49571 June 2009 Note Number 27 Output-Based Aid in Ethiopia Dealing with the "Last Mile" Paradox in Rural Electrification | Luiz Maurer and Carmen Nonay P roviding electricity to poor households in rural areas in Ethiopia is critical to improving the health and living conditions of the population, reducing poverty, and stimulating growth. A project supported by the Global Partnership on Output-Based Aid (GPOBA) is accelerat- ing the pace of connections in electrified areas and fostering energy efficiency. The scheme provides a performance-based subsidy to the state-owned utility in Ethiopia linked to the delivery of pre-agreed outputs, including the connection and provision of services on a sustainable basis. As part of the connection package, poor households receive two compact fluorescent lamps (CFLs). The project has poten- tial for scaling up with a future International Development Photo courtesy of the World Bank Association (IDA) loan to the utility. Overview of the Power Sector in the African Development Bank, and the Governments Ethiopia of India and Kuwait, which together are contributing ap- proximately US$1.2 billion. Ethiopia has one of the lowest electrification rates in The "Last Mile" Paradox Sub-Saharan Africa. Only 17­18 percent of the nation's 8,000 towns and villages are electrified, and only about The current gap between the connection and access rates 7 percent of households have metered connections. stems from the fact that many customers cannot afford About 1.6 million customers are served by the Ethiopian to pay the cost of a connection. Compared to the up- Electric Power Corporation (EEPCo), a well-run, techni- stream cost of providing electricity services (generation, cally competent utility that is state-owned and vertically transmission, and distribution), the cost of connecting integrated. EEPCo is the primary vehicle for grid-based households to the grid--the "last mile"--represents a electrification and also serves some isolated areas with small fraction of the utility's total investment (only 3 per- diesel generators. cent). However, for poor households, the connection cost The Government of Ethiopia (GoE) recognizes the constitutes a significant barrier to access. The average importance of rural electrification to reduce poverty and connection cost for customers living near the grid (usu- stimulate growth. A few years ago, the GoE launched an ally less than 150 feet) ranges from US$50 to US$100 per ambitious rural electrification program with the objec- household. In those areas where the annual income is less tive of providing access to half of rural communities in than US$500 per household, a connection costing US$75 five years, to be expanded to full coverage in ten years. EEPCo has been selected as the implementation agency, and has put in place the necessary organization and pro- cesses to deal with this challenge competently. The GoE is Luiz Maurer is a Senior Energy Specialist in the World Bank's Africa Region and Carmen Nonay is a Senior Infrastructure Specialist in working with other partners, including the World Bank, the Global Partnership on Output-Based Aid. Supporting the delivery of basic services in developing countries represents 15 percent of the average annual income Because of the last mile effect, the connection per household. rate has grown very slowly. After electricity becomes At the same time, metered grid-based electric- available in an area, only about 20 percent of cus- ity is a more affordable energy source than other tomers get metered connections in the first year. sources used by the poor. By and large, kerosene is This percentage grows very slowly, at a rate of about the fuel of choice in rural areas, used by 96 percent 10 percent per year. It takes three years for half the of households, which spend US$1.60 per month on households of an electrified town to have access to a average on kerosene. Firewood and dry-cell batter- metered connection. ies are also used to provide lighting services in some One way to address the last mile paradox is to pro- cases. The price that rural people routinely pay for vide people with the necessary microcredit to pay for these alternative fuels is more than the average price the cost of connection. EEPCo has carried out trials of electricity that EEPCo charges rural customers. that reveal that granting concessional loans to cus- Other households obtain electricity by stringing wires tomers has made the connection more affordable and to their nearest neighbors, with the neighbor charg- thus has increased the pace of connection. A few years ing them for their electricity consumption. EEPCo ago, EEPCo offered five-year, interest-free loans, later does not consider this practice theft, since energy is reduced to two-year loans. The pace of electrification being metered and billed to the neighbor. However, accelerated. Unfortunately, the concession of those this situation is far from ideal, as neighbors charge on loans consumed considerable amounts of the utility's a per lamp basis, resulting in an average equivalent working capital, which needed to be channeled to the tariff of about 30 US cents per kWh--ten times higher backbone of the rural electrification program. than the lifeline rate that EEPCo charges its metered customers. The "last mile" is thus a dual paradox. First, the The GPOBA Intervention to Address cost of connecting households represents only 3 the Last Mile Paradox percent of the total investment necessary to deliver The GPOBA project was designed to deal with the electricity to the customer. Second, while the connec- last mile paradox and to reduce the cost of electricity tion cost is a barrier for poor households, they can for poor customers. Its main objective is to accelerate afford to pay for electricity once they are connected. the pace of connections in rural towns and villages Even the poorest households, with an ability to pay with grid access, by assisting EEPCo in its program to of US$1/month, could afford the cost of the typical finance the cost of the connection fee.1A grant from electricity bill. A modest household with a consump- GPOBA covers EEPCo's costs of financing the loans tion of 15 kWh/month would spend about US$8­10 extended to poor household customers. In addition, per year at prevailing rates. In sum, people can afford the GPOBA grant will finance the cost of providing to pay for a much higher value added service such as two energy-efficient CFLs free to poor households. electric lighting, with the benefits of more light, less Those CFLs will be delivered as part of the connec- pollution, and more convenience. tion package, both to ensure an affordable bill for Table 1. Electricity Usage With and Without CFLs kW # of % Hours Monthly % Usage With Appliance installed units frequency per day consumption (Wh) without CFLs CFLs Savings Lighting 40 2.6 0.97 3.5 10,920 73.51% 18.38% 55.13% Torch 10 1.2 0.53 0.1 36 0.24% 0.24% Radio 20 1.0 0.37 1.0 600 4.04% 4.04% Cassette TV 80 1.0 0.02 1.0 2,400 16.15% 16.15% Small 10 1.0 0.41 3.0 900 6.06% 6.06% Radio Total NA NA NA NA 14,856 100.00% 44.87% 55.13% June 2009 Note Number 27 households and to promote energy efficiency. Energy outputs and service delivered, by an Independent efficiency helps poor customers lower their electric- Verification Agent. The loan to be granted by EEPCo ity bill, while reducing the utility's financial losses in will be 80 percent of the estimated cost of connection. serving the poorest customers (since lifeline rates are Customers will pay a 20 percent upfront fee when below the cost-to-serve). If customers use efficient applying for the connection. The balance of the loan lamps, such as CFLs, their total bill may decrease by will be divided into 60 equal monthly installments of 55 percent, as shown in table 1. This makes energy US$1. Repayment of the loan will be rolled into the more affordable to the poor and reduces the subsi- electricity bill. dies EEPCo currently provides to sell energy to those In the absence of reliable means-based poverty customers at lifeline rates: a win-win situation for tests, all customers would be eligible to receive the both parties. Moreover, it creates some leeway for the loan. However, to provide some form of targeting, utility to increase tariff rates and break even when it was agreed that the connection with concessional serving rural customers. financing would only be available one year after the The GPOBA intervention was originally designed village had been electrified. The assumption is that as part of the World Bank's Second Electricity Access customers who could afford to connect would have Rural Expansion Project (EAREP II), which benefits requested the connection in the first 12 months, be- from a US$130 million loan from IDA. However, fore concessional funding was made available. subsidies are being made available in all of EEPCo's grid-electrified area, not only in the EAREP II project area. Within this framework, GPOBA is providing a The Road Ahead US$8 million grant,2 which will enable EEPCo to con- This GPOBA project is about to go into effect, thereby nect and provide loans to 229,000 customers, repre- triggering disbursements. Some lessons learned and senting a population of 1.1 million people. Every time considerations for future interventions include the EEPCo provides a metered connection to a household following: customer, including two CFLs, and provides a five- year loan, it is entitled to receive a payment of US$35 · Enhance research on willingness to pay for the from GPOBA; upon verification of and opinion on, connection and desired terms and conditions Figure 1. Project Arrangements Grant Agreement Government of Ethiopia GPOBA Subsidy payment IVA Contract EEPCo EEPCo Account Connections and service Independent Verification Agent (IVA) Customers Verification of and opinion on outputs and service delivered Key: Contract Control Funds of the loan. GPOBA used a five-year loan tenor, for the first three years only. EEPCo is connect- based on the utility's previous experience. How- ing about 300,000 to 400,000 customers per ever, if the loan duration could be reduced, more year. Assuming that half of these will still need customers could be connected with the same some financial support, the funding for connec- amount of resources. Field surveys are being car- tions must be tripled, at minimum. This could be ried out to determine the most convenient terms achieved by using concessional funds from new of the loans that are able to attract new customers. projects funded by IDA to top up the GPOBA To better address the nature of the financial in- resources, for example. In the design of future centives, an experimental approach is now being electrification interventions in rural Ethiopia, conducted. It consists of a series of voucher-led the possibility of creating a performance-based experiments at the household level, to evaluate mechanism similar to the one currently funded the feasibility and the effectiveness of subsidies for by GPOBA is also being considered. GPOBA has connection fees and the purchase of CFLs. Vouch- a key demonstration effect role to play, paving the ers are distributed through a "lottery" system. way for larger accelerated connection programs in A survey of the same households will be imple- Ethiopia and in other countries. mented a few months later to measure differences in the pick-up rates between households that have received vouchers and those that have not. The survey will assess the extent to which house- holds are responsive to different levels of financial 1 The cost of connection is defined as the costs of a drop wire, an incentives for their decision to connect to the grid additional pole (if necessary), a meter, circuit breakers, and two or to buy CFLs. energy-efficient bulbs, as well as installation costs and administra- · Increase the performance-based funds, along the tive expenses. What is being called connection cost, for purposes lines of the GPOBA approach, to speed up the of the OBA intervention, excludes any upstream investment in the expansion of subtransmission or transmission networks. pace of connection. Despite its considerable size, 2 The GPOBA grant draws on funds from the United Kingdom's the GPOBA grant alone will not provide enough Department for International Development and a Multi-Donor Trust funds for the utility to connect all the customers Fund funded by the Swedish International Development Coopera- who cannot afford the connection fee. GPOBA tion Agency and the Australian Agency for International Develop- will support about 75,000 households per year, ment (AusAID). About OBApproaches OBApproaches is a forum for discussing and disseminating The case studies have been chosen and presented by the au- recent experiences and innovations in supporting the delivery thors in agreement with the GPOBA management team and are of basic services to the poor. The series focuses on the provi- not to be attributed to GPOBA's donors, the World Bank, or any sion of water, energy, telecommunications, transport, health, other affiliated organizations. Nor do any of the conclusions rep- and education in developing countries, in particular through resent official policy of GPOBA, the World Bank, or the countries output-, or performance-, based approaches. they represent. To find out more, visit www.gpoba.org eeGlobal Partnership on Output-Based Aid Global Partnership on Output-Based Aid Supporting the delivery of basic services in developing countries