69087 BACKGROUND PAPER 2 Access, Affordability, and Alternatives: Modern Infrastructure Services in Africa Sudeshna Banerjee, Quentin Wodon, Amadou Diallo, Taras Pushak, Helal Uddin, Clarence Tsimpo, and Vivien Foster FEBRUARY 2008 © 2009 The International Bank for Reconstruction and Development / The World Bank 1818 H Street, NW Washington, DC 20433 USA Telephone: 202-473-1000 Internet: www.worldbank.org E-mail: feedback@worldbank.org All rights reserved A publication of the World Bank. The World Bank 1818 H Sreet, NW Washington, DC 20433 USA The findings, interpretations, and conclusions expressed herein are those of the author(s) and do not necessarily reflect the views of the Executive Directors of the International Bank for Reconstruction and Development / The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Rights and permissions The material in this publication is copyrighted. Copying and/or transmitting portions or all of this work without permission may be a violation of applicable law. The International Bank for Reconstruction and Development / The World Bank encourages dissemination of its work and will normally grant permission to reproduce portions of the work promptly. For permission to photocopy or reprint any part of this work, please send a request with complete information to the Copyright Clearance Center Inc., 222 Rosewood Drive, Danvers, MA 01923 USA; telephone: 978-750-8400; fax: 978-750-4470; Internet: www.copyright.com. All other queries on rights and licenses, including subsidiary rights, should be addressed to the Office of the Publisher, The World Bank, 1818 H Street, NW, Washington, DC 20433 USA; fax: 202-522-2422; e-mail: pubrights@worldbank.org. About AICD This study is a product of the Africa Infrastructure Country Diagnostic (AICD), a project designed to expand the world’s knowledge of physical infrastructure in Africa. AICD will provide a baseline against which future improvements in infrastructure services can be measured, making it possible to monitor the results achieved from donor support. It should also provide a better empirical foundation for prioritizing investments and designing policy reforms in Africa’s infrastructure sectors. AICD is based on an unprecedented effort to collect detailed economic and technical data on African infrastructure. The project has produced a series of reports (such as this one) on public expenditure, spending needs, and sector performance in each of the main infrastructure sectors—energy, information and communication technologies, irrigation, transport, and water and sanitation. Africa’s Infrastructure—A Time for Transformation, published by the World Bank in November 2009, synthesizes the most significant findings of those reports. AICD was commissioned by the Infrastructure Consortium for Africa after the 2005 G-8 summit at Gleneagles, which recognized the importance of scaling up donor finance for infrastructure in support of Africa’s development. The first phase of AICD focused on 24 countries that together account for 85 percent of the gross domestic product, population, and infrastructure aid flows of Sub- Saharan Africa. The countries are: Benin, Burkina Faso, Cape Verde, Cameroon, Chad, Côte d'Ivoire, the Democratic Republic of Congo, Ethiopia, Ghana, Kenya, Lesotho, Madagascar, Malawi, Mozambique, Namibia, Niger, Nigeria, Rwanda, Senegal, South Africa, Sudan, Tanzania, Uganda, and Zambia. Under a second phase of the project, coverage is expanding to include as many other African countries as possible. Consistent with the genesis of the project, the main focus is on the 48 countries south of the Sahara that face the most severe infrastructure challenges. Some components of the study also cover North African countries so as to provide a broader point of reference. Unless otherwise stated, therefore, the term “Africa� will be used throughout this report as a shorthand for “Sub-Saharan Africa.� The World Bank is implementing AICD with the guidance of a steering committee that represents the African Union, the New Partnership for Africa’s Development (NEPAD), Africa’s regional economic communities, the African Development Bank, the Development Bank of Southern Africa, and major infrastructure donors. Financing for AICD is provided by a multidonor trust fund to which the main contributors are the U.K.’s Department for International Development, the Public Private Infrastructure Advisory Facility, Agence Française de Développement, the European Commission, and Germany’s KfW Entwicklungsbank. The Sub-Saharan Africa Transport Policy Program and the Water and Sanitation Program provided technical support on data collection and analysis pertaining to their respective sectors. A group of distinguished peer reviewers from policy-making and academic circles in Africa and beyond reviewed all of the major outputs of the study to ensure the technical quality of the work. The data underlying AICD’s reports, as well as the reports themselves, are available to the public through an interactive Web site, www.infrastructureafrica.org, that allows users to download customized data reports and perform various simulations. Inquiries concerning the availability of data sets should be directed to the editors at the World Bank in Washington, DC. Contents Acknowledgments iv Acronyms and abbreviations iv Summary v 1 The African context 1 2 Building a continental database 8 3 Reaching the goal of universal access to services 13 4 Keeping services affordable 26 5 Alternatives to modern infrastructure services 44 6 Conclusions and policy implications 58 References 60 Annexes Available separately iii Acknowledgments We are grateful to Branko Milanovic, Tito Yepes, Siobhan Murray, Cecilia Briceno-Garmendia, Tarik Chfadi, Christophe Rockmore, Rose Mungai, and Zena Angesom for their guidance and support throughout the project. Grateful thanks are due to Elvira Morella for providing exceptional research assistance in the final processing of the report. We are thankful to Doug Barnes, Judy Baker, Kenneth Simler, and Julian Lampietti, who served as our peer reviewers. They provided many insights and thoughtful comments on an earlier version of this draft. Acronyms and abbreviations AICD Africa Infrastructure Country Diagnostic CPI Consumer Price Index DHS Demographic and Health Survey GDP Gross Domestic Product JMP Joint Monitoring Program LCU Local Currency Unit LSMS Living Standards Measurement Survey LPG Liquid Petroleum Gas MDG Millennium Development Goals MICS Multi-Indicator Cluster Survey ODA Official Development Assistance OECD Organization for Economic Co-operation and Development PPI Private Participation in Infrastructure PPP Purchasing Power Parity UNICEF United Nations Children’s Fund VIP Ventilated Improved Pit WDI World Development Indicators WHO World Health Organization WSS Water Supply and Sanitation iv ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Summary A frica lags well behind other developing regions in access to infrastructure services. Limited gains made in the 1990s continued in the early 2000s, and there is now clear evidence that many countries are failing to expand services fast enough to keep up with rapid demographic growth and even faster urbanization. If present trends prevail, Africa is likely to fall even further behind other developing regions, delaying universal access for a half century or more in many countries. This report reviews recent trends in household access to infrastructure services and associated budgetary expenditures in Africa. It is based on a pooled database that draws upon the entire body of household surveys conducted in Africa in the last 15 years. The database includes 67 Demographic and Health Surveys (DHSs) conducted by the Measure DHS Program of MACRO International in the least-developed countries, as well as related surveys. Covering 32 countries, including 24 at more than one point in time, this collection of survey data provides a sound basis for analyzing historic trends in access to services. The report also draws on 30 household expenditure surveys of various kinds that provide information on the structure of the household budget, and in particular spending on infrastructure services. Our findings on water supply and sanitation are broadly consistent with those of the Joint Monitoring Program (JMP) managed by the United Nations Children’s Fund (UNICEF) and World Health Organization (WHO), although they are based on a different statistical method, and the JMP statistics include all African countries, whereas only a subset in Sub-Saharan Africa is covered here. Shrinking access to modern infrastructure services Recent trends in access suggest Figure 1 Network infrastructure services in Africa, 1990–2005 that coverage of most basic services Percentage of population with access to service (population weighted) in Africa has remained stable or increased slightly since 2000 (figure 1). Trends picked up by the DHS show modest improvements in access to all services between the early and late 1990s to early 2000s. In the case of piped water and flush toilets, coverage levels in urban areas in the early 2000s are significantly below what they were in the early 1990s: 39 percent Source: AICD DHS/MICS Survey Database, 2007. versus 50 percent for piped water, and 27 percent versus 32 percent for flush toilets. v ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA The overall trend is driven largely by declining access in urban areas, while the situation in rural areas has improved. Access to improved water sources has declined across the period in urban areas. Access to improved sanitation has held steady in urban Africa. Access to infrastructure services is more limited in Africa than in any other region of the developing world. Official estimates suggest that electricity is available to little more than 20 percent of Africa’s population, versus 33 percent in South Asia, the next-lowest region. Access to an improved water source is 56 percent (versus 78 percent in East Asia), while access to a piped water connection is just 12 percent. Access to improved sanitation, at 37 percent, is comparable to that in South Asia, but well behind the 50 percent reported for East Asia. Moreover, access to a flush toilet (connecting to a sewer or septic tank) is only 6 percent. Telecommunications is the exception to the general pattern of stasis or decline. In telephone density (landlines and cellular telephones), Africa is somewhat ahead of South Asia, with 64 versus 56 subscribers per thousand people. Landline coverage increased dramatically to reach more than 7 percent of households in the early 2000s, while cellular telephones came from nowhere to reach 10 percent of households today. Except in South Africa, almost all cellular telephones in Africa are first telephones, as opposed to second telephones for households that already have landlines. Coverage rates in urban areas are an order of magnitude higher than those in rural areas (figure 2a). In fact, Africa’s low overall access rates are partly explained by negligible service coverage in rural areas, where the bulk of the population still resides. When broader measures of improved water and sanitation are considered, the discrepancies are still large and stark. Thus, about 63 percent of the urban population has access to an improved water source, compared with about 14 percent of the rural population. Moreover, about 42 percent of the urban population has access to improved sanitation versus about 7 percent of the rural population. Figure 2 Patterns of access to modern infrastructure services in low-income countries of Africa Population-weighted average, percent, latest available year (a) By geographic area (b) By asset quintile Source: AICD DHS/MICS Survey Database, 2007. vi ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Access to modern infrastructure services is almost entirely confined to the upper-income quintiles (figure 2b). In the first three quintiles of the wealth distribution, access to modern infrastructure services is well below 10 percent, access for the fourth quintile is typically 10–40 percent, while access for the richest quintile is typically 30–50 percent. The implication is that around 80 percent of those currently connected to modern infrastructure services are in the top 40 percent of the distribution of wealth. In most countries, moreover, inequality of access has increased over time, suggesting that new connections have tended to go predominantly to more affluent segments of the population. In contrast to the general concentration of service among the wealthy, a handful of countries stand out as having reached significant levels of access to electricity (5–15 percent) among the poorest quintile. They are Gabon (17 percent), Nigeria (10 percent), South Africa (10 percent), Ghana (8 percent), and Republic of Congo (5 percent). It is striking that even among the top quintile, coverage is far from universal and highly variable across countries, ranging from around 20 percent in Chad and Central African Republic to almost 100 percent in Cote d’ Ivoire, Gabon, Namibia, South Africa, and Zimbabwe. That only a minority even of rich households has access to the full suite of modern infrastructure services poses the question of whether access rates are limited by what is locally available. The latter seems to be the case in Africa. Only 10 percent of all households have access to both piped water and electricity. Just 1 percent of households have piped water, electricity, a flush toilet, and a telephone. What is keeping access low? Despite isolated successes, the fact remains that the trendline of service coverage is static or modestly increasing for the region as a whole. A number of explanations can be identified. First, the income and urbanization levels of the country are major drivers of access to modern infrastructure services. Middle-income countries have access rates to piped water, flush toilets and telephone landlines that are three times as high as those found in low income countries, and electricity access rates that are twice as high. More highly urbanized countries have access rates to piped water, flush toilets and telephone landlines that are twice as high as those found in less urbanized countries, and electricity access rates that are three times as high. Relatively few of Africa’s countries are in the middle income, highly urbanized bracket. Second, Africa’s high demographic growth rates provide one explanation for falling levels of coverage. Demographic growth in Africa is 2.2 percent per year (compared with the next-highest rate of 2.0 percent in the Middle East and North Africa). Moreover, urban populations in Africa are growing at 3.6 percent per year (compared with the next-highest rate of 3.1 percent per year in East Asia). The analysis shows that a significant number of African countries are not increasing access rapidly enough to keep up with demographic growth, particularly in urban areas. Indeed, if historic rates of expansion continue, only a handful of countries can be expected to attain universal coverage by the year 2050. vii ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Third, decreasing household size is a second factor that frustrates coverage expansion. There is evidence that the average household size in Africa is falling over time as incomes rise. Thus, the total number of households is actually growing even faster than the total population. (The estimated rates are 3.2 percent per year for households as opposed to 2.5 percent for population.) Thus access needs to expand by 50 percent more to maintain constant coverage rates than if household size remained unchanged. Fourth, even within the group of low income countries, there is a wide diversity of performance with respect to coverage. Countries such as Ethiopia, Kenya, Madagascar, Mali stand out as already having relatively good rates of coverage for some services, in spite of their low levels of income and urbanization. Another set of low income countries stand out as having achieved relatively high growth rates increasing the number of connections by between 5 and 10 percent per year for services such as water and electricity. Successful examples include Burkina Faso, Mali, Chad, Ethiopia and Senegal (water), and Lesotho, Madagascar, and Burkina Faso (electricity). Finally, gaps in the supply of services are just part of the explanation for low access. Millions of Africans living near networked services still lack access to them, either because the services are not affordable or because consumers prefer alternatives. To identify interventions that might be capable of speeding up the rate of expansion of access, we divided the unserved urban population into two groups: (1) individuals who live close to an infrastructure network and could be reached through relatively inexpensive programs to increase service density, and (2) those who live far away from such a network and could be reached only by extending the network. Our results are surprising. Some 70–90 percent of the urban population lives in physical proximity to piped water and electricity networks, even though coverage rates are 20–40 percentage points lower than their proximity would suggest. In other words, many people who live near the network choose not to connect to it. Affordability of infrastructure services These findings suggest that affordability may be a barrier to further expansion of access. Most African households live on very modest budgets and spend more than half of their resources on food. The average African household has a budget of no more than $180 per month; urban households are about $100 per month better off than rural households. Household budgets range from around $50 per month in the lowest quintile to no more than $400 per month in the highest income quintile, except in middle-income countries, where the richest quintile has between $600 and $1,200 per month. Even the most affluent households spend about half of their monthly budget on food—among the poorest that share rises toward 65 percent. Infrastructure spending—particularly on power and transport—weighs heavily on household budgets. Spending on utilities, transport, and rubbish disposal typically absorbs 10–20 percent of the household budget, and this can rise to as much as 40 percent in some countries. Electricity and transport each absorbs 5–10 percent of the household budget in most countries. Spending on viii ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA water is typically no more than 5 percent of the household budget. Spending on telecommunications varies widely across countries. It is not unusual for infrastructure spending to absorb 40 percent of the nonfood budget of the household, and as much as 80 percent in some cases. To test the affordability of utility services priced at a level sufficient to allow the utilities to recover their costs, we calculated the percentage of urban households that would need to spend more than 5 percent of their income to purchase a subsistence level of any given utility service. The finding is that the countries fall into three groups. In most countries, between one- and two- thirds of the urban population would face difficulties in covering the cost of service.1 In eight countries, at least 70 percent of urban households would be unable to afford a monthly expenditure of $10 for water or electricity. Only in the remaining seven countries would most urban households be able to afford a monthly expenditure sufficient to allow the utility to meet its costs. Given the limited means of most African households, service providers will not be able to expand services—or even to sustain them in some cases—based solely on actual and potential revenues from customers. To connect all unserved customers to water or electricity services, the average African government would have to provide a one-time capital subsidy equal to about 1 percent of GDP for 10 years on average. Some governments would have to provide twice that amount. The cost of a recurring consumption subsidy would be slightly higher than the costs of subsidizing new connections. Some of the necessary subsidies are already being paid—but not efficiently. Existing consumption subsidies for electricity and water appear to be poorly targeted in African countries. This is because poor households tend to live in areas without electricity and water service; thus it is impossible for them to benefit from the subsidies. In addition, even where access to the network is available to the poor, many remain unconnected, often because the cost of connecting to the network and purchasing the equipment required for electricity and water use is too high. The traditional “inverted block tariff� structures used in many countries are particularly poorly targeted. First, these tariff structures spread subsidies to all households connected to the network, so that even those who consume high amounts of electricity benefit from a subsidy for the part of their consumption that falls in the lower blocks of the tariff structure. In addition, the lower blocks tend to be too generous in terms of consumption (in kWh per month) to target the poor well. And finally, the differences in unit prices between the various blocks may not be large enough. Nonpayment for infrastructure services is as a major issue, even among affluent households. Among those reporting access to piped water, electricity or telephone services, close to half did not report paying a bill during the month of the service. While nonpayment rates tend to be higher among the poorer segments of the population, 20 percent of the top quintile report not paying for electricity, and 40 percent of the top quintile report not paying for water. 1 By our best estimates, most households in most countries should be able to afford monthly charges of around $2 for any given infrastructure service, but charges of $10 a month are prohibitive for the majority. ix ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Even if subsidies could be better targeted and collection rates improved, the ability of African households to pay for infrastructure services is almost certainly not sufficient to permit providers to expand services without additional capital and operating subsidies. Alternative ways of meeting infrastructure needs With networked infrastructure services unavailable or too costly, millions of African households will continue to resort to traditional alternatives to modern infrastructure services. It is important that policy makers understand these alternatives. In some cases, promoting greater use of second-best alternatives may be a good way to expand access in an affordable way. Some second-best options are viable substitutes for networked services but even access to these second- best alternatives is still comparatively skewed toward the upper-income groups, indicating substantial room for growth in access to these forms of service. Figure 3 Patterns of access to alternative water and sanitation services Population-weighted average, percent, latest available year (a) Water (b) Sanitation Source: AICD DHS/MICS Survey Database, 2007. Among the main alternatives to household connections to piped water are standposts and water vendors, particularly in urban areas, and wells and boreholes, which predominate in rural areas. The coverage of standposts—at 15 percent for our sample and around 25 percent of the urban population—is only slightly higher than the coverage of private piped-water connections. While somewhat more equitably distributed than piped-water connections, public standposts are still regressive in their pattern of incidence. About 37 percent of African households rely on wells and boreholes for their water supply, a share that is relatively constant across the income distribution. Those with no other alternative must resort to surface water of questionable quality—this amounts to 30 percent of the population overall and about 50 percent of the poorest. In a few countries, water vendors play a significant role in urban water supply, supplying around 4 percent of the urban water market; and in Mauritania that share exceeds 30 percent. Interestingly, even though water vendors charge higher unit prices for water, those purchasing water from vendors do not necessarily spend more on buying water than those purchasing water x ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA from the public utility—they simply lower the quantity they consume. In many cases, overall spending levels are similar; where they differ those purchasing from vendors are just as likely to spend more or less per month relative to the clients of the utilities. The overall prevalence of improved latrines (such as VIP, chemical, or SAN PLAT) in Africa, at around 8 percent of the population, is scarcely higher than the prevalence of flush toilets and is equally concentrated in the upper-income segments of society. Several countries stand out as having 30–50 percent of their populations covered by flush toilets or improved latrines. Even in those countries, however, about half of the population relies on traditional pit latrines, by far the most widely used form of sanitation in Africa. In Malawi, Tanzania, and Uganda as much as 80 percent of the population is served by traditional pit latrines. As with boreholes, the share of the population using pit latrines is relatively constant across the income distribution, but, in some countries, a large share of the population lacks even that form of sanitation. In Benin, Burkina Faso, Chad, Niger, and Togo, more than 80 percent of the rural population lacks any form of sanitation. The sharing of water and sanitation facilities among multiple families is common in urban areas. At least 16 percent of urban households share their water supply facilities with other households, while more than 40 percent typically share their toilet facilities. The average African household spends 45–50 minutes per day collecting water from sources outside the household. The time spent collecting water has remained almost unchanged over the last 15 years. Most African households that lack private water connections live within one kilometer of their water source. In the case of urban households, the average distance is estimated to be just over 500 meters, while in the case of rural households the average distance is closer to one kilometer. Some 20 percent of urban households and 30 percent of rural households live more than one kilometer from their water source. The vast majority of the population cooks with traditional solid fuels and relies on kerosene for lighting. For cooking, around 80 percent of the population relies on wood, charcoal, or a substitute. Although reliance on traditional fuels is significantly higher in rural areas (close to 93 percent of households), their use in urban areas remains quite high (more than 70 percent of households in many cases). More than half of the African households dump, burn, or bury their household waste. Only 10 percent of households (but about 30 percent of urban households) have access to an advanced waste collection option such as collection by the government, a private company, or a nongovernmental organization. Conclusions and policy directions Despite the overall decline in African’s access to water and sanitation particularly in the urban areas since 2000, a significant number of countries have succeeded in expanding coverage by an annual average of 5–10 percent, a rate fast enough to make substantial coverage gains xi ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA within a reasonable time frame. Further investigation is warranted to explain what determines their superior performance. The finding that a significant share of the unserved urban population lives close to infrastructure networks but chooses not to connect suggests the need for greater efforts on the demand side—and that extending networks is not a sufficient condition for achieving higher access. The low uptake rate of services in African cities means that the financial and economic return to prior network expansion has been much lower than might be expected, leaving a relatively small customer base to cover the fixed costs of a relatively expensive network. It is therefore necessary, once the phenomenon of low uptake is thoroughly understood, to accompany further expansion with demand-side measures explicitly designed to reduce uptake barriers, such as subsidization of connection charges, which tend to be high relative to household incomes and no doubt play a role in the low uptake of available services. Urban development factors, such as insecure household tenure, may also be playing an important role, discouraging both supply and demand. Low incomes represent an absolute constraint on the rate of expansion of modern services. The average African household has little more than $30 per month to spend on all utilities and transport. Utility bills on the order of $6 per month for a service such as water or power may be affordable for most households in all but the poorest countries, but once bills reach $10 per month they are unaffordable for a substantial share of the population. The fact that most Africans rely either on alternatives to networked infrastructure services or simply do without services altogether has important implications. Given the slow rate of growth in coverage for many services in many countries, this situation is likely to persist for years. For that reason, in addition to focusing on improving the performance and expanding the ambit of formal providers of modern infrastructure services, it is important to consider what might be done to improve the lot of the unserved through alternative services. There is clearly substantial potential for second-best options such as standposts and improved latrines to reach a larger share of the population. While the results reported above provide insights into the nature of household usage of infrastructure services in Africa, they also raise many questions that cannot be immediately answered. Why is the variance in access so high across countries, even within the same income band? Why is the variance in access so high across services, and how is it that a new service such as cellular telephony made such major inroads so quickly? To find answers to many of these questions, it is necessary to dig deeper into the institutional organization and the performance of service providers in each country. Such an analysis is already underway in other components of the Africa Infrastructure Country Diagnostic. When all the work has been completed, it will be possible to revisit the findings of this study and make greater sense of the variations that have been observed. xii 1 The African context Policy makers around the world face the challenge of providing reliable and affordable infrastructure services to their people.2 One billion people do not have access to safe water; two billion people lack electricity and safe sanitation facilities; and three billion have never used a telephone. Most of these people reside in Sub-Saharan Africa (hereinafter Africa) or South Asia (Brook and Smith 2001). A strong network of public infrastructure is a precondition for national and regional economic growth and a channel through which private enterprise invests in developing countries. Infrastructure directly affects productivity and output by enlarging the size of product and labor markets (Prud’homme 2004). Infrastructure supports pro-poor growth by enhancing overall growth, removing barriers that hurt poor people, and encouraging poor people’s participation in the growth process (OECD 2006). And infrastructure stock positively affects growth, while superior quality and quantity of infrastructure reduces income inequality (Calderon and Serven 2004). In the early 1990s, when infrastructure gaps were recognized as obstacles to growth and welfare improvements, it was widely hoped that the private sector would step into the breach by investing in infrastructure. Between 1995 and 2005, the private sector invested almost $37 billion in infrastructure in Africa, according to the World Bank’s database of private participation in infrastructure (PPI), while official donors retreated from infrastructure investments. In recent years, however, the private sector’s appetite for infrastructure investments has declined because of difficulties of recovering costs and several failed and renegotiated infrastructure transactions. The Millennium Development Goals (MDGs)—together with renewed emphasis on the direct and indirect relationships of growth, equity, and infrastructure—have made investments in infrastructure a priority again. Multilateral lending institutions and bilateral donors within the Organization for Economic Co-operation and Development (OECD) have repositioned their infrastructure business, and infrastructure figures prominently in government action plans to reduce poverty and improve growth. Official development assistance (ODA) for infrastructure to Africa has been on the order of $27 billion annually for the past 10 years. In addition to traditional sources of aid, infrastructure has been attracting high volumes of finance from emerging players such as China, India, and oil-exporting nations in the Middle East. The rising tide of financial resources will be needed to bring the continent to parity with other regions of the world. Presently, Africa lags behind all other regions in coverage of water and sanitation (WSS) services. To meet the MDGs, Africa must achieve 75 percent access to improved water supply by 2015 and 66 percent access to improved sanitation by 2015—but the continent is not on target to do so. Africa is the only world region in which the share of people without access to water and sanitation increased between 1990 and 2004—by 23 percent and 30 percent respectively (JMP 2006). 2 In this study, infrastructure refers to economic infrastructure—water supply, sanitation, energy, rubbish disposal, and transport. 1 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Monitoring progress in WSS has been a significant by-product of adoption of the MDGs. WSS is the only infrastructure sector that corresponds directly with one of the goals—“halve the number of people without access to safe drinking water and basic sanitation by 2015.� The Joint Monitoring Program (JMP), sponsored by WHO and UNICEF, is entrusted with tracking access to improved WSS. Despite these recent monitoring efforts, little is known about household demand for infrastructure in Africa. This study aims to remedy this situation by contributing to the knowledge base on infrastructure coverage in Africa. The study documents access trends for infrastructure over time, looks at expenditure trends for a significant sample of countries, and analyzes the distributional incidence of subsidies for a smaller subset of countries. The systematic analysis of household surveys across a broad swathe of countries in Africa should help inform infrastructure policy decisions in several areas. By documenting access trends over time, the household surveys help to identify countries that have been relatively successful in scaling-up infrastructure coverage. By documenting affordability of infrastructure services, the household surveys can make a significant contribution to the debate about cost recovery and subsidy design for infrastructure services. This introductory section provides a brief overview of the main trends. Section 2 explains the methodological approach in greater depth. Section 3 examines the prospects for African countries to reach universal access to modern infrastructure services based on the experience of the last 15 years. Section 4 considers whether modern infrastructure services are affordable to the mass of the population, and considers the cost and efficacy of subsidy measures designed to promote affordability. Section 5 explores the alternatives to modern infrastructure services on which a large segment of the population rely. Coverage of basic infrastructure services in Africa The starting point is to put Africa’s situation in the wider context of the developing world. Official estimates from World Development Indicators (WDI) suggest that that electricity is available to just 20 percent of Africa’s population (versus 33 percent in South Asia, the next-lowest region). Access to an improved water source is 56 percent (versus 78 percent in East Asia), while access to a piped water connection is just 12 percent.3 Access to improved sanitation, at 37 percent, is comparable to that in South Asia, but well behind the 50 percent reported for East Asia. Moreover, access to a flush toilet (which includes both water-borne sewerage and septic tanks) is only 6 percent. The only exception to this pattern is telephone density (fixed plus mobile), where Africa is somewhat ahead of South Asia; with 64 versus 56 subscribers per thousand people. We find similar results using the methodology described in the next chapter. The unweighted averages of infrastructure coverage4 levels indicate that Africa lags far behind other regions in the provision of modern infrastructure services. Only South Asia comes close at the current levels. The results presented in figure 1.1 only capture the poorer countries5; it consequently underestimates the coverage in all the regions. Even among poorer countries, the coverage rates in Africa 3 The Joint Monitoring Program (JMP) of the United Nations Children’s Fund estimated that 56 percent of Africans had “access to improved water� in 2004. The comparable figure using AICD data during the period 2000–05 is 32 percent. Similarly, the JMP estimated that 37 percent of Africans had “access to improved sanitation� in 2004. The corresponding AICD figure is 18 percent. The discrepancy is explored in chapter 2. 4 The unweighted averages are simple averages—not weighted by population. 5 DHS has a mandate to collect data from the poorest countries in the world. 2 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA are much lower; the difference is particularly striking in electricity and landlines where some regions have more than three-quarter coverage. Recent access trends suggest that while coverage of some basic services in Africa have improved slightly over the last decade, this has not been the case for others. The decline in infrastructure service provision has primarily been in the urban areas and the rural Africa has experienced an upsurge in all modern infrastructure services. Figure 1.1 Comparison of network infrastructure services in Africa with other regions Percentage of population with access to service (unweighted) Piped water Flush toilet Electricity Landline telephones Source: AICD DHS/MICS Survey Database, 2007. Note: The results presented here capture only the poorer countries because the DHS surveys on which they are based have a mandate to collect data from the poorest countries in the world. The figure thus underestimates coverage in all regions by excluding richer nations. Historic trends picked up through the DHSs show modest improvements in coverage of electricity and flush toilets, but by only a few percentage points of the population, while coverage of piped water has declined slightly. Access to improved water sources has also remained stable across the period at 32 percent, while access to improved sanitation has increased from 16 percent to 18 percent in the past decade. The exception to this pattern of limited progress or decline is telecommunications, where not only landline coverage increased, but in addition cellular telephones coverage came from nowhere to reach 10 percent of households as of today. It is interesting that (with the exception of South Africa) almost all 3 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA cellular telephones in Africa are first telephones, as opposed to second telephones for households that already have landlines. Table 1.1 Coverage of network infrastructure services in Africa Population weighted averages, percent Piped water Electricity Flush toilet Landline telephone Improved water Improved sanitation 1996 1996 1990 – 2001 1990 1996 2001 1990 1996 2001 1990 – 2001 1990 1996 2001 1990 1996– 2001 –95 2000 –05 –95 –2000 –05 –95 –2000 –05 –95 2000 –05 –95 –2000 –05 –95 2000 –05 National 18 17 17 23 28 31 9 9 10 6 5 7 32 32 32 15 16 18 Urban 50 43 39 72 73 71 32 29 27 18 16 19 82 76 64 42 41 41 Rural 4 4 4 6 10 13 1 1 2 1 1 2 14 14 16 5 6 8 Note: The numbers underlying these data can be found in the cross-country annex. Source: AICD DHS/MICS Survey Database, 2007. Coverage rates in urban areas are an order of magnitude higher than those in rural areas. In fact, Africa’s low overall access rates are partly explained by negligible service coverage in rural areas, where the bulk of the population still resides. Whereas just 12 percent of the rural population has access to electricity and 3 percent to cellular telephones, and less than 5 percent have piped water, a flush toilet, or a landline, the corresponding figures for the urban population are 71 percent with access to power, 22 percent to cellular telephones, 38 percent to piped water, 28 percent to flush toilets, and 20 percent to a landline telephone. When broader measures of improved water and sanitation are considered, the discrepancies are still large. Thus, around 63 percent of the urban population has access to an improved water source versus around 14 percent of the rural population. Moreover, around 42 percent of the urban population has access to improved sanitation versus around 7 percent of the rural population. In Zimbabwe, 93 percent of urban residents have access to piped water, but rural coverage is only 4 percent, a stark example of the urban–rural divide. In Benin, Cote d’Ivoire, Gabon, Lesotho, Senegal, South Africa, Togo, and Zimbabwe, more than half of the urban population use piped supply to meet their drinking water needs; the numbers in rural areas are far lower. Africa’s national capitals and other major cities have made faster gains in expanding infrastructure coverage than have small towns and rural areas.6 Benin, Cameroon, Ethiopia, Gabon, Ghana, Guinea, Madagascar, Senegal, South Africa, and Zimbabwe, about three-fourths of the residents in capitals and other major cities have electricity. More than half of the population in the large cities of Ethiopia, Kenya and South Africa have landline telephones. In Benin, Senegal, Zimbabwe, and South Africa, more than 90 percent of large-city households are covered by piped water supply. More than three-quarters of large- city dwellers in Senegal, South Africa, and Zimbabwe have modern sewerage systems to meet their sanitation needs. Namibia’s capital Windhoek and other large cities are exceptional in their coverage—96 percent of households had piped water, 97 percent had a flush toilet, and 88 percent had electricity. Several other trends are worth noting. WSS coverage in urban areas declined in the past decade (table 1.1). In the mid-1990s, some 43 percent of urban Africans had piped supply, but by 2005, the coverage 6 We could not always distinguish capital cities from other major cities. Some surveys permit such a disaggregation, but most merge the capital city and other major cities into one category. Consequently, in our discussion of infrastructure services, the category “large cities� includes capital and other major cities. 4 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA had declined to 39 percent. Similarly, 27 percent of Urban Africans have flush toilet in the early part of this decade compared to 29 percent in late 1990s. The fact that the national coverage rate has not declined to the same extent is due to a higher proportion of the population living in urban areas over time. Electricity coverage has remained stable in urban areas and increased in the rural areas in the last 10 years. The coverage of telephone services has dramatically improved in the past 15 years, with one- quarter of urban Africans having either a cell phone or a landline. Given that all of the modern infrastructure services present major coverage gaps, it is relevant to ask whether there is a privileged minority of households that has access to all of these services, or whether different households have access to different services depending on what is locally available. The latter seems to be the reality in Africa. Only 10 percent of households have access to both piped water and electricity based on the estimates using PPP method. As additional services are added the percentage of households with access to all of them falls dramatically, down to just 1 percent of households that have piped water, electricity, a flush toilet, and a telephone (table 1.2b). In the rural areas, negligible proportion of the population has three or four modern infrastructure services (table 1.2a). Even among the rich, only a minority of households have access to the full suite of modern infrastructure services. Indeed, in a substantial number of the countries studied, more than 80 percent of the population does not have access to any of the modern infrastructure services. Table 1.2 Coverage of combinations of network infrastructure services in Africa a. Population weighted method, latest available year National Rural Urban Q1 Q2 Q3 Q4 Q5 Coverage of any one service 33 15 76 4 17 23 44 78 Coverage of any two services 17 4 47 0 2 7 19 56 Coverage of any three services 9 1 28 0 0 3 11 32 Coverage of any four services 4 0 12 0 0 1 4 16 b. Purchasing power parity method, 2002 Piped water, piped water, piped water, Piped water Electricity & Flush toilet flush toilet, Telephone Electricity, Electricity, flush toilet telephone telephone telephone Electricity Electricity Electricity and flush services All four water toilet and Total 23.6 20.0 12.1 7.2 10.7 6.0 2.1 4.0 1.6 5.5 1.1 Rural 8.8 6.1 1.5 1.1 2.1 0.5 0.2 0.3 0.1 0.2 0.1 Urban 61.4 50.2 34.4 20.2 32.6 19.5 7.5 13.1 6.2 16.6 4.4 Source: AICD Expenditure Survey Database, 2007. Income and modern infrastructure coverage are positively related in Africa (figure 1.2), as elsewhere (Komives, Whittington, and Wu 1999). Richer segments of the population have broader access to modern infrastructure services in their home or within a short distance. In the coverage ladder for water supply, rising income is associated with piped water and public standposts and a declining dependence on wells, boreholes, and surface water. In sanitation, the use of flush toilets and so-called VIP/chemical latrines is negligible among the bottom 40 percent of Africa’s population. The number of households with no sanitation facility declines steeply with increasing income, with household dependence on the traditional pit latrine increasing up to the third quintile before declining. Among the poorest 20 percent of Africa’s 5 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA population, only 4 percent use electricity for lighting purposes and 90 percent depend on “dirty fuels,� chiefly wood or charcoal, for cooking. Electricity use increases with rising income. The landline and cell- phone coverage in the highest quintile is almost similar at about 30 percent. The cell phone coverage in the lower quintiles is more evenly distributed with 2 percent of the lowest quintile in African using cell- phones. Coverage of landline telephones sharply increases above the fourth quintile with negligible coverage in the bottom two quintiles. Access to modern infrastructure services is highly concentrated in the upper-income quintiles. In the first three quintiles of the wealth distribution, access to modern infrastructure services is well below 10 percent, access for the fourth quintile is typically 10–20 percent, while access for the richest quintile is typically 30–50 percent. The implication is that around 80 percent of those currently connected to modern infrastructure services are in the top 40 percent of the distribution of wealth. Nevertheless, it is striking that even among the top quintile coverage is far from universal and highly variable across countries, ranging from around 20 percent in Chad and Central African Republic to almost 100 percent in Cote d’Ivoire, Gabon, Namibia, South Africa, and Zimbabwe. In addition, a handful of countries stand out as having reached significant levels of access to electricity (5–15 percent) among the poorest quintile. They are Gabon (17 percent), Nigeria (10 percent), South Africa (10 percent), Ghana (8 percent), and Republic of Congo (5 percent). Further analysis shows that in the vast majority of cases the distribution of infrastructure access is subject to even greater inequality than the distribution of income, and hence contributes to exacerbating inequalities in society as a whole (Diallo and Wodon, 2005). Furthermore, analysis of the distribution of new connections that have resulted from service expansion in recent years shows that these are also more inequitably distributed than income. It appears, therefore, that the benefits of access and access expansion tend to accrue to the better-off., tending to exacerbate inequalities This may be because current access rates remain low even among the wealthier segments of the population, so that this is where utilities initially concentrate their expansion efforts. The data on infrastructure coverage poses the challenge of designing home-grown solutions that can work in Africa in order to improve access much faster than is currently the case. To identify such solutions, we must first understand the interplay of supply and demand as they relate to infrastructure services. In many cases, networks are limited, denying access to many. But even if they are expanded, how many people can afford to pay for service? And if they cannot pay, how can network services be expanded without recourse to investment and consumption subsidies? In the meantime, what second-best solutions might exist between the unacceptable status quo and the distant goal of universal coverage? Should service levels and quality be seen as a continuum from which households might choose based on their means? We tackle these questions in the rest of the report. 6 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 1.2 Infrastructure coverage and income Population-weighted average, percent, latest available year (a) Water supply (b) Sanitation (c) Energy (d) Refuse collection (e) Landline telephone (f) Cellular telephone Note: The numbers underlying these graphics can be found in the cross-country annex. Source: AICD DHS/MICS Survey Database, 2007, AICD Expenditure Survey Database, 2007. 7 2 Building a continental database Household surveys have long been used to explore poverty, inequality, and the welfare of vulnerable groups. Their use in understanding access to and affordability of infrastructure is more recent but already essential. Household surveys are the only quantitative instrument that can establish relationships between the use of infrastructure services (nuanced with socioeconomic variables) and government subsidy policies (Lobo, Foster, and Halpern 2000). The inclusion of several infrastructure-related questions in the recent Demographic and Health Survey (DHS) series, conducted by the MeasureDHS Program of MACRO International in the least developed countries, is a significant boost to understanding access and use of services. With support from the World Bank and other international agencies, other income/expenditure surveys have been undertaken in the past decade that allow investigators to study patterns of household spending on infrastructure services. Although the coverage of the infrastructure sectors in these surveys is not always as comprehensive as might be wished, the scope and depth of coverage of infrastructure issues has gradually improved over time and by now represents a substantial body of knowledge—as reflected in this report. There have been previous attempts to use household surveys to understand access trends and affordability patterns for infrastructure from a cross-country perspective, particularly in Latin America and Eastern Europe. A global infrastructure study by Komives, Whittington, and Wu (1999) was one of the first attempts to use the World Bank’s living standards measurement surveys (LSMS) for 15 countries to present access trends and to evaluate the relationship between access to infrastructure and household income. Only three countries from Sub-Saharan Africa were represented in the study, however. A more recent evaluation by Estache and Wodon (2007) presents evidence on infrastructure and access and affordability trends for 10 African countries. Building on these earlier efforts, this study presents a more comprehensive picture of access and spending on infrastructure in Africa. To document access, we created a cross-national meta-database based on the DHSs and Multi- Indicator Cluster Surveys (MICS). The new database is called the AICD DHS/MICS Survey Database and is referred to throughout this report. The DHSs collect comparable information across countries on health, HIV, and nutrition. Because they are conducted every few years, it is possible to track similar indicators over time. Thirty countries in Africa have had at least one DHS conducted since 1990; 22 are covered by at least two DHS data points between 1990 and 2005. Togo, the Central African Republic, Comoros, the Republic of Congo, Gabon, and Lesotho are among those countries with only one data point during this period. In a few countries such as the Democratic Republic of the Congo, Lesotho, and Sudan where data is not available at all or only for a year, we use MICS as a substitute. Implemented by UNICEF, MICS was designed to report on the health of women and children. These surveys are closer to DHS than other survey series with respect to sampling strategy but it only covers WSS questions. Nevertheless, because of problems of comparability, we have used MICS results sparingly (annex table A1.1.1). The DHSs implemented in Africa since 1990 make it possible to analyze the following service categories: water supply, sanitation, electricity, fuels for cooking, and landlines. The DHSs are conducted in phases; there have been five phases since 1990. New questions are added in each phase, and the 8 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA questions posed in each phase are relatively well harmonized across surveys. Questions on water supply, sanitation, and cooking fuels are available since 1990, while questions on electricity, rubbish disposal, and cell phones are more recent. In fact, even now, only five countries in the sample include questions on cell phone use. This makes it difficult to track improvements over time in cell phone use. The poor coverage of infrastructure modules in the DHS is not surprising, as the objective of the surveys is to gather information on infrastructure that has direct relevance to health and nutrition. The DHSs do not collect any household income or spending information. So to establish a correlation between access to infrastructure and income, we had to construct a household welfare measure. Given the high correlation between income and ownership of assets, such a measure can be assembled using information on ownership of assets. One problem is that the asset variables available for the construction of an index vary across countries and time periods. Thus, an asset index that was completely consistent would have to ignore data available for a significant number of surveys. Following Diallo and Wodon (2005), the asset index is constructed using principal components analysis based on the maximum amount of asset information available for each country. Typically, these variables are housing attributes and use of water, electricity, and other infrastructure.7 Using the asset index, we created ranges of assets owned and asset quintiles. To document the affordability of infrastructure we used expenditure surveys. Known by different names in different countries (annex table A1.1.2), these surveys are carried out by country governments to reflect local nuances and priorities. Therefore their infrastructure modules often are not harmonized or comparable. Nevertheless, these surveys, most modeled after the LSMS, provide a wealth of information on use of and payment for infrastructure services, in addition to providing data on assets and expenditure patterns of households. These surveys contain information on coverage of rubbish disposal and cell phone services for a higher number of countries than the DHS. For this purpose, the expenditure surveys are the primary source of information for understanding rubbish disposal and cell phone use in Africa. They make it possible to draw inferences about spending patterns and the affordability of infrastructure for people at different income levels. We mined expenditure surveys for 30 African countries covering the period between 1997 and 2005. The resulting database is called the AICD Expenditure Survey Database and is referred to throughout this report. Together, the two databases—the AICD DHS/MICS Survey Database and the AICD Expenditure Survey Database—cover household surveys from 39 African countries, of which Gabon and South Africa are upper-middle-income; Angola, Cameroon, Cape Verde, the Republic of Congo, Lesotho, Morocco, and Namibia are lower-middle-income; and the rest are low-income. All the countries included in the databases are in Sub-Saharan Africa, except that the AICD Expenditure Survey Database also covers Morocco. The combined sample size of the pooled data is between 91,823 households in the early 1990s and 206,625 households in the early 2000s for the DHS/MICS Survey Database, while for the Expenditure Survey Database the total size of the combined sample is 267,711 households. Summary characteristics for the households surveyed can be found in annex table A1.1.3. 7 The asset index constructed for this report includes the following variables (the available variables differ by country): source of drinking water, type of toilet facility, type of main floor material, has electricity, has radio, has TV, has refrigerator, has bicycle, has car/motorcycle, has livestock, has farmland/other land, number of persons sleeping per room, has car/truck, has telephone, type of cooking fuel, has bed net for sleeping, shares toilet with other households/individuals, time to reach water source. 9 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA The coverage figures reported throughout this report are population weighted. In other words, they report population’s coverage of any specific infrastructure service. The expenditure figures are household weighted because infrastructure payments are usually made at the household level. Two methods are used to adequately understand the pan-African coverage of infrastructure services: population-weighted averages and averages based on purchasing power parity (PPP). The pan-African coverage figures using the former method has been primarily reported in this report. • Population-weighted averages. Average coverage for each service, by location and quintile, is weighted by the population of that category in the survey year. The drawback of this approach is that it assumes that the poorest quintile in one country is the same as in another. The advantage is that each country gets as much weight as its population and is adequately representative. • PPP averages. We transformed reported expenditures into PPP terms8 using 2002 dollars to ensure comparability among different surveys and to present a pooled pan-African picture. The PPP figures were computed by converting expenditures in local currency units (LCU) into 2002 LCUs using the local consumer price index (CPI). Thereafter, the 2002 LCUs were transformed into 2002 international dollars using the PPP conversion factor available in World Development Indicators (World Bank 2007). In this way, all Africans could be divided into five income quintiles. In practice, this means that all households in South Africa can fall into the richer pan-African quintiles, while all those in Niger can fall into the poorer pan-African quintiles. Self-reported data can present problems. Data on infrastructure spending can be fraught with inaccuracies, as the data are self-reported by the surveyed households. The questions are based on actual payments, rather than billed amounts, and it is difficult to distinguish between arrears and current payments. Furthermore, the surveys do not ask questions on metering, so it cannot be known if the household’s payments are based on its consumption or on some other form of assessment based on property values, number of rooms, or diameter of pipe. The wording of the survey questions can also be confusing. The survey may ask respondents to declare the payment they made “last month,� even though payments are not due monthly in many cases. Merging household data with utility data could provide a comprehensive picture of consumption and spending dynamics among different consumer groups. But we did not attempt to do this here since our study covers too many countries and survey years. In addition, the common field to merge the two datasets is usually the address of the household, which is often not very well documented in Africa. Therefore, merging utility data with household data is not common in researching infrastructure in Africa, although it is prevalent in other regions (Lampietti et al, 2007). To achieve comparability of different surveys across different time periods, variables had to be aggregated. The infrastructure categories used in the DHSs and expenditure surveys vary widely. Our solution was to standardize the infrastructure and socioeconomic variables relevant to this study into 8 The reference year has been selected as 2002. The reason for this selection is manifold. It is necessary that the reference year be a year with maximum number of surveys yet is close to the present day. There are 5 surveys each for 2000 and 2002 and finally, 2002 was selected as the reference year as it is closer to the present year. In addition, it allows this exercise to be comparable to the inequality study in Africa recently undertaken by the World Bank (Milanovic, 2003). The income and expenditure variables in local currency units across different years and different countries were converted from local currency to U.S. dollars using the official exchange rate in the survey year and then adjusted to the year 2002 using US CPI figures. This made it possible to compare expenditure among countries. 10 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA categories that allow comparison across surveys. Although many nuances are lost in the resulting aggregation, it is the only approach that allows cross-country comparability. The standard categories of the infrastructure variables (see annex table A1.1.4), were applied to both the DHSs and the expenditure surveys. While standardizing categories, we also used the disaggregated sources in WSS to compute access to improved and unimproved sources—an MDG indicator. “Access to safe drinking water� is defined in the MDG as the “percentage of the population using improved sources� and is monitored by the JMP, using household survey data. However, infrastructure categories are added and changed in each survey phase, which makes it difficult to track the same category over time. Therefore, in addition to the improved/unimproved categorization adopted in the JMP, this study proposes a categorization based on modern, intermediate, and basic service options (Komives, Whittington, and Wu 1999). Table 2.1 Definition of access and standardized categories of infrastructure services Main source of water supply JMP category AICD category 1 AICD category 2 Piped water into dwelling or yard Improved Improved Modern Public tap or communal standpipe Improved Improved Intermediate Wells or boreholes, hand pumps, or rainwater Improved/Unimproved Unimproved Intermediate Surface water (e.g. lake, river, pond, dam, spring) Unimproved Unimproved Basic Vendors or tanker trucks Unimproved Unimproved Basic Others (e.g., bottled water) Unimproved Unimproved Basic Main source of lighting/cooking Electricity Modern LPG or natural gas Modern Kerosene or paraffin or petrol or oil Intermediate Wood or charcoal Basic Crop residue or animal dung or leaves Basic Other Basic Toilet facility Flush toilet to network or septic tank Improved Improved Modern VIP latrine, San Plat, or chemical toilet Improved Improved Intermediate Traditional pit latrine Improved/Unimproved Unimproved Intermediate Bucket or other container Unimproved Unimproved Basic Other Unimproved Unimproved Basic No facility, nature, or bush Unimproved Unimproved Basic Rubbish disposal Collected from rubbish bin by government, private firm, or NGO Modern Rubbish pit Intermediate Rubbish heap Intermediate Thrown away, burned, buried, or dumped Basic Other Basic Source: JMP 2006 and authors. The findings presented here are broadly consistent with those of JMP, although they are based on a different aggregation method. There is no reason to expect our results to coincide with the trends in progress toward the Millennium Development Goals reported by the JMP because the methodology underlying the two sets of numbers differs significantly (table 2.1). First, the JMP statistics include all 11 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA African countries, whereas only a subset is covered here. Second, the JMP statistics are based on a survey of surveys (including assessment questionnaires sent to UNICEF field representatives), whereas our results are based solely on DHS data. Third, JMP statistics apply some standardized parameters in order to be able to separate protected and unprotected wells/boreholes to estimate ‘improved water’, and to determine to what extent traditional pit latrines can be considered ‘improved sanitation’. Our analysis reports only what can be directly supported from the DHS data. Owing to these methodological differences, there is no reason for JMP and AICD figures to be exactly the same, however, they should not be too far apart either particularly in piped supply. For instance, the urban and rural access to piped water is reported to be 40 percent and 4 percent respectively in JMP; the corresponding numbers in this study 39 percent and 4 percent. A detailed comparison of both sets of estimates can be found in annex table A.1.1.5. Information on the quality of service provision is negligible. In Eastern Europe and Central Asia, household surveys in some countries report the quality of service delivery. Elsewhere, however, surveys tell us whether services are available and affordable, but not whether they are reliable or responsive to consumer needs. Sometimes, spending patterns reflect reliability problems. For example, households and businesses incur tremendous losses from unreliable and infrequent electrical service, often obliging them to spend funds for alternative sources of energy. Or households connected to piped water supply may receive water for just a few hours a day, forcing them to spend on alternatives. Such information, where available, has significant policy implications. 12 3 Reaching the goal of universal access to services The introductory section highlighted the overall patterns of access to modern infrastructure services in Africa, including the low levels of service coverage to be found on the continent, as well as the large divergence between urban and rural access rates and across the socioeconomic spectrum. A key finding of that section was the relatively stagnant trend in access expansion; particularly since the year 2000. This section looks beyond the aggregate regional trend to uncover the diversity of experiences across different African countries and examines what this can tell us about the feasibility of attaining universal service coverage on the continent. The detailed country-by-country numbers underlying the analysis presented here can be found in Part II of the Cross-Country Annex Volume. The overall profile for each country can be found in the Country Annex Volume. Making sense of the dispersion in current access levels The broad continental trends reported in chapter 1 do not adequately convey the dispersion of experiences that exists across Africa. Figure 3.1 illustrates that the vast majority of countries have coverage rates of less than 10 percent for piped water, flush toilets, and telephone landlines. However, with respect to electricity, there is a much wider variation of coverage levels across the whole range from 0 to 100 percent; even if the modal coverage lies in the 10 to 40 percent range. What might explain this divergence of coverage across countries? It is known that coverage of modern Figure 3.1 Dispersion in current access levels across Africa infrastructure services is strongly correlated with income and urbanization. Higher incomes make services more affordable, while the greater population densities associated with urbanization help to reduce the cost of expanding services. Most of the countries in the sample are low-income countries with GDP per capita below US$1,000 per year. However, there are also a number of middle-income countries, including Cape Verde, Gabon, Lesotho, Source: AICD DHS/MICS Survey Database. Namibia, and South Africa. The degree of urbanization varies widely in Africa—from 12 percent in Uganda to 80 percent in Gabon—with the average around 35 percent. As illustrated in table 3.1 below, this pattern is clearly visible across Africa. Access to electricity is twice as high in middle income countries as in low income countries, while for landlines the rate is three times as high, and for piped water and flush toilets the difference is four times. However, there are important exceptions to this general pattern. For example, countries such as Nigeria, Uganda, and Tanzania have a relatively high income but low infrastructure coverage, whereas countries such as Zambia and Zimbabwe have a relatively low income but relatively high infrastructure coverage. With 13 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA respect to urbanization, coverage rates for all services are two to three times as high in countries with high levels of urbanization as those with low levels of urbanization. Table 3.1 Patterns of access to modern infrastructure services Latest available year Population weighted Total Rural Urban Q1 Q2 Q3 Q4 Q5 Piped water By country income Middle 44 17 69 2 16 47 69 86 Low 11 3 32 0 1 3 13 41 By urbanization level Low 7 2 33 0 0 1 5 31 Medium 17 4 42 0 2 3 20 64 High 21 7 39 1 6 17 32 49 Flush toilet By country income Middle 33 4 58 0 2 22 61 78 Low 7 2 22 0 1 2 6 29 By urbanization level Low 3 1 13 0 0 1 2 13 Medium 7 1 19 0 1 1 7 27 High 19 5 38 0 1 9 25 61 Electricity By country income Middle 55 27 81 7 28 59 86 97 Low 26 11 69 3 12 15 32 68 By urbanization level Low 11 3 56 0 0 1 4 52 Medium 15 3 48 0 1 2 11 60 High 52 30 83 8 32 45 79 94 Landline By country income telephones Middle 19 4 32 0 1 6 28 59 Low 6 2 17 0 1 1 3 25 By urbanization level Low 5 2 24 0 0 1 2 24 Medium 6 1 18 0 0 1 4 25 High 10 2 19 0 1 3 11 34 Source: AICD DHS/MICS Survey Database, 2007. Not only do these patterns hold overall, but they also hold across urban and rural service segments, and across the different quintiles of the distribution of expenditure. Thus, in more highly urbanized countries even the rural population is substantially better off in terms of coverage. Nevertheless, even in middle income and urbanized countries the benefits in terms of access are largely confined to the top three quintiles of the distribution, with the bottom two quintiles seeing relatively little benefit. Even within higher and lower income or urbanization groups, there is still significant variance in coverage. The existence of outliers within each respective income and urbanization category suggests that 14 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA other factors (for example, linked to the organization of the sector) may also be exerting an influence that leads a country to over or under-perform relative to its peer group. To make this visible, in figure 3.2 countries are ordered according to income level and urbanization and the size of the bubble represents the coverage rate for the different infrastructure services. As might be expected, there is a strong correlation between income and urbanization with most countries lining-up along a 45 degree line. The one exception is Lesotho, which has a much higher level of income than is typical for countries at similarly low rates of urbanization. Given the strong correlation between income, urbanization and service coverage, one would generally expect countries in the southwest quadrant of the graph to have smaller bubbles than countries in the northeast quadrant of the graphs. Countries that do not follow this trend are worthy of note and can thus be identified as outliers that are doing comparatively well (or poorly) in access terms given their level of income and urbanization. Some interesting patterns emerge from these figures. Cameroon and Ghana are the two highest income and most urbanized countries in the sample at the far northeast of the graphs. These countries present relatively small levels of coverage (bubbles) for piped water, flush toilets and landlines, suggesting under-performance with respect to these services. Nevertheless, for electricity their coverage rates are close to what might be expected. Occupying a central position in the graph, Senegal stands out as having levels of coverage that are relatively high and compare favorably with those of peers at similar (and even greater) levels of income and urbanization. Sitting just to the left of Senegal on the 50 percent urbanization line, Nigeria stands out as having low levels of coverage relative to comparable peers for all services except electricity. Immediately underneath, Benin stands out as being a strong performer on piped water access, though its performance on other services is less remarkable. Immediately beneath again, Zambia performs reasonably well on access to piped water and flush toilet, but its performance on access to electricity and landlines is more lack-luster. Turning to the southwestern quadrant of the graphs, Kenya stands out as being consistently the strongest performer in this group, with coverage levels substantially above its close peers across all the services. Otherwise, the countries in the southwestern quadrant perform uniformly poorly with respect to piped water and flush toilet coverage. However, with respect to electricity, countries such as Ethiopia and Madagascar do relatively well given their low income and urbanization levels. The same can be said for countries such as Ethiopia, Malawi and Tanzania with respect to their level of landline coverage. Mali also performs relatively well on water, sanitation and electricity. Occupying an isolated position in the southeast quadrant, Lesotho performs poorly for its income level on coverage of all services, with the notable exception of landline telephones. 15 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 3.2 Dispersion of coverage rates by urbanization and income categories (a) Piped water (b) Flush toilet (c) Electricity (d) Landline telephone Source: AICD DHS/MICS Survey Database, 2007. How far away is universal access? Beyond this static picture of where things stand today, it is relevant to ask how rapidly different countries are moving towards the ultimate goal of universal access to modern infrastructure services. Once again, there is a wide dispersion, both across countries and services. The overall average annual growth rates of population covered by the different services across the continent is 5 percent for electricity, 1 percent for piped water, 7 percent for flush toilet, and 12 percent for landline telephones during the period 1996–2005 (figure 3.3) It is striking that for piped water and flush toilets, around a quarter of countries do not show any evidence of positive growth, while a further third of the countries report modest growth rates of 0–4 percent per year. The strongest performers in terms of service expansion are Benin, Burkina Faso, Chad, Ethiopia, Mali and Senegal, all showing growth rates of 4–8 percent per year. A significant minority of countries are expanding flush toilet service at a rate in excess of 12 percent per annum. These are in fact a subset of those countries that are performing well with respect to piped water service expansion: Burkina Faso, Chad, Ethiopia and Mali. However, this growth is taking place from a very tiny base, and hence does not amount to a great deal in absolute terms. 16 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA The rate of expansion of electricity services is somewhat more encouraging, with almost half of the countries reporting average annual growth rates in the 4-8 percent bracket. The names of the fast expanding countries, once again, shows considerable overlap with that of the countries registering rapid expansion of piped water service: Benin, Burkina Faso, Chad, Lesotho, Madagascar, Mali, Senegal and Tanzania. The most rapid rates of coverage expansion are for landline service, where about half of the countries are expanding at over 12 percent per year, albeit from a very low base. The list of high performing countries is somewhat different in this case: Ethiopia, Ghana, Guinea, Kenya, Madagascar and Mali. The household surveys do not yet provide a time series for cellular telephones; however it is known from sector statistics that the rate of expansion for that service is even very much higher than for landlines. Considering expansion trends in urban and rural areas separately, overall we find that service expansion in urban areas has been proceeding at a slower rate than for rural areas (figure 3.3c through 3.3f). Specifically, for urban areas, the overall average annual growth rates of population covered by the different services is 3 percent for electricity, 2 percent for piped water, 5 percent for flush toilet, and 11 percent for landline telephones during the period 1996-2005. The equivalent figures for rural areas are 9 percent for electricity, 3 percent for piped water, 10 percent for flush toilet, and 19 percent for landline telephones. The strongest performers on service expansion in urban areas are Burkina Faso, Chad, Ethiopia and Mali for piped water, Burkina Faso, Chad, Ethiopia, Mali, and Senegal for flush toilets, and Benin, Chad, Mali, Tanzania for electricity. The strongest performers on service expansion in rural areas are Benin, Ethiopia, Mali, and Senegal for piped water, Burkina Faso, Guinea, Mali and Senegal for flush toilets, and Burkina Faso, Madagascar, Ethiopia, Mali for electricity. Overall, a significant percentage of the African countries surveyed are failing to ensure that service expansion keeps pace with population growth, and hence will get gradually further away from universal access until this trend can be reversed. The situation is most acute with respect to piped water and flush toilet, where close to half of the countries are expanding too slowly to keep pace with demographic growth. For electricity and landline telephones, on the other hand, around 80 percent of the countries are managing to expand coverage faster than they are expanding population. One country that stands out as falling behind demographic growth in expansion of all its modern infrastructure services is Zambia, which reports a negative growth rate for piped water, flush toilet and less than one percent growth in electricity. 17 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 3.3 Frequency distribution of average annual growth rates in service coverage (a) Overall absolute average annual growth rate 1996/05 (b) Overall average annual growth rate relative to population growth (c) Urban absolute average annual growth rate 1996/05 (d) Urban average annual growth rate relative to population growth (e) Rural absolute average annual growth rate 1996/05 (f) Rural average annual growth rate relative to population growth Source: AICD DHS/MICS Survey Database, 2007. On this basis it is possible to project the year in which each country would reach universal access for each of the modern infrastructure services, based on the assumption of continued expansion at ‘business as usual’ rates (figure 3.4). The projections indicate that under ‘business as usual’ conditions fewer than 20 percent of countries would reach universal access for piped water by 2050, while fewer than 45 percent of countries would reach universal access to electricity by the same year. However, in 18 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA approximately one third of the African countries surveyed, universal service for piped water and flush toilets (if historic trends continue) would not be reached during the current century. The projections for flush toilet and landlines are less credible in the sense that both services are currently experiencing very high growth rates from very low base levels, and these growth rates are bound to slow down as penetration increases, particularly given the high cost of these services relative to the purchasing power of the population. Even so, this Figure 3.4 Estimated year of universal coverage under business as usual analysis is probably overly optimistic in that it fails to take into account an additional trend that is further complicating the achievement of universal access for network infrastructure services: namely that of shrinking households. In addition to Source: AICD DHS/MICS Survey Database, 2007. population growth, decreasing household size frustrates coverage expansion. The average African household appears to be getting smaller as incomes rise. At work here is urbanization, declines in fertility, and greater economic resources, which allow nuclear families to disengage from extended households, because they no longer need the economies of scale provided by larger households. Because shrinking household size exerts such a strong effect on the need for new connections, countries with higher GDP per capita may not necessarily expect a smaller increase in connection needs than poorer countries, because the gains from lower population growth are more than offset by the changes in household sizes. For the African sample as a whole, the average rate of population growth is 2.5 percent, and the average increase in the number of households is 3.2 percent, so that the impact of the trend toward smaller household sizes represents almost one-third (0.7 percent) of the new connections needed to keep access rates constant (Diallo and Wodon, 2007). The full results are reported in annex table A1.1.6. In a few countries, household size has increased. Typically this occurs during hard times, as households combine forces to cope with deterioration in their living conditions. But these are exceptions. In most cases, household size has decreased between surveys. In Benin, for example, the average household size decreased from 6.0 in 1996 to 5.2 in 2001. Nevertheless, there is a wide cross-country dispersion in the relative growth rates of population versus the number of households (figure 3.5) 19 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Finally, it is relevant to Figure 3.5 Difference between average annual growth rate in number of examine to what extent coverage households versus population is converging across the different countries within Africa. If convergence is taking place, then one would expect to see countries with the lowest coverage rates experiencing the fastest rate of average annual growth in connections. If on the other hand, those countries with the highest coverage rates are also experiencing the fastest rate of average annual growth in Source: AICD DHS/MICS Survey Database, 2007. connections, then the gap between the strongest and weakest performers will only get larger over time. Figure 3.6 presents the Figure 3.6 Relationship between current coverage levels and rate of coverage growth evidence by pooling data (all services pooled) on current coverage levels and recent average annual growth rates across all countries and services. The figure illustrates that there is a small cluster of countries with low current access but high average annual growth rates, and these relate largely to flush toilet services in countries such as Chad, Ethiopia, Source: AICD DHS/MICS Survey Database, 2007. and Mali. Similarly, there is a second small cluster of countries with high current access and relatively low average annual growth rates, and these relate primarily to electricity in countries such as Cameroon, Ghana, Nigeria, and Senegal. Otherwise, the highest concentration of points can be found in the bottom left-hand corner of the graph, indicating a preponderance of countries in a stagnant situation, with both low current access rates and low average annual growth rates. There is thus no systematic evidence of convergence. For a more detailed analysis of the convergence issue, see box 3.1. 20 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Box 3.1 Evolution of access to services, 1995–2005 By plotting rates of access in 1995–2000 against 2001–05, it is possible to visualize the rate of access expansion across countries and sectors. Those countries above the 45 degree line have made significant strides in access across this period, while those on the line itself have had only stagnant coverage. In general, one might expect convergence of access rates among countries with those starting from low levels of access experiencing the highest levels of progress, and vice versa for countries starting with high levels of access. This would suggest that a regression line fitted to the data might be expected to have a positive intercept and a slope less than one. In practice, regression coefficients are slightly greater than one and intercepts are close to zero, suggesting that overall there has been a slight improvement but with no particular evidence of convergence trend across countries. However, more detailed analysis shows the intercept term for coverage of all the infrastructure services is higher for the urban population than for the rural, suggesting some degree of convergence between urban and rural coverage rates. In the graphs that follow, the coverage rate circa 1995 is plotted on the horizontal access, and the coverage rate for the same service circa 2005 is plotted on the vertical access. Points that appear above the line are experiencing an increase in service coverage, while those that appear below the line are experiencing a reduction. Points on the line are maintaining a constant access rate over time. a, Piped water supply b. Flush toilet c. Electricity d. Landline telephone What needs to be done to reach universal access? The challenge of reaching universal access is typically understood as a supply-side problem of rolling out infrastructure networks to increasingly far flung populations, entailing major investments. However, Africa’s relatively low coverage rates even in densely populated urban areas suggest that even where 21 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA infrastructure is physically present, service coverage is by no means guaranteed. Part of the access deficit therefore also seems to be related to demand-side barriers that prevent households from hooking-up to the service, even when the networks may be passing right in front of their dwellings. Demand-side barriers can take a variety of forms including high connection charges that make hook-ups unaffordable, illegal tenure that disqualifies households from connecting, and a variety of other social and economic factors that may deter households from becoming utility clients. Household survey samples are based on geographic clusters that at least for urban areas are physically small, amounting to no more than a few city blocks. It is therefore possible at least in urban areas to study the extent to which people lacking access to infrastructure live in clusters where infrastructure is available indicated by the fact that some of their immediate neighbors are hooked-up to the service. The resulting analysis gives us a sense of the degree to which low access to services is driven by supply-side issues (infrastructure networks not reaching the areas where people live) or by demand- side issues (people not connecting to available infrastructure networks). The basic concepts used to analyze Box 3.2 Coverage, access, and hook-up rates: some relationships this issue are defined in box 3.2. The main and definitions novelty is that we decompose the Coverage rate = Number of households using the service / total number of traditional measure of household coverage households into two components (as per Foster and Access rate = Number of households living in communities or clusters where service is available / total number of households Araujo, 2004 and Komives and others, Hook-up rate = Number of households using the service / Number of households 2006). The first, which we call access, living in communities where service is available gives the percentage of the population that Coverage = Access rate x hook-up rate lives in a cluster where at least one Unserved population = 100 – coverage rate household has service coverage, indicating Pure demand-side gap = Access rate coverage rate that the infrastructure is physically Supply-side gap = Unserved population pure demand-side gap proximate and that there could be an Pure supply-side gap = supply side gap x hook-up rate opportunity to connect. The second, which Mixed demand and supply side gap = supply side gap x (100 hook-up rate) we call hook-up, gives the percentage of Proportion of deficit attributable to demand-side factors only = the population living in clusters where the Pure demand side gap / Unserved population service is available that actually make a Proportion of deficit attributable to supply-side factors only = connection, and hence take-up that Pure supply side gap / Unserved population opportunity. Using these two concepts it is Proportion of deficit attributable to both demand and supply side factors only = Mixed demand- and supply-side gap / Unserved population possible to estimate the percentage of the Source: Foster and Araujo 2004. unserved population that constitutes a supply-side deficit (meaning that they are too far from the network to make a connection until further rollout takes place) versus a demand-side deficit (meaning that something other than distance from the network is preventing them from taking-up the service). The policy conclusions in each case are very different, and hence the interest in making this distinction. The solution to a supply-side deficit is to make further investments to rollout the geographic reach of infrastructure networks. The solution to a demand-side deficit is to make policy changes that help to address potential barriers to service take-up, such as high connection charges or illegal tenure. 22 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA For various reasons, it could be questioned whether absolutely everyone in a geographic cluster with some coverage really has the opportunity to connect. First, although the geographic clusters are relatively small in urban areas, the distances may still be such as to prohibit connection. Second, even though the infrastructure is present, it may not have the carrying capacity required to service all residents in a particular geographic cluster without further investment and upgrade. Third, even if a household is physically close to a network with adequate carrying capacity, they may choose not to connect simply because they have an acceptable alternative (such as a borehole) rather than due to any demand-side barriers with the service itself. Diallo and Wodon (2007b) use a statistical approach to try and correct for these problems. They simulate the maximum connection rate obtainable in any PSU based on that of the richest households in that PSU. If less than 100 percent of the richest households are connected, it suggests that something other than demand-side barriers is at work. Table 3.2 presents results for the demand-side deficit both with and without this statistical adjustment. The methodology is less applicable to rural areas because the PSUs tend to be larger and population densities much lower. Table 3.2 Proportion of infrastructure coverage deficit in urban Africa attributable to demand- and supply-side factors Proportion of deficit attributable Decomposition of coverage to demand-side factors (1) (2) (1) x (2) Percentage, population-weighted average Access Hook-up Coverage Unadjusted Adjusted Piped water By country income level Low 68 42 31 58 14 Middle 91 74 69 61 36 By urbanization level Low 76 42 33 65 20 Medium 76 56 46 63 8 High 71 49 34 55 45 Electricity By country income level Low 93 73 69 82 50 Middle 95 86 81 67 61 By urbanization level Low 87 60 53 75 15 Medium 86 58 52 76 37 High 97 85 53 81 71 Source: AICD DHS/MICS Survey Database, 2007. Note: The data pertain to urban areas only. The first point that emerges is that for piped water and electricity in urban areas of Africa access rates exceed coverage rates by between 20 to 40 percentage points. Indeed, access rates are as high as 70–90 percent; meaning that the vast majority of the urban population even in low-income countries lives in relatively close geographic proximity to existing water and electricity networks. The reason for this discrepancy is the comparatively low service hook-up rates. However, it is striking that hook-up rates for electricity at 60–90 percent are substantially higher than hook-up rates for piped water at 30–70 percent. 23 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Hook-up rates are significantly higher in middle income than in low income countries, particularly for the piped water service. As a result, the estimates of the proportion of the coverage deficit due to demand-side factors are large when no statistical adjustment is made (table 3.2). Demand-side factors account for 55-65 percent of the coverage deficit in piped water and 65-85 percent of the deficit in electricity. However, these estimates fall substantially when the above statistical adjustments are made. This is particularly true for piped water where the proportion of the coverage deficit due to demand-side factors falls to between 15- 35 percent. In the case of electricity, around 50-60 percent is still due to demand-side factors even when the statistical adjustments are made. The level of the demand-side deficit for electricity rise steeply with the rate of urbanization from 15 percent in the low urbanization countries to over 70 percent in the high urbanization countries. Examining the results at the country level, shows a very strong relationship between the level of access (that is the share of the population living in areas where the service is available) and the size of the demand-side deficit (figure 3.7). The relationship is even stronger for electricity than for piped water. Overall, there is huge variation in the size of the adjusted demand-side deficit across countries (see cross-country annex tables A.2.14 and A.2.17 for more details). For piped water, the range is from less than 5 percent in countries such as Burkina, Central African Republic, Chad, Ethiopia, Mozambique, Rwanda, Tanzania, Uganda, to over 50 percent in countries such as Republic of Congo, Cote d’Ivoire, Gabon, Senegal and Zambia. In the case of electricity, the adjusted demand-side deficit ranges from less than 10 percent in countries such as Burkina Faso, Central African Republic, Chad, Niger, and Rwanda to more than 60 percent in Cameroon, Comoros, Republic of Congo, Cote d’Ivoire, Ghana, Namibia, Nigeria, and Senegal. Thus, the relative importance of supply and demand-side considerations in policy formulation for universal access needs should be sensitive to the situation in each specific country. Figure 3.7 Country scatter-plot of demand-side deficit versus current access rates for electricity and piped water (a) Piped water (b) Electricity Source: AICD DHS/MICS Survey Database, 2007. In conclusion, this section has illustrated the diversity of infrastructure service coverage across different African countries. Overall, a strong association was found between infrastructure service 24 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA coverage and the income and urbanization level of the country. Nevertheless, even controlling for income and urbanization, some countries stand out as having much higher (or lower) levels of coverage than might be expected, and these cases merit closer study with a view to identifying the causes of this over (or under) performance. Based on historic trends, it is possible to project the time it would take to reach universal access under business as usual conditions. The results are sobering. Very few countries are expected to reach universal access in any service before 2050. In a number of countries, the coverage trend (particularly for water and sanitation) is actually negative in absolute terms or at least lower than the rate of population growth. Access to electricity is expanding more rapidly, with most countries outstripping demographic growth and a significant number expanding at 4-6 percent per annum. Finally, in order to shed light on the kinds of interventions needed to speed-up the rate of access expansion, the unserved urban population is divided between those that live physically close to an infrastructure network (and hence could be reached through relatively low cost densification programs) versus those that are physically distant (and can only be reached by more costly network rollout). While the majority of the unserved can only be reached through further roll out of infrastructure networks, there is a substantial minority that have ready physical access to the service but face some kind of demand-side barrier. This is truer to a greater extent for electricity than it is for piped water. Moreover, in some of the middle income and better-off low income countries, a majority of unserved urban customers appear to be reachable through relatively low cost (but policy-intensive) measures on the demand-side. The finding that a significant share of the unserved population in urban areas live close to infrastructure networks, suggests that affordability may be an important barrier to service uptake. This is the topic of the next section. 25 4 Keeping services affordable We have established that coverage of network infrastructure in Africa is very low. But to improve the situation, it is essential to know why. One possibility is that coverage is low because network services are not available. Another is that services are available but not affordable. Still another is that services are both available and affordable but not taken up because alternatives are preferred. This section addresses the issue of affordability in greater depth. Stipulating tariff levels high enough to allow providers of infrastructure services to recovery their costs (and thus to justify investments in expanded networks), we ask what share of the population would be able to afford those payments. Moreover, in the case that a significant share of the population cannot afford the service, we ask whether the state can afford to subsidize them, and whether it has effective means at its disposal for doing so. The detailed country-by-country numbers underlying the analysis presented here can be found in Part 3 of the Cross-Country Annex Volume. The overall profile for each country can be found in the Country Annex Volume. African household budgets Most African households live on very modest budgets and spend more than half of their resources on food. The average African household survives on no more than $180 per month; urban households are about $100 per month better off than rural households (table 4.1). Household budgets range from around US$50 per month in the lowest quintile to no more than $400 per month in the highest income quintile; except in middle-income countries, where the richest quintile has between $200 and $1,300 per month. Table 4.1 Monthly household budget Food expenditure as a share of Total household budget (2002 US$) total household budget (percent) Nat’l Rural Urban Q1 Q2 Q3 Q4 Q5 Nat’l Rural Urban Q1 Q2 Q3 Q4 Q5 Overall 177 130 241 59 97 128 169 340 55 61 48 63 64 63 60 48 Low-income countries 139 109 208 53 80 103 135 258 59 64 50 67 68 66 64 52 Middle-income countries 300 199 350 79 155 181 282 609 45 54 42 51 55 52 50 38 Source: AICD Expenditure Survey Database, 2007. The total budget envelope is important in understanding the ability of African households to pay for infrastructure. Around the monthly average household expenditure of $180, the national budget envelope on the continent ranges from $57 in Ethiopia to $539 in South Africa (2002 US$). The monthly budget of the poorest quintile ranges between $18 per household in Burundi and $160 in Morocco. The poorest quintile has a median budget of about $50; that of the richest is about $240, drawn upward by South Africa and Morocco. Nevertheless, although the amount available for each household to spend is very small, the aggregate size of the low-income market is an estimated $429 billion each year. The water and energy spending of the poorest 250 million Africans are estimated at $2.5 billion and $12 billion per year respectively (Hammond and others, 2007). 26 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA On average, Africans spend more than half of their household budget on food; not much is left over for anything else—including infrastructure. Even the most affluent households spend about half of their monthly budget on food; whereas the bottom four quintiles all devote around 60 percent of their budgets to food. The share of food in total household expenditure ranges from 28 percent in the Republic of Congo to 72 percent in Burundi. Households in Cameroon, the Democratic Republic of Congo, Madagascar, Niger, Sao Tome and Principe, Tanzania and Zambia spend more than 60 percent of their budget on food. In the poorest quintile, less than one-third of the budget is left over for nonfood items in Zambia, Sao Tome and Principe, Rwanda, Madagascar, Kenya, the Democratic Republic of Congo, and Burundi. Data on payment patterns for infrastructure services are often noisy and unreliable, and it is difficult to distinguish between no payment, missing payments, and no information on payment. In this study, only those households that reported access to the four network infrastructure services of water supply, sewerage, electricity, and telephones are included in our evaluation of spending patterns. For transport, all households are included to allow us to distinguish between those that use transport facilities and those that do not, which has implications for spending. Given that only a tiny minority of households has access to the full range of modern infrastructure services, total expenditure on the aggregate infrastructure category can be a little misleading since most households only register expenditure on some of the services. Nevertheless, it is interesting to examine the overall budget share dedicated to all infrastructure services and examine its variation across urban- rural areas and expenditure quintiles. This includes expenditure on utilities, transport and rubbish disposal. On average, total infrastructure spending absorbs about 7 percent of the household budget. For most countries, the overall infrastructure budget share falls in the 5-15 percent range, and this can rise to more than 25 percent in some countries. Thus, it is not unusual for infrastructure spending to absorb 40 percent of the nonfood budget of the household, and as much as 80 percent in some cases. Figure 4.1 Frequency distribution of overall household budget share devoted to infrastructure services (a) According to urban-rural split (b) According to budget quintiles Source: AICD Expenditure Survey Database, 2007. 27 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Absolute levels of infrastructure spending by rural households are not that different from those made by urban households. However, given lower overall household expenditure in rural areas, budget shares are considerably higher. Whereas, in most countries urban households spend less than 5 percent their budgets on all infrastructure services combined, in rural areas the budget shares going to infrastructure are well over 5 percent in most countries. The share of household budgets spent on infrastructure is similar across the quintiles, though (if anything) increasing slightly in the upper quintiles. On average, connected customers spend between 5 and 6 percent of their budget for network (piped water, electricity, and landline) infrastructure services. In some countries, however, the share is larger. Network infrastructure spending constitutes more than 10 percent of the total spending of households in South Africa, Sao Tome and Principe, Rwanda, Mozambique, Mauritania, Gabon, and Burkina Faso. The average is highest in South Africa, at 23 percent. Considering individual infrastructure services, the highest budget shares are found for cellular telephony at around 14 percent on average (for the handful of countries with available evidence). Malawi is an outlier with disproportionate spending on cell phones. Aside from Malawi, the pan-African average spending on cell phones is 9 percent. This is followed by electricity and transport, each of which absorbs 6 percent of the household budget on average (figure 4.2). Households also spend about 3 percent of the budget on LPG, primarily used for cooking. Spending on water amounts to 2 percent of household budgets on average, and rarely exceeds 3 percent in any one country. Only in Cameroon, Mauritania, and Rwanda, the water expenses are more than 5 percent of household budget. Spending on landlines amounts to 2 percent of household budgets on average but shows a very high degree of variation across countries. In general, rubbish disposal constitutes a negligible share of total household expenditure, with some notable exceptions such as Sierra Leone and Rwanda. Figure 4.2 Frequency distribution of overall household budget share devoted to specific infrastructure services Source: AICD Expenditure Survey Database, 2007. Nonpayment as an indicator of affordability The discussion so far has focused on formal utility customers that report paying a utility bill. However, to focus only on this category of users is to miss a substantial part of the African story. Household surveys provide unique insights into two other key categories of consumers (table 4.2). First, 28 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA there are those that do not have a connection but nonetheless register expenditure because they are accessing the network through some secondary source, usually a neighbors tap or power line or a public telephone facility. Second, there are those that do have a connection but do not register any expenditure, whether because they are in arrears or because the connection itself is a clandestine one. Table 4.2 Consumers of infrastructure services, differentiated by connection and payment status Consumer pays for service Consumer does not pay for service Consumer is connected or owns Customers in arrears or disconnected (e.g., for Traditional customers service nonpayment); illegal connections Neighbor’s phone, public phone, illegal Consumer is not connected or connections, neighbor’s yard tap, public tap, Unserved consumers does not own service vendors, community latrines Source: Author’s elaboration. Table 4.3 provides evidence on the relative importance of each category of consumer according to service. As we know from chapter 3, most Africans are unserved consumers. They are not connected to, nor do they pay for, formal services (table 4.3). The proportions range from 61 percent in the case of piped water, to 87 percent for cell phone service. However, what the table really serves to clarify is that for all services the traditional customers that connect and pay are actually a minority of those who use the service. Across all services, the population that connects but does not pay is almost as high as the percentage that connects and pays. Moreover, for piped water and landline telephone services the population that is unconnected but nevertheless pays to obtain the service through secondary sources is slightly higher than the one that connects and pays. Interestingly, public telephone usage seems to be largely confined to landlines, with only 1 percent of the population reporting expenditure on cellular telephones without owning a cellular telephone themselves. Nonpayment for infrastructure Table 4.3 Consumers of selected network infrastructure services, by services is a major issue, one that connection and payment status bears directly on the ability of Percent utilities and other service providers Connected Connected and Unconnected Unconnected to expand networks and improve and pay do not pay and pay and do not pay’ services by undermining their Electricity 14 11 7 67 financial strength. By comparing Piped water 13 12 14 61 the connected who do not pay Landline 5 4 6 86 against the total connected group, Cell phone 6 7 1 87 it is possible to calculate non- Source: AICD Expenditure Survey Database, 2007. payment ratios. Note: Rows and columns may not add to 100 because of rounding. Overall, around 40 percent of those connected to infrastructure services do not appear to be paying for them in any given month. The fraction is relatively similar across all of the services considered. The variation across countries is substantial however. Nonpayment rates in excess of 65 percent of customers can be found in 20 percent of countries for electricity and 30 percent for piped water (figure 4.3). 29 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Countries such as Chad, Kenya and Zambia perform consistently poorly in this respect. At the other end of the spectrum, countries such as Ethiopia and Senegal record nonpayment rates of less than 10 percent. To the extent that nonpayment is higher among the poorest, it can be taken as an indicator that households are facing affordability problems. Figure 4.4 reports nonpayment ratios by quintile for electricity and piped water services. The pattern is strikingly consistent across the two services, although nonpayment is systematically slightly worse for piped water than it is for electricity. In the first quintile, the nonpayment ratio amounts to around 60 percent of households, and this declines steadily to around 20 percent of households in the fifth quintile. This pattern indicates that nonpayment does to some extent represent an affordability issue given the decline as household budgets rise across the distribution. Nevertheless, the existence of a significant nonpayment rate, even among the richest quintiles, suggests that problems of payment culture also exist. Moreover, given that the majority of connected households are in the richer quintiles, in absolute terms the largest number of nonpaying customers also comes from the richer quintiles (even though the nonpayment ratio for this group is comparatively low). Figure 4.3 Nonpayment rate in water and electricity Figure 4.4 Nonpayment rate in water and electricity by quintile Source: AICD Expenditure Survey Database, 2007. Source: AICD Expenditure Survey Database, 2007. Can African households afford to pay full cost-recovery tariffs? Utilities will not invest in expanding their networks before establishing demand for their services— and ability to pay. From a practical and policy standpoint, therefore, it is important to know how much unserved beneficiaries can afford to pay for infrastructure. Based on a relatively small sample of African countries, Foster and Yepes (2006) conclude that 70 percent of households in Africa will have difficulty paying internationally comparable cost-recovery tariffs. This analysis is extended and deepened here to understand the limits of affordability of cost- recovery tariffs for piped water and electricity in the African context. Affordability is typically measured by the whether the share of infrastructure spending in the total household budget exceeds a set threshold (Frankhauser and Tepic 2005). There is no absolutely scientific basis for determining the value of such affordability thresholds, however based on experience with actual household expenditure patterns and results of willingness to pay surveys, certain thresholds have come to be widely used by practitioners. The WHO, for example, uses a 5 percent affordability threshold for water 30 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA and sanitation services in developing countries. The evidence presented on current expenditure patterns above suggests that households spend 5-10 percent of their budgets on infrastructure services overall, while for individual services most of the countries pay between than 2-5 percent. In the discussion that follows, 5 percent and 3 percent are used as reference affordability thresholds. The discussion can either be interpreted as referring to a single infrastructure service or to a package of infrastructure services. In order to estimate the percentage of African households likely to face affordability problems for modern infrastructure services, two elements are needed. First, some indicative values of the true cost of infrastructure services are needed as a reference point. Based on different assumptions about subsistence household consumption and the tariff applied, the absolute cost of the total monthly bill can be computed (table 4.4). For piped water service, subsistence consumption ranges between 4 cubic meters per month—based on an absolute minimum consumption of 25 liters per capita per day for a family of five—and 10 cubic meters per month—based on a somewhat more comfortable but still modest level of 60 liters per capita per day for a family of five. The indicative tariff ranges between US$0.40 to US$0.80 per cubic meter depending on whether operating or full capital cost recovery is envisaged. For electricity, subsistence consumption ranges from 25 kilowatt-hours per month (supporting use of two 100-watt light bulbs for four hours each day) to 50 kilowatt-hours per month (supporting limited use of an additional appliance, such as a radio). The indicative tariff ranges between US$0.08 to US$0.25 per kilowatt-hour, reflecting the variation that exists between relatively low cost hydro-power dominated systems and those based on diesel generators in landlocked countries with high import costs. In either case, the lower-bound monthly bill coincides at around US$2.00. The upper- bound monthly bill is around US$8 for piped water and $12 for electricity. This also suggests that at the lower bound a household could purchase both basic piped water and electricity services for around US$4 per month, which would rise to around US$20 per month at the upper bound. Table 4.4 Reference points for true cost of infrastructure services. Piped water Reference Electricity Reference Lower bound Subsistence household 4 m3 Subsistence household 25 kWh consumption consumption (kWh) Tariff (operating cost recovery) $0.40/m3 Tariff (low-cost country) $0.08/kWh US$/m3 US$/kwh Total monthly bill (US$) $2.00 Total monthly bill (US$) $2.00 Upper bound Subsistence household 10 m3 Subsistence household 50 kWh consumption consumption Tariff (capital cost recovery) $0.80/m3 Tariff (high-cost country) $0.25/kWh US$/m3 US$/kwh Total monthly bill (US$) $8.00 Total monthly bill (US$) $12.00 Second, the household survey data on budget expenditures across households is used to estimate what percentage of households would hit the 5 percent affordability thresholds at different levels of absolute expenditure. For example, a household with a monthly budget of US$100 would hit the affordability threshold of 5 percent of income once any service cost more than US$5 per month. 31 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA By pooling all African households together across countries and grouping them into a common set of quintiles based on purchasing power parity adjustments to their budgets, it is possible to report results for the continent as a whole. For the average household in each of the continental income quintiles, figure 4.5 plots the share of their budget that would be required to meet increasing levels of spending on infrastructure services. Figure 4.5 Share of average urban household budget required to purchase subsistence amounts of piped water and electricity, by continental income quintiles Current US$ Quintile Source: AICD Expenditure Survey Database, 2007. Thus, the average household in the first quintile hits the 5 percent affordability threshold around US$4 per month, which would be enough to pay for both piped water and electricity services under the lower-bound assumptions detailed above. The average household in the second quintile hits the 5 percent affordability threshold around US$7 per month, and so once again could afford only piped water and electricity under the lower-bound scenarios. The average third quintile household hits the 5 percent affordability threshold at around US$12 per month and hence could only afford one of the two services if upper bound conditions were applied. The average fourth quintile household hits the 5 percent affordability threshold at around US$20 per month, and hence would be able to pay for both piped water and electricity services even under the upper bound scenarios. Households in the fifth quintile do not face any affordability constraints within the range of service baskets considered here. From these findings one can infer that very modest consumption baskets priced at levels compatible with recovery of operating costs would appear to be affordable across the full range of household budgets in Africa (table 4.5). Nevertheless, around 60 percent of the African population cannot afford to pay full cost recovery tariffs or extend consumption beyond the absolute minimum subsistence level. These continental results mask a great deal of variation across individual countries. In particular, this is because almost all of the households in the poorer countries may be in the bottom quintile for Africa as a whole, while almost all of the households in the more affluent countries may be in the uppermost quintile for Africa as a whole. Therefore, table 4.5 provides a similar type of analysis at the country level, calculating the percentage of households in each country that would fall beyond the 5 percent affordability threshold at any particular absolute monthly cost of service. 32 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Table 4.5 Share of urban households whose utility bill would exceed 5 percent of the monthly household budget at various prices Percent Monthly bill (US$) Group $2 $4 $6 $8 $10 $12 $14 $16 Cape Verde 0 0 0 0 0 0 0 0 Morocco 0 0 0 0 0 0 0 0 Senegal 0 0 0 0 0 0 1 1 1 South Africa 0 0 0 0 1 1 1 1 Cameroon 0 0 0 0 1 2 7 17 Cote d'Ivoire 0 0 1 2 3 5 7 10 Congo, Rep. of 0 0 3 5 12 21 28 35 Ghana 0 2 7 11 30 46 55 67 Benin 0 2 4 12 33 45 60 71 Kenya 0 0 5 20 36 62 72 78 Sierra Leone 0 4 16 30 44 54 62 67 2 Sao Tome 0 2 13 29 46 64 77 81 Burkina Faso 0 4 20 34 47 62 72 78 Zambia 0 4 18 35 50 58 67 76 Nigeria 3 10 23 35 57 78 89 95 Madagascar 0 16 28 47 61 68 78 85 Niger 1 11 28 55 70 79 89 93 Tanzania 1 8 25 55 75 89 96 98 Guinea Bissau 0 6 38 65 81 89 91 93 Uganda 2 17 45 65 82 90 96 97 3 Burundi 7 29 53 72 82 90 97 100 Malawi 2 32 66 78 87 92 93 94 Congo, Dem. Rep. 9 49 79 91 98 99 100 100 Ethiopia 40 87 95 99 99 99 99 100 Low-income 5.0 18.4 32.4 44.5 59.5 72.3 79.7 84.3 Summary Middle-income 0.0 0.0 0.1 0.2 1.2 1.8 2.9 4.7 All 3.7 13.7 24.2 33.2 44.7 54.3 60.2 64.1 Source: AICD Expenditure Survey Database, 2007. The countries divide into three groups. At one extreme is Group 1 (see table 4.5) comprising Cape Verde, Morocco, Senegal, Cameroon, Cote d’Ivoire, Congo, and South Africa, where a majority of urban households can afford a monthly expenditure of US$12 (and indeed much higher). At the other extreme is Group 3, comprising Burundi, the Democratic Republic of Congo, Ethiopia, Guinea-Bissau, Malawi, Niger, Tanzania and Uganda, where the vast majority of urban households (at least 70 percent and in some cases over 90 percent of households) would be unable to afford a monthly expenditure of US$8 or $12 for water or electricity. All the remaining countries fall into Group 2 where a substantial share of the urban population—between one- and two-thirds—would face difficulties covering an upper-bound monthly expenditure. 33 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Can African governments afford to subsidize infrastructure services? The affordability of infrastructure services needs to be considered not only at the household level, but also at the level of the public finances of each country. To the extent that households cannot afford to pay cost recovery tariffs, the move towards universal access will create burgeoning liabilities for the state that must step in and meet the difference between the tariffs that the public can afford to pay and the real cost of service provision. The same analytical framework developed above can be used to estimate the aggregate value of these subsidies of the countries concerned. This gives us a sense of whether the strategy of subsidizing services to reach universal coverage is itself an affordable strategy at the country level. Once again, there is no absolute scientific method to determine the affordability threshold at the country level; nevertheless, it is possible to get a sense of when costs reach a level that is manifestly unattainable. Two types of subsidies for electricity and piped water are considered, based on highly simplified assumptions for illustrative purposes. First, a one time capital subsidy of US$200 per unserved household to cover the costs of connection of all these households spread over a period of 10 years. Second, an ongoing subsidy of US$2 per month respectively for water and electricity to ensure that these services remain affordable to households once connected. The operating subsidy is calculated assuming the households can afford at least the minimum subsistence quantities of electricity and water. With regard to the capital subsidy for water or electricity, we find that overall this would cost around 1 percent of African GDP for the countries included in the sample. Around 60 percent of the countries would face costs in excess of 1 percent of GDP. In particular, this policy would cost more than 2 percent of GDP in Ethiopia, Malawi, Congo, Dem. Rep. and Sudan. The highest burden on fiscal resources will be for the Democratic Republic of Congo, which has to spend about 18 percent of GDP on piped water connections. The spending envelopes are similar for piped water and electricity. Nevertheless, in more affluent countries, such as Gabon, the cost of this policy would amount to no more than 0.02 percent of GDP. With regard to the ongoing subsidy of $10 per month for the unserved customers, we find that the burden on government resources would be similarly onerous. For 40 percent of the countries, providing a monthly subsidy of $2 for electricity or water would amount to spending between 1 and 2 percent of GDP. The maximum burden would be on the Democratic Republic of Congo, followed by Ethiopia, Malawi, Niger, and Sudan, which would have to spend more than 2 percent to maintain a sustainable consumer base for water and electricity services. Like the capital subsidy, this operating subsidy comes out to be 1.1 percent of African GDP for water or electricity for the countries included in the sample (figure 4.6). 34 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 4.6 Subsidy needed to maintain affordability of water and electricity a. Capital subsidy b. Operating subsidy Source: AICD DHS/MICS Database, 2007. But do utility subsidies really work in practice? Irrespective of the arguments advanced above, concerns about affordability are already a pervasive reality in the water and power sector in Africa. The evidence amassed through the household surveys provides a rare opportunity to evaluate how effectively existing subsidies perform, and to glean lessons for improved subsidy design in the future. When combined with information about the nature of tariff structures for power and water services in Africa, it becomes possible to unravel the pattern of subsidy incidence across different households. Customers receive substantial subsidies in most African countries as residential electricity and water tariffs tend to be below utility costs. The working assumption is that the price per kWh in the highest bracket of consumption in the tariff schedule can be used as a first approximation of the cost of providing the service (actually, the estimates of targeting performance are not very sensitive to that assumption). As shown by Angel-Urdinola and Wodon (2006), a simple framework can be used not only to analyze the targeting performance of electricity and water subsidies in about 20 African countries for which data are available, but also to understand what affects targeting performance through so-called access (who uses 35 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA electricity and/or water) and subsidy design factors (who benefits from subsidies and by how much among users).9 The targeting performance indicator used in the analysis, denoted by (Omega), is simply the share of the subsidies received by the poor divided by the proportion of the population in poverty. In other words, a value of one for implies that the subsidy distribution among the poor is proportional to their share in the overall population. If the poor account for 30 percent of the population, then a neutral targeting mechanism would allocate 30 percent of the subsidy to the poor. A value (lower) greater than one implies that the subsidy distribution is (regressive) progressive, since the share of benefits allocates to the poor is (lower) larger than its share in the total population. For instance, suppose that 30 percent of the population is poor and that they obtain 60 percent of the subsidy benefits. In such case, would equal to two, meaning that the poor are receiving twice as much subsidies as the population on average. As shown in Figure 2, in none of the countries is the targeting indicator superior to one, and for both electricity and water, it is often well below one. While there are some comparability issues between countries, the message is clear: utility subsidies tend to be very poorly targeted, with on average the poor benefiting only from a fourth to a third of what a household randomly selected in the population would get. While most indicators of targeting performance are silent as to why subsidies are targeted the way they are (they only give an idea a whether the subsidies reach the poor or not and to what extent), the framework used here allows for analyzing both “access� and “subsidy design� factors that affect targeting performance. Access factors are those related to the availability of electricity and water service in the area where a household lives and to the household’s choice to connect to the network when service is available. These access factors have a strong influence on targeting performance but are usually difficult to change in the short run. Subsidy factors are more susceptible to policy design, such as changes in tariff structures affecting who is targeted to receive the subsidies, as well as the rates of subsidization and the quantities of electricity and water consumed by the households that benefit from the subsidies. It turns out that most electricity and water subsidy mechanisms are poorly targeted, essentially because most of the poor lack access to the electricity and water network and therefore cannot benefit from electricity and water subsidies, but also because the existing tariff structures are not designed in a way to target subsidies to the poor. 9 See also Komives and others 2005. 36 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 4.7 Overall targeting performance ( ) of utility subsidies, African countries Source: Wodon and others (2007a, 2007b). 37 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA This can be seen clearly in figure 4.8, which decomposes the value of the targeting indicator into access and subsidy design factors, so that = (access factors) x (subsidy design factors). The value of access factors are on the horizontal axis, and the value of the subsidy design factors are represented on the vertical axis. The curves added to the graphs represent combinations of values for the access and subsidy design factor that result in the same value for . The further a country is located to the upper right of the graphs, the better the targeting performance, since again omega is the product of the access and subsidy design factors. The two variables used to compute the access factors are first, whether a household lives in an area served by the electricity and water network, and second whether when a household is in such an areas, it is actually connected or not to the network, i.e., whether the household actually takes up the service. The value of the access factors is simply the rate of connection among the poor to the network (which depends on access and uptake when there is access) divided by the rate of connection in the population as a whole. As is clear in figure 4.8, and as expected, the access factors are much lower than one for all countries, simply because the poor have much lower connection rates than the population as a whole, on average. The second variable affecting the value of the targeting parameter is the Subsidy Design Factors, which take into account who benefits from subsidies among households connected to the network, and how large the subsidies are. What this Subsidy design factor represents is the ratio of the average benefit from the subsidy among all poor households that are connected to the network, divided by the average benefit among all households connected to the network, whether poor or non-poor. Surprisingly, in many countries, the subsidy design factors are also below unity, thereby also limiting targeting performance. The main explanation is that while the rate of subsidization of the poor (i.e., the discount versus the full cost of providing electricity and water for the utility) is often larger than for the population as a whole that is connected to the network, the quantities consumed by the population as a whole tend to be larger than those consumed by the poor, so that the overall subsidy received by the poor is lower on average than that received by the population as a whole. Several clear messages emerge from the empirical analysis of the targeting performance of electricity subsidies presented above. Consumption subsidies for electricity and water appear to be poorly targeted in African countries. Several reasons explain this poor targeting performance. First, access factors are important in determining the potential beneficiaries of consumption subsidies. As poor households tend to live in areas without electricity and water service, it is impossible for them to benefit from the subsidies. In addition, even when there is potential access to the network where the poor live, many among the poor remain not connected to the networks, either because they live still to far from the electric lines or water pipes, or because the cost of connecting to the network and purchasing the equipment to required to use electricity and water is too high. In order to compensate for the negative impact of access factors on targeting performance, good subsidy design mechanisms are required. Unfortunately, the traditional Inverted Block Tariff structures that prevail in many countries tend to be poorly targeted. First, these tariff structures spread subsidies to all households connected to the network, since even those that consume high amounts of electricity benefit from a subsidy for the part of their consumption that belongs to the lower level blocks of the tariff structure. In addition, the lower blocks tend often to be too high in terms of consumption (in kWh per month) to target the poor well. And finally, the differences in unit prices between the various blocks may not be large. 38 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 4.8 Access factors and subsidy design factors affecting targeting performance (a) Electricity (b) Water Source: Wodon and others (2007a, 2007b). 39 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA One possible alternative is to provide connection as opposed to consumption subsidies, assuming that the generation or production capacity is sufficient to expand the network. Figure 4.9 provides the potential targeting performance of connection subsidies under the three scenarios. First, we assume that connection subsidies will be distributed in the same way as existing connections. This is a pessimistic assumption from a distributional point of view since it tends to favor better off households, but it could be realistic if access rates to the network are low. Second, we assume that new connections could be distributed randomly among households that are currently not connected, but live in a neighborhood where connections are available. Third, we assume that new connection subsidies could be randomly distributed among all households that do not currently have access. This is a very optimistic assumption given that many of these households do not live in neighborhoods where access is available. As shown in figure 4.9, and as expected, the value of Omega is largest under the assumption that new connections benefit households that are selected randomly from the population without access. In all countries, for both water and electricity, Omega is larger than one under this assumption. Yet the assumption is not realistic. The second scenario assumes that households that benefit from new connections are selected from non-served households living in those areas where there is already access to the network. The values of Omega, while often lower than one, are still much better than those for consumption subsidies. In the third scenario, targeting performance remains poor. Thus, if connection subsidies could be designed in order to reach the majority of households not connected today but living in areas where service is provided, the targeting performance of those subsidies would be better than that of consumption subsidies. In addition, connection subsidies help in reducing the cost of service for users (as compared to street vendors for water for example), while also bringing in positive externalities in areas such as education and health. Finally, it is often argued that any removal of utility subsidies would be detrimental to power. Again, the household survey evidence provides an opportunity to test this hypothesis. For most countries, over the population as a whole electricity and water spending represents only a tiny fraction of total consumption, often well below 1 percent. Among households connected to the network and consuming electricity and/or water, the fraction is much higher, typically between three and five percent. This is turn is directly related to the impact of a proportional increase in electricity and water tariffs on poverty. For simplicity, relative poverty measures can be used whereby the poverty line in each country is set at half the mean level of per capita consumption. At the national level, the impact of a 50 percent increase in tariffs, or even of a doubling of the tariffs, is truly marginal, with in many countries the national estimates of the shares of the population living in poverty changing by barely one-tenth of a percentage point. Among households with a connection to the network, the impact is larger, but still fairly limited. Indeed, there is rarely an increase in the share of households in poverty larger than one or two percentage points, and in addition, because the households that benefit from a connection tend not to be poor as compared to other households, the increase in poverty starts from a very low base. Thus, in general, it can be said that the impact on poverty of an increase in tariffs is small in most cases. This does not mean that such an impact does not have a negative impact on those hit by it. But it means that if subsidies were reduced, and the funds were used in a different, more pro-poor way, there would be a gain for poverty reduction that could be substantial. 40 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 4.9 Potential targeting performance of connection subsidies under various scenarios a. Electricity b. Water Source: Wodon and others (2007a, 2007b). 41 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA In conclusion, African households (including the wealthier ones) survive on relatively modest household budget, and spend at least half of that budget on food. Infrastructure spending is an important component of the non-food budget. Overall, African households spend 7 percent on all infrastructure services. The average budget share for households reporting some kind of access to individual infrastructure services is as follows: 6 percent for energy and transport, 2 percent for water; the corresponding figure for cellular telephones is 14 percent although based on a limited sample of countries. Given that few households enjoy multiple infrastructure services, the budget share for all infrastructure services is relatively low compared to that for individual services. Two tests of affordability of services are used. The first relates to patterns of non-payment. The results show that around 40 percent of those with power and water connections report not having paid for the service in the month preceding the survey. This ratio ranges from 60 percent in the first expenditure quintile to 20 percent in the uppermost expenditure quintile. The second test of affordability is based on an examination of the distribution of expenditure across households. A significant proportion of African households can afford to pay very modest consumption baskets priced at levels compatible with operating cost recovery. Nevertheless, except in the richest countries, a substantial share of the population cannot afford to pay full capital cost recovery tariffs. The implication that a significant percentage of the population would require some kind of government subsidy for access to modern infrastructure services raises questions about the feasibility of this. A very simple simulation exercise is performed to gauge the aggregate magnitude of subsidies potentially required to support a universal access policy. We find that even if a one-time capital subsidy is provided to the unserved population, affording it on a long-term basis through an ongoing subsidy of $10 per month is a costly proposition. For majority of Governments in Africa, enabling sustainable service provision will cost less than 1 percent of GDP. In addition, historic experience with public subsidies shows that these have been woefully inadequate at reaching the poor, partly due to low service coverage among the poor, but also reflecting certain design flaws in the corresponding tariff structures. 42 5 Alternatives to modern infrastructure services The coverage of network infrastructure in Africa is so low and universal access still so far away that the vast majority of Africans will continue to depend on alternate sources to meet their demand for the foreseeable future. These alternatives can be grouped according to the level of service that they provide (table 5.1). In most cases four levels can be distinguished going from modern infrastructure services, to intermediate alternatives, to basic alternatives, to nothing at all. The level of service provided by the intermediate options, although it falls well short of the modern level, is nonetheless typically a substantial improvement on the basic service both in terms of convenience and avoidance of the worse health risks. Even the basic level of service, though still far from adequate, offers some advantages over not having any service at all. Table 5.1 Different levels of service for infrastructure Modern service Intermediate service Basic service No service Electricity connection Street lighting Kerosene/candles No artificial light Piped water connection Shared connections, stand-posts Surface water n.a. or boreholes Sewerage Shared flush toilets, or VIP, Traditional pit latrines No sanitation Chemical, SAN PLAT latrines Landline or cellular telephone Public telephone Letter mail No post or telecommunications n.a. = not applicable. These observations highlights the importance—from a policy standpoint—of identifying measures to move people up across all rungs of the ladder, without focusing exclusively on moving people into private connections. Accordingly, this section turns the spotlight on the three lower service levels and asks what can be learnt in policy terms from the household survey evidence about these alternatives. The detailed country-by-country numbers underlying the analysis presented here can be found in Part IV of the Cross-Country Annex Volume. The overall profile for each country can be found in the Country Annex Volume. Alternatives to piped water Boreholes are by far the most widely used form of water supply in Africa (37 percent overall). However, the dominance of boreholes is driven largely by the rural areas. In the urban areas they account for no more than 24 percent of supply, and utilities—whether through private taps (38 percent) or standposts (25 percent)—are the dominant source of water. The overall coverage of stand-posts is only slightly higher than the coverage of private piped water connections. While somewhat more equitably distributed than the latter, public stand-posts are still regressive in their pattern of incidence. In urban areas, standpost coverage is significantly lower than private taps, whereas in rural areas the opposite is true. Those with no other alternative must resort to surface water of questionable quality—this amounts to 30 percent of the population overall and about 50 percent of the poorest. 43 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Table 5.2 Patterns of access for alternatives to piped water Percentage of households using Population weighted average alternative Piped water Standposts Well/Boreholes Surface water Vendors By time period (national) • Early 1990s 18 15 37 41 1 • Late 1990s 17 15 38 31 1 • Early 2000s 17 16 41 33 2 By location (latest available year) • Rural 4 10 43 41 1 • Urban 38 25 24 7 4 By quintile (latest available year) • First 0 6 44 49 1 • Second 3 11 46 39 1 • Third 7 13 42 34 2 • Fourth 18 20 35 23 2 • Fifth 46 23 20 7 2 By income group (latest available year) • Low-income countries 11 14 40 32 2 • Middle-income countries 44 22 13 18 1 By subregion (latest available year) • East Africa 10 15 30 43 0 • West Africa 12 13 52 18 4 • South Africa 29 17 32 20 0 • Central Africa 14 15 21 48 0 Source: AICD DHS/MICS Survey Database, 2007. In most countries there is a wide diversity of water sources in use (figure 5.1a). Typically, piped water, standposts, surface water, and vendors each account for well under 20 percent of the population. The balance is made up by boreholes, whose contribution varies widely from 20 to 80 percent of the population. The countries most reliant on boreholes are Burkina Faso, Chad, Malawi, Mali, Niger and Uganda. In Ethiopia, as much as two thirds of the population report surface water as their primary source. The overall average annual growth rate for alternative water services has been 1 percent for piped water, 3 percent for standposts, 3 percent for boreholes, -6 percent for vendors, and 0 percent for surface water (figure 5.1b). In most countries, the modal growth rate in standposts and boreholes is 0 to 4 percent per annum, which is slightly lower than the modal growth rate for piped water in the 4 to 8 percent range. This suggests that the share of piped water connections is increasing over time. While most countries have experienced very low or negative growth rates in the prevalence of surface water usage, in a handful of cases reliance on this form of water has been expanding rapidly. 44 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 5.1 Country frequency distribution for alternatives to piped water a. Current coverage level of alternatives b. Absolute average annual growth rate in prevalence of water alternatives, 1996–2005 Source: AICD DHS/MICS Survey Database, 2007. In order to examine the extent to Figure 5.2 Scatter plot of average annual growth rate of piped water which countries may be shifting service against alternatives between alternative forms of water service over time, a cross-plot is made between the average annual rate of growth of piped water and the average annual rate of growth of the alternatives, that is standposts, boreholes and surface water (see figure 5.2). The analysis groups the countries into a number of clusters. In the first cluster, which lies well below the 45 degree line, are found countries such as Cameroon, Chad, Mozambique, and Senegal that have rapid expansion of piped water but Source: AICD DHS/MICS Survey Database, 2007. that register substantially slower progress on standposts. In the second cluster, which lie well above the 45 degree lines, are found countries such as Guinea, Madagascar, Mali, Rwanda, and Zambia that register much faster expansion in standposts than in piped water, and Ethiopia, Lesotho, and Rwanda that registers much faster expansion in boreholes than in piped water. In the third cluster are found countries such as Burkina Faso, Ghana, Ethiopia that are registering simultaneous rapid expansion of piped water and standposts. In the third cluster, which lies close to the origin, are found countries such as Malawi, Nigeria, and Lesotho that are not experiencing rapid expansion in any area. Overall, there is significant correlation between the rate of expansion of piped water and that of standposts (0.6). On the other hand, there is a substantial negative correlation between the rate of expansion of wells/boreholes and piped water (–0.5) 45 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA On average across countries, the percentage of households without access to piped water in their dwelling or yard fell slightly from 86 percent in the first half of the 1990s to 82 percent in the second half of the 1990s, only to rise back to 87 percent in the first half of the 2000s. Fetching water from outside the home is an activity dominated by women and girl children. Blackden and Wodon (2006) compute that out of 6 million hours spent in Ghana in 1992 on fetching water; more than two-thirds was spent by women. Traveling outside the home to fetch water has a cost in time that might be spent on education or other productive purposes. Providing African households with reasonable access to water would bring significant gains in productivity, health, and welfare. Most households that lack private water connections live within one kilometer of their water source. In the case of urban households, the average distance is estimated to be just over 500 meters, while in the case of rural households the average distance is closer to one kilometer. Some 20 percent of urban households and 30 percent of rural households live more than one kilometer from their water source. Patterns of access vary from country to country, but, on average, a much higher proportion of urban households have ready access to water than do rural households (figure 5.3). For instance, 53 percent of rural households in Tanzania live more than two kilometers from their water source. On the other extreme are households in South Africa, Nigeria, and Madagascar, where fewer than 2 percent of rural households live more than two kilometers away. In Sierra Leone, no household has water in its dwelling, either in rural or urban areas. Even in urban areas, water can be far away. In urban Mauritania, for example, 66 percent of households live more than two kilometers away from their water source. In urban Ghana and Sierra Leone the corresponding figure is 53 percent. In comparison, less than 5 percent of households in urban areas in the Republic of Congo, the Democratic Republic of Congo, Ethiopia, Morocco, Niger, Nigeria, South Africa, Uganda, and Zambia live more than two kilometers from their water source. The household surveys allow us to Figure 5.3 Distance of households from water source in selected measure changes in the time households countries spend fetching water. Since 1990, the average time spent fetching water for household consumption has remained virtually unchanged at 45–50 minutes (roundtrip). In some countries more time is spent at the task. Households in Ethiopia, Mozambique, Tanzania, and Uganda, spend more than one hour each day fetching water for household consumption. In Ethiopia, Mozambique, Tanzania, and Uganda, moreover, the amount of time has increased over the years. These are also Source: AICD Expenditure Survey Database, 2007. countries where more than 90 percent of households fetch water from outside their dwelling. 46 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA In some countries, water vendors play a significant role in urban water supply, catering to those that lack any other alternative (figure 5.4).10 The presence of these operators signals a failure of the formal market to supply water. In several countries water vendors operating with tanker trucks or through small piped systems supply around 5 percent of the urban water market (Burkina Faso, Chad, Niger, Nigeria, and Tanzania). More generally, the percentage of the urban population covered by water vendors is highly skewed across countries. Whereas in two thirds of the countries surveyed, vendors accounted for less than one percent of the urban population; in a small minority of countries vendors account for more than 20 percent of the urban population. The two salient cases are Niger (21 percent) and Mauritania (32 percent). Comparing the situation in the 1990s with that in the early 2000s reveals that the market share of vendors has changed markedly in a number of countries. Thus, vendors’ market share has fallen substantially in Chad (27 to 16 percent) and Rwanda (3 to 0.1 percent). At the same time, vendors’ market share has increased substantially in Nigeria (8 to 11 percent) and Tanzania (0.3 to 6 percent) These findings suggest that vendors constitute a fairly flexible segment of the market that reacts quite rapidly to changes in broader market conditions, be they positive (for example emergence from conflict) or negative (for example pressures of urbanization). Figure 5.4 Dependence on water vendors in urban Africa, 1990–2005 a. Frequency distribution of urban population coverage 2000s b. Evolution of urban population coverage from 1990s to 2000s Source: AICD DHS/MICS Survey Database, 2007 In a study of 10 cities in Sub-Saharan Africa, Collingnon and Vezina (2000) find that $5–40 million is generated in each of these local markets, amounting to 1–3 percent of the cities’ total domestic product. Consumers of vendor water are not necessarily poor. The rich may also count on water vendors as their primary source or to supplement another source. However, the poor are more likely to depend on them, as they are least served by the piped water supply. Kariuki and Schwartz (2004) compare the prices charged by vendors to formal network prices. They show that small-scale piped network operators charge 1.5 times the formal network price, point sources charge 4.5 times, and mobile distributors can charge up to 12 times as much as the formal utility tariff. In a recent survey of Nairobi, Accra, and Dar es Salaam, McGranahan and others (2006) find that the price 10Small-scale service providers are also active in the electricity market, where they may be independent of the network, developing their own power, or dependent on the formal network, and reselling to consumers via a mini- grid, mobile distributors, or fixed-location vendor (Kariuki and Schwartz 2005). 47 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA of piped water ranges from $0.5 to $1.5 per cubic meter, whereas small water enterprises charge between $4 and $6. Further data on prices, quantities, and expenditures relating to small water enterprises are available in comprehensive case studies (Whittington 1989; Collignon and Vezina 2000); through social assessments linked to World Bank project preparation in cities in Lesotho, Mozambique, and Zambia; and in detailed studies commissioned by the U.K. Department for International Development on Ghana, Kenya, Sudan, and Tanzania. Interestingly, even though water vendors charge higher unit prices for water, those purchasing water from vendors do not necessarily spend more on buying water than those purchasing water from the public utility, but simply adjust the quantity consumed. Data from the AICD expenditure survey database for 10 countries in Africa reveal that payments to water vendors in many cases exceed amounts paid to the water utility (table 5.3). In Ghana the poorest quintile of households pays a very high price to vendors. Table 5.3 Ratio of payment of vendors to piped supply by consumer groups Congo, Cape Congo, Dem. Mauri- Mozam- Sierra Verde Rep. Rep. Ghana tania Morocco bique Nigeria Rwanda Leone By location National 0.6 0.9 1.1 2.3 0.7 0.8 0.3 0.8 0.5 0.3 Rural 1.1 1 1.2 8.7 2.2 1.9 0.1 0.4 0.6 0 Urban 0.5 0.9 1.1 1.8 0.5 0.7 0.5 1.2 0.7 0.4 By quintile First 1 0.9 1.1 27.5 2.6 1.2 0.4 0.7 n.a. 1 Second 0.9 0.8 1.1 5 1 1 0.4 0.7 0.9 0.4 Third 0.9 0.9 1.1 4.7 0.6 1.1 0.7 0.8 1.9 0.2 Fourth 0.6 0.9 1.1 3.6 0.8 0.8 0.4 1 1.5 0.2 Fifth 0.5 0.9 1.1 1.5 0.7 0.8 0.3 0.8 0.7 0.5 Source: AICD Expenditure Survey Database, 2007. Note: In Rwanda, there is no payment for piped water in quintile 1, therefore the ratio between piped water and vendor water payments is undefined. n.a. = not applicable. Alternatives to modern sanitation services Traditional pit latrines are by far the most prevalent form of sanitation in Africa accounting for 51 percent of the population; a share that remains remarkably constant between urban and rural areas and across the socioeconomic spectrum of households. In Malawi, Tanzania, and Uganda as much as 80 percent of the population is served by traditional pit latrines. The second most prevalent situation is to have no sanitation at all (31 percent), although these cases are heavily skewed towards rural areas (41 percent), and the bottom two quintiles of the expenditure distribution (40 to 50 percent). In urban areas, 92 percent of the population has some form of sanitation, however rudimentary (table 5.4). In Benin, Burkina Faso, Chad, Niger, and Togo, more than 80 percent of the rural population lacks any form of sanitation. At the other end of the spectrum are Republic of Congo, Comoros, Cameroon, and Rwanda, where fewer than 10 percent of any households are without access to some form of sanitation. In most countries, the share of the population served by a flush toilet or an improved latrine is well below 20 48 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA percent of the total (figure 5.5a). The difference is made up to varying degrees by traditional pit latrines or no sanitation facilities. The overall average annual growth Table 5.4 Patterns of access flush toilet and alternatives rate for sanitation services has been 7 Percentage of households, population weighted average percent for flush toilets, 11 percent for VIP/Chemical/ Traditional VIP/SAN PLAT/Chemical toilets, -0.3 By time period (national) Flush toilet SANPLAT latrine No facility percent for traditional pit latrines, and Early 1990s 9 6 50 46 0.3 percent for no facility. In a number Late 1990s 9 7 49 37 of countries (Ghana, Kenya and Early 2000s 10 9 52 34 Tanzania), the proportion of households By location unserved by modern sanitation actually Rural 2 5 52 41 rose in the past decade (figure 5.5b). Urban 28 14 49 8 However, a more typical situation for By quintile most countries was to be experiencing First 0 0 50 49 modest average annual growth rates of 0 Second 1 2 54 41 to 4 percent for traditional pit latrines. A Third 4 6 57 32 few outliers are experiencing double Fourth 12 11 54 23 digit growth in the expansion of Fifth 34 19 40 6 improved latrines (Benin, Burkina Faso, By country income group Kenya, Mali and Zambia). Low 7 8 52 33 Middle 33 8 41 13 In order to examine the extent to By subregion which countries may be shifting between East Africa 4 4 56 35 alternative forms of sanitation over time, West Africa 12 8 48 33 a cross-plot is made between the average South Africa 23 11 36 28 annual rate of growth of flush toilets and Central Africa 3 13 65 18 the average annual rate of growth of the Source: AICD DHS/MICS Survey Database, 2007. alternatives; that is VIP and traditional latrines (figure 5.6). The analysis groups the countries into three main clusters. In the first cluster, which lies well below the 45 degree line, are found countries such as Chad, Mali and Senegal that have rapid expansion of flush toilets but that register substantially slower progress on alternative services. In the second cluster, which lies well above the 45 degree lines, are found countries such as Kenya, Malawi, Mozambique, Rwanda, Tanzania, and Zambia that register much faster expansion in alternative forms of sanitation than in flush toilets. Finally, countries such as Madagascar sit close to the origin on the 45 degree line, indicating that there has not been much progress either in modern or alternative forms of sanitation. Overall, however, there is slightly negative correlation between the rate of expansion of flush toilets and that of improved latrines (-0.1) and between that of VIP and traditional latrines (-0.5). On the other hand, there is a substantial positive correlation between the rate of expansion of flush toilets and traditional latrines (0.3). It is very striking that this classification of countries is almost identical to that found for the water sector in figure 5.2, suggesting that the balance of effort across modern and more basic alternatives is common across services in any given country. 49 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 5.5 Country frequency distribution of alternative forms of sanitation a. Current coverage level of alternatives b. Absolute average annual growth rate in prevalence of alternatives, 1996–2005 Source: AICD DHS/MICS Survey Database, 2007. Figure 5.6 Scatter plot of average annual growth rate of flush toilet service against alternatives Source: AICD DHS/MICS Survey Database, 2007. Sharing of water and sanitation facilities Sharing of water and sanitation facilities among multiple families is common in urban areas. At least16 percent of households share their water supply facilities with other households, while more than 40 percent typically share their toilet facilities (figure 5.7). The household surveys focus only on formal communal provision through standposts and these figures underestimate the informal sharing of installations through household reselling which is now considered as significant source of primary water supply in Africa. In addition, the sharing of water in public or communal wells/boreholes is also not captured as it is difficult to disaggregate public from private wells/boreholes. Indeed, in Benin, Burkina 50 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Faso, Republic of Congo, Ghana, Guinea, and Madagascar, more than half of households share toilet facilities, while in Lesotho about half share their source of water. In Ghana and other countries of West Figure 5.7 Current country frequency distribution for Africa compound housing is a common way of percentage of population sharing water and toilet facilities living. In Ghana, for example, 80 percent of the population shares their dwelling with other families. Under such circumstances it is common for water and sanitation facilities to be shared. Shared forms include a neighbor’s yard tap, community or public taps, standposts, public wells and boreholes, public hand pumps, and community toilets. Sharing with other households not only implies that household members lose time in accessing these facilities, but also that they Source: AICD DHS/MICS Survey Database, 2007. may pay more than in the formal network. Owners and operators of yard taps, standposts, and toilets may charge exorbitant amounts that depend on several factors, including distance to and quality of alternate sources and the availability and cost to connecting to the formal network. Alternative fuels for cooking and lighting The majority of Africa’s population cooks with traditional solid fuels and relies on candles or kerosene for lighting. For cooking, around 83 percent of the population relies on traditional solid fuels, such as wood and charcoal. Though this trend is prevalent across the population, it is more pronounced among the poor. In Ethiopia, Guinea, Lesotho, Mozambique, Tanzania, Zambia, Zimbabwe, all the surveyed households in the bottom quintile reported using wood or charcoal for cooking. In all other countries except South Africa, more than 90 percent of households in the bottom quintile depend on dirty fuels to meet their cooking needs. The situation is somewhat better in Gabon and South Africa (where some 30 percent of people still use traditional fuels), as well as Lesotho and Senegal (around 60 percent). Elsewhere, around 90 percent of the population cooks with traditional fuels. Although reliance on traditional fuels is significantly higher in rural areas (close to 100 percent of households), their use in urban areas remains quite high (more than 70 percent of households in many cases). Kerosene is the most common lighting fuel in countries with sparse electricity coverage. In fact, for lighting purposes, kerosene is the only fuel that can be considered as an alternative to electricity. There are a few exceptions. In Guinea Bissau, 30 percent of households depend on natural gas or liquefied propane gas for lighting. Not surprisingly, dependence on kerosene is prevalent in rural areas and in the poorest households. An overwhelming proportion of rural households in Kenya, Sierra Leone, Tanzania, and Uganda use it, as do more than three-quarters of the poorest households in Benin, Burkina Faso, Cameroon, Republic of Congo, Ghana, Kenya, Madagascar, Niger, Senegal, Sierra Leone, Tanzania, and Uganda. In some low-income countries, poor households also use charcoal and wood for lighting. This practice is found in Burundi, Democratic Republic of Congo, Ethiopia, Mozambique, and Rwanda. The 51 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA use of kerosene is not limited to the poor population. In Benin, Chad, Ethiopia, Kenya, Niger, Tanzania, and Uganda, more than 60 percent of the richest households use it. Table 5.5 Patterns of access to electricity and alternatives Percentage of households, population weighted average Electricity Kerosene LPG Wood/charcoal (lighting) (cooking) (cooking) (cooking) By time period (national) Early 1990s 23 Late 1990s 28 9 3 80 Early 2000s 31 8 3 74 By location Rural 12 3 1 93 Urban 71 21 8 58 By quintile First 4 1 0 98 Second 14 3 0 95 Third 20 4 2 90 Fourth 38 9 4 79 Fifth 72 23 13 49 By country income group Low 26 7 3 87 Middle 55 15 8 52 By subregion East Africa 12 3 2 92 West Africa 38 13 4 80 South Africa 33 8 3 71 Central Africa 31 4 14 80 Source: AICD DHS/MICS Survey Database, 2007. The overall average annual growth rate for energy services has been 5 percent for electricity, -5 percent for kerosene, 4 percent for LPG, and 2 percent for wood or charcoal. Africans primarily use wood and charcoal for cooking and most of the countries are growing between 0-4 percent in the use of these fuels. The only other alternatives for cooking are kerosene and LPG which are sparingly used. The only countries where there is kerosene used for cooking by more than 10 percent of households is in Lesotho and Nigeria. LPG, which is a clean fuel, is used for cooking by 60 percent of the households in Gabon. The other countries where fuel is relatively important are Senegal and Mauritania. Most of the countries have exhibited a growth rate of less than 4 percent in LPG coverage in the past ten years. 52 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Figure 5.8 Country frequency distribution of alternative forms of energy (a) Current coverage level of different alternatives (b) Absolute average annual growth rate 1996/05 Source: AICD DHS/MICS Survey Database, 2007. Alternatives to modern rubbish disposal More than half of the African households dump, burn, or bury their household waste. Only 30 percent of urban households have access to an advanced waste collection option such as collection by the government, a private company, or a nongovernmental organization (table 5.6). Only 1 percent of rural Africans have their household waste collected by a formal authority. In rural areas, burning or burying is the most dominant mode of disposal, while in urban areas dumping is another widely used method (figure 5.9). Figure 5.9 Current country frequency distribution for modes of rubbish disposal a. Urban b. Rural Source: AICD Expenditure Survey Database, 2007. In Ghana, Rwanda, Madagascar, and Sierra Leone, more than 90 percent of the households dump their household garbage. Only in Morocco did more than half of the households surveyed report that their 53 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA waste was collected. In urban Morocco, waste is collected from 86 percent of households. In urban Chad, Cameroon, Cote d’Ivoire, Republic of Congo, Mauritania, and Nigeria waste is collected from more than one-quarter of households. Not surprisingly, the situation is much worse in rural Africa. There is no collection system in rural Congo, Rwanda, or Ghana, where everyone dumps their household garbage. More than 90 percent of households in rural Benin, Cameroon, Cote d’Ivoire, Mauritania, Madagascar, and Sierra Leone dispose of their household waste in informal, unregulated dumps. Collected Pit/Heap Burned or buried By location Rural 1 29 57 Urban 30 26 40 By quintile First 7 27 55 Second 7 29 54 Third 8 29 52 Fourth 11 29 50 Fifth 18 28 46 By country income group Low 8 31 50 Middle 37 0 59 By subregion East Africa 3 22 52 West Africa 8 12 23 South Africa 5 36 52 Central Africa 6 36 47 Alternatives to personal telephones The main alternative to owing a telephone remains using public telephone. There is no direct survey evidence on the use of public telephones. However, the limited evidence available suggests that use of public telephones is not particularly widespread and confined mainly to landlines. This insight comes from identifying households that report telecommunications expenditure but do not report owning a telephone. On this basis, it was found that around 6 percent of the population reported expenditure on a landline telephone, even though they did not own one, while less than 1 percent of the population reported expenditure on a cellular telephone even though they did not own one. While this indirect evidence cannot be regarded as firm, it does suggest that the percentage of the population using public telephone services is much lower than the percentage of the population owning a private telephone. It also suggests that public telephone services are being provided primarily through landline telephones, as opposed to retailing of air time on mobile networks. Nevertheless, at present there are very few household surveys that incorporate questions on cellular telephony, so the evidence on this remains very partial. In conclusion, the vast majority of African households relies and for the immediate future will continue to rely on second best alternatives to modern infrastructure services. The growth rate of second best alternatives has been at least as high as that for modern infrastructure services. Thus, while the 54 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA population with access to piped water has been growing at 1 percent per year, the population with access to standposts has been growing at 3 percent per year. In the case of sanitation, the overall average annual growth rate of the population with access has been at 7 percent for flush toilet and 12 percent for VIP latrines. Extending services through shared connections and public access What is striking from the household surveys is the extent to which formal or informal sharing of connections and public forms of access provide a way of leveraging the benefits of the network across a broader section of the population. For one thing, the percentage of the urban population living within range of infrastructure networks is substantially higher than the coverage rate for private household connections (table 5.7). The difference between these two rates provides a first order estimate of the potential for communal use of network services that ranges between 20 and 40 percent depending on the service. Thus, in the case of water, at least 25 percent of urban households share public standposts, compared with a potential level of 35 percent. While it is well known that household resellers have emerged as a significant alternate provider of water supply, it is not possible to identify the share of households that depend on household resellers from the household surveys, because that share is hidden among households who use piped water supply as their primary water source. In the case of sanitation, 50 percent of households report sharing sanitation facilities, compared with a potential level of 38. In the case of electricity and telephones, 22 percent of urban households live in electrified areas even if they do lack electricity connections, as such they can (or at least potentially could) benefit from services such as public street lighting. In the case of telephony, 48 percent of the urban population lives in areas where telephone service is available, and therefore can (or at least potentially could) benefit from public telephony or resale of telephone services. Table 5.7 Importance of communal modes of access in urban areas Percentage of households Rate for potential Coverage rate communal use (C) Access rate (A) (A – C) Evidence of sharing Electricity 71 93 22 None; no data on street lighting Piped water 38 73 35 25 percent of urban households share their connection through public tap or standposts Flush toilet 31 69 38 50 percent of urban households share their facilities Telephone 20 68 48 Evidence on public telephone usage too sketchy to provide firm conclusions. 55 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA 6 Conclusions and policy implications Africa lags well behind other developing regions in infrastructure access; the limited gains of the 1990s appear to have prevailed in the early 2000s. The overall picture on access to infrastructure services in Africa is a very sobering one. There is clear evidence that many countries are failing to expand services fast enough to keep ahead of rapid demographic growth and even faster urbanization. As a result, if present trends continue, Africa is likely to lag even further behind other developing regions, and universal access will be more than 50 years away in many countries. However, the wide diversity of performance across countries suggests that there are valuable lessons to be learned. These aggregate statistics conceal substantial variation in performance across countries; even within the low and middle income brackets. A significant number of countries have succeeded in expanding the population served with water, electricity and sanitation by an annual average of 5-10 percent, which is fast enough to make substantial coverage gains within a reasonable time frame. Further investigation is warranted to explain what determines the superior performance of these countries. Moreover, the very positive experience of cellular telephony in the last decade highlights the possibility of making rapid progress under the right circumstances. The rapid expansion of cellular telephony to come from almost nowhere in the late 1990s to reaching a substantial number of African households as of today provides considerable food for thought. Much of the explanation lies in factors that are unique to cellular technology, including the relatively low fixed investments, the novel and high value nature of the service, and the commercial innovation in terms of low entry charges and prepayment facilities. While not all of these things can be directly applied to other infrastructure services, they nonetheless provide pointers in terms of directions for change that could help to support faster access in other services. These include lowering capital costs, reducing up-front connection charges, and providing alternative more flexible payment methods to the traditional monthly bill. The finding that a significant share of the un-served urban population lives close to infrastructure networks suggests the need for greater efforts on the demand side. The low take-up rate of infrastructure services in urban areas indicates that rolling out networks is a necessary but far from sufficient condition for achieving higher access. It also suggests that the financial and economic return of network rollout in African cities has been much lower than might be expected, leaving a relatively small customer base to cover the fixed costs of a relatively expensive network. It is therefore necessary to improve our understanding of the low uptake phenomenon, and second to accompany network rollout with demand- side measures explicitly designed to reduce uptake barriers. An important issue to explore is the magnitude of connection charges, which in Africa tend to be high relative to household incomes. Urban development factors, such as insecure household tenure, may also be playing an important role in blocking infrastructure uptake. The fact that the majority of the African population rely either on self-supply or simply do without services altogether, also has important implications. Although formal electricity and water service providers play an important role in urban areas—reaching about half the population with electricity and around three quarters with a combination of private and public taps—overall they reach only about 30 56 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA percent of the population. The remaining two-thirds of African households either makes do without safe water, sanitation and lighting or supply themselves from boreholes, traditional pit latrines and candles. Given the slow rates of coverage growth, this situation is likely to persist for some time to come. However, most policy efforts focus on improving the performance and expanding the ambit of the formal infrastructure providers. While this is valid, it would also be important to consider what measures, if any, could be taken to improve the lot of this large segment of the population pending the expansion of modern infrastructure services. “Second best options� such as stand posts and improved latrines still have a long way to go in reaching a substantial share of the population. The coverage of “second best options� for water and sanitation is surprisingly low, and remains relatively skewed toward the upper income echelons. There is clearly substantial potential for these services to be expanded further. The low levels of household income represent an absolute constraint on the rate of expansion. The average African household survives on a modest budget of US$180 per month, of which more than half is absorbed by food expenditures. Typically, about 20 percent of household budgets (or 40 percent of the household non-food budget) is spent on infrastructure services, mainly power and transport. In absolute terms, this translates to little more than US$30 per month on all utilities and transport. Utility bills of the order of US$6 per month for a service such as water or power is therefore likely to be affordable for most households in all but the poorest countries. However, utility bills of the order of US$10 per month start to become unaffordable for a substantial share of the population. At the same time, connection charges of the order of US$100-200 would clearly be beyond the reach of all but the wealthiest households. Finally, the results presented in this paper perhaps raise as many questions as they answer, demanding further investigation. While the results reported above provide many insights in to the nature of household usage of infrastructure services in Africa, they also raise many questions that cannot be immediately answered. Why is the variance in access so high across countries, even within the same income band? Why is the variance in access so high across services, and how can a new service such as cellular telephony made such major inroads in so brief a period? Why do many households fail to connect to modern infrastructure services, even when these are physically close at hand? In order to find answers to many of these questions, it is necessary to dig deeper into the supply side, including the institutional organization and the performance of the service providers in each of the countries. Such an analysis is already underway in other components of the broader Africa Infrastructure Country Diagnostic study. When all the work has been completed, it will be possible to revisit the findings of this study and make greater sense of the variations that have been observed. 57 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA References Anand, P.B. 2006. “An assessment of progress with respect to water and sanitation: Legacy, synergy, complacency, or policy.� UN-WIDER Research Paper 2006/01. Angel-Urdinola, D., M. Cosgrove-Davies, and Q. Wodon. 2006. “Rwanda: Electricity Tariff Reform.� In A. Coudouel, A. Dani, and S. Paternostro (eds.), Poverty and Social Impact Analysis of Reforms Lessons and Examples from Implementation. Washington, DC: World Bank. Bayliss, K., and T. McKinley. 2007. “Privatizing basic utilities in Sub-Saharan Africa: The MDG impact.� Research Brief No. 3. International Poverty Center, Brasilia Blackden, C. M., and Q. Wodon. 2006. “Gender, Time Use, and Poverty in Sub-Saharan Africa.� World Bank Working Paper 73. The World Bank, Washington DC Briceno-Garmendia, C., and I. Klytchnikova. 2006. “What do household surveys tell us about infrastructure and what should they tell?� Processed, World Bank, Washington DC. Brook, P., and W. Smith. 2001. “Improving access to infrastructure services to the poor: Institutional and policy responses.� Background Paper for World Development Report 2001, World Bank, Washington DC Calderon, C. and L. Serven. 2004. The effects of infrastructure development on growth and income distribution. Policy Research Working Paper 3400, Washington DC: World Bank Collignon, B., and M. Vezina. 2000. “Independence water and sanitation providers in African cities: Full report of a ten-country study.� Water and Sanitation Program. Washington DC. Diallo A.B., and Q. Wodon. 2007a. "Growth and welfare under demographic transitions and economies of scale.� Forthcoming Demographic Research Diallo, A., and Q. Wodon. 2007b. " Demographic Transition Towards Smaller Household Sizes and Basic Infrastructure Needs in Developing Countries�, Economics Bulletin, 15(11): 1–11 Diallo, A., and Q. Wodon. 2005. “A note on access to network-based infrastructure services in Africa: Benefit and Marginal Incidence Analysis.� Processed, World Bank, Washington DC Estache, A. 2006. “Africa’s infrastructure: Challenges and opportunities.� Paper presented at the high- level seminar: Realizing the Potential for Profitable Investment in Africa, Tunis, Tunisia ———. 2004. “Emerging infrastructure policy issues in developing countries - a survey of the recent economic literature.� Policy Research Working Paper 3442, World Bank, Washington DC. Estache, A., and Q. Wodon. 2007. Infrastructure and poverty in Sub-Saharan Africa. Forthcoming, Directions in Development Series, Processed, World Bank Fay, M. 2000. “Financing the Future: Infrastructure Needs in Latin America, 2000-05.� Policy Research Working Paper No.2545, World Bank, Washington DC. 58 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Fay, M., and T. Yepes. 2003. “Investing in Infrastructure: What is needed from 2000-2010.� Policy Research Working Paper 3102, World Bank, Washington DC Foster, V. and M.C. Araujo. 2004. “Does infrastructure reform work for the poor? A case study from Guatemala.� Policy Research Working Paper 3185, World Bank, Washington DC. Foster, V., and T. Yepes. 2006. “Is cost recovery a feasible objective for water and electricity? The Latin American experience.� Policy Research Working Paper 3943. The World Bank, Washington DC. Hammond, A., W.J. Kramer, R.S. Katz, J.T. Tran, and C. Walker. 2007. “The Next four billion: Market size and business strategy for the bottom of the pyramid.� World Resources Institute, Washington DC Hoogeveen, J.G. 2007. Tanzania: Electricity tariffs for the rich. Policy Notes, Africa Region, World Bank, Washington DC. JMP (Joint Monitoring Program). 2006. Meeting the MDG drinking water and sanitation target: The urban and rural challenge of the decade. WHO and UNICEF. Geneva and New York. Kakwani, N.C., A. Wagstaff, and E. Van Doorslaer. 2000. “Socioeconomic inequalities in child mortality: Comparisons across nine developing countries.� Bulletin of the World Health Organization. Vol. 78(1): 19-29. Kariuki, M and J. Schwartz, 2005. “Small scale private service providers of water and electricity: A review of incidence, structure, pricing and operating characteristics.� Policy Research Working Paper 3727, World Bank, Washington DC Komives, K. D. Whittington, and X. Wu. 1999. “Infrastructure coverage and the poor: A global perspective.� Policy Research Working Paper 2551, World Bank, Washington DC. Komives, K., V. Foster, J. Halpern, and Q. Wodon, with support from R. Abdullah (2005), Water, Electricity, and the Poor: Who Benefits from Utility Subsidies? World Bank, Washington, DC. Lobo, A., V. Foster,, and J. Halpern. 2000. “Better household surveys for better design of infrastructure subsidies.� Viewpoint Note 213, World Bank, Washington DC McGranahan, G, C. Njuru, M. Albu, M. Smith, and D. Mitlin. 2006. “How small water enterprises can contribute to MDGs. Evidence from Dar es Salaam, Nairobi, Khartoum, and Accra.� Loughborough University, Leichestershire. Milanovic, 2003. “Is inequality in Africa really different?� Policy Research Working Paper 3169, World Bank, Washington DC OECD, 2006. “Promoting pro-poor growth: Infrastructure.� OECD, Paris. Prud’homme, R. 2004. “Infrastructure and Development.� Annual Bank Conference on Development Economics, Washington DC. Whittington, D, D. T. Lauria, D. A. Okun, and X. Mu 1989. "Water Vending Activities in Developing Countries: A Case Study of Ukunda, Kenya." International Journal of Water Resources Development, 59 ACCESS, AFFORDABILITY, AND ALTERNATIVES: MODERN INFRASTRUCTURE SERVICES IN AFRICA Wodon, Q., 2007a, Editor, “Electricity Tariffs and the Poor: Case Studies from Africa.� World Bank, Washington, DC. ———. 2007b, Editor, “Water Tariffs, Alternative Service Providers and the Poor: Case Studies from Africa.� World Bank, Washington, DC. 60