Overview

Africa’s Infrastructure: A Time for Transformation

he Africa Infrastructure Country Diag- dis economies of scale in production and nostic is an unprecedented attempt to high profit margins caused by lack of T 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 • The cost of addressing Africa’s infrastruc- fi eldwork across Africa, the following main ture needs is around $93 billion a year, about fi ndings 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 fi nanced, with the central gov- access. ernment budget being the main driver of • Africa’s diffi cult economic geography pre- infrastructure investment. sents a particular challenge for the region’s • Even if major potential effi ciency 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, refl ecting 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 fi nancial operational effi ciency. market constraints. For one set of countries, power emerges as the most limiting factor by far, cited by more than half the fi rms in more Finding 1: Infrastructure than half the countries as a major business Contributed over Half of Africa’s obstacle. For a second set, ineffi cient function- Improved Growth Performance ing of ports and associated customs clearance is equally signifi cant. Defi ciencies 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 benefi ted 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, signifi cant 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 defi cit 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 fi rm 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 signifi cant 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 (fi gure 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, (fi gure O.1, panel b).

Table O.1 Africa’s Infrastructure Deficit Sub-Saharan Finding 3: Africa’s Difficult Africa Other low-income low-income Economic Geography Presents a Normalized units countries countries Challenge for Infrastructure Paved-road density 31 134 Development Total road density 137 211 Main-line density 10 78 Relative to other continents, Africa is char- Mobile density 55 76 acterized by low overall population density Internet density 2 3 (36 people per square kilometer), low rates of urbanization (35 percent), but relatively rapid Generation capacity 37 326 rates of urban growth (3.6 percent a year), a Electricity coverage 16 41 relatively large number of landlocked coun- Improved water 60 72 tries (15), and numerous small economies. Improved sanitation 34 51 A further complication is that the continent Source: Yepes, Pierce, and Foster 2008. experiences particularly high hydrological Note: Road density is measured in kilometers per 100 square kilometers of arable land; telephone density in lines per thousand variability, with huge swings in precipitation population; generation capacity in megawatts per million popula- across areas, seasons, and time, which climate tion; electricity, water, and sanitation coverage in percentage of population. change is likely to exacerbate.

Figure O.1 Access to Household Services

a. Stagnant trends b. Rural-urban divide 40 80

30 60

20 40 % of population 10 % of population 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

Source: Banerjee, Wodon, and others 2008. 4 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION

Africa’s atomized nation-states are refl ected 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 fl oods 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 artifi cial 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 fi ber-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 fi ber- 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- defi cient 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 fi lling the in genuinely higher costs, and sometimes in resulting gap (Banerjee, Wodon, and others 2008; high profi ts. 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 benefi t 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 (fi gure 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 profi t 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 confl icts that destroyed power payments are counted. Profi t margins, by con- infrastructure in fragile states. Africa’s fi rms 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 fi gure that rises to to 160 percent. The underlying cause is limited 20 percent for informal fi rms 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 refl ect a mixture of cost for some countries constitute a large pro- and profi t 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 signifi cant 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

Figure O.2 Underlying Causes of Africa’s Power Supply Crisis

MAURITANIA CAPE VERDE MALI NIGER SENEGAL CHAD SUDAN ERITREA

GAMBIA BURKINA DJIBOUTI FASO GUINEA-BISSAU GUINEA NIGERIA BENIN ETHIOPIA SIERRA LEONE CENTRAL GHANA CÔTE

D'IVOIRE AFRICAN REPUBLIC LIBERIA TOGO CAMEROON DEMOCRATIC SOMALIA EQUATORIAL GUINEA REPUBLIC OF UGANDA CONGO KENYA SAO TOME AND PRINCIPE GABON RWANDA

REP. OF CONGO BURUNDI TANZANIA SEYCHELLES

COMOROS MAYOTTE Main cause or trigger ANGOLA

MALAWI natural causes (droughts) ZAMBIA oil price shock ZIMBABWE NAMIBIA system disrupted by conflict MOZAMBIQUE MAURITIUS

high growth, low investment/structural issues BOTSWANA MADAGASCAR

SWAZILAND SOUTH AFRICA LESOTHO

Source: Eberhard and others 2008.

Finding 6: Africa’s Infrastructure • Interconnect capitals, ports, border cross- Spending Needs at $93 Billion ings, and secondary cities with a good- a Year Are More than Double quality road network. Previous Estimates by the • Provide all-season road access to Africa’s Commission for Africa high-value agricultural land. • More than double Africa’s irrigated area. Meeting Africa’s infrastructure needs calls for • Meet the MDGs for water and sanitation. a very substantial program of infrastructure Raise household electrifi cation rates by 10 investment and maintenance: • percentage points. Develop an additional 7,000 megawatts • Provide global systems mobile voice signal a year of new power generation capacity • and public access broadband to 100 percent (about half through multipurpose water of the population. storage schemes). • Enable regional power trade by laying Implementing such an ambitious program to 22,000 megawatts of cross-border transmis- address Africa’s infrastructure needs would cost sion lines. around $93 billion a year (about 15 percent of • Complete the intraregional fiber-optic the region’s GDP). Some two-thirds of this total backbone network and continental subma- relates to capital expenditure, and the remain- rine cable loop. ing one-third to operation and maintenance Africa’s Infrastructure: A Time for Transformation 7 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 fi nancial crisis of 2008 can be mission for Africa report in 2005. That fi gure 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, refl ecting growth are those for the power sector. However, Africa’s particularly large defi cits. 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 signifi cant 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 signifi cant 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 infl ation, 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 confl ict (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 (fi gure O.3). The recent confl icts 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- 25 alty payments they received during the recent 20 commodity boom provide a ready source of

% of GDP 15 fi nance. Yet resource rich-countries actually 10 lag nonfragile low-income countries in their 5 infrastructure stocks and spend less on infra- 0 structure. They have been devoting their added wealth not to infrastructure development but to

countries countries resource-richcountries countries paying off debts. The governance challenges in a middle-income fragile low-income Sub-Saharan Africa resource-rich environment may thus prevent the nonfragile low-income transformation of wealth into infrastructure. capital operation and maintenance Meeting the infrastructure needs of the middle-income countries looks to be much Source: Briceño-Garmendia, Smits, and Foster 2008. Note: Figures refer to investment (except public sector) and more manageable. These countries should include recurrent spending. Public sector covers general govern- be able to meet their infrastructure spending ment and nonfinancial enterprises. needs with 10 percent of GDP. They are also much stronger in asset maintenance and insti- or dilapidation of their (already modest) tutional effi ciency. Their more urban popula- national infrastructure platforms. In the Dem- tions also facilitate network rollout. ocratic Republic of Congo, about 50 percent of infrastructure assets need rehabilitation. The fragile states’ infrastructure spending needs Finding 8: A Large Share of Africa’s are especially large, particularly when mea- Infrastructure Is Domestically sured against the size of their economies. Such Financed countries would, on average, need to devote 37 percent of their GDPs to infrastructure Existing spending on infrastructure in Africa is spending to build a solid infrastructure plat- higher than previously thought, amounting to form. With their diffi cult environments, they $45 billion a year when budget and off-budget attract relatively little external fi nancing, cap- spending (including state-owned enterprises turing only 10 percent of overseas development and extrabudgetary funds) and external fi nan- assistance and 6 percent of private capital fl ows ciers are taken into account. The latter include allocated to infrastructure. In addition to their the private sector, offi cial development assis- huge fi nancing burden, the fragile states do not tance, and fi nanciers that do not belong to use their current resource envelope well; they the Organisation for Economic Co-operation underspend on maintenance and have ineffi - and Development (OECD). As much as two- cient service providers. thirds of this overall spending is domestically Nonfragile low-income countries need to sourced: $30 billion of annual spending is allocate, on average, about 23 percent of their fi nanced by the African taxpayer and infra- GDPs to build and sustain a basic infrastruc- structure user, and a further $15 billion is from ture platform, a level diffi cult to envisage in external sources (table O.4). practice. Therefore, these countries will have to The public sector remains the dominant make diffi cult choices about the prioritization source of fi nance for water, energy, and transport of their infrastructure investments, and most in all but the fragile states. Public investment is of them have a long way to go in improving the largely tax fi nanced and executed through cen- effi ciency of operating existing infrastructure. tral government budgets, whereas the operating The resource-rich countries are, in principle, and maintenance expenditure is largely fi nanced much better placed to meet their infrastruc- from user charges and executed through state- ture spending needs, though in practice they owned enterprises. Current levels of public Africa’s Infrastructure: A Time for Transformation 9

Table O.4 Infrastructure Spending on Addressing Sub-Saharan Africa’s Infrastructure Needs $ billions annually Operation and maintenance Capital expenditure Infrastructure Public Public Non-OECD Private Total sector sector sector ODA fi nanciers sector Total spending ICT 2.0 1.3 0.0 0.0 5.7 7.0 9.0 Power 7.0 2.4 0.7 1.1 0.5 4.6 11.6 Transport 7.8 4.5 1.8 1.1 1.1 8.4 16.2 WSS 3.1 1.1 1.2 0.2 2.1 4.6 7.6 Irrigation 0.6 0.3 — — — 0.3 0.9 Total 20.4 9.4 3.6 2.5 9.4 24.9 45.3 Source: Briceño-Garmendia, Smits, and Foster 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; ODA = official development assistance; OECD = Organisation for Economic Co-operation and Development; WSS = water supply and sanitation. — Not available.

fi nance 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 (fi gure O.4). In abso- 7.0 lute terms, however, spending remains very low, no more than $20–$30 per capita a year 6.0 (Briceño-Garmendia, Smits, and Foster 2008). 5.0 Looking only at investment, one fi nds that offi cial development assistance, private partici- 4.0 pation in infrastructure, and non-OECD fi nan- 3.0 % of GDP ciers together exceed domestically financed public investment (Briceño-Garmendia, Smits, 2.0 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 signifi cant, capital fl ows are provided by offi cial develop- countries countries countries ment assistance and, to a lesser extent, non- middle-income OECD fi nanciers, such as China, India, and the fragile low-incomeSub-Saharan Africa nonfragile low-incomeresource-rich countries Arab states. The focus differs markedly in each capital operation and maintenance case. Offi cial development assistance makes an important contribution to water and transport, Source: Briceño-Garmendia, Smits, and Foster 2008. particularly in fragile states. Non-OECD fi nance is signifi cant in energy and rail, especially in resource-rich countries. Private participation in further—to the tune of $17 billion a year. This infrastructure is heavily concentrated in ICT. is Africa’s major infrastructure effi ciency gap (Briceño-Garmendia, Smits, and Foster 2008). First, some countries are allocating more Finding 9: After Potential Efficiency resources to some areas of infrastructure than would appear to be warranted (Briceño- Gains, Africa’s Infrastructure Garmendia, Smits, and Foster 2008). This Funding Gap Is $31 Billion a Year, “excess expenditure” amounts to $3.3 billion Mostly in the Power Sector a year overall. The largest share of this excess expenditure relates to public spending on ICT Addressing a wide range of ineffi ciencies could infrastructure that the private sector could pro- make the existing resource envelope go much vide, particularly in middle-income countries. 10 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION

Although some of this “overspending” may cost of sound preventive maintenance. For be justifi ed 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 identifi ed 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 $2.4 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 ineffi ciency in distribution in theory increase by 30 percent without any losses, undercollection of revenues, and over- increase in spending, simply by addressing the staffi ng (fi gure 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 effi cient 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 signifi cant 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 (fi gure 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 confl ict. The reha- in developing countries. For ICT, countries bilitation backlog refl ects a legacy of under- retaining state-owned incumbents are often funding maintenance, a major waste given incurring signifi cant losses from overstaffi ng 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 ineffi ciencies can easily 50 exceed the current turnover of the utilities by several multiples. For power, these losses are 40 also material at the national level, absorbing 0.5 percent of GDP on the Sub-Saharan Afri- 30 can average, or $3.4 billion annually (Brice- ño-Garmendia, Smits, and Foster 2008). For 20 water, the absolute value of the ineffi ciencies is smaller, with the average amount accounting 10 for 0.2 percent of GDP, or $1 billion a year. Fifth, underpricing of infrastructure services % of assets in need rehabilitation 0 is substantial. Although African infrastructure charges are high by international standards, so

irrigation railways are the infrastructure costs. Even relatively high main roads rural water rural roads urban water rural average overall average tariffs can fail to cover more than the operat- power generationnonrural average ing costs. The revenues uncollected because of Source: Briceño-Garmendia, Smits, and Foster 2008. underpricing of power and water amount to Africa’s Infrastructure: A Time for Transformation 11

Figure O.6 Hidden Costs of Utility Inefficiencies

a. Power b. Water 5 5

4 4

3 3 % of GDP 2 % of GDP 2

1 1

0 0

countries countries countries countries countries countries middle-income middle-income fragile low-incomeSub-Saharan Africa fragile low-incomeSub-Saharan Africa resource-rich countriesnonfragile low-income resource-rich countriesnonfragile low-income unaccounted losses collection inefficiencies labor redundancies

Source: Briceño-Garmendia, Smits, and Foster 2008.

as much as $4 billion a year on aggregate, an irrigation. There is no signifi cant funding gap implicit subsidy for infrastructure consum- for ICT or transport. ers, and that is without taking into account Looking across countries, the dollar sizable subsidies to large industrial customers amount of the funding gap split evenly across that cannot be so readily quantifi ed (Briceño- income groups. Although the largest fi nancing Garmendia, Smits, and Foster 2008). Because gaps relate to capital investment, shortfalls in of the very regressive access to infrastructure funding for operation and maintenance are services in Africa, about 90 percent of those substantial, particularly in fragile states. If who have access to piped water or electricity the infrastructure fi nancing gap is expressed services belong to the richest 60 percent of the as a percentage of GDP, the level of diffi culty population (see fi gure O.9, panel a; Banerjee, involved in closing the gap becomes immedi- Wodon, and others 2008). Thus, better-off ately apparent. The burden associated with the households largely capture any subsidy to resi- infrastructure fi nancing gap is insurmountable dential services. In fact, targeting is so defi cient for fragile states. They would need to spend an that a completely random process for allocat- additional 25 percent of GDP on infrastruc- ing subsidies across the population would per- ture to eliminate their infrastructure defi cits. form three times better at reaching the poor. Relative to the size of economies, by far the The overall funding shortfall for meeting largest fi nancing gaps are in the energy, trans- Africa’s infrastructure needs is given by the port, and water sectors of fragile states (fi gure difference between estimated infrastructure O.7, panel b). spending needs and a potential resources enve- As shown, the size of the funding gap for lope that includes existing spending and the low-income countries in particular is prob- potential effi ciency gains. Even if all these effi - ably more than they could conceivably raise ciency gains could be fully realized, a funding through available funding channels. For this gap of about $31 billion a year would remain particularly challenging group of countries, (table O.5). This gap can be addressed only by additional measures may need to be taken. raising additional fi nance or alternatively by One option is to extend the time horizon adopting lower-cost technologies or less ambi- for the proposed investment program. Simu- tious targets for infrastructure development. lations suggest that low-income countries Looking across sectors, about 60 percent of could achieve the proposed investment targets the funding gap relates to power (fi gure O.7, within a period of 20 years without increas- panel a). The remainder relates to water and ing existing spending envelopes, as long as they 12 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION

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 Effi ciency 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 ineffi ciencies 3.4 1.2 — 2.4 1.0 n.a. 8.0 Gain from tariff cost recovery 2.3 — — 0.1 1.8 n.a. 4.2 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.

Figure O.7 Infrastructure Funding Gap by Sector and Country Type

a. Sector b. Country type 4 30 25 3 20 2 15 % of GDP % of GDP 10 1 5 0 0

power irrigation transport water supply countries resource-richcountries countries nonfragile low- middle-income informationcommunationtechnology and and sanitation fragile low-income income countries capital operation and maintenance

Source: Briceño-Garmendia, Smits, and Foster 2008.

fully exploit effi ciency 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 ineffi ciencies 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 (fi gure O.8). The entire population. focus also varies. In telecommunications, the Africa’s Infrastructure: A Time for Transformation 13

Figure O.8 Scores against Indicators of Institutional Best Practice

a. Utilities b. Transport 100 100 90 90 80 80 70 70 60 60 50 50

percent 40 percent 40 30 30 20 20 10 10 0 0

rail water tele- power ports reform regulation governance communications

Source: Vagliasindi and Nellis 2009. emphasis has been on implementing sector businesses delivers cash fl ows high enough to reform, and in water on improving the gover- fi nance 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 fi ve 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 signifi cant fi nance 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 confi ned to cies. Performance contracts with incentives South Africa; traffi c volumes elsewhere are and independent external audits have become not enough to make such endeavors fi nancially 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 fi nance. 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 effi ciency, 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 benefi cial 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 benefi t 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 effi ciency incumbent unambiguous. Weak regulatory autonomy and Power capacity constraints undermine the credibility of independent regulators. Most African regu- Power generation 34 independent power Few cancellations but Likely to continue, projects provide 3,000 frequent given huge unsatisfi ed latory agencies are embryonic, lacking funding MW of new capacity, renegotiations; power demands and limited and in many cases qualifi ed 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 fi ndings, one can make the fol- contracts cancelled frameworks before completion lowing 10 key recommendations: Transport • Addressing Africa’s infrastructure effi ciency 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 fl agrant ineffi ciencies 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 traffi c, 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 Institutional reform should also go beyond Roads 10 toll-road projects, No cancellations Limited because only • 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- traffi c threshold, almost ous defi ciencies in the budgetary process. all in South Africa Reforms are needed to get full value from Water • 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

• Infrastructure’s social policy needs to be infrastructure ineffi ciencies. As African coun- rethought, placing more emphasis on recov- tries begin to feel the pinch of the global fi nan- ering costs from those who can afford it and cial crisis, and as other sources of funding begin on recasting subsidies to accelerate access. to dry up, measures to improve the effi ciency • Achieving universal access will call for of using existing resources become particularly greater attention to removing barriers that attractive. Such measures provide an addi- prevent the uptake of services and offer- tional internal source of fi nance at a relatively ing practical and attractive second-best low monetary cost. Of course, in some cases, solutions. signifi cant investments may be needed before • Closing Africa’s infrastructure fi nancing effi ciency gains can be captured (for example, gap is critical to the region’s prosperity, and reducing distribution losses in power or water). the global fi nancial crisis has only made In other cases, the economic context of the crisis infrastructure more relevant. may simply increase the political cost of taking such measures, such as raising cost recovery or laying off excess employees. Recommendation 1: Address Africa’s Potential effi ciency gains take a wide variety Infrastructure Effi ciency Gap as a of forms, which are developed in the recom- Pressing Policy Priority mendations that follow. Briefl y, they include The fi ndings presented underscore the magni- the following areas: tude of ineffi ciency with which Africa spends its current infrastructure resources. Of Africa’s • Safeguarding maintenance expenditure to overall infrastructure spending needs of about avoid wasting resources on the repeated $93 billion a year, as much as $17 billion could rehabilitation of existing assets, which could be met simply by using existing resources save $2.6 billion a year in avoidable capital more effectively. expenditure for the roads sector alone Reaping this effi ciency dividend has to be • Reforming institutions to improve the oper- a major policy priority for the region, and ational performance of utilities and other efforts to scale up infrastructure fi nance need service providers that are currently wasting to be made in the context of genuine com- $6 billion a year on ineffi ciencies such as mitments to address effi ciency. Pouring addi- overstaffi ng, undercollection of revenues, tional funding into sectors characterized by and distribution losses high levels of ineffi ciency makes little sense. • Addressing defi ciencies in the public expen- However, postponing increases in fi nance until diture framework, where $3.3 billion a year effi ciency improves is not a valid option: the of infrastructure resources appear to be cost to economic growth and human develop- poorly allocated across sectors and low bud- ment is simply too high. Rather, development get execution prevents $1.8 billion a year of partner efforts to secure additional resources public investment funds from being spent for infrastructure fi nance must be matched by Modernizing administrative and regulatory government efforts to improve their effi ciency • frameworks to reduce bottlenecks that pre- in using such resources. Parallel progress is vent services from being provided effectively needed on both fronts. across existing infrastructure networks and Moreover, investment fi nance is needed in impose substantial costs on infrastructure some cases to allow ineffi ciencies to be captured users (for example, where roads must be rehabilitated before they return to a “maintainable” condition • Reaping the economies of scale and coor- or when meters must be installed to improve dination benefi ts associated with regional revenue collection). These kinds of effi ciency- integration, which in the case of power related investments deserve to be prioritized alone can be as high as $2 billion a year because of the high returns they typically bring. • Securing the highest returns from new The current global financial crisis only infrastructure investments by using them strengthens the motivation for addressing to secure economies of agglomeration and 16 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION

to facilitate the development of productive of multiyear performance-based contracts for activities along key economic corridors roads has further contributed to the effi cacy • Rethinking infrastructure social policy and effi ciency of road maintenance. These to place more emphasis on cost recovery fi ndings illustrate that a combination of fund- from those who can afford to pay, and ing mechanisms, institutional capacity, and redirecting the current $4 billion a year of contractual incentives is needed to overcome subsidies to accelerate access among lower- the maintenance challenge. income groups Donors have traditionally eschewed fund- • Reducing the costs of meeting key infra- ing maintenance, arguing it is more sustainable structure targets by adopting lower-cost for funding directly from country budgets. The technologies that provide reasonable lev- argument is a good one. However, the willing- els of service at a price that is affordable to ness of donors to fund asset rehabilitation both consumers and the government. can create perverse incentives for countries to neglect maintenance, because governments face a choice between raising taxes today to Recommendation 2: Make Greater fi nance maintenance or simply waiting a few Efforts to Safeguard Maintenance years to obtain subsidized donor capital for Spending reconstruction. In low-income, low-capacity The traditional neglect of maintenance expen- environments where maintenance is unlikely diture needs to be reversed by rethinking to be forthcoming, donors may be well advised maintenance as asset preservation. One-third to take this choice explicitly into account in of Africa’s infrastructure assets need rehabilita- project design, rather than simply assume that tion, indicating that historic neglect of main- maintenance will happen. One way of doing tenance is endemic. For fragile states and for so is to choose more capital-intensive, low- rural infrastructure, the share of assets needing maintenance technologies. Even if they rep- rehabilitation is much higher. The shortfall of resent a higher investment cost in the short $0.6 billion a year in road maintenance spend- run, overall life-cycle costs may be lower if ing is costing Africa $2.6 billion a year in avoid- reconstruction can be avoided or postponed. able capital expenditures. In fact, $1 spent on As donors move toward sectorwide budget maintenance can be a savings of approximately support, they will have a greater opportu- $4 to the economy. nity to ensure that maintenance spending is Thus, Africa’s infrastructure fi nancing gap adequately supported in the budget envelope. is not only about raising investment capital; a In any case, as a general principle, the estab- substantial part of it relates to maintenance. Yet lishment of a sound framework for fi nancing maintenance offers one of the highest returns maintenance should be a prerequisite for the to infrastructure spending, so it may be more funding of major capital programs. helpful to think of maintenance as a kind of investment in asset preservation. Recommendation 3: Tackle The road sector shows that maintenance Ineffi ciency through Institutional can be improved through suitable institutional Reform reforms. Since the mid-1990s, the majority of Since the mid-1990s, the institutional agenda African countries have established road funds has broadened and deepened (Vagliasindi and as a means of channeling road user charges Nellis 2009). In the 1990s, the emphasis of to network maintenance. Countries with institutional reform was on sector restructur- road funds do signifi cantly better at raising ing and private participation, transplanting adequate maintenance funds as long as the to Africa experiences from other parts of the fuel levies paid into these funds are set high developing world. This approach yielded dra- enough to provide material fi nancing. More- matic results in telecommunications, but else- over, countries with both road funds and road where the benefi ts were more limited and the agencies do signifi cantly better in safeguard- experiences more problematic. Even so, private ing the quality of their road networks. The use fi nance to African infrastructure came from Africa’s Infrastructure: A Time for Transformation 17 nowhere to provide a fl ow of funds comparable enterprises by their supervisory agencies, in scale to overseas development assistance. whether line ministries or ministries of fi nance. A more nuanced, less dogmatic perspec- Transparency and accountability of state- tive on the private sector has emerged. This owned enterprises depend on solid systems perspective values private fi nancing in mobile of fi nancial management, procurement, and telephony, power generation, and ports, while management information. Today, basic oper- recognizing its limits in roads, rail, power, and ational and fi nancial data on fi rm perfor- water (see table O.6). Even for infrastructure mance are not produced, reported, or acted where the proven appetite for private fi nance on. Without information or, perhaps worse, is very limited, the potential contribution of without action on what information is pro- the private sector to tackling costly manage- duced, better outcomes cannot be expected. ment inefficiencies (undercollected utility Key measures include auditing and publishing revenues, low labor productivity, or neglected fi nancial accounts and using comprehensive road maintenance) remains valuable. Indeed, cost-based accounting systems that allow the the effi ciency gains from such performance functional unbundling of costs and a clearer improvements are themselves a signifi cant sense of cost centers. After this foundation is source of sector fi nance. Moreover, the concept in place, contracting mechanisms can improve of private participation has undergone signifi - performance—within the public sector or with cant expansion. More emphasis has fallen on the private sector. the local (not international) private sector and Public sector performance contracts need on hybrid models that experiment with differ- strong performance incentives. Initial attempts ent ways of allocating responsibilities between to improve African state-owned enterprises public and private partners. through performance contracts with their line Another important way in which the insti- ministry or other supervisory agency were tutional reform agenda has broadened is the minimally effective. Recent efforts in water greater focus on the quality of governance for (Uganda), however, have had a much more enterprises that remain state owned (Vagli- positive effect. The key feature of these con- asindi and Nellis 2009). The recognition that tracts is to incorporate incentives for good the private sector will never be a ubiquitous managerial (and staff) performance and, more service provider has come with the realiza- rarely, sanctions for failure to reach targets. tion that state-owned enterprises are here to Creating effective performance incentives stay. Therefore, it is necessary to recommit to in the public sector can be challenging, mak- the diffi cult process of reforming state-owned ing management contracts with the private enterprises. sector a relevant option. Either expatriate or Renewed efforts on state-owned enterprise local management teams can be contracted reform should favor governance over technical with, each of which offers advantages. Clarity fi xes. Fortunately, better governance of state- about what a contract can and cannot achieve, owned enterprises can improve performance. particularly given its short time horizons, is Past efforts at improving utility management essential. At best, a management contract can focused too heavily on technical issues at the improve performance in a handful of rela- expense of corporate governance and account- tively manageable aspects of effi ciency, such ability. Future state-owned enterprise reforms as revenue collection and labor productivity. It seem justifi ed as long as they focus on deeper cannot solve defi ciencies in the broader insti- institutional issues. Key measures include tutional framework; ideally, these should be greater decision-making autonomy for the addressed beforehand. Nor can a management board of directors, more objective selection contract raise investment fi nance or deliver criteria for senior managers, rigorous disclo- major effects on service quality that require sure of confl icts of interest, and more trans- substantial investments or lengthy gestations. parent, merit-based recruitment processes. In principle, regulation can do much; but in Parallel efforts can strengthen fi nancial practice, regulation has proved diffi cult. Regu- and operational monitoring of state-owned lators have been set up across Africa, precisely 18 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION

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 signifi cant. 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 (fi nally) 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 Institution Reform Agenda in more competitively priced power. Much of the emphasis of recent reforms has Because domestic public spending fi nances 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 fi nance that focuses too narrowly planning, participate in the formulation of the on specifi c project interventions. So donor public budgets, and execute investments. How- resources are best channeled programmatically ever, defi ciencies 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 specifi c 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 refl ect the high profi t 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 fl eet 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 profi t margins for steep increases in budget allocations to the sec- the trucking industry without lowering trans- tor followed water’s identifi cation 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 traffi c 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 fi nance 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 affl ict 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 traffi c, further impedes container refl ects the lack of reliable, well-maintained traffi c. Making the most progress are dry and locomotives, but it also refl ects the absence of liquid bulk exports, where many port facilities clear contractual incentives to service traffi c 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 benefi ts 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 traffi c 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 fi rms 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 effi ciency 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 to fragmentary national boundaries preventing infrastructure involves a high level of trust achievement of scale economies. between countries, not least because of the In ICT, power, ports, and airports, regional implied dependence on neighbors for key collaboration essentially provides scale econ- resources, such as energy and water. For example, omies that reduce the cost of service. Most if regional power trade were pursued fully, 16 African countries are simply too small to African countries would import more than half develop infrastructure cost-effectively on their power needs. A large share of that power their own. In ICTs, regional collaboration in would come from fragile states, such as the continental fi ber-optic submarine cables can Democratic Republic of Congo and Guinea. reduce Internet and international call charges Regional institutions are needed to facili- by half, relative to national reliance on satel- tate agreements and implement compensa- lite communications. In power, 21 countries tion mechanisms. Some countries have more have national power systems below the mini- to gain from regional integration than others mum effi cient scale of a single plant. By shar- do. As long as regional integration provides ing large-scale, cost-effective energy resources a substantial economic dividend, one should across countries, regional trade can reduce be able to design compensation mechanisms electricity costs by $2 billion a year. The traf- that make all participating countries better off. fi c fl ows to most of Africa’s national ports and Benefi t sharing was pioneered through inter- airports are too low to provide the scale econ- national river basin treaties, such as that for omies needed to attract services from major Senegal, and could be applied to other regional international shipping companies and airlines. infrastructure more broadly. Africa has an Regional collaboration in multicountry hubs extensive architecture of regional political and can help overcome this problem. technical bodies, but they have overlapping In road and rail corridors and transbound- memberships, limited technical capacity, and ary river basins, collaborative management of limited enforcement powers. Nor do they cur- these regional public goods reduces the cost. rently have the capacity to implement cross- Many of Africa’s infrastructure assets and natu- border compensation mechanisms. ral resources are regional public goods that cut Moving on regional projects that deliver across national frontiers and can be effectively quick wins is important. Because of the daunt- developed and maintained only through inter- ing investment agenda, better sequencing and national collaboration. Road and rail corridors priority setting for regional projects are needed. need to be managed collaboratively to smooth Political, economic, and spatial approaches transport and trade services to Africa’s 15 land- have all been widely discussed. Regional proj- locked countries, avoiding the extensive border ects range from bilateral cooperation on a delays that slow international road freight to transmission line or border post to vast and 10 kilometers an hour. Africa’s 63 international complex interventions, sometimes with a con- river basins call for cooperative water resource tinental reach. Given the size of the challenges, management and coordinated investments to starting small with projects that deliver tangible increase basin yields of food, power, and other high returns and building incrementally on ini- economic opportunities, while strengthening tial successes may be advisable. environmental sustainability and mitigating Regulatory harmonization needs to go hand the effects of droughts and fl oods. in hand with physical integration. Unless regu- Reaping these benefits poses numerous latory frameworks and administrative proce- institutional challenges. Among them are dures are harmonized to allow the free fl ow of mobilizing political will, developing effective services across national boundaries, physical regional institutions, setting priorities soundly, integration of infrastructure networks will harmonizing regulatory procedures, and facili- be ineffective. Making progress on regulatory tating project preparation and fi nance. reform has a relatively low monetary cost, but Notwithstanding the economic case for it can have a very high return. A good example regional integration, the mobilization of politi- is the Yamoussoukro Decision: opening the cal will faces considerable obstacles. Regional skies for air transportation across Africa, it has 22 AFRICA’S INFRASTRUCTURE: A TIME FOR TRANSFORMATION

led to greater freedom in the negotiation of spatially uncoordinated manner. In Africa— bilateral agreements. too often—the limited infrastructure available Greater efforts are needed to facilitate prepa- is thinly spread out, preventing such synergies ration of complex regional projects, which are from being captured. particularly costly and time-consuming to pre- The urbanization process calls for a regional pare. That is especially true when projects are development perspective on infrastructure that large in relation to the size of the host economy looks at each city and its rural hinterland as and when they essentially depend on fi nancing an integrated economic unit. Africa is urban- from downstream benefi ciaries. They also stretch izing fast, creating change that is predictable the donor fi nancing systems that are more typi- and benefi cial for both urban and rural areas. cally geared toward national investments. Prosperity and density go together, as changes in productivity require agglomeration econo- Recommendation 7: Take a Spatial mies, larger markets, and better connectiv- View of Infrastructure Development ity. Concentration and urbanization trigger Priorities prosperity in both urban and rural areas, and Infrastructure networks are inherently spatial, well-functioning cities facilitate the transition both refl ecting and underpinning the spatial from subsistence agriculture by providing a distribution of economic activity. Infrastruc- large market for rural products and support- ture plays a key role in enabling cities to benefi t ing nonfarm activities. The debate of rural from economies of agglomeration. Transport versus urban development should therefore networks interconnect urban centers with each be replaced by the understanding that rural other and with international trading networks, and urban development are closely linked and providing the basis for exchange between the mutually dependent—and that economic inte- urban and rural economies. Energy, water, and gration of rural and urban areas is the only way ICT all enhance productivity within urban and to produce growth and inclusive development. rural spaces. Therefore, infrastructure plans In urban areas, defi ciencies in land policies and priorities should be strategically informed and planning have become a huge impediment by a clear understanding of the spatial dis- to extending infrastructure services. African tribution of economic activity and potential. cities are growing fast, but with insuffi cient A clear example of this approach is the Spatial infrastructure and poor institutions, most Development Initiative of the New Partnership new settlements are informal and not cov- for Africa’s Development (NEPAD). ered by basic services. Urban planning should The spatial lens is a useful basis for pri- be strengthened to reduce sprawl, enhance oritization of infrastructure investments and densifi cation, prevent development in pre- provides insight into cross-sectoral links. carious environmental zones, and provide Looking at infrastructure through a spatial the appropriate balance between public and lens allows identifi cation of the key bottle- private land to safeguard key trunk networks. necks along various trading corridors, which Property rights must be clearly defi ned so that are typically the highest-return interventions. land markets can function. Cities frequently Cross-sectoral links also become more appar- lack the fi nancial basis to develop the infra- ent through a spatial view, shedding light on structure critical to their success. The local the need for coordinating interventions across tax base, though potentially large, is typically infrastructure sectors and between infrastruc- unexploited, leaving municipalities reliant on ture and client economic sectors. An emerg- central government transfers, which are too ing literature suggests that because of synergy often inadequate or unpredictable. effects, the returns from bundling multiple Large agricultural sectors and rural econo- infrastructure interventions in a particular mies remain central to economic growth and spatial area (Torero and Escobal 2005) or along poverty reduction in Africa. Yet the access of a given spatial corridor (Briceño-Garmendia rural populations to infrastructure is extremely and Foster 2009a, 2009b) are higher than low. Rural roads and irrigation systems are those from making the same investments in a together perhaps the most pressing of rural Africa’s Infrastructure: A Time for Transformation 23 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 in expanding rural access, with one in two 60 rural Africans now in range of a global systems mobile signal. This platform can contribute to 40 agricultural productivity through simple text- 20 message extension services, through bulletins % of households on agricultural market prices and meteoro- 0 Q1 Q2 Q3 Q4 Q5 logical conditions, and as a vehicle for fi nan- budget quintile cial transactions. The possibilities are only just piped water electricity beginning to be explored.

b. Affordability curve Recommendation 8: Rethink 100 Infrastructure Social Policy 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 beyond the 5 percent services is costing Africa $4.2 billion a year in affordability threshold % of households falling forgone revenues. In addition, because of ineq- 0 uitable access to infrastructure services, these 2 4 6 8 10 12 14 16 monthly utility bill (US$) subsidies are highly regressive, largely bypass- ing the poor (fi gure O.9). The underrecovery low-income countries middle-income countries of costs impairs the fi nancial 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 (fi gure 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 affl uent 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 affl uent 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 affl u- 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 benefi t 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 fi scal 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 densifi cation 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 fi ne-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 affl uent. 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- diffi cult 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 effi ciency measures, a substantial infrastruc- India to Argentina and Mexico, have used ture fi nancing gap of $31 billion a year remains. infrastructure-based fi scal stimulus in times of Such a large shortfall looked daunting even economic crisis. If well targeted to addressing before the onset of the global fi nancial 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 fi nance 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 benefi t from porting African infrastructure. Offi cial devel- such a program, the continent does not have opment assistance fl ows almost doubled, from the means to fi nance 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 fl ows 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 fi nancing infrastructure (ODI 2009). within a framework of South-South coop- Any increase in donor fi nance for African eration. Their commitments rose from almost infrastructure should pay particular attention nothing in the early 2000s to fi nance 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 fi nance 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 fi rst 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 fi nance for infrastructure. Given the magni- billion a year could widen further as public tude of the fi nancing gap that these countries budgets are squeezed, external capital fl ows 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 fi scal compres- tries have the potential to raise a signifi cant sion to balance the public books came from amount of local fi nance 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 signifi cant volumes of Calderón, César. 2008. “Infrastructure and infrastructure fi nance, Nigeria being the most Growth in Africa.” Working Paper 3, Africa salient example (Irving and Manroth 2009). Infrastructure Country Diagnostic, World Bank, Washington, DC. However, most of this fi nance takes the form of relatively short-maturity commercial bank Carruthers, Robin, Ranga R. Krishnamani, and 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 Bofi nger, 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.