-
holes, which, depending on their location, can
The Millennium Development be more or less convenient than standposts.
Goal for Water—Elusive for Many The water delivered can be of good qual-
ity, though that depends on the local aquifer
Whereas the rest of the world is on track to and protection from contamination. Surface
achieve the MDG for water supply,1 Sub- water is at the bottom of the ladder, because
Saharan Africa reports that only 58 percent of its quality is in most cases questionable, and
its population enjoys access to safe drinking it is seldom convenient. Although the objec-
water vis-à-vis a target rate of 75 percent to tive may be to provide universal piped-water
be reached by 2015 (WHO/UNICEF 2006).2 access to the population, it may not be feasible
Progress has been modest, with access increas- or affordable in the short run. An important
ing by only 9 percentage points between 1990 first step is to move people away from surface
and 2006, or less than 1 percentage point a water to one of the lower rungs of the water
year. To meet the target, growth should stand supply ladder.
at over 2 percentage points a year. As a result,
Sub-Saharan Africa lags all other regions,
including South Asia, whose performance was Differing Patterns of Urban and
broadly comparable to that of Sub-Saharan Rural Access
Africa in the past but which has moved at a
much faster pace in recent years. In rural areas, reliance on surface water remains
Some countries are closer to meeting the prevalent. The share of the population relying
MDG targets than others. According to the on surface water fell quite steeply in the 1990s,
most recent WHO/UNICEF Joint Monitoring from 50 percent to just over 40 percent, where
Programme data for 2006, five African coun- it has remained through 2005 (table 16.1).
tries have already met the MDG target: Burkina Boreholes are the main source of improved
Faso, Ghana, Malawi, Namibia, and South water, accounting for a further 40 percent of
Africa. Moreover, an additional 12 countries the population. Access to piped water and
had reasonable prospects of meeting the tar- standposts is very low, barely increasing over
get by 2015 if they continued to make steady the last 15 years. Indeed, in many countries,
progress: Cameroon, the Central African less than 1 percent of the rural population
Republic, the Comoros, Côte d’Ivoire, Eritrea, receives piped water. Strikingly, in countries
Water Supply: Hitting the Target? 301
with higher levels of urbanization, access to Table 16.2 Services Provided by Utilities in Their Service Areas
piped water and standposts in rural areas is percentage
substantially higher. Population already enjoying Population gaining access
access to utility water to utility water each year
In urban areas, rapid population growth
has caused piped-water coverage to fall mark- Resale of
Country Piped neighbor’s Piped
edly over the last 15 years. However, at close to type water Standposts water Total water Standposts Total
40 percent, it is still the single-largest source of Low
water supply. Standpost coverage has similarly income 30 19 15 68 1.9 1.0 2.9
declined, whereas that of boreholes has risen, Middle
so that each of these types of service reaches income 89 10 0 99 4.5 −0.2 4.5
about 24 percent of the urban population. Averagea 44 14 6 86 2.5 0.9 3.2
The lower coverage of standposts compared Source: Banerjee, Skilling, and others 2008.
Note: The figures are based on utility data for utility service areas. The coverage is higher than that
with piped water is particularly striking, given obtained from household surveys because utility service areas do not cover the entirety of urban areas
their relatively low cost and the pressure to and because for some countries, data were available only for utilities in larger cities.
a
Average is population weighted.
expand services rapidly. Reliance on surface
water, at 7 percent of the urban population,
has hardly changed.
Utilities are the central actors respon- of the unconnected households. Household
sible for water supply in urban areas. Over- resale, while prevalent, is often illegal, although
all, about two-thirds of the urban populace Côte d’Ivoire illustrates the potential benefits
depends on utility water. In the middle- of legalization (box 16.1).
income countries, utilities are essentially the Utilities report providing around 20 hours
only players, reaching about 99 percent of the of service daily, and just over 80 percent of their
urban population, the vast majority through samples pass chlorine tests (table 16.3). They
private piped-water connections. In low- typically produce just over 200 liters per capita
income countries, only 49 percent of urban served, though the amount for middle-income
residents benefit from utility water, fewer countries is about twice that for low-income
than half through private piped connections countries. If the utilities’ total water produc-
(table 16.2). For the rest, informal sharing of tion could be evenly distributed to the entire
connections through resale between neigh- population residing in the utility service area, it
bors (15 percent of the urban population) is would amount to 74 liters per capita a day, just
almost as prevalent as formal sharing through about adequate to meet basic human needs.
standposts (19 percent of the urban popula- Urban households that do not benefit
tion). In Maputo, Mozambique, one-third from utility water rely on several alternatives.
of unconnected households purchase water The rapid expansion of boreholes in urban
from their neighbors, and in Maseru, Lesotho, areas has already been noted. Water vendors,
household resellers provide water to 31 per- another alternative, may retail water from
cent of the population, including almost half utilities, boreholes, or surface sources, either
302 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
BOX 16.1
Table 16.3 Quality of Services Provided by Utilities in Their Service Areas coverage much more rapidly than all the
Availability of utility water Quality of water supply utility-based alternatives together. Within the
Liters per Liters per Percentage purview of the utility, access to standposts
capita available capita available Hours/day of of samples seems to be growing faster than access to
for residents in for utility continuous passing
Country type service area customers service chlorine test
piped water. But the combined growth rates
of the various improved forms of water sup-
Low income 74 149 19 83
ply in urban areas (less than 1 percent a year)
Middle income 272 277 24 99
still fall short of population growth (more
a
Average 167 224 21 83 than 4 percent a year).
Source: Banerjee, Skilling, and others 2008. Access to improved water sources is highly
a
Average is population weighted.
inequitable across the income distribution
(figure 16.2). Piped water and standposts are
from trucks and carts, or sometimes through heavily concentrated among the more affluent
their own private distribution networks. segments of the population, typically in urban
Water vendors account for only 3 percent of areas. The poorest 40 percent of the population,
the African urban market, rising to 7 percent by contrast, depends on surface water and on
for West Africa. In some countries, how- wells and boreholes in almost equal measure.
ever, their contribution to the urban water Piped supply covers only 10 percent of African
supply is much larger: Nigeria (10 percent), households in the bottom 60 percent of the
Chad (16 percent), Niger (21 percent), and population. For the middle-income countries,
Mauritania (32 percent). In 15 large cities in access to piped water and standposts among
Africa, the cost of vendor water, particularly the poorest quintiles is substantially higher
when transported directly to the household, than in the low-income countries.
can be 2 to 11 times more expensive than hav-
ing a household connection (table 16.4). This
strong willingness to pay for vendor water is a Financing the MDG
potential revenue source that the utilities are
typically unable to capture. The overall price tag for reaching the MDG
The dynamics of service expansion reveal a target for access to water is estimated at $16.5
similar overall pattern in both urban and rural billion (roughly 2.6 percent of Africa’s GDP),
areas: the absolute number of people depend- which is somewhat higher than previously
ing on surface water continues to grow, a grim thought (Mehta, Fugelsnes, and Virjee 2005).
statistic in its own right (figure 16.1). Across Capital investment needs based on minimum
the board, wells and boreholes are expanding acceptable asset standards and accounting for
Water Supply: Hitting the Target? 303
Table 16.4 Average Price for Water Service in 15 Largest Cities, by Type of Provider
House Small piped Stand Household Water Water
connection network post reseller tanker vendor
Average price ($/cubic meter) 0.49 1.04 1.93 1.63 4.67 4.00
Markup over house connection (%) 100 214 336 402 1,103 811
Source: Keener, Luengo, and Banerjee 2008.
a. Urban b. Rural
2.0 2.0
access (% of population)
annualized increase in
access (% of population)
annualized increase in
1.5 1.5
1.0 1.0
0.5 0.5
0 0
piped supply standposts wells and surface water
boreholes –0.5
piped supply standposts wells and surface water
boreholes
Source: Banerjee, Skilling, and others 2008.
Figure 16.2 Coverage of Water Services, by Table 16.5 Estimated Annual Financing Needed to Meet the Water MDG
Budget Quintile
$ billions annually % of GDP
60 Country type O&M Capital Total O&M Capital Total
50 Sub-Saharan
% of households
30 Low-income
fragile 0.98 2.41 3.39 2.55 6.27 8.81
20
Low-income
10 nonfragile 1.91 4.36 6.27 1.73 3.95 5.68
0
poorest second third fourth richest Middle income 1.19 1.19 2.38 0.44 0.44 0.88
quintile Resource rich 1.47 3.12 4.59 0.66 1.40 2.06
piped water public standpost Source: Briceño-Garmendia, Smits, and Foster 2008.
surface water well water Note: O&M = operation and maintenance. Row totals may not add exactly because of rounding errors.
services is presented together here. The more addressed. Improving cost recovery of water
detailed discussion of sanitation can be found utilities could reduce the gap by $1.8 billion a
in chapter 17 of this volume.) African countries year, addressing operating inefficiencies would
are already devoting significant resources to bring an additional $0.9 billion a year, and
meeting the water and sanitation MDG targets. raising capital budget execution could recoup
Domestically funded spending accounted for some $0.2 billion a year. A larger part of these
over half of total spending. For capital invest- gains is for the low-income countries, whose
ment, donors have had a predominant role, par- utilities should focus equally on improving
ticularly in low-income countries, where they cost recovery and on reducing operating inef-
finance most of the investment (table 16.6). ficiencies. For middle-income countries, much
Financiers outside the Organisation for Eco- of the gain would come from reducing operat-
nomic Co-operation and Development have ing inefficiencies.
also played a role in the low-income coun- Even if all of these efficiencies could be
tries, whereas private finance has been neg- addressed, the water sector would still face
ligible. However, household self-finance of a sizable financing gap of $9.3 billion a year
on-site sanitation facilities—such as latrines— (1.5 percent of GDP; table 16.7). The gap for
is estimated to be substantial. capital requirements is more than twice that
Public investment in water and sanitation for operation and maintenance, suggesting that
is almost exclusively the domain of the central the MDG challenge is mainly about expanding
government. Within the public sector, general access to improved water sources and rehabili-
governments3 carry out most capital spend- tating existing assets in poor condition.
ing, whereas public enterprises tend to execute Although the costs of investment appear
most spending for operation and mainte- high, the health dividends of meeting the MDG
nance. Therefore, the institutional effective- target are substantial. Every dollar invested in
ness of the line ministries is just as important water supply generates economic benefits of at
as the institutional effectiveness of the utilities least $1.50 (Hutton and Haller 2004). Access
in ensuring that resources are well used. Thus, to improved water brings a variety of benefits,
better public expenditure management, good particularly improved health and reduced time
selection of projects, and clear strategic guid- spent collecting water. Serious illnesses trans-
ance for investment should all be considered mitted through unsafe water, such as infectious
integral aspects of the sector reform agenda. diarrhea, are a leading cause of infant mortal-
The existing resource envelope for Africa’s ity. Waterborne illnesses can be a substantial
water supply sector would go considerably economic burden, causing adult deaths and
further if various serious inefficiencies— workdays lost and affecting children’s health
amounting to $2.7 billion a year—could be and education.
Household members, primarily women and the majority of those gaining access to
children, face a substantial opportunity cost in improved water in utility service areas each
time to fetch water. More than 20 percent of year. A survey of 51 water utilities shows that
the population in Cameroon, Ghana, Maurita- about 2 percent of the population in utility
nia, Niger, and Tanzania must travel more than service areas gains access to a formal utility
2 kilometers to the primary water supply. Rural water service each year—1.5 percent to piped
dwellers tend to travel farther than urban dwell- water and 0.5 percent to standposts. Given the
ers. Therefore, the time savings from accessing higher unit cost of piped water, utilities are not
water from a nearby source are enormous even maximizing the effect on coverage from their
when valued at a discounted wage rate. limited investment budgets. Indeed, utilities
could double the value for money of their
investment programs (dollars per capita gain-
Using Appropriate Technologies ing access to improved water) if they weighted
their investment programs toward standposts
For many countries, the costs of meeting the rather than private connections.
water MDG look prohibitive. Although the cost Standposts provide safe water in urban
of meeting the MDG amounts to 2.3 percent areas at about one-third the per capita cost of
of GDP on average, for a handful of countries a private tap. Therefore, it is striking that the
(including Benin, the Democratic Republic coverage of public standposts in urban Africa
of Congo, Kenya, and Madagascar), the cost lags so far behind that of private taps and that
would be in excess of 7 percent of GDP, well their distributional incidence is skewed toward
beyond what could be feasibly attained. So the better-off households.
these countries must choose between achieving Several institutional challenges prevent
the MDG at a lower rung of the water supply standposts from being more widely adopted
ladder with lower-cost technologies and sub- (Keener, Luengo, and Banerjee 2008). Stand-
stantially postponing their achievement of the posts describe a range of public water supply
goal. Concentrating all coverage expansion in arrangements, from an unattended public
lower-cost technologies, such as standposts and access tap to a kiosk with a human operator.
boreholes, would lower the cost of meeting the Despite their low investment cost, standposts
MDG to 1.6 percent of GDP on average. For are in practice often beset by poor main-
Benin, Kenya, and Madagascar, the cost could tenance and high retail prices. The service
be reduced to less than 4 percent of GDP, which provided by the standpost may be very low
would be more affordable. The Democratic because the utility’s claims about the number
Republic of Congo would be the only country served by a single standpost are highly vari-
still experiencing an overall cost above 7 per- able and often unrealistic, ranging from just a
cent of GDP. few customers to 5,000, with an average of
This strategy, however, runs contrary to cur- about 700. Official utility data indicate that
rent practice, which sees piped water serving about three-quarters of standposts are in good
306 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
working order, but more detailed surveys sug- 20 times in some extreme cases (for example,
gest that fewer than half may be functioning in Kinshasa in the Democratic Republic of
normally at any one time. Congo). Community engagement can increase
The utility manages about half the stand- the accountability of kiosk operators, but it
posts, but a growing number are being can also lead to corruption and mismanage-
delegated to private or community manage- ment in countries that lack social cohesion or
ment. Although this arrangement helps ensure oversight by supporting institutions.
sustainability, price markups to the final Even so, a handful of countries are making
consumer can be large, and local elites can headway with the expansion of standposts.
capture the service. Many utilities originally Each year in Rwanda, an additional 3 percent
offered standpost services free of charge and of the population resident in the service area of
later moved toward charging preferential rates the national utility Electrogaz gains access to
for standpost water, typically half the price standposts (box 16.2). Similarly, Lusaka Water
charged for a private connection. When del- and Sewerage Company in Zambia adds 3 per-
egated management is introduced, however, cent of its resident population to this category
markups are added to cover the salaries of the every year. Another intermediate approach
standpost operators, often with significant sometimes adopted is to install yard taps,
profit markups. As a result, standpost prices shared by four or five households. They are
rise to 3 times the utility rate, and as much as lower in cost than private taps, and they avoid
BOX 16.2
Standposts in Kigali, Rwanda post costs $0.55 per cubic meter to supply
when distribution losses are fully accounted
The water production capacity of Electrogaz, for but brings in only $0.42 in revenue.
the main utility in Kigali, is inadequate to meet Kigali has roughly 240 standposts, of
network demand. The lack of bulk supply which an estimated 193 (80 percent) were
causes rolling shortages throughout the city, operating in December 2008. Utility offi-
often forcing residents with private connec- cials estimate that about 60,000 people use
tions to seek water from public sources, such standposts, though this figure includes con-
as standposts. sumers who use them only when their pri-
The financial sustainability of standposts in mary source is unavailable. According to the
Kigali can be estimated from the tariff paid by total water volume recorded by public stand-
standpost operators ($0.42 per cubic meter), post meters, the standposts could supply only
the total cost of Electrogaz’s production ($0.36 48,500 people with 20 liters per capita a day.
per cubic meter), the rate of unaccounted-for This figure would be more indicative of the
water in distribution and selling (35 percent and upper bound of the population that primarily
5 percent, respectively), and the volume and depends on standposts (about 6 percent of
price of water sold at the standposts. Three the city’s population).
hypothetical operators selling 100 jerricans The utility’s limited production capacity has
each per day at $0.02, $0.03, and $0.05 per affected both the level of peak demand at
jerrican would earn estimated monthly net standposts and the cost of production. Obser-
incomes of about $314, $949, and $1,584, vations and interviews with consumers indicate
respectively (the 2007 GDP per capita was that prices have often been higher in areas
$341), making this an attractive business prop- when and where water service has been cut
osition for them. However, from the utility’s and have been lower after periods of precipita-
perspective, the combination of a low tariff tion that increase the availability of other supply
and the 35 percent rate of unaccounted-for options, such as rainwater and natural springs.
water in distribution creates a loss-producing
scenario: each cubic meter supplied to a stand- Source: Keener and others forthcoming.
Water Supply: Hitting the Target? 307
some of the issues associated with standposts, a set of complex financial arrangements with
which serve hundreds of customers. the central government that prevents optimal
resource allocation, financial sustainability, and
economically efficient use of water resources
The Challenge of Cost Recovery (Briceño-Garmendia, Smits, and Foster 2008).
As a result, the utility management postpones
By underpricing water, the sector forgoes at basic investment and rehabilitation decisions
least $1.8 billion a year in revenues (0.3 percent to “make it through the day” financially. Thus,
of GDP). Underpricing of water—a common although these policies appear to be socially
problem throughout the world—is also wide- benign, by debilitating the financial position
spread across Africa. In the worst cases, such of the utility, they ultimately lead to delayed
as Côte d’Ivoire, Madagascar, and Senegal, investment and hence hold back service expan-
underpricing can result in utilities’ capturing sion to reach the unserved population.
less than 40 percent of the revenues they need Average water tariffs in Sub-Saharan Africa
and can lead to an economic burden of around are about $0.67 per cubic meter, two-thirds of
0.7–0.9 percent of GDP (figure 16.3). The asso- a cost-recovery threshold of just over $1.00 per
ciated utility deficit is often concealed within cubic meter. Tariffs are already relatively high
by international (even developing-country)
standards (box 16.3). In 2004, average water
Figure 16.3 Economic Burden of Water Underpricing, tariffs were about $1.00 per cubic meter in
by Country members of the Organisation for Economic
Co-operation and Development, around
Senegal $0.35 in middle-income countries, and as low
Côte d’Ivoire as $0.10 in South Asia (GWI 2005). Tariffs of
Madagascar about $0.40 per cubic meter are considered
Malawi more than adequate to cover operating costs
Ghana in most developing-country contexts. In low-
Lesotho income African countries, however, operating
Niger costs are as high as $0.60 per cubic meter on
South Africa average, reflecting inadequate selection of tech-
Congo, Dem. Rep. of
nologies, low population density, country risk
Zambia
premiums, and the high cost of inputs. Current
tariffs cover operating costs on average but do
Mozambique
not contribute much toward capital costs.
Namibia
In most cases, the state or donors have
Ethiopia
subsidized capital costs almost entirely. Subsi-
Kenya
dies to residential consumers are highly regres-
Benin
sive (Banerjee, Wodon, and others 2008). Across
Rwanda the bottom half of the income distribution,
Sudan barely 10 percent of households have access to
Nigeria piped water. Indeed, more than 80 percent of
Burkina Faso households with piped water come from the
Tanzania top two quintiles of the income distribution.
Uganda Because poorer households are almost entirely
Cape Verde excluded, they cannot benefit from subsidies
0 0.2 0.4 0.6 0.8 1.0
embedded in prices for piped water. In many
% of GDP cases, targeting performance is further exacer-
bated by poor tariff design, with widespread
Source: Briceño-Garmendia, Smits, and Foster 2008.
Note: The economic burden of underpricing is defined as the use of minimum charges and rising block tar-
difference between the average effective tariff and the full cost- iffs that provide large lifeline blocks of highly
recovery tariff multiplied by the total volume billed. It is the total
revenue shortfall attributable to underpricing. subsidized water to all consumers.
308 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
BOX 16.3
% of utilities
14 13 70
12 60 24%
45%
10 50
8 7 40 20% 49%
6 30
4 3 20 25%
29%
2 1 1 1 10 16% 9%
0 0
4 10 40
+ um
ge
ed
f
rif
ea
ea
IB
ar
ap
ta
T
ion im
lin
lin
e+
ch
sh
ck
pt min
+
g
ed
blo
U
ar
ge
fix
ch
ar
ing
as
fix
ed
re
ns
fix
inc
Full capital cost recovery is generally Figure 16.4 Affordability of Cost-Recovery Tariffs in Low-Income Countries
affordable in the middle-income countries 100
affordability threshold
tion. Assuming household subsistence con- 70
sumption of 10 cubic meters per month (or 60
about 65 liters per capita a day), a monthly 50
utility bill under full cost-recovery pricing of 40
$1 would be around $10. Based on an afford- 30
ability threshold of 5 percent of household 20
income, full cost-recovery tariffs would prove 10
affordable for 40 percent of the population 0
2 3 4 6 8 10 12 14 16
in low-income countries (figure 16.4). With
monthly utility bill (US$)
about 10 percent of the national population
cost of recovery bill, low bound
already enjoying a direct water connection, an
cost of recovery bill, medium bound
additional 30 percent of the population could
cost of recovery bill, upper bound
be added to service and be able to afford it. percentage of low-income country population unable to afford to pay
The majority of the remaining 60 percent of
the population would be able to afford bills of Source: Adapted from Banerjee, Wodon, and others 2008.
Note: Tariffs are described as affordable if a monthly bill based on subsistence consumption absorbs
around $6 a month that would result either no more than 5 percent of the household’s budget.
from operating cost recovery or from full cost
recovery at a more modest consumption of
about 6 cubic meters per month (or 40 liters Congo, Ghana, and Zambia, operating inef-
per capita a day). ficiencies can create an economic burden of
0.7–1.0 percent of GDP (figure 16.5).
Another source of inefficiency is overem-
Improving Utility Performance ployment. State-owned enterprises can be
through Institutional Reform social buffers to (very inefficiently) transfer
rents or resources to the population. African
The operational inefficiencies of water util- utilities have an average of five employees per
ities, including revenue undercollection, dis- 1,000 connections, more than twice the two
tribution losses, and labor inefficiencies, cost employees per 1,000 connections frequently
the region $0.9 billion a year (or as much as used as the international benchmark for devel-
0.15 percent of GDP). Operating inefficien- oping countries (Tynan and Kingdom 2002).
cies are rampant among Africa’s water utili- Operating inefficiencies have been imped-
ties, divided roughly evenly between revenue ing expansion. Inefficiencies not only drain the
undercollection and distribution losses. Aver- public purse but they also seriously undermine
age collection ratios are relatively high in Africa the performance of utilities.
at 90 percent but still short of the best practice One casualty of insufficient revenue is
of 100 percent. Average distribution losses in maintenance. The rate of bursts per kilome-
Africa are 35 percent, far above the common ter of water main provides some indication of
norm of 20 percent, with all countries affected the condition of the underlying infrastructure,
to some degree. The overall magnitude of these and hence the extent to which it is adequately
inefficiencies can be quantified by compar- operated and maintained. Among African util-
ing the revenues available to the utility with ities, a huge variation occurs between low- and
the revenues available to an ideal utility that middle-income countries, with bursts ranging
is able to charge cost-recovery tariffs, collect from five per kilometer in the latter to just
all its revenues, and keep distribution losses over one per kilometer in the former. Utility
to the technical minimum (Ebinger 2006; managers must often choose among paying
Saavalainen and ten Berge 2006). In the worst salaries, buying fuel, or purchasing spares.
cases, such as the Democratic Republic of Frequently, they must cannibalize parts from
310 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Figure 16.5 Economic Burden of Water Utility Utilities with fewer employees per connection
Operational Inefficiencies, by Country manage to have, on average, 85 percent of the
water supplied with adequate chlorine, in con-
Congo, Dem. Rep. of
trast to 75 percent of the remaining utilities.
Ghana
Conversely, utilities with higher hidden costs
Zambia tend to deliver slightly higher-quality water.
Malawi Institutional reforms hold the key to improv-
Cape Verde ing utility performance. Good institutional
Uganda frameworks pay off in lowering the inefficiency
Sudan
of utilities. Utilities that have decentralized
or adopted private sector management have
Mozambique
hidden costs that are substantially lower than
Burkina Faso
those that have not (figure 16.7). In addition,
Rwanda unbundling has a considerable effect on utility
Madagascar efficiency; however, unbundling is rare in Africa
Kenya and is concentrated exclusively in middle-
South Africa income countries, whose superior performance
Nigeria
can be explained by many other reasons. Con-
versely, higher levels of regulation and gover-
Lesotho
nance, as well as corporatization, are associated
Ethiopia
with higher efficiency.
Namibia Nevertheless, introducing reforms is easier
Benin said than done. In recent years, many African
Tanzania countries have initiated water sector reforms
Niger to improve performance (figure 16.8). This
Côte d’Ivoire
reform agenda has had two major thrusts:
encouraging private participation and improv-
Senegal
ing governance from within.
0 0.2 0.4 0.6 0.8 1.0
The first thrust has focused on experiment-
% of GDP
ing with private participation. Overall, 26
unaccounted-for losses
collection inefficiencies
private sector transactions have occurred in
African water, affecting the majority of coun-
Source: Briceño-Garmendia, Smits, and Foster 2008. tries in one way or another. Most have been
Note: Utility inefficiency is defined as the revenue lost from under-
collection of water billed, as well as the value of technical and
lease contracts (or affermage). Experience
nontechnical losses of water on the distribution network valued at with private sector participation is dispro-
the full cost-recovery tariff.
portionately concentrated in the francophone
countries of West Africa (Côte d’Ivoire, Guinea,
other working equipment. The investment Niger, and Senegal), with some exceptions
program is another major casualty. (Mozambique and Uganda). Another distinc-
Service expansion (measured as a percentage tive feature of the African experience has been
of the population resident in the utility service the use of concessions for joint power and
area that every year gains access to either piped water utilities, as in Gabon and Mali.
water or standposts) is significantly higher for The rate of cancellation of private sector con-
more efficient utilities. In particular, utilities tracts for water supply in Africa has been much
with low hidden costs achieve an average annual higher than elsewhere. Some 29 percent of pri-
increase in coverage of more than 3 percent of vate contracts for water supply in Africa have
the resident population, essentially twice as been terminated prematurely, a much higher
much as that registered by utilities with high rate than in any other developing region. As a
hidden costs (figure 16.6). result, the number of active private operators
For similar reasons, more efficient utili- has shrunk to just a handful, with four in South
ties deliver better-quality water (figure 16.6). Africa and one each in Cameroon, Cape Verde,
Water Supply: Hitting the Target? 311
Figure 16.6 Effect of Utility Inefficiency on Access Expansion and Water Quality
85
% of sample passing
(% of population)
3.0
chlorine test
80
2.0
75
1.0
70
0 65
employees per connection hidden costs employees per connection hidden costs
Figure 16.7 Hidden Costs and Institutional Figure 16.8 Overview of Reforms Affecting Urban
Frameworks Utilities
hidden costs (% of revenue)
200 100
80
% of countries
150
60
100
40
50 20
0 0
ion
cts
ion
g
nc
nc
nc
lin
ion
ce
ion
ion
ra
en
at
at
ge
lin
rie
na
nd
an
nt
at
tiz
liz
lat
at
em
ya
nd
er
pe
bu
rn
co
liz
tiz
ra
ra
gu
ov
bu
ag
ex
or
un
ve
ra
nt
po
ra
ce
Eg
re
lat
an
un
go
nt
ce
po
an
or
PS
h
ce
gu
rm
SO
de
Ec
or
hig
h
rm
de
hig
re
Ec
cto
h
SO
rfo
hig
SO
se
pe
te
iva
yes no
pr
BOX 16.4
their market share of only 9 percent. Half of to improve internal processes and corporate
those gains were made in Côte d’Ivoire alone governance mechanisms, more so than in other
(Marin 2009). infrastructure sectors. In particular, a growing
Anglophone and francophone countries number of utilities are using performance con-
have taken two distinct approaches to sector tracts, for instance, in Lesotho, Uganda, and
regulation. About half of the countries (mainly Zambia, although not all of them incorporate
anglophone) have established distinct regula- the penalties, performance-based remunera-
tory agencies for the water sector, although a tion, and third-party monitoring that make
significant number of them have not adopted these mechanisms truly effective. Uganda’s use
private sector participation. Conversely, a num- of performance contracts to underpin sub-
ber of francophone countries with private par- stantial improvements in sector performance
ticipation have adopted regulatory frameworks is particularly noteworthy (box 16.5).
contractually, without establishing an indepen-
dent regulatory agency. No evidence appears to
support the superiority of either of these two Reforms in the Rural Space
approaches to regulation. Neither does any evi-
dence indicate that the creation of a regulatory Africa remains a predominantly rural conti-
entity has led to discernible improvements in nent. About 400 million rural inhabitants have
utility performance. Even where explicit regu- no form of utility-provided water. Rural cover-
latory frameworks have been established, they age of piped-water supply and standposts has
typically meet only about half of the corre- barely risen in the past 15 years, and most of the
sponding good-practice criteria. gains have come from rural inhabitants mov-
The second thrust of the reform agenda has ing up the water supply ladder from surface
focused on improving the governance of the water to wells and boreholes. Until improved
sector from within. This approach was based water sources serve more of the rural popula-
on the recognition that the standard infra- tion, MDG goals will continue to be elusive.
structure reform prescriptions were not always In rural areas, the central challenge is to
as relevant or as easily applied to the water sec- reduce reliance on surface water through a sus-
tor as to other areas of infrastructure, and that tainable network of water access points, which
service provision would remain dominated by are most typically boreholes. About half of the
public enterprises in the near future. The focus sampled countries are reducing the share of
of these reforms has been on moving toward the rural population reliant on surface water
corporatization of state-owned enterprises and and in the best cases (Lesotho, Mozambique,
on decentralizing responsibilities to lower levels and Uganda) are managing to shift 2–3 per-
of government. In addition, some measures cent of the rural population away from this
have been taken to improve the governance option each year. In the other half of the coun-
of state-owned enterprises, aimed at adopting tries sampled, however, the share of the rural
commercial principles and modern manage- population reliant on surface water is actually
ment methods. About 80 percent of Africa’s increasing. In Chad, Kenya, and Rwanda, an
larger water utilities have now been corpora- additional 1 percent of the rural population
tized, thereby laying the foundation for a more is forced to rely on surface water, whereas
commercial form of management. Close to in Burkina Faso that number rises to 3 percent,
half of the countries sampled have decentral- and it is nearly 10 percent in the Democratic
ized their water utilities since the mid-1990s to Republic of Congo.
bring responsibility closer to local communities; Even in countries that are successful in
however, all the francophone countries studied expanding rural access to improved water
retain centralized organization of the sector. sources, sustainability is still a concern. A recur-
Africa’s state-owned water utilities typically ring problem in rural water systems is the lack
fulfill only about half of the good-practice cri- of technical or financial capacity to maintain
teria for enterprise governance. Since the mid- assets. Decentralization and the breakdown of
1990s, some serious efforts have been made community management arrangements have
314 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
BOX 16.5
put greater strain on local governments to at least one in two (the Democratic Republic
manage services. This situation leads to rapid of Congo, Madagascar, Malawi, Nigeria, and
deterioration of rural water points to where Tanzania). Nevertheless, in the best-case sce-
they no longer provide the intended service, narios, the share needing rehabilitation can be
and populations are forced to return to relying as low as 10–20 percent (Benin and Uganda).
on surface water. On average, one in three rural Inadequate maintenance of rural water
water points needs rehabilitation, and for a sig- systems reflects both institutional weaknesses
nificant number of countries, the share rises to and inappropriate technology choice. Besides
Water Supply: Hitting the Target? 315
cy
y
ts
reform agenda for rural water comprises five
nd
lic
lic
oin
en
fu
po
po
ag
lp
er
er
ry
er
at
ve
ru
lw
at
lw
co
lw
of
ra
re
ru
ra
ap
ru
st
ru
co
surprising given their strong track record on focus of the agenda has shifted toward a more
institutional reforms. For moderate reform- pluralistic view of public and private sector
ers, the results can go either way (box 16.6). roles. The reform agenda also needs to move
The degree of reform also affects whether beyond utilities to encompass line ministries
rural water points are maintained adequately. and the entire public expenditure framework
The percentage of rural water points needing that underpins, and too often hinders, sector
rehabilitation tends to be lower for countries investment programs. Room for improvement
with more advanced rural reform processes.4 exists in cost recovery so that scarce subsidy
Thus, Benin and Uganda score high on sector resources are redirected to promote access
reform as well as on maintaining rural water among the poorest. For the majority that
points. The converse is true for the Democratic do not enjoy access to piped water, greater
Republic of Congo and Malawi. thought needs to be given to how standposts
can become more effective for urban water
supply and how to get the best out of small-
Policy Recommendations scale independent providers. The burgeoning
use of wells and boreholes for supply in urban
On the basis of this diagnostic, the water sec- areas demands urgent attention from policy
tor evidences a number of key areas for policy makers, both to improve their understanding
attention. The institutional reform agenda of this phenomenon and to develop suitable
remains as relevant as before, even if the regulatory tools.
BOX 16.6
Continue to Pursue Institutional even if the private sector can sometimes help
Reforms execute it. A key lesson from Africa’s experience
Institutional reforms are the key to improving with lease contracts is the difficulty of achiev-
water sector performance. Countries pursuing ing seamless coordination on investment plans
institutional reforms create more efficient and between the contractor and the public holding
effective sector institutions and promote more company. In addition, incorporating clear con-
rapid expansion of higher-quality services. The tractual incentives for efficiency improvements
potential dividend is large because address- (for example, by basing the contractor’s rev-
ing utility inefficiencies alone could make a enues on ideal rather than actual performance
substantial contribution to closing the sector parameters) is important (Marin 2009).
funding gap in many countries. The new agenda places greater emphasis on
Although the majority of African countries broader reforms within governing state-owned
have embarked on the sector reform agenda, enterprises. Given the limited scope of private
few have completed it. The glass is still half participation, state-owned utilities remain cen-
full, but the experience of those countries that ter stage. Without addressing the typical defi-
are furthest ahead provides some guidance for ciencies that afflict such enterprises—including
the region. numerous and conflicting objectives, political
In rural areas, a few critical interventions can interference, and lack of transparency—the
make a difference. Establishing a clear sector sector will have difficulty exiting low-level
policy, creating a strong central capability for equilibrium. Three key areas for attention are
sector financing and project implementation, (a) internal process improvements, (b) increased
moving to greater cost recovery, and develop- managerial autonomy, and (c) more stringent
ing a system to monitor the condition of rural performance monitoring. Incorporating mea-
water points are all measures that, when imple- sures to streamline corporate processes, such as
mented as a package, can boost performance. procurement, financial management, and per-
In urban areas, the story is more complex. formance management, is essential for strength-
The traditional reform agenda of the 1990s has ening the application of commercial principles
not fully proved its relevance to the sector. Per- and accountability mechanisms. Measures to
haps surprisingly, no clear evidence indicates broaden the board of directors, to increase the
that regulation has made a positive contribution use of external audit and independent audit
to sector performance across the board. Private of accounts, and to incorporate independent
sector participation, although controversial in members from beyond the public sector would
implementation, has in many cases proved a help depoliticize decision making and consoli-
useful tool for improving operational perfor- date the arm’s-length relationship. Adopting
mance and efficiency (Marin 2009). Expecta- performance-based monitoring arrangements
tions that the private sector would finance that mimic private sector contracts is also of
new infrastructure for water utilities have not interest, but only to the extent that they create
been met, with negligible private capital flows credible incentives by incorporating meaning-
that are dwarfed by public and donor finance. ful rewards and penalties at the personal and
However, the private sector has contributed to corporate levels and are subject to third-party
expanding access, though typically with public monitoring.
funding (Marin 2009).
The new reform agenda for water retains a Improve Efficacy of Public Expenditure
role for private participation. Lease contracts The bulk of investment in the water sector is by
may be the form of private participation best line ministries through the budgetary process,
suited to African water utilities. By transferring often with external support. The existing pat-
more responsibility to the private sector than terns of spending clearly show that although
in a management contract, they have provided utilities are instrumental in delivering services,
greater scope for operational improvements. In the general government (using either domes-
contrast to concessions, recognition is explicit tic or external capital) continues to make
that investment will need to be publicly funded, most of the investment decisions. Therefore,
318 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
a solid public investment appraisal system and appraising, prioritizing, planning, and pro-
strong public spending management are pre- curing investment projects has an equally
requisites for improving both urban and rural important role. Donors can support countries
water supply. in developing good project identification and
Major bottlenecks hold back the disburse- appraisal tools that systematically consider
ment of public investment funds. Capital bud- the technological alternatives for expanding
get execution ratios for public investment in access and that examine the importance of
water are relatively low, 75 percent on average. spending on maintenance and rehabilitation
In many instances, the capacity to disburse against new investment.
budgetary resources in a timely fashion is the
binding constraint, rather than their availability Experiment with Different Institutional
(Prevost 2009). In Tanzania, steep increases Models for the Unconnected
in budget allocations to the sector followed The modest role played by standposts in urban
its identification as a priority in the country’s water supply is striking. In most countries, gov-
poverty reduction strategy; however, disburse- ernment and utility attention continues to focus
ments increased at a much slower pace, so on expansion of piped-water connections. This
there was no immediately discernible effect on battle is being lost, however, because of a com-
access (Van den Berg 2009). bination of rapid urbanization and financially
Key aspects of the public expenditure debilitated utilities. Standposts are very limited
framework must be addressed. The budgeting in the African urban water scene, are expand-
process needs to move toward a medium-term ing relatively slowly, and remain concentrated
framework and make stronger links between among the more affluent segments of the
sector objectives (such as MDG targets) and population. Simple simulations suggest that if
resource allocations. Clear sector plans that utilities were to shift their existing investment
detail specific activities and their associated budgets from piped connections to standposts,
costs should underpin the budget process. The the rate of service expansion could potentially
careful incorporation of maintenance needs double. However, as long as urban households
into medium-term sector planning tools is crit- are inconvenienced by higher payments and
ical to prevent asset rehabilitation. Administra- longer water collection times, standposts are
tive processes that delay the release of budgeted not necessarily going to be a superior solution
funds also need to be overhauled. At the same even if they are a cheaper alternative to private
time, procedures for procurement, disburse- piped connections. In low-income countries,
ment, financial management, and accountabil- resale of water by neighbors through informal
ity should be modernized and streamlined. standpost arrangements is almost as prevalent
Donor resources are best channeled pro- as formal standposts.
grammatically as budgetary support or through The key to this paradox could lie in the
sectorwide projects. Given the sector’s strong problematic institutional arrangements associ-
dependence on external funds, a solid public ated with standposts in African cities. Utilities
expenditure management system for African charge little or nothing for standpost water,
countries also requires that donors improve the and standpost revenues constitute a negligible
predictability of their support and make prog- portion of the revenue base. Therefore, utili-
ress on streamlining and harmonizing admin- ties have little financial incentive to expand the
istrative procedures. In that sense, a focus on service. Standpost operators, where they exist,
multidonor initiatives that pool funds to pro- often charge large markups that make the
vide general budgetary support for a sector- service prohibitively expensive and that may
wide program of interventions is preferable. generate significant revenues never captured
Technical assistance to the sector has tra- by the utility. The quality of service provided
ditionally been understood as improving the by standposts can be very low because of the
management practices of utilities. However, high rates of malfunction and the very large
technical assistance to support line ministries (sometimes implausibly large) numbers of
in improving the framework for identifying, people expected to rely on each one.
Water Supply: Hitting the Target? 319
Solving this conundrum demands seri- compared with the formal one. Increasing
ous attention. The way forward is not yet water production capacity and improving
clear, but it calls for intensive experimenta- the efficacy of the distribution network can
tion with alternative network designs and significantly affect the welfare of the uncon-
institutional setups. Standposts cover a wide nected as well as the connected, because it
range of communal arrangements or del- drives down the premium on alternative
egated management models, some of which sources of water supply (Keener, Luengo, and
may be more promising than others. One Banerjee 2008).
option would be to increase the density of
standposts to increase competition, with an Increase Cost Recovery with Careful
immediate effect on convenience and price of Social Policies
water supply. Yard taps, which provide com- Underpricing water is contributing to the
munal access to a smaller group of four or five financial weakness of utilities, slowing access
contiguous households, lower costs but only expansion, and holding back the quality of
partially address the problem of maintenance service. Given that utility customers are drawn
and management. Whatever the approach, from the upper end of the income distribution,
an important component of the solution will the result is a highly regressive incidence of
be to ensure a fairer distribution of revenues subsidies to the sector. A large (and generally
between utilities and standpost operators or poor) segment of the urban population is pay-
other secondary water retailers. The experi- ing multiples of these prices to access utility
ences of the handful of low-income countries water indirectly, and in many cases more than
that have achieved more than 20 percent urban the utility cost-recovery price.
coverage of standposts—notably Côte d’Ivoire, Countries need to make progress toward
Rwanda, and Senegal—deserve some study. further cost recovery while considering the
The popularity of the household resale economic circumstances of their populations.
option could also be exploited by making it The key principle is to verify the affordability
an explicit part of the utility’s rollout strategy. of water tariffs with reference to household
Household resale of water through yard taps budgets, rather than simply to assume that
appears to be a widely used option in many they will be unaffordable. Although the pur-
African cities. Survey evidence highlights a chasing power of African households is quite
variety of reasons why residents may find this limited, the analysis confirms that operating
approach preferable to official standposts. cost recovery is a perfectly feasible objective
Neighbors can offer more convenient operating for just about all African countries. Tariffs that
hours and better water pressure; because they recover full capital costs also look to be afford-
are nearby, less time is required to collect water. able for the richest 40 percent of the popula-
In addition, they offer more flexible payment tion in low-income countries, where today’s
mechanisms than either public standposts piped water coverage is concentrated. Thus,
or a private connection (Keener, Luengo, and little economic justification exists for today’s
Banerjee 2008). Therefore, increasing recogni- subsidies. Countries would be better served
tion should be given to this water supply modal- by recovering full costs from their existing
ity, removing any legal barriers and potentially customer base and using the resulting cash
considering the creation of such household- flow to accelerate access expansion in poor
based water retail enterprises as an integral neighborhoods. In the longer term, however,
component of utilities’ expansion plans. as access to piped water increases, low-income
Ultimately, investing in utility production countries will need social tariffs that provide
and distribution of water is the best policy for water priced at operating cost recovery levels
keeping down the costs of alternatives. Within for a minimum level of consumption to the
cities, water markets are strongly connected in substantial share of their population that can-
the final price offered to the consumer. The not afford full capital cost-recovery tariffs.
more disrupted the formal piped system, the Government entities need to become bet-
higher the price will be in the informal sector ter customers. Government entities can easily
320 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
In addition to growing groundwater reli- Sarah Keener, Manuel Luengo, Dennis Mwanza,
ance, African cities are characterized by heavy Eustache Ouayoro, Heather Skilling, Caroline
use of low-grade in situ sanitation, mainly in the van den Berg, Quentin Wodon, Guillermo
form of unimproved latrines (see chapter 17 of Yepes, and Yvonne Ying, and received extensive
support from the Water and Sanitation Pro-
this volume). Deployment of latrine sanitation
gram network in Africa.
at excessive population densities or with lack of
1. The MDG for water supply commits countries
proper latrine operation can lead to increased by 2015 to halving the percentage of their pop-
groundwater contamination that can affect the ulations without access to an improved water
entire urban aquifer providing the groundwater source relative to the baseline situation in 1990.
supplies (Xu and Usher 2006). For the purposes of the MDG, “improved water”
Furthermore, extensive unregulated use of includes access to piped water or standposts, as
groundwater by private actors may prevent the well as some types of wells and boreholes that
most rational and efficient exploitation of the are adequately protected.
resource for public water supply. In particular, 2. The Joint Monitoring Programme of the World
it prevents cities from reaching economies of Health Organization and the United Nations
Children’s Fund systematically tracks access to
scale in groundwater exploitation and from
improved water supply and sanitation, but the
following the principle of conjunctive surface
data constraints are immense. Systematic infor-
water and groundwater use that allows ground- mation and data about suppliers’ characteristics
water to play its natural role as a backup supply and institutional environments are poor when
in times of drought (Foster 2008). they exist. Often, even the well-performing ser-
Developing an improved understand- vice providers are unrecognized outside their
ing of the benefits and risks of groundwater immediate environments, and lessons learned
use in fast-growing African cities and towns are not widely shared. Under the auspices of
and of how those benefits and risks vary with the Africa Infrastructure Country Diagnostic,
the hydrogeological setting is urgent (Foster, a limited effort has been made to use a spe-
Tuinhof, and Garduño 2008). This objective cially designed questionnaire on institutional
environment, governance structure, and tech-
should begin with a city-level appraisal of (a)
nical and financial performance to collect data
the quantity and quality of available urban
covering 51 utilities in 23 countries. This novel
groundwater resources; (b) the drivers, dynam- database covering the period from 1995 to 2005
ics, and patterns of usage; and (c) an assess- is paired with household survey data of various
ment of the vulnerability of urban aquifers years from 1990 to 2005.
to pollution from the land surface. Creating a 3. General government includes central and local
groundwater-monitoring framework and pro- governments and special funds when off the
mulgating appropriate construction and oper- budget.
ation protocols for wells and in situ sanitation 4. Rural water indexes are negatively—if admit-
facilities (mainly latrines) would help safeguard tedly weakly—correlated.
groundwater quality, but guidelines for safe use
of groundwater sources should accompany this References
framework. Appropriate governance arrange- Banerjee, Sudeshna G., Vivien Foster, Yvonne Ying,
Heather Skilling, and Quentin Wodon. 2008.
ments also need to be established, recognizing “Cost Recovery, Equity, and Efficiency in Water
the broad reach of groundwater resources, and Tariffs: Evidence from African Utilities.” Working
should involve water utilities, public health Paper 7, Africa Infrastructure Country Diagnos-
authorities, and municipal agencies, including tic, World Bank, Washington, DC.
a suitable channel for public consultation. Banerjee, Sudeshna G., Heather Skilling, Vivien
Foster, Cecilia Briceño-Garmendia, Elvira
Morella, and Tarik Chfadi. 2008. “Ebbing Water,
Notes Surging Deficits: Urban Water Supply in
The authors of this chapter are Sudeshna Ghosh Sub-Saharan Africa.” Background Paper 12,
Banerjee, Elvira Morella, Cecilia Briceño- Africa Infrastructure Country Diagnostic,
Garmendia, and Vivien Foster, who drew on World Bank, Washington, DC.
background material and contributions from Banerjee, Sudeshna G., Quentin Wodon, Amadou
Tarik Chfadi, Piers Cross, Alexander Danilenko, Diallo, Taras Pushak, Helal Uddin, Clarence
322 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Tsimpo, and Vivien Foster. 2008. “Access, Afford- Service Providers and Public Stand Posts in
ability, and Alternatives: Modern Infrastructure Africa.” World Bank, Washington, DC.
Services in Africa.” Background Paper 2, Africa Keener, Sarah, Manuel Luengo, and Sudeshna
Infrastructure Country Diagnostic, World Bank, Banerjee. 2008. “Provision of Water to the Poor
Washington, DC. in Africa: Informal Water Markets and Experi-
Briceño-Garmendia, Cecilia, Karlis Smits, and ence with Water Standposts.” Working Paper 13,
Vivien Foster. 2008. “Financing Public Infra- Africa Infrastructure Country Diagnostic, World
structure in Sub-Saharan Africa: Patterns, Issues, Bank, Washington, DC.
and Options.” Background Paper 15, Africa Marin, Philippe. 2009. Public-Private Partnerships
Infrastructure Country Diagnostic, World Bank, for Urban Water Utilities: A Review of Experi-
Washington, DC. ences in Developing Countries. Washington, DC:
Cardone, Rachel, and Catarina Fonseca. 2003. Public-Private Infrastructure Advisory Facility
“Financing and Cost Recovery.” Thematic Over- and World Bank.
view Paper 7, IRC International Water and Sani- Mehta, Meera, Thomas Fugelsnes, and Kameel
tation Centre, Delft, the Netherlands. Virjee. 2005. “Financing the Millennium Devel-
Collignon, Bernard, and Marc Vézina. 2000. Inde- opment Goals for Water Supply and Sanitation:
pendent Water and Sanitation Providers in Afri- What Will It Take?” International Journal of
can Cities: Full Report of a Ten-Country Study. Water Resources Development 21 (2): 239–52.
Washington, DC: UNDP–World Bank Water Muhairwe, William T. 2009. “Fostering Improved
and Sanitation Program. Performance through Internal Contractualisa-
tion.” Paper presented at World Bank Water
Ebinger, Jane. 2006. “Measuring Financial
Week, Washington, DC, February 17–20.
Performance in Infrastructure: An Application
to Europe and Central Asia.” Policy Research NWSC (National Water and Sewerage Corpo-
Working Paper 3992, World Bank, ration). 2006. “Corporate Plan 2006–2009.”
Washington, DC. NWSC, Kampala, Uganda.
Foster, Stephen S. D. 2008. “Urban Water Supply Oyo, Anthony. 2006. “Spare Parts Supplies for
Security in Sub-Saharan Africa: Making Best Handpumps in Africa: Success Factors for Sus-
Use of Groundwater.” Paper presented at the tainability.” Field Note 15, Water and Sanitation
Africa Groundwater and Climate Conference, Program, Africa Region, Nairobi.
Kampala, Uganda, June 24–28. Prevost, Christophe. 2009. “Benin Rural Water Pub-
lic Expenditure Review: Findings, Impacts and
Foster, Stephen S. D., Albert Tuinhof, and Hector
Lesson Learned.” Paper presented at World Bank
Garduño. 2008. “Groundwater in Sub-Saharan
Water Week, Washington, DC, February 17–20.
Africa: A Strategic Overview of Development
Issues.” In Applied Groundwater Studies in Saavalainen, Tapio, and Joy ten Berge. 2006. “Quasi-
Africa: IAH Selected Papers in Hydrogeology, Fiscal Deficit and Energy Conditionality in
vol. 13, ed. Segun Adelana and Alan MacDonald, Selected CIS Countries.” Working Paper 06/43,
9–21. London: Taylor & Francis. International Monetary Fund, Washington, DC.
GWI (Global Water Intelligence). 2005. “The GWI Tynan, Nicola, and Bill Kingdom. 2002. “A Water
2005 Water Tariff Survey.” Global Water Intel- Scorecard: Setting Performance Targets for
ligence 6 (9). Water Utilities.” Viewpoint Note 242, World
Bank, Washington, DC.
Harvey, Peter A., and Robert A. Reed. 2007. Van den Berg, Caroline. 2009. “Public Expen-
“Community-Managed Water Supplies in diture Review in the Water Sector: The Case
Africa: Sustainable or Dispensable?” Commu- of Tanzania.” Paper presented at Water Week,
nity Development Journal 42 (3): 365–78. World Bank, Washington, DC, February 17–20.
Hutton, Guy, and Laurence Haller. 2004. Evaluation WHO/UNICEF (World Health Organization/
of the Costs and Benefits of Water and Sanitation United Nations Children’s Fund). 2006.
Improvements at the Global Level: Water, Sanita- MDG Assessment Report. Geneva and
tion and Health Protection of the Human Environ- New York: WHO/UNICEF Joint Monitoring
ment. Geneva: World Health Organization. Programme for Water Supply and Sanitation.
Iliyas, Mohammed, Dozie Eneh, and Igiri Oside. Wodon, Quentin. 2007. “Water Tariffs, Alternative
2009. “Public Expenditure Review in the Service Providers and the Poor: Case Studies
Rural Water and Sanitation Sector for Cross River from Africa.” World Bank, Washington, DC.
State—Nigeria.” World Bank, Washington, DC. Xu, Yongxin, and Brent Usher, eds. 2006. Ground-
Keener, Sarah, Sudeshna G. Banerjee, Nils Junge, water Pollution in Africa. London: Taylor &
and Geoff Revell. Forthcoming. “Informal Water Francis/Balkerma.
Chapter 17
T
arget 7 of the Millennium Development but coverage of unimproved latrines is grow-
Goals (MDGs) for sanitation access calls ing much faster than coverage of any of the
for halving by 2015 the percentage of the improved alternatives.
population in 1990 without improved sanita- Although a bleak picture overall, some
tion. At the present pace, Africa will unlikely countries have expanded or upgraded sanita-
meet the target either at the regional or (with tion, each year moving as much as 3 percent
few exceptions) at the country level. The Joint of their population up the “ladder” to better
Monitoring Programme (JMP) of the World forms of sanitation. Ethiopia has done so with
Health Organization and the United Nations unimproved latrines; Burkina Faso, Madagas-
Children’s Fund is charged with assessing the car, and Rwanda with improved latrines; and
state of sanitation and progress toward the Senegal with septic tanks.
MDG target. The JMP’s latest data show only To meet the MDG target for sanitation, coun-
modest improvement, from 26 percent of the tries need to spend an estimated 0.9 percent of
population with access in 1990 to 31 percent GDP a year, of which 0.7 percent is for invest-
in 2006 (United Nations 2008; box 17.1), and ment and 0.2 percent for operation and main-
many countries face difficulty in making prog- tenance. A few countries already invest in new
ress (Water and Sanitation Program 2006). sanitation facilities up to the recommended
Today, about 30 percent of the population level, but many do not. Households pay for
practices open defecation (40 percent in rural most of the investment bill, but whether they
areas) and about half of the population, urban are spending enough on operation and main-
and rural, rich and poor alike, relies on unim- tenance is doubtful. Based on limited evidence,
proved latrines, a heterogeneous collection of governments contribute only a small fraction
facilities with poorly understood health effects. of investments. The health benefits attaching
Flush toilets, mostly connected to septic tanks to these investments are considerable, includ-
rather than sewers, remain a luxury, as do ing significant reductions in the incidence of
improved latrines, which have made headway diarrhea, intestinal worms, and trachoma.
in only a handful of countries. The prevalence The challenges and policy options differ
of open defecation has finally started to fall, substantially across and within countries.
323
324 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
BOX 17.1
supply problem and the need for training and
local market development. Both responses are
relevant, but by starting from the supply side,
What Is Improved policy makers can minimize the need for a sub-
Sanitation? sidy and promote cost-reducing innovations.
Where septic tanks predominate, the chal-
The improved sanitation category in
lenge is providing access to improved sanita-
the data from the Joint Monitoring Pro-
tion to a larger, lower-income population. As
gramme of the World Health Organiza-
tion and the United Nations Children’s
population densities increase and water con-
Fund includes both flush toilets and sumption rises, Africa’s burgeoning cities will
improved latrines. It also includes half eventually need to develop more extensive
of the traditional latrines, an adjustment sewerage networks. Thus, reducing the cost
that is made because this modality can- of networks through technological innova-
not be disaggregated exactly between tion is critical.
improved and unimproved sanitation.
The household analysis presented
here, based on Demographic and Health
Surveys and Multiple Indicator Cluster The State of Sanitation in Africa
Surveys, does not adopt this kind of
adjustment. Instead, the analysis disag-
gregates the different types of improved
Rungs on a Ladder
and unimproved sanitation to allow a
One can think of the different types of sanita-
richer discussion of sanitation options tion as rungs on a ladder, with each rung hav-
and issues. Notwithstanding these ing a higher investment cost and greater health
methodological differences, findings are benefits than the one below (figure 17.1). The
broadly consistent with those reported bottom of the ladder is open defecation, a
by the JMP. practice harmful to public health. The first
rung is unimproved latrines, which comprise
Source: Banerjee and others 2008.
various kinds of pits that vary greatly in their
efficacy but provide at best only basic sani-
tary protection.1 The next rung is improved
latrines, including a variety of engineered facil-
ities such as SanPlat and Ventilated Improved
Individual countries or urban and rural Pit (VIP) latrines, and basic pits with slabs.2
regions thereof may face three patterns of When appropriately used, these provide ade-
access to sanitation; each requires appropriate quate sanitary protection at reasonable cost.
policy responses. The final rung of the ladder is the flush toilet,
Where open defecation remains prevalent, which may be connected to either a septic tank
people must be encouraged to use latrines, if or (where it exists) the sewerage network. From
available. Key factors in changing behavior a health perspective, the most critical move-
are a community’s commitment to cultural ment is from no service (open defecation) or
change and peer pressure. Relatively modest unimproved service (unimproved latrine) to
government expenditure to promote hygiene an improved or sanitary service. Once the basic
education can raise awareness. level of sanitary protection is reached, returns
Where unimproved latrines are prevalent, in health benefits diminish with each higher
they should be upgraded to improved models. rung on the sanitation ladder.
However, upgrading faces constraints on both Unimproved latrines are the most preva-
the demand and supply sides. The significant lent sanitation option in Africa, but under-
cost of improved facilities suggests a demand standing the health benefits they can deliver
problem and the need for capital subsidy. Lack is difficult. Classifying unimproved latrines is
of domestic construction capacity suggests a complicated by the variety of installations under
Sanitation: Moving Up the Ladder 325
Figure 17.1 The Sanitation Ladder Figure 17.2 Percentage of the Population Sharing
Toilet Facilities
open cost fixed-point defecation
defecation 50
flush toilet
45
improved 40
not latrine
unimproved
% of countries
35
acceptable latrine 30
25
20
health benefits 15
10
Source: Morella, Foster, and Banerjee 2008. 5
0
0–20 20–40 40–60 60–80
% of population sharing facilities
that despite their lower cost, improved latrines groups highlights a crucial issue: the most vul-
remain something of a luxury, with little suc- nerable populations are failing to benefit from
cess in penetrating the middle of the income efforts to improve sanitation.
distribution. More important, the minimal Not only are unimproved latrines the most
presence of improved sanitation across poorer prevalent facilities in Africa, but also their use
is the fastest growing. In recent years, they
have been used by an additional 2.8 percent
Table 17.1 Patterns of Access to Sanitation in Africa
percentage of population of the population each year in urban areas
and an additional 1.8 percent in rural areas,
Open Traditional Improved Septic
Area defecation latrine latrine tank which is more than twice the rate of expan-
Urban 8 51 14 25 sion of septic tanks and improved latrines
combined (figure 17.4). Growth in the use of
Rural 41 51 5 2
unimproved latrines is concentrated among
National 34 52 9 10
the poorer quintiles, whereas growth in the
Source: Morella, Foster, and Banerjee 2008.
use of improved latrines and septic tanks is
concentrated among the richer quintiles.
Figure 17.3 Coverage of Sanitation by Budget Because the MDG target focuses on the
Quintile
two most improved sanitation options, the
100 expanding use of unimproved latrines does
not always fully register in policy discussions.
80
Meanwhile, the prevalence of open defecation
% of households
2.0 2.0
(% of population)
(% of population)
1.0 1.0
0 0
–1.0 –1.0
septic improved unimproved open septic improved unimproved open
tank latrine latrine defecation tank latrine latrine defecation
Figure 17.5 Moving Up to the Bottom Rung of the Sanitation Ladder: Côte d’Ivoire and Ethiopia, 1990–2005
(% of population)
4.0 4.0
2.0 2.0
0 0
–2.0 –2.0
–4.0 –4.0
–6.0 –6.0
septic improved unimproved open septic improved unimproved open
tank latrine latrine defecation tank latrine latrine defecation
a. Madagascar b. Rwanda
8.0 8.0
annual change in coverage
6.0 6.0
(% of population)
(% of population)
4.0 4.0
2.0 2.0
0 0
–2.0 –2.0
–4.0 –4.0
–6.0 –6.0
septic improved unimproved open septic improved unimproved open
tank latrine latrine defecation tank latrine latrine defecation
Figure 17.7 Mainstreaming Septic Tanks: Senegal, across urban and rural areas within countries
1990–2005 that can help structure policy alternatives
8.0 (figure 17.8). Urban areas tend to follow one
annual change in coverage
Figure 17.8 Characterizing Patterns of Access to Sanitation across Urban and Rural Areas
national sanitation policies, and most countries available evidence report negligible public
have an accepted definition of sanitation and investment on sanitation of 0.02 percent of
a hygiene promotion program. But only seven GDP, on average, although serious measure-
countries have policies that include cost recov- ment problems mean a large portion of public
ery, and only eight have a sanitation fund or a investment in sanitation is likely not separately
dedicated budget line (in some cases, funded coded from water supply (Briceño-Garmendia,
exclusively by donors, as in Chad and Ethiopia, Smits, and Foster 2008). Nevertheless, house-
or by a combination of the government, sector holds appear to be footing the bulk of the
levies, and donors). Côte d’Ivoire has the only investment bill. In countries with very low
fund financed entirely by sector levies. current spending, whether households alone
will be able to increase investment up to the
level needed is uncertain. In addition, noth-
Households Foot the Bill ing is known about their spending on opera-
Building on earlier work (Mehta, Fugelsnes, tion and maintenance, which is estimated to
and Virjee 2005), one can estimate the overall require an additional 0.2 percent of GDP in
price tag for reaching the sanitation MDG at the future. Public spending on operation and
$6 billion a year, or roughly 0.9 percent of maintenance appears to have already reached
Africa’s GDP (Morella, Foster, and Banerjee this level, but evidence is limited, and spending
2008). Capital investment needs based on refers mostly to sewer networks. Operation
minimum acceptable asset standards and and maintenance of on-site sanitation remain
accounting for both new infrastructure and a household responsibility, and facilities are
rehabilitation of existing assets can be con- notoriously poorly maintained.
servatively estimated at $4.5 billion a year Although the costs of meeting the MDG
(0.7 percent of the region’s GDP). The main- sanitation target are high, so is the associ-
tenance requirements are $1.5 billion a year ated health dividend (Hutton and others
(0.2 percent of the region’s GDP). 2007). Sanitation reduces the risk of intestinal
No reliable data exist on sanitation worms, diarrhea, and trachoma, and it is very
expenditures because individual households important—more than access to safe water—
undertake so much of the expense. However, in fighting hookworm infection (Esrey and
recent investment can be estimated from others 1991). Access to adequate sanitation
household surveys, using the number of house- reduces diarrhea incidence by an estimated
holds acquiring access in successive years and a 36 percent. Trachoma incidence was reduced
standard unit cost. Because this method treats by 75 percent in Gambian villages solely by
all new or improved facilities as newly built, it controlling flies that breed when excreta are
may overestimate the cost of increasing access not safely disposed of (Emerson and others
mainly by upgrading unimproved latrines. 2000). One study estimates that reaching the
This approach suggests that, on average, Afri- MDG for both water and sanitation in Africa
can countries are investing about 0.5 percent of would prevent 172 million diarrhea cases a
GDP in new sanitation facilities, quite close to year, saving $1.8 billion in treatment costs
the recommended investment level. Half of the (Hutton 2000).3
countries appear to invest less than 0.7 percent
of GDP, which is the level needed to meet the
sanitation access MDG. Some countries, par- Challenges and Policy Options
ticularly Madagascar and Rwanda, appear to
have made rapid progress, investing as much as One of the strongest findings emerging from
1.0 percent of GDP. At the other end of the this review is how much the sanitation chal-
spectrum, Kenya, Lesotho, Namibia, and lenge differs across countries and between
Zambia spend less than 0.2 percent of GDP. urban and rural areas within the same country.
How much of the estimated total spend- Decisions about where to focus policy efforts
ing on sanitation comes from the public purse along the sanitation ladder should be informed
is hard to pin down. The few countries with by access patterns. Recommendations will
330 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
therefore distinguish the different groups in latrines without accompanying hygiene edu-
identified in figure 17.8. Judicious and low- cation led to only 37 percent of men using the
cost public interventions can leverage house- facilities despite 100 percent coverage (World
hold spending for construction of latrines. Bank 2002). Hygiene education is critical
The ultimate objective should be to pro- regardless of the type of sanitation challenge a
vide universal access by expanding service and country faces; safe disposal of feces and hand
reducing open defecation as much as possible. washing with soap protect health in all sani-
Policy makers are often tempted to focus on tation settings. Promoting hygiene can start a
rungs of the sanitation ladder well above the virtuous cycle that builds demand for better
realities of their societies: for example, chan- sanitation, raising awareness of the benefits of
neling limited public resources into sewer net- sanitation and establishing codes of conduct
works that serve only a few thousand people and new life standards.
while overlooking the urgent need to lift mil- Incorrect use of latrines can dramatically
lions more people away from open defecation. reduce or even reverse their health benefits. A
Policy attention needs to focus on moving facility is sanitary and safe not only because
people up from the lowest rungs of the ladder. of the technology and material used but also
More expensive options should be left to house- because of good practices and behaviors,
holds with the resources to take them up. such as keeping the facility contained and
African countries may face high preva- clean. An improved latrine that is not cor-
lence of open defecation, especially in rural rectly used and emptied still poses high risks
areas; dominance of unimproved latrines; or of environmental contamination and disease.
significant development of septic tanks that Thus, rolling out a physical investment pro-
reach a small share of the population, mainly gram without accompanying promotion of
wealthier urban residents. The policy options hygiene makes little sense. Moreover, effec-
for each issue are presented as separate cases tive hygiene promotion alone may stimulate
below, and countries may need to use different self-financed household investment in bet-
combinations of these approaches. The first ter facilities. Too often, these “soft” aspects
priority is to stimulate demand for sanitation of sanitation are overlooked, and priority is
and behavior change where open defecation given to the “hard” aspects, such as installing
prevails. The second is to ensure an adequate and upgrading infrastructure.
supply of improved sanitation options in Changing behavior requires sustained
settings dominated by unimproved latrines, communication and public education at the
before evaluating the need for policy interven- community level. Understanding the moti-
tions on the demand side of the market. The vations that interest people in hygiene and
third is to expand access to improved sanitation sanitation is important. Health is one consid-
across larger shares of the population, which in eration, but not necessarily the foremost in
high-density settlements requires making sew- people’s minds: convenience, dignity, and social
erage more affordable. status may be more important. Adapting hygiene
and sanitation promotion programs to cultural
Stimulate Demand for Sanitation and and institutional norms and intensely market-
Behavior Change Where Open ing them to stimulate communitywide involve-
Defecation Prevails ment are critical. Peer pressure—to improve
Unlike other infrastructure services, demand one’s status—can also help. When a commu-
for sanitation cannot be assumed. Populations nity recognizes preferred behaviors, pressure to
accustomed to open defecation may require conform arises, and social structures and lead-
a substantial change in cultural values and ers begin to contribute. A successful example is
behavior to use a fixed-point facility. Without the Southern Nations Regional Health Bureau’s
such change, people may not use latrines at all sanitation advocacy campaign launched in
or they may use them in a way that undermines 2003 in southern Ethiopia. It increased latrine
the potential health benefits. A study in south- coverage from 13 percent of the population to
ern India showed that large public investment 78 percent in just two years (box 17.2). Once
Sanitation: Moving Up the Ladder 331
BOX 17.2
a culturally appropriate formula is found, dra- of moving people away from open defecation.
matic change can be achieved with modest The problem is rather about improving facili-
public spending focused on promoting sanita- ties. The debate centers on whether the main
tion rather than on financing hardware. impediment to upgrading latrines comes from
the supply side or the demand side.
Ensure Adequate Supply before Using standardized unit costs from Senegal’s
Addressing Demand in Settings sanitation sector, one can estimate the percent-
Dominated by Unimproved Latrines age of a household’s monthly budget that would
Where unimproved latrines prevail, the cen- be absorbed by the up-front investment in dif-
tral problem is how to upgrade them to more ferent types of sanitation facilities (table 17.2).
hygienic facilities so that the full health benefits Although unimproved latrines are affordable
of fixed-point defecation can be achieved. Coun- across all income levels, improved latrines
tries where unimproved latrines are widely used cost much more than a month’s household
have already overcome the behavioral challenge income in Senegal—even for households
332 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
in the highest income group. These findings prevalence of improved latrines is low, even in
are consistent with the patterns of access to middle-income countries, except in a handful
sanitation observed across the socioeconomic of cases. Second, 40–50 percent of the popula-
spectrum in Africa. The fact that half of Afri- tion use unimproved latrines, even among the
can households have invested in unimproved highest-income groups who may be able to pay
latrines in the absence of any subsidy suggests for more advanced facilities.
that these modest investment costs are afford- A weak private sector dominated by small
able across the income spectrum. The fact that entrepreneurs at the local level compounds the
improved latrines are found only among the supply problem. Latrine construction demands
wealthiest households indicates that affordabil- skills not widely available, and small enterprises
ity may well be an issue in this case. In addition, often do not have the resources to develop new
in urban areas, poor dwellers in slum settings skills or adopt new technologies.
often do not own their land or house and so Supply bottlenecks should be tackled first.
have fewer incentives to invest in improving Otherwise, subsidy resources may be wasted
their living conditions. on households that could have financed the
The appropriate policy response to these facilities on their own. Allowing the local mar-
demand-side constraints likely entails a ket to develop also provides space for inno-
public subsidy for the additional capital vation that can lower the cost of improved
costs associated with a standard package of latrines. Technological innovation is needed
improved facilities. However, a subsidy may to secure greater health benefits with cheaper
have drawbacks, including distorting demand variants that are tailored to a locality. This
and markets. Subsidies can reduce the demand innovation should be grounded in a better
of households with the ability to pay. More- understanding of the most prevalent designs
over, suggesting a standard package may make for unimproved latrines in any given local-
poor households feel entitled to such a facility, ity and should explore how relatively minor
regardless of whether it is the most appropriate changes in these designs could help achieve
for their circumstances and geographic loca- health benefits.
tion. Widespread adoption of a standard could Policies need to address supply-side limita-
also discourage innovations that may lower tions. Government support is best channeled
costs. Therefore, in the African context, many toward (a) conducting research and devel-
other policy measures likely need to be taken oping products, (b) marketing latrines, and
before subsidies become relevant. (c) opening supply channels for key inputs.
From the supply side, poor knowledge in Training small service providers and provid-
the construction sector about required designs, ing access to credit can also help. The National
a lack of skilled construction workers, and a Sanitation Program in Lesotho, established
shortage of materials can explain the low prev- 20 years ago, is dedicated to sanitation pro-
alence of improved latrines. Access patterns motion and private sector training. House-
already provide some clues that supply-side holds directly employ private latrine builders
issues are a real constraint in Africa. First, the trained under the program. The program has
Sanitation: Moving Up the Ladder 333
are also needed. Cost recovery has proved to of sanitation functions, even within the public
be a limited incentive because the only tariffs sector, prevent one entity from leading, and
in sanitation are those on sewerage, and they sanitation issues fall between the cracks.
apply only to the minority of the popula- A key policy issue is therefore to identify and
tion served by that network. Moreover, most empower a clear sanitation champion within
African utilities responsible for providing the public sector. In Senegal, the decision to
wastewater services also supply water, and the take sanitation seriously was expressed through
lack of accounting separation between these the creation of a dedicated sanitation utility.
services makes it likely that water pays for Senegal was also the first country to establish
sanitation. Burkina Faso has taken an innova- a government body at the national level—
tive approach by levying a sanitation tax on the Ministry for Prevention, Public Hygiene
the water bill, which is then used to subsidize and Sanitation (recently reorganized as the
access to improved sanitation facilities in Ministry of Urban Affairs, Housing, Urban
Ouagadougou (box 17.3). Water, Public Hygiene and Sanitation)—to
coordinate sector activity. Although creat-
Needed—a Champion for the ing a ministry in the central government may
Sanitation Sector not always be necessary, Senegal provides an
Given that on-site sanitation, rather than water- important lesson in singling out one entity
borne sewerage, will likely continue to domi- with a clear mandate to lead.
nate sanitation in Africa, households rather
than governments will remain center stage. Measuring Progress
Even so, the government’s role in promoting Although the JMP has made strides in moni-
demand and addressing supply bottlenecks toring progress toward the MDG target for
remains. Too often, dispersion and duplication sanitation, no commensurate effort has
been made to create detailed and frequent
country-level monitoring and evaluation sys-
tems critical to guiding policy interventions.
BOX 17.3 Most countries have no system, and in the
countries that are developing a system, pro-
viding a clear picture of the sector is not yet
Burkina Faso’s Sanitation Tax
possible. Moreover, monitoring and evalu-
The on-site sanitation problems in Ouagadougou are specifically ation systems rarely measure the effect of
addressed in the Sanitation Strategic Plan being implemented by the improved sanitation on health.
national public utility in charge of water supply and sanitation. At the country level, better monitoring and
A sanitation marketing approach has enhanced construction ser-
evaluation systems could be built by ensuring
vices offered to households by small providers and stimulated house-
more coordination at the ministerial level, for
hold demand for improved sanitation facilities. Some 700 masons
and social workers have been trained since the beginning of the
instance, between the ministry in charge of sani-
program. tation and the ministry in charge of health. In
Burkina Faso’s national utility offers to provide part of the mate- addition, a larger role should be played at the
rial free to households—equivalent to about a 30 percent subsidy— local level, especially by the decentralized techni-
with the households financing the rest. The utility finances the subsidy cal departments, in collecting data and monitor-
through a small sanitation tax on the water bill. ing progress. This would require more capacity
This example shows that on-site sanitation corresponds to a and resources from the central government.
strong demand from urban dwellers, with more than 60,000 pieces A limitation of the JMP’s framework is
of sanitation equipment subsidized so far—latrines as well as gray the inability to discriminate among the levels
water–removal systems. It also demonstrates the importance of a
of sanitary protection provided by different
local financing mechanism. Donors have contributed to the mecha-
variations within the large class of unim-
nism, but only modestly. Most of the funds come from the tax on the
water bill.
proved latrines, which will continue to domi-
nate African sanitation. Unimproved latrines
Source: Reproduced from Water and Sanitation Program 2008. include a heterogeneous collection of instal-
lations, only some of which can be regarded
Sanitation: Moving Up the Ladder 335
as improved sanitation. Unfortunately, the are estimated at $0.50 or less. As for economic
JMP’s household survey instruments that losses from lost time at work and school and
track progress toward the MDG target cannot from death, Hutton relies on the concept of
discriminate among the qualities of installa- minimum wage rates for his estimates, adjusted
to reflect the varying productivity of the dif-
tions within the unimproved latrine category.
ferent countries (for each country, the value
As a result, the data on progress in sanitation
of the minimum wage rate must be no larger
in Africa are blurriest precisely where most than the local gross national product per capita
of the progress is taking place. The precision and no smaller than the manufacture added
of household survey instruments should be value). Both health and economic benefits
improved in this respect. are presented, assuming that all interventions
Tracking the intermediate goal of increas- were implemented within 2000. To account for
ing the share of households making use of population growth, the projected population
some kind of sanitation facility, even if it is figures for 2015 are used.
an unimproved latrine, may also be relevant,
given that this is where Africa has been making References
the greatest progress. Banerjee, Sudeshna G., Quentin Wodon, Amadou
Diallo, Taras Pushak, Helal Uddin, Clarence
Tsimpo, and Vivien Foster. 2008. “Access, Afford-
Notes ability and Alternatives: Modern Infrastructure
The authors of this chapter are Elvira Morella, Services in Sub-Saharan Africa.” Background
Sudeshna Ghosh Banerjee, and Vivien Foster, who Paper 2, Africa Infrastructure Country Diagnos-
drew on background material and contributions tic, World Bank, Washington, DC.
from Piers Cross, Pete Kolsky, Marianne Leblanc, Briceño-Garmendia, Cecilia, Karlis Smits, and
Eustache Ouayoro, and Ede Perez. Vivien Foster. 2008. “Financing Public Infra-
1. Unimproved latrines refer to various kinds of structure in Sub-Saharan Africa: Patterns, Issues,
pits for excreta disposal. Well known in Africa, and Options.” Background Paper 15, Africa
Asia, and Latin America, they normally consist Infrastructure Country Diagnostic, World Bank,
of a simple pit covered with logs, not usually Washington, DC.
roofed, and sometimes with no walls. Emerson, Paul M., Sandy Cairncross, Robin L.
2. Improved latrines—comprising SanPlat, VIP Bailey, and David C. Mabey. 2000. “Review of
latrines, and basic pits with slabs—ensure more the Evidence Base for the ‘F’ and ‘E’ Components
hygienic separation of excreta from the immedi- of the SAFE Strategy for Trachoma Control.”
Tropical Medicine and International Health 5 (8):
ate living environment. Improved versions have
515–27.
walls and a roof and may include a ventilation
pipe or a cover plate for the squat hole. The col- Esrey, Steven A., James B. Potash, Leslie Roberts,
lection chamber may vary from an unlined pit and Clive Shiff. 1991. “Effects of Improved
Water Supply and Sanitation on Ascariasis,
to a composting chamber. The superstructure
Diarrhea, Dracunculiasis, Hookworm Infection,
may be a crude shelter or an attractive brick Schistosomiasis and Trachoma.” Bulletin of the
or thatch construction with or without a vent World Health Organization 69 (5): 609–21.
pipe and with or without a seat. SanPlat latrines
Hutton, Guy. 2000. Considerations in Evaluat-
are slightly elevated for ease of use in the dark.
ing the Cost-Effectiveness of Environmental
They can be located close to the house and have Health Interventions. Geneva: World Health
a fitted lid to prevent odors and keep away flies. Organization.
VIP latrines consist of the normal pit but are
Hutton, Guy, U-Primo E. Rodriguez, Lydia
fitted with a screened vent pipe.
Napitupulu, Pham Ngoc Thang, and Phyrum
3. Hutton (2000) uses health care unit costs from Kov. 2007. Economic Impacts of Sanitation in
the World Health Organization to estimate Southeast Asia: A Four-Country Study Conducted
the cost of treating diarrhea, to which he adds in Cambodia, Indonesia, the Philippines and
other expenses (such as transport) incurred by Vietnam under the Economics of Sanitation
the patient. A number of assumptions are made Initiative (ESI). Jakarta, Indonesia: World Bank,
regarding treatment (such as number of visits Water and Sanitation Program.
or length of hospitalization). As a result, the Mehta, Meera, Thomas Fugelsnes, and Kameel
mean cost per case of diarrhea for the patient Virjee. 2005. “Financing the Millennium
is $10–$23, and the additional costs per visit Development Goals for Water Supply and
336 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION
Sanitation: What Will It Take?” Water Resources Countries. Joint Report of the African Ministers’
Development 21 (2): 239–52. Council on Water, African Development Bank,
Melo, Jose Carlos. 2005. The Experience of Condo- the European Union Water Initiative, the United
minial Water and Sewerage Systems in Brazil: Case Nations Development Programme, the Water
Studies from Brasilia, Salvador and Parauapebas. and Sanitation Program–Africa, and the World
Lima, Peru: Water and Sanitation Program— Bank. Nairobi, Kenya: Water and Sanitation
Latin America, World Bank. Program–Africa.
Morella, Elvira, Vivien Foster, and Sudeshna G. ———. 2008. “Can Africa Afford to Miss the
Banerjee. 2008. “Climbing the Ladder: The State Sanitation MDG Target? A Review of the
of Sanitation in Sub-Saharan Africa.” Back- Sanitation and Hygiene Status in 32 Coun-
ground Paper 13, Africa Infrastructure Country tries.” Joint Initiative of the African Ministers’
Diagnostic, World Bank, Washington, DC. Council on Water, African Development Bank,
World Bank, and Water and Sanitation
United Nations. 2008. The Millennium Program, Washington, DC.
Development Goals Report 2008. New York:
United Nations. World Bank. 2002. “Urban Environmental
Strategic Sanitation Planning: Lessons from
Water and Sanitation Program. 2006. Getting Bharatpur, Rajasthan, India.” Field Note 23771.
Africa on Track to Meet the MDGs on Water and Water and Sanitation Program–South Asia,
Sanitation: A Status Overview of Sixteen African New Delhi.
Index
Boxes, figures, notes, and tables are indicated with b, f, n, or t following the page number.
freight tariffs, 220 Commission for Africa, 7, 31, 43, 44, 58, 77
ICT investment, 147, 148b, 148t Commission Internationale du Bassin Congo–Oubangui–
infrastructure, 48 Sangha, 204–205
multimodal transport network, 204–205 Common Market for Eastern and Southern Africa, 32, 156,
roads, 211 159b, 264–265, 264t
rural road accessibility, 222 commons approach to radio spectrum use, 177
Central African Power Pool, 52, 53–54t, 185, 187 community-driven development
Central African Republic irrigation, 296, 296b
airports and air transport, 261 sanitation improvement, 330, 331b
drinking water access, 300 service expansion, 97–98
irrigation, 290, 290f water resource management, 279
multimodal transport network, 204–205 Comoros
transit corridor development, 207, 208b drinking water access, 300
Central Africa Transport and Transit Facilitation ICT investments, 58
Project, 208b competition
Central East African Railways, 204 mobile phone market, 113, 169, 170, 170f
Chad ports and shipping, 256
airports and air transport, 261 railways, 239
broadband access, 173 Comprehensive Africa Agricultural Development Program,
power sector underpricing, 191 288, 292
power trade, 149, 150f concessions
private sector participation, 115 airports and air transport, 266
road maintenance, 214, 215, 216 port sytems, 253, 254t
sanitation, 329 railways, 238–240, 242–243, 242f, 245
transit corridor development, 207, 208b water utilities, 310
water supply, 313 condominial sewerage, 333
children Congo. See Democratic Republic of Congo; Republic of Congo
infant mortality, 45 connection charges, 90, 98–99, 98b, 308b
piped water and, 94 connection subsidies, 98b
water resources and, 45–46, 304–305 construction costs, 136, 141n9, 216, 225
China Cooperative Development of Operational Safety and Continuing
financing from, 79, 79t, 80, 82, 83f, 85n3 Airworthiness Program, 268
infrastructure expansion, 48 corporatization of state-owned water utilities, 313
irrigation, 289 corridor associations, 151
land use planning, 139 corruption, 108, 151, 205–206, 207b
manufacturing, 145 cost recovery
national infrastructure programs, 146 improvement of, 70–71, 72t
power sector investment by, 200 power sector, 191–192f, 191–194, 197–198
public sector expenditures, 67 railways, 245
urban infrastructure financing, 137 sanitation, 334
urban population density, 134 for subsistence consumption, 91–92, 91f
China Export-Import Bank, 85n3 tariffs and, 99–100, 99f
civil aviation authorities, 268, 269 water supply, 307–309, 307f, 308b, 319–320
Clean Development Mechanism, 149 Côte d’Ivoire
climate change airports and air transport, 266
agriculture and, 287, 288 electricity access, 193
food security and, 293–294, 294t infrastructure as constraint on economic growth, 2
water resource management and, 53–54, 272, 279 irrigation, 290
coastal countries local capital markets and, 81
population density, 134–135, 135t municipal budgets, 134
urbanization, 127, 127t piped water access, 88
urban vs. rural income level differentials, 128, 128t power outages, 184
Code-Division Multiple Access (CDMA), 168 private sector participation, 112, 115
Commercial Reorientation of the Electricity Sector Toolkit railways, 204, 232, 233, 238
(CREST), 196b road funding, 213
340 Index
irrigated areas expansion, 54, 55t IPPs. See independent power producers
power sector, 52–53, 53–54t irrigation, 287–297
transport network, 55–56, 57t challenges, 294–296
water and sanitation, 54–55, 55t community-driven development, 296, 296b
water security, 53–54 current state of, 289–290, 289t, 290f
urban areas, 128–130 economic impact of, 288–289
costs of, 130–132 impact of, 293–294
financing, 133–138 institutional reforms, 295, 295b
Infrastructure Consortium for Africa, 32, 78, 158, 160n3 investment needs, 36, 54, 55t, 290–293, 291b, 291t,
Infrastructure Institutional Scorecard, 106–107b, 108–109f, 109t 292f, 293t
inland waterways, 204 operating inefficiencies, 74t
institutional competence and capacity, 105–123. See also specific project analysis, 295–296
infrastructure sectors public expenditures on infrastructure, 70–71t
independent regulators, 120–121, 121b, 121t sector performance data collection, 40
private sector participation, 110–117, 111f, 111t, 116b, 117t strategic planning, 294–295
recommendation for, 16–19 underdevelopment of, 274, 276f, 289–290
reform efforts, 12–14, 13f, 106–110 water resource management and, 280
regional integration, 143, 155–156
road infrastructure, 214f Japan, ODA from, 78
scorecard, 106–107b, 108–110f, 109t Joint Monitoring Programme (WHO/UNICEF), 300, 321n2,
state-owned enterprises, 117–120, 118b, 119b, 190b 323, 324b, 334–335
water and sanitation, 274
Intergovernmental Panel on Climate Change, 272 Kager Basin Organization, 282
International Air Transport Association, 268, 269 Kano–Katsina–Maharadi initiative, 157
International Benchmarking Network (IBNET), 39 Kenya
International Civil Aviation Organization, 268, 269 airports and air transport, 261, 266, 267
International Container Terminal Services, 252 external governance and, 110
International Convention for the Safety of Life at Sea, 255 fiber-optic cable network, 171–172
International Development Association (IDA), 78, 82, 83f, governance, 108
158, 159 hydrologic variability, 273
International Energy Agency, 40 informal settlements, 136–137, 139
International Finance Corporation, 148b irrigation, 290, 290f
International Food Policy Research Institute, 35b, 40 multimodal transport network, 204
international ICT connectivity, 171, 176 power sector, 189b
International Monetary Fund (IMF), 36, 37, 77b, 185 power trade, 150f
international phone service, 167, 167f, 171t PPI investments, 79
International Ship and Port Facility Security Code, 255 private sector participation, 113, 189b
international shipping market, 255–256 railways, 244
International Telecommunication Union, 39 sanitation, 325, 329
International Union for Conservation of Nature, 35b state-owned enterprises, 118
Internet access submarine cable project, 148b
ICT sector and, 165, 167f telecommunications, 170, 179n1
investment needs, 57, 172–173, 172t transport infrastructure, 150, 151t
price premium on, 50–52 urban infrastructure, 96b, 137
pricing for access, 167, 167f, 171, 171t water MDG and, 305
regional cooperation opportunities, 146–148 water resource management, 275, 276
Interstate Aviation Committee, 263 water supply, 313, 315
intraregional trade, 152–153 Kenya Airways, 151, 260, 262
investment climate surveys, 45 Kenya Electricity Generating Company, 189b
investment spending Kenya Power and Lighting Company, 188, 189b
irrigation needs, 290–293, 291b, 291t, 292f, 293t kerosene lamps, 94, 97t
maintenance vs., 8, 8f, 60 Korea. See Republic of Korea
power sector backlog, 185–186 Kreditanstalt für Wiederaufbau (Germany), 32
railways, 236–238 Krugman, Paul, 144
water security needs, 279–283 KwaZuluNatal program, 226
Index 345
railways, 238 private sector participation, 113, 114b, 115, 157, 257b
sanitation, 325, 329 railways, 233
telecommunications sector inefficiency, 168 road network, 213, 216, 222
utility payment collection, 100 telecommunications industry privatization, 170
utility tariffs, 99 traffic volumes, 212
water supply, 315 transport infrastructure, 151t
water tariffs, 308b undermaintenance, 73
national flag carriers, 267 urban planning, 138
National Freight Logistics Study (South Africa Department of water supply, 314, 315, 316b, 320
Transport), 209 Nigeria Airways, 261
National Geospatial-Intelligence Agency (U.S.), 35b Nigerian Ports Authority, 257b
National Ports Authority (South Africa), 256 Niger River basin, 272, 279
National Sanitation Program (Lesotho), 332 Niger River Commission, 153, 282
National Transportation Safety Board (NTSB, U.S.), 268 Nile Basin Initiative, 282, 283f
National Transport Safety Commission (Ghana), 226 Nile River basin, 272, 279
National Water and Sewerage Commission (NWSC, nonfragile low-income states
Uganda), 119b access issues, 89–90
negative concessions, 209, 210n2 in AICD typology, 48, 51b
neighborhood effect of urban areas, 126 capital budget for infrastructure, 69t, 76, 76t
Netherlands and irrigation investment, 295b economic return on infrastructure investment, 70, 71t
New Partnership for Africa’s Development (NEPAD), 32, 154, financing needed to meet water MDG, 302–305, 303t
288, 292 funding gap, 8, 66–67, 67t, 75, 75t
Niger infrastructure, 2–3, 45, 48
affordability issues, 92 institutional competence and capacity, 107
airports and air transport, 261 local capital markets and, 81t
broadband access, 173 mobile phone coverage, 166
connection subsidies, 90, 98b non-OECD financing, 68f, 79t
governance, 108 ODA to, 67, 68f, 78t
irrigation, 290, 292 power sector spending needs, 185, 185t, 186, 186t
ODA, 79 PPI investment, 68f, 80, 80t
power sector public sector expenditures, 67, 68f, 69f, 70–71t
inefficiencies, 187 road network spending needs, 219t
trade, 149, 150f spending estimates, 43, 58–59
underpricing, 191 urbanization, 127, 127t
private sector participation, 112, 115 utility water supply, 301–302, 302t
road network, 212, 216 water resources cost recovery, 309, 309f
spending estimates, 59 water resources funding gap, 302–305, 305t
utility tariffs, 90, 99, 99f non-OECD finance sources
water resources as human development constraint, 46, 305 cost of capital, 82, 83f
water supply, 315 funding gap, 66, 67, 76, 78–80, 79t
water utilities, 310, 311 infrastructure investment, 8–9, 9t
Niger Basin Authority, 153, 282, 283, 283f power sector, 186, 200, 201f
Nigeria water resources, 304
airports and air transport, 260 North America’s water resource management, 274, 275t
capital budget for infrastructure, 76
drinking water access, 300 Oak Ridge National Laboratory, 35b
electricity access, 194 Obasanjo, Olusegun, 154
fiber-optic cable network, 171–172 OECD countries
informal settlements, 139 telecommunications sector performance, 168
irrigation, 290, 290f, 292, 296b water tariffs, 307
land use policies, 138 off-grid power technologies, 182
non-OECD financing, 78, 79 Office du Niger, 295b
ports and shipping, 250, 253, 257b official development assistance (ODA)
power sector governance, 196b cost of capital, 82, 83f
PPI investments, 79 funding gap, 66, 67, 75–78, 78t
348 Index
water resource management, 46, 280, 282, 305 transportation sector, 203–210
water supply, 314, 315 alternative technologies for, 84
Tanzania Investment Centre, 136b capital budget for, 68
Tanzania Railways Corporation, 233 competition, 207–208
tariffs financing, 133
cost recovery and, 70–71, 99–100, 99f human development constrained by, 46
increasing block tariffs, 101, 308b infrastructure as constraint on economic growth, 45
power sector, 191 institutional competence and capacity, 108, 109
price premium of, 49–52 local capital markets and, 81
rail freight, 235 logistics systems for, 205–206, 205f, 207b
road freight, 5, 50, 151 multimodal transport, 204–205
subsidies of, 92–93, 93–94f non-OECD financing, 68f
telecommunications, 170, 170f ODA for, 68f
water, 299, 307–309, 308b, 319–320 operating inefficiencies, 74t
TEAMS (The East Africa Marine System, Kenya), 148b, 148t PPI investment, 68f
telecommunications. See also information and communications private sector participation, 116, 117t, 209
technologies public expenditures on infrastructure, 68f, 69f, 70–71t
access to, 88–89f regional integration, 145, 147f
cost recovery, 5 roads. See roads
economic impact of, 45, 45f, 47 sector performance data collection, 39
independent regulators and, 120, 121t social obligations and, 209–210
institutional competence and capacity, 107, 108, 109f, 110f spending needs, 36, 55–56, 57t
landline. See landline telephones transit corridor development, 206–207, 207b, 208b
operating inefficiencies, 66, 72 urban vs. rural network quality, 129, 130f
price premium on, 50–52 Transports Internationaux Routiers (TIR, EU), 206, 208b
private sector participation, 111–113, 111f, 113f trucking regulation, 206–207, 224–225
regional cooperation opportunities, 146–148, 148b
sector performance data collection, 39 Uganda
state-owned enterprises, 118 affordability issues, 92
Tema Development Corporation, 138 deficit-financed public investment, 77b
thermal power, 149, 200 drinking water access, 300
time allocation and basic infrastructure access, 95b hydroclimatic variability, 272, 273f
TIPAC (Transit Inter-États des Pays de l’Afrique Centrale), 208b hydropower project, 201
TIR (Transports Internationaux Routiers, EU), 206, 208b informal settlements, 139
Togo, urban infrastructure financing in, 137 mobile phone coverage, 178
toll roads, 213 multimodal transport network, 204
tour de role dispatching, 203, 210n1, 220, 224 paved road density, 96b
trade power sector inefficiencies, 187
by landlocked countries, 206 PPI investment, 80
power sector, 53, 187, 198–200 private sector participation, 112, 113
regional integration and, 149, 150–153, 150f, 151t, 159b railways, 233, 244
transportation network and, 153, 204–205 road maintenance, 216
traffic volumes sanitation, 325
air transport, 260–261, 260f, 261t standposts, 95
maritime, 250–251, 250–251t, 251f state-owned enterprises, 118, 119b, 314b
railways, 233–234, 234f telecommunications industry privatization, 171
roads, 212 undermaintenance, 73
Trans-Africa Highway Network, 34, 152–153, 212 water supply, 313, 315, 316, 320
transboundary river basins, 153–154, 272–273, 274f, 279 state-owned enterprise reforms, 314b
transit corridor development, 206–207, 207b, 208b water utilities, 310, 313
Transit Inter-États des Pays de l’Afrique Centrale (TIPAC), 208b underwater communication cables, 146, 149t, 174
Transnet, 256 unimproved latrines, 94, 96f, 97t, 321–327, 328f
transparency Union of African Railways, 237
private sector participation, 114–115, 117 Union of Producers, Transporters and Distributors of
state-owned enterprises, 195 Electric Power in Africa, 40
Transportation Department (U.S.), 268 United Arab Emirates, submarine cable project, 148b
354 Index
United Kingdom and development assistance, 32 economic development and, 272–276, 277
United Nations Children’s Fund (UNICEF), 300, 321n2, financing, 133, 302–305, 303t
323, 324b funding gap, 66, 75, 75t, 302–305, 305t
United States and development assistance, 78, 295b groundwater, 320–321
Universal Safety Oversight Program (International Civil Aviation human development constrained by, 45–46, 274–275
Organization), 268 independent regulators and, 120, 121t
Urban Affairs, Housing, Urban Water, Public Hygiene and institutional competence and capacity, 108–109, 109–110f,
Sanitation Ministry (Senegal), 334 115–116, 116b, 282–283
urban areas, 125–142 management of, 4, 272–275, 274–275f, 275t
agricultural productivity near, 126–127, 127t, 140n2 MDG for, 299, 300, 302–305, 303t, 321n1
capital budgets, 133–137, 134f non-OECD financing, 68f, 304
electrification, 198 ODA for, 67, 68f, 304, 304t
as growth engine, 126 operating inefficiencies, 10–12, 11–12f, 12t, 66, 72, 73, 74t,
growth of, 127–128, 127t, 136–137 310–311f, 310–313, 312t
housing planning, 134–136 poverty reduction and, 274, 277, 278f
informal settlements, 136–137, 139 PPI investment, 68f, 304, 304t
infrastructure, 2–3, 3f, 128–130, 129–130f, 131t private sector participation, 111–112, 111f, 115–116, 116b,
costs of, 130–132, 131t, 133–134f 117t, 310–313, 311f, 312b
financing, 133–138, 135t public sector expenditures, 68f, 69f, 70–71t, 304, 304t
investment needs, 132–133 reforms, 309–313, 311f, 315t, 317
land regulations, 134–136, 136b, 138–139 regional integration of resources, 153–154, 279,
mobile phone coverage, 166, 166f 282–283, 283f
principles of efficient urbanization, 139–140 resellers, 301, 302b
public transportation, 94, 96b rural access, 129, 129–130f, 300–302, 301–302t, 313–316,
roads, 220–221, 223–224 315f, 315t, 316b
rural links to, 126–130, 140n1 sector performance data collection, 39–40
sanitation, 325–326, 326f, 326t, 327, 328f security. See water security
spatial planning, 134–136 spending needs, 36, 54–55, 55t, 58, 280–281, 281t
territorial planning and, 138–139 standposts for, 305–306, 306b
transportation networks and, 56 state-owned enterprises, 119b, 120f, 120t, 314b
water supply access, 300–302, 311f, 320–321 subsidies, 92–93, 93–94f
user charges, 68, 70–71, 72t tariffs, 92–93, 299, 307–309, 308b, 319–320
U.S. Geological Survey, 35b technologies for, 305–307, 318–319
utilities. See also specific sectors transboundary river basins, 153–154, 272–273, 274f, 279
connection subsidies, 90, 98–99, 98b, 100–101, 101t, 308b urban access, 129, 129–130f, 300–302, 301–303t, 320–321
institutional competence and capacity, 108 water security
nonpayment issues, 91, 91f challenges, 276–277
operating inefficiencies, 10–12, 11–12f, 11f, 12t, 66, 71–74, economic development and, 276f, 277, 277b
73t, 310–313, 311f, 312t investment needs, 279–283
distribution losses, 72–73 spending needs, 53–54
overemployment, 72 storage infrastructure, 279–282
undercollection of bills, 73 Water Utilities Partnership, 39
undermaintenance, 73–74, 73f weather forecasting, 275–276, 279
prepayment meters, 100, 100f wells, 94, 96f, 97t, 300
water supply and, 301, 301t, 302–303t, 306b, 318–319 urban vs. rural access to, 300–301, 301t, 303f, 303t, 320
West Africa. See also specific countries
Ventilated Improved Pit (VIP) latrines, 324, 335n2 airports and air transport, 263, 268, 268f
Voice over Internet Protocol (VoIP), 175 air transport hubs, 151, 260
voice services, 166–167, 172, 172t backup generators, 184
capital markets, 81
water resources, 271–285, 299–322. See also irrigation economic growth, 44
alternative technologies for, 84, 96f, 97t electricity access, 193
capital budget for, 68 freight services profit margins, 50, 220
climate resilience, 279 ICT investment, 147, 148b, 148t
cost recovery, 307–309, 307f, 308b, 319–320 infrastructure, 48
Index 355
maritime traffic growth, 250, 250–251t, 251f Yemen ports and shipping, 253
multimodal transport network, 204 Ying, Yvonne, 95b
port system, 252, 255
railway concessions, 238–239 Zambezi River basin, 272, 280
regional integration, 156 Zambia
roads, 211 border crossing delays, 159b
water utilities, 310 broadband access, 173
West African Economic and Monetary Union, 156, 264–265, 264t governance, 108
West African Gas Pipeline, 154 hydrologic variability, 273
West African Power Pool, 52, 53, 53–54t, 159–160, 185, 187 irrigation, 290, 290f
Master Plan, 34 land use policies, 138
WHO/UNICEF Joint Monitoring Programme, 300 mobile phone coverage, 172
WiMAX technology, 57, 58 power sector underpricing, 191
wireless technologies, 178 railways, 204, 230, 235, 239, 244
Wodon, Quentin, 95b road network, 213, 214, 215, 222–223
women and water resources, 45–46, 94, 304–305 sanitation, 329
World Bank shipping, 251
AICD and, 32, 37, 40 transit corridor development, 207
AICD GIS data from, 35b utility inefficiencies, 309
Country Policy and Institutional Performance Assessment, 51 water resource management, 276, 282, 306
irrigation investment by, 295b water supply, 315, 320
ODA from, 78 water utilities, 313
regional project criteria, 158 Zambia–Malawi–Mozambique growth triangle, 157
World Development Report 2009 (World Bank), 127, 141n4, 145 Zimbabwe
World Health Organization (WHO), 300, 321n2, 323, border crossing delays, 159b
324b, 335n3 drinking water access, 300
World Meteorological Organization, 275–276 railways, 230, 232, 238, 244
transit corridor development, 207
Yamoussoukro Decision (1999), 151, 159, 259, 264, 268 water resource management, 281
yard taps, 319 water supply, 320
ECO-AUDIT
Environmental Benefits Statement
The World Bank is committed to preserving Saved:
endangered forests and natural resources. The • 31 trees
Office of the Publisher has chosen to print • 10 million Btu of
Africa’s Infrastructure: A Time for Trans- total energy
formation on recycled paper with 30 percent • 2,957 lb. of net
postconsumer fiber in accordance with the greenhouse gases
recommended standards for paper usage set • 14,242 gal. of waste
by the Green Press Initiative, a nonprofit water
program supporting publishers in using fiber • 865 lb. of solid waste
that is not sourced from endangered forests.
For more information, visit www.greenpress
initiative.org.
Sustainable infrastructure development is vital for Africa’s prosperity. And now is the time
to begin the transformation. This volume is the culmination of an unprecedented effort
to document, analyze, and interpret the full extent of the challenge in developing Sub-
Saharan Africa’s infrastructure sectors. As a result, it represents the most comprehensive
reference currently available on infrastructure in the region. The book covers the five main
economic infrastructure sectors—information and communication technology, irrigation,
power, transport, and water and sanitation.
Africa’s Infrastructure: A Time for Transformation reflects the collaboration of a wide array of
African regional institutions and development partners under the auspices of the Infrastruc-
ture Consortium for Africa. It presents the findings of the Africa Infrastructure Country
Diagnostic (AICD), a project launched following a commitment in 2005 by the international
community (after the G8 summit at Gleneagles, Scotland) to scale up financial support for
infrastructure development in Africa. The lack of reliable information in this area made it
difficult to evaluate the success of past interventions, prioritize current allocations, and
provide benchmarks for measuring future progress, hence the need for the AICD.
Africa’s infrastructure sectors lag well behind those of the rest of the world, and the gap
is widening. Some of the main—policy-relevant—findings highlighted in the book include
the following: infrastructure in the region is exceptionally expensive, with tariffs being
many times higher than those found elsewhere. Inadequate and expensive infrastructure
is retarding growth by 2 percentage points each year. Solving the problem will cost over
US$90 billion per year, which is more than twice what is being spent in Africa today.
However, money alone is not the answer. Prudent policies, wise management, and sound
maintenance can improve efficiency, thereby stretching the infrastructure dollar. There
is the potential to recover an additional US$17 billion a year from within the existing
infrastructure resource envelope—simply by improving efficiency. For example, improved
revenue collection and utility management could generate US$3.3 billion per year. Regional
power trade could reduce annual costs by US$2 billion. And deregulating the trucking
industry could reduce freight costs by one-half. So, raising more funds without also tackling
inefficiencies would be like pouring water into a leaking bucket.
Finally, the power sector and fragile states represent particular challenges. Even if every
efficiency in every infrastructure sector could be captured, a substantial funding gap of
$31 billion a year would remain. Nevertheless, the African people and economies cannot
wait any longer. Now is the time to begin the transformation to sustainable development.
ISBN 978-0-8213-8041-3
SKU 18041