WPS4881 P olicy R eseaRch W oRking P aPeR 4881 Is Low Coverage of Modern Infrastructure Services in African Cities due to Lack of Demand or Lack of Supply? Quentin Wodon Sudeshna Banerjee Amadou Bassirou Diallo Vivien Foster The World Bank Human Development Network Development Dialogue on Values and Ethics Unit & Africa Region Vice Presidency Sustainable Development Department March 2009 Policy ReseaRch WoRking PaPeR 4881 Abstract A majority of sub-Saharan Africa's population is not as opposed to supply-side issues are fairly different, connected to electricity and piped water networks, and it is important to try to measure the contributions of even in urban areas coverage is low. Lack of network both types of factors in preventing better coverage of coverage may be due to demand or supply-side factors. infrastructure services in the population. This paper Some households may live in areas where access to piped shows how this can be done empirically using household water and electricity is feasible, but may not be able to survey data and provides results on the magnitude of pay for those services. Other households may be able to both types of factors in explaining the coverage deficit of afford the services, but may live too far from the electric piped water and electricity services in urban areas for a line or water pipe to have a choice to be connected to it. large sample of African countries. Given that the policy options for dealing with demand This paper--a product of the Development Dialogue on Values and Ethics in the Human Development Network and of the Sustainable Development Department in the Africa Region Vice Presidency--is part of a larger effort in the Network and Region to document the access to, and affordability of basic infrastructure services. Policy Research Working Papers are also posted on the Web at http://econ.worldbank.org. The corresponding author may be contacted at sbanerjee@ worldbank.org. The Policy Research Working Paper Series disseminates the findings of work in progress to encourage the exchange of ideas about development issues. An objective of the series is to get the findings out quickly, even if the presentations are less than fully polished. The papers carry the names of the authors and should be cited accordingly. The findings, interpretations, and conclusions expressed in this paper are entirely those of the authors. They do not necessarily represent the views of the International Bank for Reconstruction and Development/World Bank and its affiliated organizations, or those of the Executive Directors of the World Bank or the governments they represent. Produced by the Research Support Team Is Low Coverage of Modern Infrastructure Services in African Cities due to Lack of Demand or Lack of Supply? Quentin Wodon, Sudeshna Banerjee, Amadou Bassirou Diallo, and Vivien Foster Keywords: infrastructure; electricity; water; Africa JEL codes: O12, R21, R31 1. Introduction Many households are not connected to network-based infrastructure services such as electricity and piped water in sub-Saharan Africa (Komives et al., 2003; Anand, 2006; Banerjee et al., 2007), even in urban areas (Clarke and Wallsten, 2003). Yet it is not clear whether this is due mainly to demand-side or supply-side factors. On the demand-side, because most of the population is poor or near-poor, some households may simply not be able to afford to pay for piped water and electricity services even when connection to the network is feasible because the households live near an electric line or a water pipe. The lack of affordability of the service, or more generally of demand for the service, may be due to different reasons. A key reason could be that tariffs are too high for the households, or that connection charges are too high for getting access to the network (Franceys, 2005; Kayaga and Franceys, 2007). Other demand-side issues may relate to lack of land titles or illegal tenure, which makes it difficult for the utility company to accept the household as a client. Still another demand-side issue (from the point of view of the household) could be related to poor quality of service, so that some households may prefer to use alternative ways of satisfying their water and electricity needs rather than by using a network connection, at least when such alternatives such as a private well, a neighbor's tap, or a public stand-post are available. On the supply-side, many households simply live in urban neighborhoods that do not have access to piped water or electricity. In addition, even when there is access somewhere in the neighborhood, many households may still live too far from the electric line or water pipe to have a chance to be connected to it. In addition, even if some households would like to be connected, there may be a lack of capacity within the utility company to provide such connections, for example due to lack of manpower or other resources (on ways to conduct an analysis of the investments needed in infrastructure, see for example Fay and Yepes, 2003). In some cases, a policy may be in place in the utility company not to extend the network, because the utility already faces capacity constraints to properly serve existing consumers. Indeed, in many sub-Saharan countries, power and water cuts are frequent, as the generation and production capacity of the utilities is limited and insufficient to meet the existing demand. There may also be financial factors affecting the capacity or willingness of the utilities to expand their network, especially if tariffs are too low to permit capital cost recovery. As noted among others by Estache et al. (2002; see also Estache, 2004; Komives et al., 2005; and Estache and Wodon, forthcoming), the policies that need to be implemented in order to promote higher coverage rates are very different depending on the nature of the obstacles to increase coverage. If the main obstacle is a lack of demand due for example to a lack of affordability, utilities or governments may consider implementing special tariffs or subsidies for the poor, whether this is done for reducing the cost of the consumption of households once they are connected, or for reducing the cost of connecting itself. If the main problem is a lack of supply, the first line of answer lies in finding the necessary resources in order to expand the network to those who do not have access. Given that the policy options for dealing with demand as opposed to supply-side issues are fairly different, it is important to try to measure the contributions of both demand and supply-side obstacles to better coverage of infrastructure services. The aim of this paper is to show how this can be done empirically in a simple way using household survey data. The importance of assessing the role of demand as opposed to supply-side issues has been recognized by Foster and Araujo (2004, hereafter F&A) in their study of the impact of infrastructure reforms on the poor in Guatemala. These authors proposed a nice and simple statistical method for assessing the contribution of pure demand-side problems, pure supply-side problems, and combined demand and supply-side problems to coverage deficits. If a household living in an area with access to piped water or electricity service was not connected, this was taken as a sign that the service was not affordable for the household (pure demand-side problem). In practice, the authors assessed whether households lived in an area with access simply by checking if any other household living in the same primary sampling unit of the survey had access. Indeed, household survey samples rely on geographically defined primary sampling units which tend to be well delimited areas, especially in an urban setting. To the extent that the primary sampling units in urban areas are small (about 15-20 households per primary sampling units who tend to live in specific neighborhoods), access by one household in the primary sampling unit could be considered as indicating potential access for all the households in that primary sampling unit. F&A then defined the magnitude of supply-side problems as the part of the lack of coverage that was not due to the pure demand-side problem mentioned above. In addition, they decomposed supply-side problems into two components. The authors noted that even if there were access to the service in neighborhoods currently without access, some households would still not connect to the network. They therefore argued that in areas without access, there was for some households a combined or mixed problem with both demand and supply-side problems. Next, for those households who would probably connect to the network if there were access in their neighborhood to the service, the authors argued that there was a genuine pure supply-side problem. Overall, the authors thus decomposed the lack of coverage of the network in the sum of a pure demand-side problem, a pure supply-side problem, and a combined demand and supply-side problem. Others, including Angel-Urdinola et al. (2006), Angel-Urdinola and Wodon (2007) and Komives et al. (2005; forthcoming) have expanded on the work of F&A in order to analyze factors determining not only who benefits or not from a connection to the network, but also who benefits (or is likely to benefit) from various connection or consumption subsidies for modern infrastructure services. However, a weakness with the simple statistical approach used by F&A lies in the fact that there are limitations in the surveys used to assess empirically the magnitude of demand-side and supply-side problems, and that this may lead to biases in the estimates of demand as opposed to supply-side problems. As already mentioned, some households may live in an area where there is access to the service, but may still be located too far from the electric line or water pipe to be able to be connected (or perhaps the capacity of the electric line or water pipe may be designed to support a specific and limited number of households). Under the simple empirical procedure for estimating demand-side and supply-side problems proposed by F&A, these households would be considered as suffering from a demand-side problem, while the true nature of the issue may be a supply-side constraint. To some extent, this type of biases can be dealt with by using regression techniques. In this paper we suggest how this can be done, and we show that using an econometric as opposed to a statistical approach to the estimation can make a significant difference in the results. The rest of the paper is structured as follows. In section 2, we describe and formalize in simple mathematical notations the methodology used by F&A for assessing the relative role of demand and supply-side problems to explain lack of coverage of modern infrastructure services. Results obtained with this methodology for African countries in the case of urban coverage of piped water and electricity are then provided. The next section presents our alternative econometric approach to assessing the magnitude of demand and supply-side constraints to coverage, as well as the results obtained from this alternative method. A conclusion follows. 2. Statistical approach In this section, we start by presenting in mathematical notation the approach proposed by F&A for assessing demand and supply-side problems limiting coverage of network services. Denote by C the percentage coverage or connection rate of a service in the population. This is the number of households using the service divided by the total number of households (with appropriate survey-based household weights). Next define the access rate (A) as the number of households living in communities or primary sampling units where service is available divided by the total number of households. Finally, denote by U the take-up or hook-up rate which is the number of households actually using the service (i.e., connected to the network) divided by the number of households living in communities where service is available. The coverage rate is the product of the access and take-up rates (C=AxU). The share of the population not served by the network is 1-C. The objective is to assess whether the unserved population is not served due to a demand-side problem (the service is available, but not taken up by the households, probably because it is not affordable, but perhaps also because it is of low quality) or a supply-side problem (the service is simply not available). F&A define the pure demand-side gap (PDSG) as: PDSG = A - C = A × (1 - U ) (1) This definition implies that when there is access in the areas where the households live, if a household does not take-up the service, it is symptomatic of a demand issue. Thus, lack of demand is responsible for all of the difference between the neighborhood access rate and the actual coverage rate. Next, the authors define the supply-side gap as follows: SSG = (1 - C ) - PDSG = (1 - A × U ) - A × (1 - U ) = 1 - A (2) In other words, the supply gap is the difference between the neighborhood access rate and the coverage rate. Said differently, the sum of the pure demand-side gap, the supply-side gap, and the coverage rate is equal to one: PDSG + SSG + C = 1 (3) However, in areas that are not covered by the network, and are responsible for the supply gap above, it is likely that even if supply were available, some households would not take up the service due to affordability issues. If one assumes that the take-up rate in non-served areas would be similar to the take-up rate in areas where there is service now, the additional coverage that we would obtain by providing access to these areas would be equal to the supply-side gap times the take-up rate where there is access. This is defined as the pure supply-side gap: PSSG = SSG × U = (1 - A) × U (4) The difference between the pure supply-side gap and the supply-side gap can then be deemed to represent a combined demand and supply-side gap, since first there is no access to the service, and second even if there were access, some households would not be connected. F&A defined this as the mixed demand and supply-side gap, defined as follows: MDSSG = SSG × (1 - U ) (5) Given the above definitions, the proportion of the deficit in coverage that is attributed to demand-side factors is defined as the ratio of the pure demand-side gap to the unserved population. The proportion of deficit attributable to supply-side factors is the ratio of the pure supply-side gap divided by the unserved population. Finally, the proportion of deficit attributable to both demand and supply-side factors is the ratio of the mixed demand and supply- side gap divided by the unserved population. The sum of the three proportions is equal to one. The results from the decomposition are presented in tables 1 and 2 for urban areas in sub- Saharan African countries (it is not as useful to do the same work for rural areas, because access is very limited there in most countries, so that lack of coverage is principally a supply-side issue). The data used are from the latest Demographic and Health Survey (DHS) completed in each of the countries. Most of the countries have data after the year 2000. A household is deemed to have access to piped water or electricity if the household lives in an area (which is the primary sampling unit of the survey to which the household belongs) where at least one household has access. We discuss here the Africa averages, leaving the discussion of country- specific results for later. All Africa averages are provided both with population weights (in which case countries such as Nigeria play a larger role due to their larger population), and without weights. The data suggest that access at the neighborhood level is fairly widespread for both water (73 percent of households have access, see table 1, and this increases to 79 percent when no population weights are used) and electricity (93 percent of households have access, see table 2, and this is slightly reduced to 89 percent without weights) in African cities. Take-up rates are lower, at 48 percent for piped water (49 percent without weights), and 75 percent for electricity (61 percent without weights). This means that the coverage rate for piped water on average is 38 percent (41 percent without weights), and for electricity it is a much higher 71 percent (56 percent without weights). Conversely, the share of households not currently served is 62 percent for piped water (59 percent without weights), and 29 percent for electricity (44 percent without weights). The proportion of the deficit in coverage attributable to demand-side factors is large for piped water, at 59 percent on average for the region when countries are population-weighted, and at 68 percent when we use a straight average for all countries. For electricity, the corresponding figures are 79 percent, both with and without country population weights. The proportion of the deficit in coverage that is attributable only to supply-side factors is much lower, at 15 percent to 18 percent for piped water depending on whether country weights are used, and at 12 percent to 15 percent for electricity. The combined demand and supply-side problems account for 18 to 23 percent of the coverage deficit for piped water, and 6 to 9 percent for electricity on average for all the countries in the sample. Clearly, these results suggest that demand-side factors may be much larger than supply-side factors in explaining lack of infrastructure coverage in African cities. 3 Econometric approach As mentioned in the introduction, a key weakness of statistical approach presented in the previous section is that all households not connecting to the network where there is access are assumed to suffer from a demand-side problem, which may lead to an overestimation of the proportion of deficit coverage that is attributed to demand-side factors. In this section, we propose an alternative econometric method to try to better identify demand and supply-side problems. The idea is simple. We estimate for each country a regression of the determinants of the take-up of the household as a function of the following variables: a set of dummies for the quintile of wealth to which the household belongs, and the leave-out take-up rate in the primary sampling unit where the household lives. The index of wealth is estimated using factor analysis because we do not have household income or expenditure data in the DHS. The variables used for the factor analysis are allowed to differ between countries depending on the data available in each survey so as to maximize the information used. In practice, the variables used include housing variables, variables on the access to various types of provision for basic infrastructure services (there is a slight issue of endogeneity here, since we are modeling take-up of utility services, but it is minor given the many other variables included in the index), and variables on a range of assets owned. The regressions on take-up of service are estimated only on the samples of households who live in neighborhoods where there is access, and the estimation follows a simple probit procedure. The regressions are not presented here, as there are many of them, but they are rather straightforward. The leave-out access rate is meant to capture the general conditions of the neighborhood (including factors such as the average distance from the water pipes or electic lines), while the wealth index quintiles are used to deal with the affordability issue. Once the regressions have been estimated, we simulate what the access rate would be if all households living in areas where there is access would be lifted in terms of wealth from wherever they are in the distribution of wealth to the top wealth quintile. That is, we simulate what the take-up rate would be for all households living in primary sampling units where there is access based on what the behavior of the households would be if they were in the top quintile, which corresponds implicitly to an assumption of no affordability problem, since the households in the top quintile should be able to afford the cost of piped water and electricity services. When aggregating the results for urban areas as a whole, we denote by U* the alternative take-up rate obtained in this way (U*>U). We then define the adjusted pure demand-side gap (APDSG) as: APDSG = A × (U * -U ) (6) This definition means that we consider as a demand-side or affordability issues the difference between the simulated take-up rate when all households are given the wealth of the richest households in the country and the observed take-up rate. We next define the adjusted supply-side gap as follows: ASSG = (1 - C ) - APDSG = (1 - A × U ) - A × (U * -U ) = 1 - AU * (7) The adjusted supply-side gap is thus the difference between full coverage and the coverage that would be achieved taking into account first the current level of availability of the network in areas (the A variable), and second the take-up rate expected when there is no affordability issue. As before, the sum of the adjusted pure demand-side gap, the adjusted supply-side gap, and the coverage rate is equal to one: APDSG + ASSG + C = 1 (8) The third step is to decompose the adjusted supply-side gap into two components. First, the adjusted pure supply-side gap is defined as follows: APSSG = ASSG × U * = (1 - AU *) × U * (9) Finally, the adjusted mixed demand and supply-side gap is defined as follows: AMDSSG = ASSG × (1 - U *) = (1 - AU *) × (1 - U *) (10) The proportions of the deficit in coverage due to demand-side, supply-side, and combined problems can then be computed using the above adjusted definitions, with the sum of the three proportions still being equal to one. The results are provided in tables 3 and 4. The findings are fundamentally reversed versus what was obtained with the simple statistical decomposition. The proportion of the deficit in coverage attributable to demand-side factors is now small for piped water, at 19 percent (population weighted data) to 23 percent (unweighted data). For electricity, the corresponding figures are 39 percent (unweighted data) to 52 percent (population weighted data). By contrast, the proportion of the deficit in coverage that is attributable only to supply-side factors is now much larger, at 41 percent to 42 percent for piped water depending on whether country population weights are used or not, and at 37 percent to 39 percent for electricity. The combined demand and supply-side problems account for 35 to 39 percent of the coverage deficit for piped water, and 11 to 21 percent for electricity on average for all the countries in the sample. Given that the combined supply and demand factors reflect first a supply issue (these are urban areas where the network is not available today), it is clear that supply appears to be a larger constraining factor than demand in terms of explaining coverage deficit in urban areas in Africa. Beyond these average results for the continent as a whole, it is also useful to provide graphical representations of the results for different countries. This is done in Figures 1 through 6. In each Figure, we have a scatter plot with the neighborhood access rate in the country in urban areas on the horizontal axis, and the estimates along the econometric method for the proportions of deficit coverage due respectively to demand-side factors, supply-side factors, and combined factors on the vertical axis. The curves through the scatter plots have been simply fitted in Excel for visual purposes. Clearly, pure demand-side factors are much more important in countries where access is already high, as expected. Pure supply-side factors appear not to depend as much on access rates. This may at first seem surprising, but one should remember that even in countries where access is high, there are significant neighborhoods or parts of neighborhoods that remain unserved. In addition, supply-side factors are expressed in the Figures in percentage terms of the lack of coverage, so that the share of the unserved population due to supply-side issues should not necessarily be smaller where neighborhood access and thereby supply are higher. As to combined demand and supply-side factors, they are lower in percentage terms where there are higher access rates, essentially because when the access rate is higher, demand-side issues tend to show up more in the pure demand-side component of the decomposition than as mixed problems. Overall, given substantial differences in the nature of the obstacles to coverage between countries at different levels of neighborhood access, when thinking of policy options, it is clearly important to look at the specific estimates obtained for a given country. In fact, ideally, it would be even better to look at a lower level of disaggregation, for example for the capital city as opposed to other urban areas, if the data so permit. Census data would be useful, as the type of work conducted here is based on variables that are typically available in censuses. 4. Conclusion As part of the efforts needed to reach the Millennium Development Goals, many countries in sub-Saharan Africa are aiming to improve coverage of network-based infrastructure services such as piped water and electricity in the population. Yet in order to inform policies necessary to do so, it is important to first understand whether lack of coverage is due primarily to demand-side or affordability issues, or to a lack of supply. Indeed, some households may live in areas where access to piped water and electricity is available, but may not be able to pay for those services. Other households may be able to pay for the services, but may live too far from the electric line or water pipe to be able to connect to it. In this paper, using DHS data for a large sample of African countries, we have relied on two different methods for decomposing the lack of coverage observed in urban areas into three components: pure demand-side problems, pure supply-side problems, and mixed demand and supply-side problems. The results obtained with the statistical method suggest that for Africa as a whole, demand-side problems are prominent on average. But the results obtained from the sounder econometric method suggest that for piped water, lack of supply appears to be the main issue, and for electricity, supply-side problems loom as large as demand-side problems. At the country level, whether one is confronted mostly with demand- or supply-side problems depends in large part on the underlying access rate to the services at the neighborhood level. Because we have been dealing in this paper with data from a large number of countries, we have only tried to provide a broad snapshot of the issues. The method used here could easily be refined in order to be applied for policy work with more depth for any given country. For example, one could check the robustness of the econometric simulations to alternative estimation techniques, or alternative specifications of the regressions. One could also rely on census data in order to obtain estimates of demand as opposed to supply-side problems for smaller geographic areas. The results obtained from survey or census data could also be combined with additional information from willingness to pay studies, or focus group discussions. Data from household surveys with information on the service cuts imposed on households for non-payment of their utility bills would also provide additional information in order to assess the magnitude of supply as opposed to demand-side problems. Finally, changes over time in the estimates obtained with repeated cross-sections of data would also be very useful to assess how the mix of demand and supply-side issues evolves when, for example, neighborhood access is being improved. 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, Helsinki. Angel-Urdinola, D., M. Cosgrove-Davies, and Q. Wodon (2006): "Rwanda: Electricity Tariff Reform", in A. Coudouel, A. Dani and S. Paternostro, editors, Poverty and Social Impact Analysis of Reforms Lessons and Examples from Implementation, World Bank, Washington, DC. Angel-Urdinola, D., and Q. Wodon (2007): Do Utility Subsidies Reach the Poor? Framework and Evidence for Cape Verde, Sao Tome, and Rwanda, Economics Bulletin, 9(4): 1-7. Banerjee, S., A. B. Diallo, and Q. Wodon (2007): "Measuring Trends in Access to Modern Infrastructure in Sub-Saharan Africa: Results from Demographic and Health Surveys", Findings, No. 281, World Bank Africa Region, Washington, DC. Clarke, R. G., and S. J. Wallsten (2003): Universal Service: Empirical Evidence on the Provision of Infrastructure Services to Rural and Poor Urban Consumers," in P. Brook and T. Irwin, Infrastructure for Poor People: Public Policy for Private Provision, World Bank, Washington, DC. Estache, A. (2004): "Emerging infrastructure policy issues in developing countries - a survey of the recent economic literature." Policy Research Working Paper 3442, The World Bank, Washington DC Estache, A., V. Foster, and Q. Wodon (2002): Accounting for Poverty in Infrastructure Reform: Learning from Latin America's Experience, World Bank, WBI Development Studies, Washington, DC. Estache, A., and Q. Wodon, forthcoming, Infrastructure and poverty in Sub-Saharan Africa, Directions in Development Series, The World Bank, Washington, DC. 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. Franceys, R. (2005): Charging to Enter the Water Shop? Determining the Charges and Costs of Urban Connections for the Poor, Final Report to DFID, Institute of Water and Environment, Cranfield University. Kayaga, S., and R. Franceys (2007): Costs of urban utility water connections: Excessive burden to the poor, Utility Policy (15): 270-277. Komives, K. D. Whittington, and X. Wu (2003): "Infrastructure coverage and the poor: A global perspective," in P. Brook and T. Irwin, Infrastructure for Poor People: Public Policy for Private Provision, 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, Directions in Development, Washington, DC. Komives, K., J. Halpern, V. Foster, Q. Wodon, and R. Abdullah (2007): "Utility Subsidies as Social Transfers: An Empirical Evaluation of Targeting Performance," Development Policy Review, 25(6): 659-679. Figure 1: Demand side problems and access rate (water) 100% 90% Demand side problem (%) 80% 70% 60% 50% 40% 30% 20% 10% 0% 30% 40% 50% 60% 70% 80% 90% 100% Access rate (%) Source: Authors using DHS data. Figure 2: Supply side problems and access rate (water) 100% 90% 80% Supply side problem (%) 70% 60% 50% 40% 30% 20% 10% 0% 30% 40% 50% 60% 70% 80% 90% 100% Access rate (%) Source: Authors using DHS data. Figure 3: Combined demand and supply side problems and access rate (water) 100% 90% Combined problem (%) 80% 70% 60% 50% 40% 30% 20% 10% 0% 30% 40% 50% 60% 70% 80% 90% 100% Access rate (%) Source: Authors using DHS data. Figure 4: Demand side problems and access rate (electricity) 100% 90% Demand side problem (%) 80% 70% 60% 50% 40% 30% 20% 10% 0% 50% 60% 70% 80% 90% 100% Access rate (%) Source: Authors using DHS data. Figure 5: Supply side problems and access rate (electricity) 100% 90% Supply side problem (%) 80% 70% 60% 50% 40% 30% 20% 10% 0% 50% 60% 70% 80% 90% 100% Access rate (%) Source: Authors using DHS data. Figure 6: Combined demand and supply side problems and access rate (electricity) 100% 90% 80% Combined problem (%) 70% 60% 50% 40% 30% 20% 10% 0% 50% 60% 70% 80% 90% 100% Access rate (%) Source: Authors using DHS data. Table 1: Estimation without controls of demand and supply-side obstacles for access to piped water, Africa ­ Urban areas (%) Proportion of Proportion of Proportion of Mixed deficit deficit deficit Take- Pure demand attributable attributable attributable to up rate Pure supply- and to demand- to supply- both supply Access given Coverage Unserved demand- Supply- side supply- side factors side factors and demand- Country Rate access rate Population side gap side gap gap side gap only only side factors Benin 81 75 60 40 21 19 14 5 52 36 12 Burkina Faso 87 38 33 67 54 13 5 8 81 7 12 CAR 39 16 6 94 33 61 10 51 35 10 55 Cameroon 80 30 24 76 56 20 6 14 74 8 18 Chad 68 32 22 78 47 32 10 22 60 13 28 Comoros 81 53 43 57 38 19 10 9 66 18 16 Republic of Congo 91 51 46 54 45 9 4 4 84 8 8 Côte d'Ivoire 96 67 65 35 32 4 2 1 90 7 3 Ethiopia 88 55 48 52 40 12 6 5 78 12 10 Gabon 96 57 55 45 41 4 2 2 91 5 4 Ghana 72 47 34 66 38 28 13 15 58 20 22 Guinea 78 36 28 72 50 22 8 14 70 11 19 Kenya 78 64 50 50 28 22 14 8 57 28 16 Lesotho 94 53 50 50 44 6 3 3 89 6 5 Madagascar 65 26 17 83 48 35 9 25 58 11 31 Malawi 85 38 32 68 53 15 6 9 78 8 14 Mali 75 39 29 71 46 25 10 15 65 14 22 Mauritania 75 37 28 72 48 25 9 16 66 12 22 Mozambique 55 36 20 80 35 45 16 29 44 20 36 Namibia 91 87 79 21 12 9 8 1 56 39 6 Niger 89 35 31 69 58 11 4 7 85 5 10 Nigeria 53 29 15 85 37 47 14 33 44 16 40 Rwanda 56 29 16 84 40 44 13 32 47 15 37 Senegal 98 78 77 23 22 2 1 0 93 5 1 South Africa 94 93 88 12 6 6 5 0 52 45 3 Tanzania 65 34 22 78 43 35 12 23 55 15 30 Togo 93 55 51 49 42 7 4 3 86 8 6 Uganda 65 22 14 86 51 35 8 27 59 9 32 Zambia 78 60 46 54 31 22 13 9 58 25 17 Zimbabwe 100 93 93 7 7 0 0 0 100 0 0 Simple average 79 49 41 59 38 21 8 13 68 15 18 Weighted average 73 48 38 62 34 27 10 18 59 18 23 Source: Authors using DHS data. All variables are expressed as percentages (%). Table 2: Estimation without controls of demand and supply-side obstacles for access to electricity, Africa ­ Urban areas (%) Proportion of Proportion of Proportion of Mixed deficit deficit deficit Take- Pure demand attributable attributable attributable to up rate Pure supply- and to demand- to supply- both supply Access given Coverage Unserved demand- Supply- side supply- side factors side factors and demand- Country Rate access rate Population side gap side gap gap side gap only only side factors Benin 83 61 51 49 32 17 11 7 65 21 13 Burkina Faso 92 58 54 46 38 8 5 3 83 10 7 CAR 57 19 11 89 46 43 8 34 52 9 39 Cameroon 94 82 77 23 17 6 5 1 74 21 5 Chad 77 26 20 80 58 23 6 17 72 7 21 Comoros 100 54 54 46 46 0 0 0 100 0 0 Republic of Congo 98 52 51 49 47 2 1 1 96 2 2 Côte d'Ivoire 100 90 90 10 10 0 0 0 100 0 0 Ethiopia 99 87 86 14 13 1 1 0 92 7 1 Gabon 100 91 91 9 9 0 0 0 100 0 0 Ghana 98 79 77 23 21 2 2 1 90 8 2 Guinea 89 72 63 37 25 11 8 3 69 23 9 Kenya 80 64 51 49 29 20 13 7 59 27 15 Lesotho 87 32 28 72 59 13 4 9 82 6 12 Madagascar 80 65 52 48 28 20 13 7 58 27 15 Malawi 84 40 34 66 50 16 6 9 76 10 14 Mali 81 51 41 59 40 19 9 9 68 16 16 Mauritania 85 60 51 49 34 15 9 6 69 18 12 Mozambique 80 37 30 70 50 20 7 13 71 11 18 Namibia 93 80 75 25 18 7 6 1 72 22 5 Niger 94 43 41 59 53 6 3 4 90 5 6 Nigeria 98 86 84 16 14 2 2 0 86 12 2 Rwanda 72 37 27 73 45 28 10 17 62 14 24 Senegal 99 82 82 18 17 1 0 0 97 2 1 South Africa 95 91 86 14 8 5 5 0 60 37 3 Tanzania 83 47 39 61 45 17 8 9 73 13 14 Togo 96 46 44 56 51 4 2 2 92 3 4 Uganda 93 51 47 53 46 7 3 3 87 7 6 Zambia 84 59 50 50 34 16 9 6 69 19 13 Zimbabwe 100 90 90 10 10 0 0 0 100 0 0 Simple average 89 61 56 44 33 11 5 6 79 12 9 Weighted average 93 75 71 29 22 7 4 3 79 15 6 Source: Authors using DHS data. All variables are expressed as percentages (%). Table 3: Estimation with controls of demand and supply-side obstacles for access to piped water, Africa ­ Urban areas (%) Adjusted Adjusted Adjusted proportion Adjusted proportion proportion of deficit Adjusted mixed of deficit of deficit attributable Take-up Adjusted Adjusted demand attributable attributable to both rate pure Adjusted pure and to demand- to supply- supply and Access given Coverage Unserved demand- supply- supply- supply- side factors side factors demand- Country Rate access rate Population side gap side gap side gap side gap only only side factors Benin 81 85 60 40 9 31 26 5 22 67 12 Burkina Faso 87 40 33 67 2 65 26 39 3 39 58 CAR 39 24 6 94 3 90 22 68 3 23 73 Cameroon 80 49 24 76 15 61 30 31 20 39 41 Chad 68 34 22 78 2 77 26 51 2 33 65 Comoros 81 69 43 57 13 44 31 14 23 53 24 Republic of Congo 91 93 46 54 38 15 14 1 71 26 2 Côte d'Ivoire 96 99 65 35 31 5 5 0 87 13 0 Ethiopia 88 57 48 52 2 50 28 21 4 55 42 Gabon 96 98 55 45 39 6 6 0 86 13 0 Ghana 72 66 34 66 14 52 35 18 21 52 27 Guinea 78 42 28 72 5 67 28 39 6 39 54 Kenya 78 70 50 50 5 45 32 14 10 63 27 Lesotho 94 68 50 50 14 36 24 11 28 49 23 Madagascar 65 32 17 83 4 79 25 54 5 31 65 Malawi 85 44 32 68 5 63 27 35 8 40 52 Mali 75 48 29 71 6 64 31 34 9 43 48 Mauritania 75 50 28 72 10 62 31 31 14 43 43 Mozambique 55 42 20 80 3 77 32 44 4 40 55 Namibia 91 95 79 21 7 14 13 1 31 65 4 Niger 89 40 31 69 4 65 26 39 6 37 57 Nigeria 53 38 15 85 4 80 30 50 5 36 59 Rwanda 56 34 16 84 3 81 27 54 3 33 64 Senegal 98 92 77 23 13 10 9 1 57 39 4 South Africa 94 98 88 12 5 8 8 0 37 62 1 Tanzania 65 38 22 78 3 75 29 46 4 37 59 Togo 93 67 51 49 11 38 25 13 22 52 26 Uganda 65 24 14 86 1 85 20 65 1 23 76 Zambia 78 98 46 54 29 24 24 1 55 44 1 Zimbabwe 100 96 93 7 3 4 4 0 44 54 2 Simple average 79 61 41 59 10 49 23 26 23 41 35 Weighted average 73 58 38 62 8 54 24 30 19 42 39 Source: Authors using DHS data. All variables are expressed as percentages (%). Table 4: Estimation with controls of demand and supply-side obstacles for access to electricity, Africa ­ Urban areas (%) Adjusted Adjusted Adjusted proportion Adjusted proportion proportion of deficit Adjusted mixed of deficit of deficit attributable Take-up Adjusted Adjusted demand attributable attributable to both rate pure Adjusted pure and to demand- to supply- supply and Access given Coverage Unserved demand- supply- supply- supply- side factors side factors demand- Country Rate access rate Population side gap side gap side gap side gap only only side factors Benin 83 83 51 49 18 31 26 5 36 53 11 Burkina Faso 92 62 54 46 3 43 27 16 7 57 35 CAR 57 33 11 89 8 81 27 54 9 30 61 Cameroon 94 98 77 23 15 8 8 0 65 34 1 Chad 77 28 20 80 1 79 22 57 2 27 71 Comoros 100 82 54 46 28 18 15 3 61 32 7 Republic of Congo 98 86 51 49 33 16 13 2 68 27 4 Côte d'Ivoire 100 99 90 10 10 1 1 0 93 7 0 Ethiopia 99 90 86 14 3 11 10 1 20 72 8 Gabon 100 99 91 9 9 1 1 0 94 6 0 Ghana 98 95 77 23 15 8 7 0 67 31 2 Guinea 89 78 63 37 6 31 24 7 16 66 18 Kenya 80 72 51 49 6 42 31 12 13 63 24 Lesotho 87 42 28 72 8 64 27 37 12 37 51 Madagascar 80 86 52 48 16 32 27 5 34 56 10 Malawi 84 48 34 66 7 59 29 31 10 43 47 Mali 81 62 41 59 9 49 31 19 16 52 32 Mauritania 85 84 51 49 20 29 24 5 41 49 10 Mozambique 80 51 30 70 11 59 30 29 16 43 41 Namibia 93 99 75 25 17 8 8 0 69 31 0 Niger 94 49 41 59 6 54 26 27 9 45 46 Nigeria 98 98 84 16 12 4 4 0 73 26 1 Rwanda 72 46 27 73 6 67 31 36 8 42 49 Senegal 99 100 82 18 17 1 1 0 95 5 0 South Africa 95 100 86 14 8 6 6 0 58 42 0 Tanzania 83 55 39 61 7 54 30 25 11 49 40 Togo 96 66 44 56 19 37 24 12 34 44 22 Uganda 93 58 47 53 6 46 27 20 12 51 37 Zambia 84 84 50 50 21 29 24 5 42 49 9 Zimbabwe 100 99 90 10 8 2 1 0 85 15 0 Simple average 89 74 56 44 12 32 19 14 39 39 21 Weighted average 93 87 71 29 11 18 12 6 52 37 11 Source: Authors using DHS data. All variables are expressed as percentages (%).