Document of The World Bank Report No.:42413 PROJECT PERFORMANCE ASSESSMENT REPORT LAO PEOPLE'S DEMOCRATIC REPUBLIC SOUTHERN PROVINCES RURAL ELECTRIFICATION PROJECT (CREDIT 3047-LA) February 4, 2008 Sector, Thematic and Global Evaluation Division Independent Evaluation Group (World Bank) Currency Equivalents (annual averages) Currency Unit = Lao Kip 2004 US$1.00 K 10,646 Abbreviations and Acronyms CAS Country Assistance Strategy DOE Department of Electricity EdL Electricité du Laos EPDP Ethnic People's Development Plan ERR Economic Rate of Return ESMAP Energy Sector Management Assistance Program FIRR Financial Rate of Return GoL Government of Lao PDR GTZ Deutsche Gesellschaft für Technische Zusammenarbeit ICR Implementation Completion Report IED Innovation Energie Développement IEG Independent Evaluation Group IEGWB Independent Evaluation Group (World Bank) JICA Japan International Cooperation Agency MEM Ministry of Energy and Mines MIH Ministry of Industry and Handicrafts NORAD Norwegian Agency for Development Cooperation OGS Off-Grid Promotion Support Office in DOE O&M Operation and Maintenance PAD Project Appraisal Document PGI Provincial Grid Integration Project RAP Resettlement Action Plan REP Rural Electrification Project PESCO Provincial Energy Services Company PPAR Project Performance Assessment Report PRSP Poverty Reduction Strategy Paper PVS Photo-voltaic Systems SHS Solar Home Systems SPRE Southern Provinces Rural Electrification Project UNDP United Nations Development Program VEM Village Electricity Manager VOPS Village Off-grid Promotion Scheme WTP Willingness to Pay Fiscal Year Government: January 1 ­ December 31 Director-General, Evaluation : Mr. Vinod Thomas Director, Independent Evaluation Group (World Bank) : Ms. Cheryl Gray Manager, Sector, Thematic and Global Evaluation Division : Mr. Alain Barbu Task Manager : Mr. Howard White IEGWB Mission: Enhancing development effectiveness through excellence and independence in evaluation. About this Report The Independent Evaluation Group assesses the programs and activities of the World Bank for two purposes: first, to ensure the integrity of the Bank's self-evaluation process and to verify that the Bank's work is producing the expected results, and second, to help develop improved directions, policies, and procedures through the dissemination of lessons drawn from experience. As part of this work, IEGWB annually assesses about 25 percent of the Bank's lending operations through field work. In selecting operations for assessment, preference is given to those that are innovative, large, or complex; those that are relevant to upcoming studies or country evaluations; those for which Executive Directors or Bank management have requested assessments; and those that are likely to generate important lessons. To prepare a Project Performance Assessment Report (PPAR), IEGWB staff examine project files and other documents, interview operational staff, visit the borrowing country to discuss the operation with the government, and other in-country stakeholders, and interview Bank staff and other donor agency staff both at headquarters and in local offices as appropriate. Each PPAR is subject to internal IEGWB peer review, Panel review, and management approval. Once cleared internally, the PPAR is commented on by the responsible Bank department. IEGWB incorporates the comments as relevant. The completed PPAR is then sent to the borrower for review; the borrowers' comments are attached to the document that is sent to the Bank's Board of Executive Directors. After an assessment report has been sent to the Board, it is disclosed to the public. About the IEGWB Rating System IEGWB's use of multiple evaluation methods offers both rigor and a necessary level of flexibility to adapt to lending instrument, project design, or sectoral approach. IEGWB evaluators all apply the same basic method to arrive at their project ratings. Following is the definition and rating scale used for each evaluation criterion (additional information is available on the IEGWB website: http://worldbank.org/ieg). Outcome: The extent to which the operation's major relevant objectives were achieved, or are expected to be achieved, efficiently. The rating has three dimensions: relevance, efficacy, and efficiency. Relevance includes relevance of objectives and relevance of design. Relevance of objectives is the extent to which the project's objectives are consistent with the country's current development priorities and with current Bank country and sectoral assistance strategies and corporate goals (expressed in Poverty Reduction Strategy Papers, Country Assistance Strategies, Sector Strategy Papers, Operational Policies). Relevance of design is the extent to which the project's design is consistent with the stated objectives. Efficacy is the extent to which the project's objectives were achieved, or are expected to be achieved, taking into account their relative importance. Efficiency is the extent to which the project achieved, or is expected to achieve, a return higher than the opportunity cost of capital and benefits at least cost compared to alternatives. The efficiency dimension generally is not applied to adjustment operations. Possible ratings for Outcome: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Risk to Development Outcome: The risk, at the time of evaluation, that development outcomes (or expected outcomes) will not be maintained (or realized). Possible ratings for Risk to Development Outcome: High Significant, Moderate, Negligible to Low, Not Evaluable. Bank Performance: The extent to which services provided by the Bank ensured quality at entry of the operation and supported effective implementation through appropriate supervision (including ensuring adequate transition arrangements for regular operation of supported activities after loan/credit closing, toward the achievement of development outcomes. The rating has two dimensions: quality at entry and quality of supervision. Possible ratings for Bank Performance: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Borrower Performance: The extent to which the borrower (including the government and implementing agency or agencies) ensured quality of preparation and implementation, and complied with covenants and agreements, toward the achievement of development outcomes. The rating has two dimensions: government performance and implementing agency(ies) performance. Possible ratings for Borrower Performance: Highly Satisfactory, Satisfactory, Moderately Satisfactory, Moderately Unsatisfactory, Unsatisfactory, Highly Unsatisfactory. Contents PRINCIPAL RATINGS....................................................................................................................V KEY STAFF RESPONSIBLE..........................................................................................................V PREFACE......................................................................................................................................VII SUMMARY .....................................................................................................................................IX 1. INTRODUCTION................................................................................................................ 1 The role of the World Bank and other donors............................................................... 2 2. PROJECT PREPARATION, DESIGN, IMPLEMENTATION AND OUTCOMES.............. 3 Preparation........................................................................................................................ 3 Objectives and components............................................................................................ 4 Overview of implementation............................................................................................ 5 Financial performance of EdL......................................................................................... 7 Technical performance of the grid: system losses and outages................................ 9 Off-grid component........................................................................................................ 10 Resettlement, environmental and social issues ......................................................... 12 Monitoring and evaluation............................................................................................. 14 3. IMPACT OF RURAL ELECTRIFICATION AND ECONOMIC ANALYSIS...................... 14 Coverage and distribution of rural electrification in the Southern Provinces......... 14 Uses of electricity........................................................................................................... 17 4. RATE OF RETURN ANALYSIS....................................................................................... 18 The financial rate of return (FIRR) ................................................................................ 18 The economic rate of return.......................................................................................... 18 This report was prepared by Howard White, who assessed the project in March 2007. Soon-Won Pak provided administrative support. 5. RATINGS AND LESSONS LEARNED............................................................................ 20 Outcome: Moderately satisfactory ............................................................................... 20 Risk to development outcome: Low............................................................................. 21 Bank performance: Satisfactory................................................................................... 21 Borrower performance: Moderately Satisfactory........................................................ 22 Lessons learned ............................................................................................................. 22 ANNEX A. BASIC DATA SHEET ................................................................................................. 25 ANNEX B. ANALYSIS OF RE I BASELINE DATA (2004) .......................................................... 29 ANNEX C. ECONOMIC RATE OF RETURN................................................................................ 37 ANNEX D. LIST OF PEOPLE MET .............................................................................................. 41 v Principal Ratings ICR* ICR Review* PPAR Outcome Satisfactory Satisfactory Moderately Satisfactory Institutional Substantial Substantial ------ Development Impact** Risk to ------ ------ Low Development Outcome Sustainability*** Likely Likely ------ Bank Satisfactory Satisfactory Satisfactory Performance Borrower Satisfactory Unsatisfactory Moderately satisfactory Performance * The Implementation Completion Report (ICR) is a self-evaluation by the responsible Bank department. The ICR Review is an intermediate IEGWB product that seeks to independently verify the findings of the ICR. **As of July 1, 2006, Institutional Development Impact is assessed as part of the Outcome rating. ***As of July 1, 2006, Sustainability has been replaced by Risk to Development Outcome. As the scales are different, the ratings are not directly comparable. Key Staff Responsible Project Task Manager/Leader Division Chief/ Country Director Sector Director Appraisal Veronique Bishop Yoshihiko Sumi Ngozi N. Okonjo-Iweala Completion Jie Tang Junhui Wu Ian C. Porter vii Preface This project performance audit report (PPAR) covers the Southern Provinces Rural Electrification Project (SPRE, Credit 30470-LA for US$34.7 million), which was approved in March 1998 and closed in June 2004. The PPAR was prepared by the Independent Evaluation Group (IEG, formerly OED), based upon the Implementation Completion Report (ICR), project documents, and interviews with government officials, civil society and private sector representatives and staff of the Bank and other agencies. Fieldwork was undertaken in March 2007, which benefited from the support of Ms. Mingngakham Pangnasak and Mr. Morten Larsen of the World Bank office in Vientianne in arranging the mission. Mr. Anousak Phongsavath and Mr Oudom Kasisingdeth arranged fieldwork for the mission, and accompanied it to the field. The mission is grateful to others who spared their time to share their views, and for the cooperation of the Government of Lao PDR in undertaking this study. This project was selected for review as an input to the IEG study The Welfare Impact of Rural Electrification: A Reassessment of the Costs and Benefit, delivered in June 2007. Baseline data were made available to IEG which allowed a more in depth analysis of impact issues than is usually the case in a PPAR. Following standard IEG procedures, the draft of this PPAR was sent to the borrower for comments before finalization, but none were received. ix Summary The Southern Provinces Rural Electrification Project (SPRE), approved in 1998, followed on from earlier Bank projects providing electricity to rural areas of Lao PDR through the electricity utility, Electricité du Laos (EdL). There was also an off-grid component, implemented by the Department of Electricity (DOE). In summary, the project exceeded its physical goals for both the grid and off-grid components, but both suffered from institutional shortcomings. SPRE had two objectives: (i) expand rural electricity service in seven central and southern provinces of Lao PDR, where economically justified, through grid extension and off-grid electrification; and (ii) strengthen EdL's capacity to plan and implement electrification investments and to operate on a commercial basis. After a slow start, the project exceeded its physical goals, connecting 721 villages compared to the target of 520. Cost savings allowed for additional civil works, including a third sub-station. However, system losses remained high (at around 21% in 2004 versus a target of 16%), being highest in project provinces. The off-grid component provided electricity to 6,097 homes compared to the target of 4,600. Performance of the grid extension component was overshadowed by continuing financial difficulties for EdL, stemming from both tariffs below cost recovery levels and arrears mainly from public agencies. By the end of the project in 2004, the performance of most financial indicators was below expectations. However, subsequent implementation of the financial recovery plan is moving EdL toward a financially sustainable situation. The off-grid component suffered from technical problems, so that many households are receiving less than one hour a day of electricity rather than the 3-4 hours they would get were the system working satisfactorily. These technical problems are a manifestation of an underlying structural problem in the program incentives, which are oriented toward planning and installation of connections to new villages, with maintenance being financially unrewarding. However, a private company is providing solar home systems to many communities despite a much lower level of subsidies than those available through the Bank project, so there are also issues of technical capacity and delineating roles between subsidized and unsubsidized services. SPRE was highly relevant: government has set a target of 90 percent electrification by 2020, and rural electrification is part of the Bank's Country Assistance Strategy. The project achieved its first objective of supplying energy services, but not the second of strengthening EdL. Efficiency is measured by the ERR, which was 13 percent (lower than the 61 percent in the ICR given technical errors in the latter's calculation). The overall outcome rating is moderately satisfactory, because of EdL's unsatisfactory financial performance for most of the project's duration and the technical problems discussed above. Bank performance was satisfactory, and that of the borrower moderately satisfactory (again reflecting EdL's poor financial performance). The risk to development outcome is modest since measures implemented toward the end of, and x subsequent to, project completion have addressed the financial sustainability issue, and technical issues are being addressed under the follow-on Rural Electrification Project. The lessons learned from the review are: · Consumers are willing to pay tariffs at cost recovery levels. The willingness to pay analysis shows that consumers are willing to pay for electricity at levels exceeding supply costs. Lao PDR has been able to implement sizeable tariff increases in a short space of time with no adverse social impact or notable demand reductions (indeed it can be argued that any reduction in usage by already connected customers would help extend supply to new locations). · But poorer households remain unconnected. Even after the grid has been in a village for more than 10 years, some 20 percent of households remain unconnected. Smart (i.e. efficiently targeted) connection subsidies for late connectors would help achieve government's 90 percent coverage target. · Explicit attention needs to be paid to technical efficiency. Technical problems of system losses and outages reduce financial performance and undermine project benefits. Explicit components are needed to tackle such issues. · Off-grid components promote social equity in electrification coverage... but are not necessarily the most cost effective strategy. The off-grid component has helped reach remote communities which would not otherwise have got electricity for a decade or more. However, the rate of return to these investments is lower than that to grid extension, or of subsidizing connections for late connectors, and it is not proven that the beneficiaries from off-grid programs are poorer than those benefiting from grid extension. · In countries where the private sector is relatively undeveloped, substantial support may be required to provide off-grid services in poor areas. At the same time subsidies should not crowd out purely commercial actors. A sensible balance would direct project supported companies to the more remote areas unlikely to be attractive to purely commercial firms, but a subsidy scheme needs to ensure that such activities are sufficiently attractive. Vinod Thomas Director-General Evaluation 11 Annex A 1 1. Introduction 1.1 Electricity in Lao PDR is largely generated by hydropower, with a limited use of diesel. Off-grid schemes are being developed in less accessible areas. The first grid system was developed around the Nam Ngum dam, completed in 1971 with assistance from a group of eight donors, including the World Bank. A large part of the electricity generated by Nam Ngum was exported to Thailand. Exports to Thailand continue to account for a large, though falling, share of the electricity generated, and have been supplemented by exports to Vietnam. However, the country's grid has developed in a fragmented manner, with four principal grids in different parts of the country (Northern Grid, Central Grids I and II and Southern Grid). Since the grid is not national, regional consumption needs are also met by imports from Thailand, making Lao PDR both an exporter and importer of electricity. 1.2 The Government of Lao PDR (GoL) places high priority on rural electrification, with the goal of electrifying 90 percent of the country's households by 2020 (of which 75 percent will be on grid) as well as increasing hydropower exports to neighboring countries. Interim targets are for 70 percent coverage by 2010 and 80 percent by 2015. Growth in coverage since the nineties has been high, averaging over 10 percent a year which is more than sufficient to meet these targets (Figure 1), and several new dams are under construction to increase power generation. Off-grid connections are needed if electricity is to be available in less accessible communities, including those in mountainous regions and on river islands. 1.3 The Government's Figure 1 National electrification rate strategy of electrification, adopted in 1998, gave priority to 50 45 provincial capitals and areas of ) 40 tourist potential. The strategy was %( 35 written following the first steps tear 30 towards commercialization and 25 strengthening of the power sector, 20 15 which are in turn reflected in four noitacfiitrcelE10 policy goals set out in the GoL's 5 "Power Sector Policy Statement" 0 issued in 2001: (i) Maintain and 1990 1992 1994 1996 1998 2000 2002 2004 2006 expand an affordable, reliable, Source: Rural Electrification, Project Appraisal Document, Table 4, p. and sustainable electricity supply 20. No. of households extrapolated back from 1995 at assumed growth in Lao PDR; (ii) Promote rate of 2.7% p.a. hydropower exports to earn revenues to meet GoL development objectives; (iii) Enhance the legal and regulatory framework to underpin power sector development; and (iv) Strengthen institutional structures, clarify responsibilities, streamline administration, and foster economic and social development. A subsequent workshop identified the following priorities: (i) expand electrification, (ii) complete commercialization of EdL, and (iii) develop a financing strategy for domestic and export power developments. 2 1.4 Rural electrification falls under the Ministry of Energy and Mines (MEM, formerly the Ministry of Industry and Handicrafts, MIH), with generation by Electricité du Laos (EdL) and services provided by provincial electric power authorities. In the latter half of the nineties steps were taken to commercialize of EdL as part of Government's strategy to commercialize the energy sector. Along with a reorganization of the utility into profit centers, EdL entered into a performance contract with the Government, laying out performance targets and committing GoL to take steps toward improving EdL's financial viability, including through an increase in tariffs for domestic power consumption. The utility nevertheless continues to depend on export revenue, which has historically been used to cross-subsidize domestic residential consumption. As an earlier IEG assessment pointed out, this represents an extremely regressive use of natural resource rents since it is the better off who benefit most from electricity subsidies.1 The domestic tariff structure also cross- subsidized residential consumption from revenue from commercial users. EdL has a sizeable debt from the substantial investments of recent years, the burden of which became severe following the large devaluation of the domestic currency in the aftermath of the East Asian crisis in 1997. Successive Bank projects have sought to restore EdL to a sounder financial footing through higher tariffs. The role of the World Bank and other donors 1.5 The Bank has been involved in the country's electrification from the outset, contributing toward the construction of hydropower facilities, and now includes rural electrification in the Country Assistance Strategy. Over the last 20 years Bank support has focused on the central and southern provinces, first through the Southern Provinces Electrification Project (SPE), then the Provincial Grid Integration Project (PGI)--which had a substantial rural electrification component --and the Southern Provinces Rural Electrification Project (SPRE); see Table 1. The latter project is being followed up by the Rural Electrification Project (REP I, but called SPRE II during preparation). This new project will support off-grid services throughout the country, though investments in on-grid services will continue to be confined to the seven southern provinces. Despite the introduction of off- grid components, on-grid has taken the bulk of the finances (Table 2). Table 1 Recent World Bank support to rural electrification in Lao PDR FY Cost (US$ Connections Outcome (Board- million) rating Close) IDA Total Southern Provinces 1987- Villages: 147 Electrification Project 1995 26 31 H/holds: 8,354 n.a. Provincial Grid Villages: 569 Integration 1993-99 36 49 H/holds: 40,100 Satisfactory Southern Provinces Villages: 721 Rural Electrification 1998- H/holds: on-grid 51,805 Project (SPRE) 2004 35 42 Off-grid 4,910 Satisfactory 10 Villages: on-grid 540, off-grid 200 Rural Electrification 2006- (GEF: H/holds: on-grid 42,000, off-grid (Phase 1) 2010 4) 36 10,000 n.a. 1. PPAR for Provincial Grid Integration, Report No. 24021, April 16, 2002. 3 Table 2 Share of Bank project 1.6 Consistent with the Bank's emphasis on the budgets to on-grid and off-grid financial viability of government-owned utilities, the Amount (US$ Percent conditions attached to the Bank's loans have sought millions) on-grid tariff increases to cost recovery levels, especially On-grid Off-grid since the Asian crisis worsened EdL's financial PGI 30.0 0.0 100.0 position in the years after 1997.2 However, an earlier SPRE 31.9 2.1 93.7 project evaluation by IEG concluded that although RE 26.4 2.4 91.8 the Provincial Grid Integration Project had met its Source: project documents physical targets it had not achieved its institution building objective, failing to improve the poor financial position of EdL.3 The report argued that these problems were evident at the time of project preparation and actions to address them should have been included in the conditions for effectiveness, noting also that the same argument applied with even greater force to SPRE. 1.7 The other main donor supporting the sector in recent years has been the Asian Development Bank, which has supported projects in the Northern Provinces and Vientiane Plain. In addition, UNDP supported the Vientiane Plain Rural Electrification Projects I and II while NORAD co-financed ADB's project for Vientiane Plain and Northern Provinces. Export credits from India and China are of growing importance, and both countries have projects in the southern provinces. 2. Project preparation, design, implementation and outcomes Preparation 2.1 The Southern Provinces Electrification and Provincial Grid Integration projects established the basic infrastructure for rural electrification in the central and southern provinces, drawing largely on domestically generated energy. In principle, the Southern Provinces Rural Electrification Project (SPRE) was a straightforward extension of these previous projects, requiring a relatively small amount of high-tech investment: two new substations and less than 60 km of high voltage sub-transmission lines were included in the design. The bulk of the project cost was taken up by distribution extension (poles and lines) reliant upon the existing transmission infrastructure. Technical assistance was provided by the Swedish utility, SwedPower, which designed the proposed investments in line with agreed least-cost principles. However, preparation was delayed. Appraisal was originally planned for December 1996, but eventually took place in December 1997. The project was submitted for Board approval in March 1998 and became effective in August of that year. 2. Since 1993 the Bank has, in accordance with the new energy strategy, promoted increased private sector participation in electricity generation and distribution, which in most countries has been focused on generation. Privatization is not a political option in Lao PDR at present, so the Bank has pursued commercialization. 3. PPAR for Provincial Grid Integration, Report No. 24021, April 22, 2002. 4 2.2 Three factors explain the delay. The first was the resettlement policy. The Bank argued that there was little capacity and understanding of this issue on the government side, and the consultants involved in project design were not convinced of the need to address resettlement. At an internal Bank review meeting, it was decided that the issue should not be pressed, in order to retain government support for the project. But the correspondence trail after that meeting shows that resettlement was later put back on the agenda. A mission was fielded in September 1996 with the result that a Resettlement Action Plan (RAP) was in place by the time of project appraisal. However, problems relating to this issue resurfaced during implementation. 2.3 The project was also delayed by two related factors--the commercial status and financial performance of EdL. The Bank wanted a Performance Contract between the government and EdL setting out their respective responsibilities and giving the utility more independence. This contract, prepared with Bank assistance, was signed in January 1998. EdL's domestic supply operated at a substantial loss, with the utility having increasing problems servicing its debt. The performance contract agreed on a schedule of tariff increases to address these issues. There was also an average tariff increase of 63 percent in October 1997. 2.4 The original project design was built around grid extension. But government's electrification targets could not be met by grid extension alone. Therefore, at appraisal, an off-grid component was added to the project, to be operated on a commercial basis by EdL. Objectives and components 2.5 As stated in the Project Appraisal Document (p.2), SPRE had two objectives: (i) Expand rural electricity service in 7 central and southern provinces of Lao PDR, where economically justified, through grid extension and off-grid electrification; and (ii) Strengthen EdL's capacity to plan and implement electrification investments and to operate on a commercial basis. 2.6 The project had three components: (i) grid extension; (ii) off-grid electrification; and (iii) institution building. 2.7 Grid extension: The Distribution Extension component was to increase electricity service in seven southern Lao provinces: Bolikhamsai, Khammouane, Savannakhet, Saravane, Champassak, Attopeu, and Sekong. It was planned to finance the construction of about 52 km of high-voltage (115 kV) subtransmission lines, 2x20 MVA of 115/22 kV of transformer capacity, 1200 km of medium-voltage (22 kV) lines, 900 km of low-voltage (380 kV) lines, 34 MVA of 220/380 kV distribution transformers, and 50,000 consumer meters. The project would also pilot the use of low-cost single wire earth return (SWER) systems in rural areas. 5 2.8 Off-grid electrification: The Off-Grid component was to test methods of providing electricity service in approximately 46 rural communities which will not receive grid access for at least 10 years. The pilot was to: (i) establish local institutional, financial, and technical capacity for sustainable implementation of rural diesel and renewable energy systems; (ii) implement small diesel, micro-hydro, photovoltaic systems (and others according to demand from operators); and (iii) identify and develop the objectives, institutional arrangements, and scope for a national off-grid rural electrification program. The aim was to make off-grid project development more sustainable by: (i) maximizing the involvement of the private sector and beneficiaries; (ii) operating the off-grid systems on a cost-recovery basis; (iii) defining and strengthening appropriate institutional and financial mechanisms; and (iv) developing designs to minimize costs. 2.9 Institutional Development: This component was intended to finance technical assistance and equipment for project management, commercialization and sector regulation and development. Project management assistance was to be provided to EdL, including (i) engineering consultants to provide hands-on assistance and training in project management, procurement, construction supervision, and monitoring; and (ii) equipment including vehicles for supervision and maintenance, computers and other office equipment. Financial management and commercialization support would be provided to EdL to finance the services of (i) the Resident Financial Management Advisor; (ii) the procurement and installation of billing and collections system; and (iii) short-term expert advice to EdL's General Manager as determined during the course of the project. This was expected to include advice with regard to corporate planning, budgeting, and training. Technical assistance was also to be provided to MIH and EdL to assist in regulatory and technical matters, including the formulation of a strategy for the power sector that would support implementation of the Electricity Law and the preparation of a sector development action plan and a hydropower development strategy. Overview of implementation 2.10 Implementation lagged in the first two years, with only 36 percent of the expected amount disbursed by the second year of the project (Figure 2). There Figure 2 SPRE disbursement profile were initially delays in preparing 40 bidding documents so no 35 disbursements at all took place 30 during the first six months of the sn 25 project. But the main concern was illio m 20 EdL's finances, which worsened 15 considerably following the East $SU 10 Asian crisis and the devaluation of 5 the kip from less than Kip 1,000 per 0 US$ to over 7,000 per US$ by 1999. 98 99 00 01 02 03 04 05 Within a year of the project's start, Appraisal estimate Actual the Bank considered suspending the credit unless actions were taken to resolve these financial issues. Source: ICR 6 Proposals made by the government in March 2000 were considered insufficient to remove the need for suspension. That month a supervision mission proposed a financial restructuring plan for EdL, threatening suspension unless there were signs of progress by June 2000. In April, the government wrote assuring the Bank that measures would indeed be taken to redress the situation. Its proposed plan of action included formation of a Tariff Review Committee (rather than specific proposals on tariffs). This was accepted by the Bank in October. But by mid-2001 the Bank was expressing concern that implementation of the plan had stalled. A program of tariff increases was finally introduced from May 2002 (see paragraph 2.15 below). 2.11 These problems did not hinder progress with grid extension, although there were initial procurement delays, and heavy rains slowed progress with the works. Eventually, more villages than planned were connected to the grid--721, compared to the appraisal estimate of 520, with 51,805 household connections against the planned 50,000 (Table 3). Cost savings on the planned activities allowed additional works to be undertaken. Specifically, US$3.9 million was spent on (i) upgrading of the 115/22 kV Paksan substation with 2x16 MVA transformers; and (ii) constructing a new 115/22 kV outdoor substation at Thakhek with 2x30 MVA transformers and the erection of a 3 km double circuit steel lattice 115 kV transmission line. The success of these activities demonstrates both the existing capacity of EdL in implementing grid extension, and the supportive role played by the institutional development activities to support these investments. Table 3 Physical outputs Component Target Actual* Unit 115 KV lines 52 53.4 km MV (22 kV) lines 1,200 1,554 km LV (380 V) lines 900 1,566 km 115/22 kV transformer capacity 40 132 MVA 22/0.38 kV transformer capacity 34 44.9 MVA Household connections 50,000 51,805 - * Latest estimate or actual output. Includes components not included at appraisal, procured from credit cost savings. 2.12 The off-grid component also got off to a slow start. The hope that the pilot program would be commercially viable proved over-optimistic, imposing an additional financial burden on EdL, the implementing agency. Hence the credit agreement was amended, allowing the Department of Energy to take responsibility for this component. As with grid extension, physical targets were exceeded, with 6,097 households receiving off-grid power compared to the appraisal target of 4,600. However, there have been problems with the maintenance of these systems (see paragraph 2.24 below). 2.13 In short, during project implementation, quantitative targets were amply exceeded, but there were institutional problems with both on and off-grid programs. For the on-grid program, the central concern was the performance of EdL--its weak financial position and its failure to cope with system losses and outages. The off-grid program failed to ensure adequate maintenance, reflecting flaws in the incentive structure built into program design. 7 Financial performance of EdL 2.14 EdL became an autonomous agency in 1997, charged with operating on a more commercial basis. But a review of financial performance two years later found it to be unsatisfactory. Although operating profits seemed to be satisfactory, the return on capital was low (and overstated, since fixed assets had not been re-valued). Also, residential consumers were heavily cross-subsidized by exports. Growing domestic demand reduced exports, these falling from over 80 percent of generated power in the early 1980s to less than 60 percent by the late 1990s (Figure 3). In addition, EdL was unable to finance the investment program required for continued service expansion, reducing its ability to service current and future loans. Finally, there were serious payment arrears, notably from government agencies (but also irrigation users). These problems were exacerbated by the devaluation of the kip, since most materials were imported and the loans were denominated in foreign currency. 2.15 With the assistance of the World Bank and the Asian Development Bank, the government took steps to strengthen EdL's performance. The financial restructuring plan of 2001 partially converted government debt into equity, and retroactively adjusted the terms of other loans. It was also agreed that, from January 2002 onwards, tariffs should be increased by 2 percent per month for a period of 36 months. During 2002, tariffs regained their pre- crisis dollar value (Figure 4). However, the government suspended the tariff increases in 2004, pending the results of a tariff review study.4 It then rejected the recommendations of the study so further work was done to amend the study's conclusions. The government was then invited to choose one of Figure 3 Share of exports in generation two deadlines for moving all 90 900 tariffs toward levels sufficient to cover operation 80 Exports as % generation 800 (left axis) and maintenance (O&M) 70 700 costs and achieve a 4 percent 60 600 real return on fixed assets. It ) %( Wh)G( accepted the 2011, rather 50 500 are than the 2009 deadline. shtrop 40 Total Exports 400 (GWh) (right axis) portsxel ta Tariff rises, in line with the Ex To 30 300 agreement with the Bank and 20 200 ADB, began again in July 2005 (i.e. after SPRE had 10 100 closed), with further rises 0 0 1980 1985 1990 1995 2000 2005 being built into the Action Plan for Financial Source: EdL Sustainability of November 2005 (prepared as a precondition for SPRE II, since renamed REP). This index links the tariff to the consumer price index and the exchange rate, in order to ensure that the real tariff rises by the agreed rate. Tariffs are now well established at a historical high. 4. Electrowatt-Ekono Ltd. and Fichtner EdL Tariff Study: Final Report, December 2004. 8 Figure 4 Real electricity tariffs 600 5.00 4.50 500 4.00 prices) US cents 3.50 400 2004,h (right axis) 3.00 h ip/kWk(f /kWs 300 2.50 cent rifat Real kip 2.00 SU (left axis) 200 veragea 1.50 ealR 1.00 100 0.50 0 0.00 1992 1994 1996 1998 2000 2002 2004 Sources: Tariff data from EdL, price and exchange rates from IMF Statistical Annex (various years). 2.16 Under the financial plans for EdL, it was also agreed that existing arrears would be offset against taxes. However, new arrears continue to accrue which are not covered by the agreement, though three measures are in place to limit such arrears (earmarked budget allocation for bill payment, an energy audit of government customers and an agreement that government customers can be disconnected for non-payment). Continuing measures are being taken to address this situation, but it is recognized that it may take some years for it to be fully resolved. Table 4 SPRE financial indicators 1996 2004 Accounts receivable turnover declines from 4.7 4.7 ~5 months in 1996 to 2 months by 2000 Ensure that government arrears to EdL do not n.a. 19 months exceed 60 days from 1999 onwards System losses from 30% in 1996 to 20% in 2000 and 30 20.7 16% in 20041 Return on net re-valued assets in operation of 8% n.a. 4.7 Debt service coverage equal to at least 1.51 1.2 1.2 Maintain a self-financing ratio of 20% from 1998 20 31 onward and 30% from 2000 onward1 Note: 1/ Condition from GoL-MEM agreement. Source: ICR and EdL 9 2.17 Although measures are now in place to strengthen EdL's finances, progress during SPRE was only moderately satisfactory, with all but one of the performance targets being missed at the end of the project (Table 4). Only the target for the self-financing ratio was met. However, it is not clear that this is a good measure of financial performance for a utility undertaking substantial system expansion which is mostly financed by foreign loans and grants. Such aid-financed expansion will drive down the self-financing ratio, but this does not necessarily reflect weak underlying performance (the verdict on which will hinge on whether the user charges for the new capacity are sufficient to cover O&M and debt servicing). Technical performance of the grid: system losses and outages 2.18 In 1996 system losses were very high: the cost of delivery was 30 percent higher than the cost of generation. At appraisal a target was set to reduce system losses to 20 percent by 2000, and the Performance Contract stated that they should be reduced to 16 percent by 2004. But losses remained close to 21 percent in 2004, though they had been lower in earlier years. EdL's Annual Report for 2005 states that, "system loss is high and on the rise" (p.23). The problem is worse in SPRE provinces, where system losses averaged 20.2 percent in 2004 and 20.6 percent in 2005, compared to 15.6 percent and 12.5 percent respectively for non- project provinces (ibid.). Continued high losses will further weaken EdL's financial position. 2.19 Both EdL engineers and Bank staff agree that system losses would almost inevitably have remained high under SPRE since it largely extended distribution systems on the basis of existing infrastructure. This results in long distribution networks using medium and low voltage lines--the source of the system losses. These engineers agree that as the Box 1 The problem of electricity outages: findings from network becomes denser--that is, a study of electricity supply to the private sector as more substations are added-- losses will decline. Also, REP 1 Manufacturing firms identified problems with electricity services has a specific component to reduce as the most important obstacle to business--more than 40 percent system losses which is paying for of firms identified this as a major or severe obstacle (more than the upgrading of equipment and any other issue). The estimated cost of these problems was equal wires; and additional substations to 5% of sales, double that in neighboring countries. ADB and are already being provided by other the World Bank commissioned a follow-up report, which surveyed 34 manufacturing firms. Twelve of these reported more agencies (e.g. two have been built than 10 outages in the previous rainy season. EdL data confirm in Champasak province with this seasonal pattern. Other service-related problems were also Chinese funding). By April 2007, identified, notably EdL's slowness in providing connections to EdL had already attained the enterprises. system loss target for REP of 17 Source: Tokoyo Electric Power Company (2006) "Private Sector percent. Adjustment and Investment Climate Assessment in Lao PDR: Electricity Supply Survey," report for ADB and World Bank. 2.20 Power cuts are a further problem, according to two private sector surveys (Box 1). IEG field visits confirmed that outages are a common occurrence in the rainy season; households covered by the REP 1 baseline survey reported a median of 10 outages per year, averaging three hours each. Provincial data for high and medium voltage lines confirm that cuts peak during the rainy season. IEG's impression is that the problem is 10 not being seriously addressed by EdL--or by the Bank. For example, there is no performance indicator on outages. Off-grid component 2.21 A World Bank (ESMAP) study on the appropriate institutional arrangement for the delivery of off-grid services was conducted under the auspices of SPRE.5 This study proposed that private sector operators be brought in at the provincial level, to be trained and supplied by EdL. However, the proposal was never implemented and, ultimately, responsibility shifted to the Department of Energy (Ministry of Industry and Handicrafts), with the Bank and GoL agreeing that EdL had no expertise on private sector promotion and could ill afford the financial burden associated with this component. Meanwhile, JICA had helped finance a pilot off-grid project in Vientiane Province, implemented by DOE; but it was unwilling to fund scaling up. It was this pilot which served as the basis for the SPRE- financed off-grid component (employing some of the same staff), although the model bears strong similarities to the ESMAP proposal. 2.22 The program was run by the Off-Grid Promotion Support Office (OGS) in DOE. In each province a private company was to be appointed as the Provincial Energy Services Company (PESCO). In consultation with DOE, EdL and the provincial authority, the PESCO identified those villages which would not be linked to the grid for at least ten years (this threshold is sometimes reduced to five years). From this list, the PESCO selects villages for the planning exercise: an assessment of ability to pay based on an inventory of land, livestock, and other assets drawn up in consultation with the village head. If the PESCO decides to proceed, a village meeting is held at which a Village Electricity Manager (VEM) is appointed, who is responsible for day-to-day maintenance and collecting the monthly payment. This payment is divided up between the Village Fund (10 percent), VEM, (20 percent) and PESCO (20 percent), with the remaining one-half channelled to the reflow account of OGS to maintain materials supply. For the village to be eligible for the program at least 80 percent of households must sign up at the planning stage. 2.23 The options are for diesel powered small generator, pico-hydro, or solar home systems (SHS). Wind power is considered not economically viable in Lao PDR and biomass systems are currently under development. To date, all villages have opted for SHS, except one which chose a pico-hydro system. The household can choose different panel sizes and pay an installation fee (the lowest being US$16) and then monthly payments under hire purchase arrangement over five or ten years. The cheapest panel costs Kip 10,000 (US$1) a month for ten years (Table 5). The solar power is routed through a controller to a car battery for recharging, which can then power light bulbs, radio, and TV for up to 4 hours per day. 5. ESMAP, Lao PDR: Institutional Development for Off-grid Electrification, June 1999. 11 Table 5 Solar Home System options and payments under OGS Installation Monthly Payments for Fee (kip) solar kit (kip/month) 5 year 10 year repayment repayment period period 20W 1 or 2 lamps for 2-3 hours on 160,000 20,000 10,000 evenings following sunny days. Radio-cassette. B+W TV, 1 hour. 30W 1 or 2 lamps for 3-4 hours on 190,000 30,000 15,000 evenings following sunny days. Radio-cassette. B+W TV, 3 hours. 40W 2 or 3 lamps for 3-4 hours most 220,000 40,000 20,000 evenings except when no sun for two days. Radio-cassette, TV, other 12 V appliances. 50W 3 or 4 lamps for 3-4 hours most 250,000 50,000 25,000 evenings. Radio-cassette, TV, other 12V appliances. Source: Village Energy and Electricity: Best Practice in Lao PDR, February 2004, Ministry of Industry and Handicrafts, Vientiane. 2.24 Over 6,000 households were connected to off-grid supply under SPRE, exceeding the appraisal target by one-third. However a survey undertaken in Vientiane Province in 2006 revealed a number of operational problems, with over 80 percent of SHSs not working properly or at a low level of service.6 Most controllers were no longer working: 40 percent were absent, and most of the others had been bypassed so the panel was connected directly to the battery. The resultant power fluctuations shorten battery life (which was anyway only 2-3 years), as does excess use by connecting too many lights or appliances. As a result nearly all batteries were past their useful life, with over half the batteries being more than four years old. Most households have not replaced these batteries (which cost US$20-50), but continue to charge the battery getting just 30 minutes to one hour of electricity each day. Consequently, project benefits are much lower than expected. 2.25 These problems are partly technical and partly incentive-based. There is disagreement about the technical flaws. The rival private scheme (Sunlabob; see next paragraph) uses five-year solar batteries, arguing that these are superior. OGS (now called the Village Off-grid Promotion Scheme, VOPS) program managers argue the greater cost of these batteries is not warranted. The second technical issue is the quality of training and capacity of the field staff recruited under the project (the PESCO and VEM). Sunlabob provides its engineers a higher level of training. But it is the incentive structure which puts the greatest constraint on maintenance. Payments to the provincial company, the PESCO, are 6. IED, SHS Inventory, Diagnosis and Troubleshooting: pilot survey report from Vientiane Province, October 2006. 12 geared toward planning and installation: the PESCO receives US$2 for each household which signs up and US$1 for installation. By contrast, the cost involved in collecting the monthly fee means that the PESCO barely breaks even (source: IED). Overall, being a PESCO (or a VEM) appears not to be that financially attractive. One PESCO withdrew and VOPS is having trouble recruiting in new provinces, though three new PESCOs were eventually established under REP in early 2007. 2.26 A further issue is the channel through which the service is delivered. Although the PESCO is a private company, the scheme is subsidized by Bank resources. But at the same time there is a private company that delivers SHS which has won several international awards, including one from the Bank's Development Marketplace. There is clearly not a level playing field. Sunlabob has received assistance (e.g. GTZ financed its initial training) but not on the scale of the subsidies to OGS/VOPS. In addition, Sunlabob has to pay other duty and taxes from which the government program is exempt. As a result the minimum monthly payment for a Sunlabob system is Kip 35,000, compared to Kip 10,000 for OGS/VOPS. It is telling that no Sunlabob trained engineers have applied to become PESCOs, suggesting that there is not money in it. 2.27 Is it appropriate for the Bank to subsidize a government program in direct competition with a private company? This is a question which need be confronted even if it is recognized that government's electrification targets are very unlikely to be met by relying on the private sector alone. The response under the Bank's follow-on Rural Electrification Project has been to make VOPS more commercially viable. To begin with, management of the scheme was contracted out. Although DOE retains oversight, a French company, IED, is responsible for implementation, allowing a greater degree of flexibility. IED has been discussing with the ministry the possibility of raising tariffs closer to commercial rates. While ensuring that operations are sufficiently financially attractive for the VEM and PESCO, steps have been taken to encourage VOPS to focus its activities on the less accessible and poorer regions that are less likely to attract commercial suppliers ­ hence leaving a market space for Sunlabob rather than acting in direct competition. This approach supports the government's goal of a 90 percent electrification rate by 2020. Resettlement, environmental and social issues 2.28 SPRE did not require large-scale land acquisition or resettlement; there were only three substations constructed, all on land already owned by the government. A small length of high-voltage transmission line was built. These are meant to have a 30-metre channel cleared on either side of them, requiring land acquisition, some forest clearance, and a very small number of homes to be resettled. Distribution lines, which nearly all follow the line of the road in order to keep down installation costs, have a very small "footprint." However, government is meant to provide compensation per square meter of land used where appropriate. 13 2.29 The Resettlement Action Plan (RAP), prepared by the government before appraisal, was approved by the Bank. Some problems arose during implementation--there are differences of opinion about their nature and extent. The Bank's Implementation Completion Report, in line with supervision reports, states that some provincial authorities were unwilling to reimburse owners for land acquired for public works, as this was not normal practice. Where there was reimbursement it was not at the agreed rates. The report also notes that some provincial authorities were unaware of the RAP. On the other hand, EdL states that the disagreement arose over scrub land. They agree that the Plan allows for compensation but argue that it was not appropriate to compensate households for land in which the household had not invested. Following the visit of a resettlement specialist (and a revision of the Plan) a compromise was reached. Compensation was paid for scrub land, but the money went to the village fund rather than to the household. The independent evaluation report of the resettlement component, commissioned by EdL, takes a middle Box 2 Voluntary resettlement in response to line. It notes that EdL and the electrification: Don Khor village, Champasak provincial authorities were new to the Province idea of a RAP, accepting that there may have been misunderstandings at Don Khor village is situated on the banks of the Mekong, but first and pointing to scrub land as the some village families live on an island in the river. It will be uneconomic to extend the grid to the island, so in 2004 the main bone of contention. But, the PESCO supplied solar power which was taken up by all 34 report goes on to argue, EdL did draw island households. In 2006, the grid arrived in the mainland up the maps that would be needed to part of the village. Less than one year later 10 of the 34 assess the terms of compensation and households had relocated to the mainland. it did acquire the necessary expertise Source: IEG field visit in resettlement planning. 2.30 For the reasons outlined in paragraph 2.28, the environmental impact of the project was slight, though not nil. The category B rating was therefore appropriate. The procedure is that once trees are identified for clearance, EdL notifies the Forestry Department which then sells the concession to a commercial company. According to the Forestry Law an equivalent area should be replanted elsewhere, but this does not appear to be monitored or enforced (although there are other schemes to encourage investment in tropical hardwoods, notably teak). 2.31 A final issue was the failure to enforce the safeguard for indigenous people. Many of those living in project areas are from minority ethnic groups. The Bank's Operational Directive (OD 4.20, now replaced by OP 4.10) required that an Ethnic People's Development Plan (EPDP) be prepared. No such plan was prepared, the appraisal report making no mention of the ethnic composition of the population in the project area. An EPDP has been prepared for the follow-on project, REP I, and this sets out a consultation process, in order to ensure that ethnic identity is respected. 2.32 The benefits from rural electrification are not to be denied, and are documented later in this report. However there is one issue not referred to in Bank discussions, which is that of involuntary resettlement. The government's position on resettlement admits of various interpretations. Government has in the past attempted involuntary resettlement, especially of people in remote areas. It is currently promoting a policy of "village clusters" to 14 facilitate service delivery. Clustering might entail some measurement of resettlement, potentially violating the terms of the new resettlement policy worked out with external agencies which does not permit involuntary resettlement. In some respects, government is complying. Communities on the far side of the new Theun 2 dam were encouraged by the government to resettle. But when they refused to do so the government agreed to provide them with off-grid electrification. It remains to be seen whether the resettlement policy will be honored in the long term. Rural electrification may still potentially provide a platform for involuntary resettlement. But at present rather the contrary is the case, with households voluntarily moving in order to be closer to the grid (Box 2). Monitoring and evaluation 2.33 Project monitoring was oriented toward EdL performance, in particular monitoring of compliance with the performance contract. Information was also collected on the overall outputs from the project (connections). This report has already made some suggestions on the indicators used: specifically that system loss data should also be collected specifically for the project provinces; that outages should be included as a performance indicator; and that the self-financing ratio is not necessarily an appropriate measure of financial performance at times of aid-financed expansion. Overall, the monitoring system was satisfactorily incorporated into EdL's operations, was well focused and served the project well. 2.34 At appraisal there were no plans for a comprehensive evaluation; no survey was undertaken. There had been a baseline survey under the earlier Southern Provinces Electrification Project and government had felt that the benefits were so self-evident that there was no need for an end-line survey in that case. However a baseline survey was undertaken for REP I in 2004. EdL commissioned an evaluation of the RAP (see paragraph 2.29). 2.35 Monitoring and evaluation of the off-grid component was weak, so that the extent of technical problems was not appreciated until the survey carried out by IED in 2006 (paragraph 2.24). A systematic monitoring system has been introduced under VOPS. 3. Impact of rural electrification and economic analysis 3.1 The economic analysis in this PPAR is based on data collected in the project areas in 2004 as the baseline for the follow on project, REP. This analysis underpins the ERR analysis, which addresses the efficiency question. But the data also shed light on patterns of electricity coverage and usage. Coverage and distribution of rural electrification in the Southern Provinces 3.2 The background documents for the baseline survey present data on the number of electrified households (Table 6). The first four columns show data for villages electrified by the three World Bank projects (though only the earlier years of SPRE) and other projects. The electrification rate is 31 percent in the seven provinces as a whole and 41 percent in 15 project areas (i.e. including those not yet reached under SPRE and REP 1). These figures rise to 43 percent and 69 percent respectively once the additional households electrified under SPRE by project close are included in the total. At the time of the survey, Bank-supported projects accounted for 60 percent of the electrification in rural areas of the seven provinces, rising to 69 percent including the additional households electrified under SPRE. Table 6 Breakdown of households and villages Villages with electricity Villages not yet electrified Other Not SPE PGI SPRE1A projects SPRE1B REP 1 electrified Total Villages 191 572 86 480 375 540 2969 5213 Households Electrified* 19661 46633 3508 46546 0 0 1264 117612 Non-electrified** 5110 11418 4258 14340 32917 48600 145015 261658 Total 24771 58051 7766 60886 32917 48600 194879 379270 Source: REP baseline survey sampling document. * 1264 households in "not electrified" row are off-grid. **To be electrified or villages not yet electrified. 3.3 Just over three-quarters (77 percent) of households in electrified villages are electrified, with this rate being lowest in most recently connected communities (Figure 4). For most communities which were connected more than three years ago, 80 to 85 percent of households have a grid connection. Connection rates are, as expected, lower for poor households: 65 percent of poor households in electrified villages are connected to the grid. This bias in connections is shown in Figure 5, which shows the Lorenz curve for grid connections and electricity consumption. The bottom 40 percent of the population (ranked by a wealth index, see Annex B) have just one quarter of the grid connections and account for just 15 percent of consumption of electricity supplied by the grid. Figure 4 Electrification rates by year of Figure 5 Lorenz curves for grid connection community grid connection and electricity consumption 1.20 1.0 0.9 1.00 All households noi 0.8 umpt 0.80 tear nsoc/s 0.7 cationifi 0.60 tionce 0.6 The poorest 40% of the population account for 25% of ectrlE Poor households only onnc 0.5 di grid connections and 15% of Grid connection 0.40 gr grid electricity consumption erahs 0.4 0.20 evit 0.3 Grid consumption 0.2 mula 0.00 Cu 1988 1990 1992 1994 1996 1998 2000 2002 2004 0.1 Year in which community was electrified 0.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 Cumulative share of population Source: REP 1 baseline data Source: REP 1 baseline data 3.4 However, the gap in the electrification rate between the poor and non-poor has closed over time. In the early nineties the electrification rate for the poor was only one-third 16 of that for the population as a whole, but then grew rapidly from 1996 to reach 80 percent by 2004 (see figure 6). This increase came about as electrification reached poorer villages, as shown by the poverty rate in communities by year of grid connection. 3.5 These results raise a question about which communities are connected. The Bank projects have begun from the main population centers, which are to the west of the country along the Mekong River, and then spread out from there, following the road in the interest of least cost. Regression analysis of the survey data confirms that villages nearer to roads and the provincial headquarters are more likely to be connected (see Annex B). More remote villages ­ those in mountainous areas or on river islands (see Box 2 above) ­ are less likely to be connected. Figure 6 Electrification rates for villages Figure 7 Community poverty rate by year of electrified 1990 - 2004 grid connection 0.6 90 Electrification rate (ER) Annual average 80 0.5 ER: poor only 70 Ratio oit 60 e 0.4 ra atr 50 and yt gea 40 verop 0.3 ent rceP age Three year 30 moving average verA0.2 20 10 0.1 0 1988 1990 1992 1994 1996 1998 2000 2002 2004 0.0 1990 1992 1994 1996 1998 2000 2002 Year village electrified Source: REP 1 baseline data Source: REP 1 baseline data 3.6 Thus far, EdL has been increasing coverage by extensive growth (that is extending the grid to more communities) rather than intensive growth (that is expanding coverage to households in already connected communities). As shown in Figure 4, even in communities connected 10-15 years ago, there remain around one-fifth of households which are not connected. These are poorer households who cannot afford the connection fee of around US$100 (Kip 1 million). This pattern was evident under SPRE, when there was a substantial increase in the number of connected communities ­ 721 versus 520 planned at appraisal, a 39 percent increase, but a much smaller increase of less than 4 percent in the number of connected households (from 50,000 to 51,805). Analysis of the survey data shows that the vast majority of houses that will connect do so in the first year, with most of the remainder doing so in the following two years (Figure 8). Reaching the poorer households, who can be designated as "late connectors" will require some subsidy if government's target of 90 percent coverage is to be met.7 Finding a way of doing this would be consistent with the Bank's energy strategy in which the first of four pillars is directly benefiting the poor. 7. Economic theory suggests that monopoly suppliers should practice price discrimination if they can segment the market. The market for electricity is readily segmented, since most households take the connection 17 Figure 8 Connection rate by years since grid reached village 90 ... then it takes 7 years for the next 10% to 80 connect Another 10% connect in All households 70 the next two years... 54% connect in first year 60 Poor households te ra 50 tiona rifictc 40 Ele 30 20 10 0 0 2 4 6 8 10 12 14 Years since grid connection Note: Graphs are based on fitted values from Cox (survival) regression. Sample is electrified households in electrified villages only. Source: IEG analysis of REP baseline data Uses of electricity 3.7 Lighting is the most common use for electricity: all electrified households get their light from this source. However, the majority of electrified households also have appliances that are less common in non-electrified households, notably TVs and fans, both of which are owned by over 70 percent of electrified households. Grid households watch 2.5 hours of TV a day on average, compared to less than 15 minutes for non-grid households. A significant minority of electrified households (27 percent) have rice cookers. But in general electrified households still use fuelwood for cooking purposes, though on average electrified household use less than do non-electrified ones. 3.8 There is less evidence of the use of electricity for productive purposes and it is not clear how many new enterprises have been created as a result of electrification. Less than 10 households covered by the REP baseline survey have separate workshops using electricity, and there is no evidence that these were set up as a result of electrification. There is one type of electricity-related enterprise that can be identified: the community at "the end of the line" operates battery charging centers immediately and are unlikely to wait. The connection subsidy could thus become available three years after the grid first reaches the community. Such a connection subsidy would qualify as a "smart subsidy" in recent terminology. 18 3.9 for areas not yet connected, though this business shifts once grid extension takes place. 4. Rate of return analysis The financial rate of return (FIRR) 4.1 The financial rate of return is relatively straightforward. On the benefit side are the revenues earned from new electricity sales. On the cost side are the investment costs to create the new capacity and the O&M expenses for that capacity. Given the assumptions stated in the ICR, a financial rate of return of 6.5 percent is obtained (Annex C). This is the real return, so it is more than sufficient to service the credit, which government on-lends to EdL at 7 percent (a real interest rate of zero at the current inflation rate). Since the tariff increases are index linked, this return should be protected --unless average real tariffs are allowed to fall below cost recovery at some point in the future. In addition, there is no explicit treatment of system losses in the analysis. It should be assumed that the data build in the existing level of losses, at 20 percent in 2004. Achieving a reduction in these losses to 12 percent (a 3 percent reduction in system losses a year) will increase the financial rate of return to 7.0 percent. The financial rate of return to off-grid investments is 15.9 percent (ICR calculation), reflecting the fact that this is a subsidized program. The economic rate of return 4.2 The ICR estimates the rate of return to the grid extension component to be 60.5 percent, compared to the appraisal estimate of 22.8 percent. This higher rate of return is attributed to a higher willingness to pay (WTP) than anticipated at appraisal. The WTP estimate is based on the survey data from 2004. These data were also used as the appraisal estimate for REP; the appraisal report presents the calculations in more detail. The latter analysis estimates that the WTP for grid connection for lighting alone is US$72 per month per household (equal to US$866 per year although the SPRE ICR takes a lower figure of US$584). This is a remarkably high figure, rather in excess of the total income of many households in the area, though it is commonly recognized that households are generally willing and able to devote up to 10 percent of their income to energy needs. The high estimate comes from an error in the calculation of the WTP. 4.3 The most common means of calculating WTP is to obtain two data points on the demand curve: the price and quantity of consumption with electricity and with the main fuel source (usually kerosene) of non-electrified households. The demand curve is interpolated between these two points, preferably by assuming a constant elasticity of demand, so that the area under the demand curve can be calculated as the willingness to pay. The Lao survey allowed observation of four, rather than two, data points, corresponding to those using candles and lamps, those using car batteries, those using a mixture of these sources, and those with access to electricity. As shown in Figure 9, joining these four points gives a reasonably shaped demand curve. 19 4.4 However, the Figure 9 Demand curve for lighting WTP calculation did not generate the demand 5.0 (Qo,P0) curve from joining the 4.5 four points. Rather it 4.0 estimated the WTP from ) 3.5 three separate linear mkl/pki 3.0 demand curves, each 2.5 (Q1,P1) created by joining the 000( 2.0 Excess estimate respective unelectrified ceirP1.5 option to the electricity 1.0 (price, quantity) point, (Q2,P2) 0.5 resulting in a substantial (Qe,Pe) 0.0 over-estimation of WTP. 0 100 200 300 400 500 Table 7 shows the Quantity (klm/month) derivation of the PAD/ICR estimate of Source: derived from REP PAD, p.69 US$72, and estimates using the same underlying data assuming a single demand curve (US$23) and a single demand curve in which interpolation between the points assumes constant elasticity (US$6). This final estimate of US$5.60 seems far more plausible and is the figure used here. Table 7 Different estimates of the WTP for lighting (US$/household/month) Multiple Single linear Single demand curve demand demand with constant elasticity curves curve segments Candle and lamp users 90.6 24.68 7.18 Candle, lamp and battery users 44.4 19.69 2.92 Car/motor cycle battery users 21.2 18.89 2.16 Total (weighted average benefit) 72.13 22.78 5.56 Source: IEG calculations. 4.5 The revised estimate can be checked for plausibility in a number of ways. Compared to the ICR figure of US$585, the PPAR figure is well less than 10 percent of household income rather than nearly all of it. As the ICR itself says a number of times, many households can only afford to pay US$1-2 a month, though the survey shows that on average unelectrified households pay US$3, and IEG calculation suggest that they would be willing to pay US$5.60, which is higher but plausibly so. Other sources of data can help triangulate IEG's WTP estimate. First the 2004 survey data used to estimate WTP in the ICR also contained two direct questions on WTP. The first of these was whether or not the household would be willing to pay for electrification if the cost were 1,000,000 kip (approximately 20 US$100),8 to which the majority of households (57 percent) replied in the negative. A second question asked those not willing to pay 1,000,000 kip how much they would pay, to which the median answer was 500,000 (US$50). A second source of information comes from the break during project, during which supplies of project-supplied light bulbs ran out. Anecdotal evidence is that many households chose not to replace the bulbs through commercial purchases (costing around US$7, compared to the project-supplied price of US$3.50), preferring to go six months without electricity than pay this additional amount. 4.6 The revision of WTP has, of course, affected the ERR quite considerably. Using IEG's estimate of the WTP decreases the ERR for the grid extension component from 60.5 (ICR estimate) to 12.6 percent (Table 8). The return to the off-grid component is adversely affected by not only the lower willingness to pay estimate, but also because actual hours of electric power fall considerably short of that planned because of the technical problems outlined above. The combination of these factors reduces the ERR for the off-grid component to 1.5 percent (see Annex C). In consequence the ERR for the whole project is 12.1 percent. Table 8 Estimates of the Economic Rate of Return PAD ICR PPAR Grid extension 22.8 60.5 12.6 Off-grid component 14.01 26.0 1.5 Whole project 21.62 59.1 12.1 Notes: 1/ PAD presents separate ERRs for different off-grid options, that for solar home systems is reported here as all but one village have adopted this technology. 2/ Not reported in PAD, the figure given is an average, weighted by the cost of the respective components. 5. Ratings and lessons learned Outcome: Moderately satisfactory 5.1 Relevance: The project was highly relevant, and remains so, being consistent with both the 1996 and 2005 Country Assistance Strategies. In the latter, electrification is listed under improving infrastructure to support regional integration and private sector development. Government is strongly committed to electrification, having adopted a target of 90 percent electrification coverage by 2020. Rural electrification is seen as a supporting sector for growth in the 2004 Poverty Reduction Strategy Paper (PRSP), and electrification explicitly listed in the priority sectors for poverty reduction. The off-grid component, added at appraisal, has helped electricity reach more remote communities which would otherwise have been excluded. However, the design did not explicitly address how to encourage connections amongst poor households. 8. Although the exchange rate fluctuates, many prices in Lao are quoted in dollars but paid in kip at the exchange rate of US$1=10,000 kip. 21 5.2 Efficacy: The IEG definition of efficacy is where the project objectives are met, which they were for the physical aspects of the project (with targets being exceeded), showing also the success of the project in meeting ID objectives related to technical support. But the objective of EdL operating on a commercial basis was not met in the lifetime of the project, though the basis was laid for this to be achieved at a later date. Similarly, the off-grid component achieved its physical objectives. But the aim of providing these services on a cost recovery basis was not achieved, and it has proven difficult to stimulate private sector interest in the project ­ despite the fact that there is a private sector firm also in operation. 5.3 Efficiency: The ERR for the grid extension component was at the threshold, and that of the off-grid component well below it, with an overall average return also at the threshold. In aggregate, the project was efficient at providing electricity to rural households because of the focus on grid extension. However, the off-grid component can be considered justified, despite the low ERR, as it was a pilot which will provide lessons for the future. 5.4 Taking into account relevance, efficacy and efficiency with respect to the two project objectives, the overall outcome rating is Moderately Satisfactory (see Table 9 for summary). Table 9 Summary of outcome ratings Objective System expansion Institutional development Relevance High: Rural electrification a GoL High: EdL needs to be put on priority, and contained in CAS sounder financial footing for program to be sustainable Efficacy High: Physical targets exceeded Low: Targets not met during project, though basis laid for subsequent improvements Efficiency Moderately satisfactory for grid, Low as efficacy was low. failure to reduce system losses. Off- grid ERR below threshold. Overall outcome Satisfactory Unsatisfactory Risk to development outcome: Low 5.5 The principal source of risk is the financial sustainability of EdL's operations. The Financial Recovery Program has put in place a program, agreed with government, to tackle this problem. With the agreed program of tariff increases, tariff rates will be at a level to achieve cost recovery and a return on capital. There has not been social resistance to these increases, partly since tariffs had initially fallen substantially in real terms, and remain around the average for the region. Bank performance: Satisfactory 5.6 As stated in the ICR, project preparation was sound. Agreement of the performance contract between EdL and government was a condition for negotiations, though in the event 22 proved insufficient to ensure adequate steps being taken, which only finally happened once the project was underway and the Financial Recovery Program was proceeding. Bank staff kept up pressure on this issue; but without compromising the pace of overall project implementation. Borrower performance: Moderately Satisfactory 5.7 Borrower performance was sound with respect to grid extension--where a sizeable expansion was accompanied by significant cost savings; and establishment of a successful off-grid program, albeit with some technical problems. However, overall borrower performance is rated as only moderately satisfactory on account of the time taken to resolve EdL's financial problems, leaving most financial performance indicators unmet at the close of the project, and the technical problems in both grid and off-grid components. This rating is lower than that in the ICR. However, the ICR's rating, which indicates a more positive view of the financial performance issue, appears based on developments which have largely taken place since SPRE closed. The assessment in this PPAR is based on actions taken during the lifetime of the project. Lessons learned 5.8 Consumers are willing to pay tariffs at cost recovery levels. The willingness to pay analysis shows that consumers are willing to pay for electricity at levels exceeding supply costs. Lao PDR has been able to implement sizeable tariff increases in a short space of time with no adverse social impact or notable demand reductions (indeed it can be argued that any reduction in usage by already connected customers helps extend supply to new locations). Tariffs have already reached historical highs, and will reach full cost recovery for short run marginal cost (O&M) plus a return on capital by 2011 ­ the WTP calculations show that these levels will be acceptable to consumers. 5.9 But poorer households remain unconnected. Even after the grid has been in a village for more than 10 years some 20 percent of households remain unconnected. Smart connection subsidies for late connectors would help achieve government's 90 percent coverage target. 5.10 Explicit attention needs to be paid to technical efficiency. Technical problems of system losses and outages reduce financial performance and undermine project benefits. Explicit components are needed to tackle these issues. 5.11 Off-grid components promote social equity in electrification coverage... but are not necessarily the most cost effective strategy. The off-grid component has helped reach remote communities which would not otherwise have got electricity for a decade or more. However, the rate of return to these investments is lower than that to grid extension, or of subsidizing connections for late connectors, and it not proven that the beneficiaries from off-grid programs are poorer than those benefiting from grid extension. 5.12 In countries where the private sector is relatively undeveloped substantial support may be required to provide off-grid services in poor areas. At the same time subsidies should 23 not crowd out purely commercial actors. A sensible balance would direct project supported companies to the more remote areas unlikely to be attractive to purely commercial firms, but the subsidy scheme needs to ensure that such activities are sufficiently attractive. 25 Annex A Annex A. Basic Data Sheet LAO SOUTHERN PROVINCES RURAL ELECTRIFICATION PROJECT Data (amounts in US$ million) Appraisal Actual or Actual as % of estimate current estimate appraisal estimate Total project costs 39.3 40.8 105 Loan amount 34.7 34.4 99 Cofinancing 0.74 0.7 100 Cancellation - 0.2 - Cumulative Estimated and Actual Disbursements FY99 FY00 FY01 FY02 FY03 FY04 FY05 Appraisal estimate 2.2 7.7 15.2 23.8 31.7 34.7 34.7 (US$M) Actual (US$M) 1.0 2.8 9.7 17.5 24.5 31.9 34.4 Actual as % of appraisal 45 36 64 74 77 92 99 Date of final disbursement: FY2006 Project Dates Original Actual Project concept note - 07/12/1996 Appraisal - 02/15/1998 Board approval - 03/17/1998 Effectiveness 08/29/1998 08/12/1998 Mid-term review 6/30/2001 04/11/2002 Closing date 6/30/2004 12/31/2004 Staff Inputs (staff weeks) Stage of project cycle Actual/Latest Estimate No. Staff weeks US$ ('000) Identification/Preparation 42.5 250.0 Appraisal/Negotiation 14.7 86.3 Supervision 82.9 487.9 ICR 6.3 32.5 Total 146.4 856.6 26 Annex A Mission Data Stage of Project Cycle No. of Persons and Specialty Performance Rating Implement Developm Month/Year Count Specialty ation ent Progress Objective Identification/Preparation 2/25/97 4 MISSION LEADER (1); SR POWER ENGINEER (1); FINANCIAL ANALYST(1); RENEWABLE ENERGY SPEC (1) Appraisal/Negotiation 12/16/97 4 MISSION LEADER (1); FINANCIAL ANALYST(1); ENERGY ECONOMIST (1); ENVIRONMENTAL SPEC (1) 02/11/98 1 TASK TEAM LEADER Supervision 10/13/98 3 ENERGY ECONOMIST (1); S S FINANCIAL ANALYST (1); RENEWABLE ENERGY SPECIALIST (1) 02/12/99 2 ENERGY ECONOMIST (1); S S FINANCIAL ANALYST (1) 05/24/99 2 TASK MANAGER (1); ENGINEER S U (1) 11/11/99 3 TASK MANAGER (1); FINANCIAL S U ANALYST (1); POWER ENGINEER (1) 03/17/00 4 TASK TEAM LEADER (1); S U RESETTLEMENT SPECIALIST (1); RENEWABLE EGY.SPECLST. FINANCIAL ANALYST (1) 09/28/00 2 TASK TEAM LEADER (1); S U FINANCIAL ANALYST (1) 02/02/01 3 TASK TEAM LEADER (1); S U RENEWABLE ENERGY SPEC. (1); RURAL ENERGY SPEC. (1) 04/11/02 5 TASK TEAM LEADER (1); S S HYDROPOWER ENGINEER (1); POWER ENGINEER (1); FINANCIAL ANALYST (1); RENEWABLE ENERGY SPEC. (1) 02/14/03 6 TASK TEAM LEADER (1); S S SR.OPERATIONS OFFICER (1); FINANCIAL ANALYST (2); ENVIRONMENT SPECIALIST (1); RESETTLEMENT/LAND ACQ. (1) 10/03/03 4 SR. POWER ENGINEER (1); S S FINANCIAL SPECIALIST (1); RESETTLEMENT SPECIALIST (1); ALTERNATIVE ENERGY SPEC. 05/11/04 7 TASK TEAM LEADER (1); S S 27 Annex A POWER ENGINEER (1); ALTERNATIVE ENERGY SPEC. (1); RESETTLEMENT SPEC. (1); ENERGY SPEC. (1); FINANCIAL ANALYST (1); PROCUREMENT ASSISTANT (1) 10/04/04 3 TASK TEAM LEADER (1); S S PROCUREMENT SPEC. (1); ENERGY SPECIALIST (1) ICR 01/24/05 3 TASK TEAM LEADER (1); S S PROCUREMENT SPEC. (1); ENERGY SPECIALIST (1) Other Project Data Borrower/Executing Agency: FOLLOW-ON OPERATIONS Operation Credit no. Amount Board date (US$ million) Rural Electrification Project 1 36.27 04/27/2006 28 Annex A 29 Annex B Annex B. Analysis of RE I baseline data (2004) Description of survey The survey was carried out in the seven provinces covered by SPRE I and REP 1. The sample was divided into four groups (Tables 1 and 2). Group 2 comprised households electrified under earlier projects (SPE, PGI or SPRE), group 3 was non-electrified households in the same community and group 4 was made up of households in villages to be electrified under REP 1. Finally Group 1, considered the baseline, was made up of households in villages scheduled for electrification under SPRE but not yet reached. Electrification here refers to grid connection. The sample included a small number using solar home systems (SHS) or pico-hydro, and a substantial number obtaining electricity from car batteries. Table B. 1 Survey design Electrified village Non-electrified village Electrified G2: 1,000 (actual: 1,043) households household Non-electrified G3: 1,000 (actual: 818) households G1: 500 (actual: 428) households household G4: 1,000 (actual: 915) households Table 2 shows the number of villages and households under the various projects. The sample design was also stratified for province and project but in a way that was self-weighting. Therefore survey weights need only be used to join the different groups together into a single sample. Doing so gives a sample which is representative of the areas surveyed, i.e. past and current project areas. The sample is not representative of rural areas in the seven provinces since no data were collected from non-project villages. These data cannot be used for analyzing inter-community targeting, nor for estimating access to electricity. Table B. 2 Breakdown of households and villages Villages with electricity Villages not yet electrified SPRE Other Not SPE PGI 1A project SPRE1B REP 1 electrified Total Villages 191 572 86 480 375 540 2969 5213 Households Electrified* 19661 46633 3508 46546 0 0 1264 117612 Non- electrified** 5110 11418 4258 14340 32917 48600 145015 261658 Total 24771 58051 7766 60886 32917 48600 194879 379270 Source: REP baseline survey sampling document. * 1264 households in "not electrified" row are off-grid. **To be electrified or villages not yet electrified. The total electrification rate in the seven provinces is 30 percent as shown in Table B. 2 and B. 3, whereas it is 40 percent in Bank project areas alone if including those villages not yet reached under SPRE and REP 1 ­ a discrepancy which illustrates that the sample is not representative of the provinces as a whole. Table B. 3 also shows that, at the time of the survey, Bank-supported projects accounted for 60 percent of the electrification in rural areas of the seven provinces surveyed. 30 Annex B Table B. 3 Electrification rates in all seven provinces, by project (2) (1) Households (1) as % of (3) electrified by electrified (1) as % total project households households SPE 19,661 17 5 PGI 46,633 40 12 SPRE1A 3,508 3 1 Other 46,546 40 12 Total 116,348 100 30 Source: Table 2 Sources of energy Table B. 4 shows the sources of energy. The main points are as follows: (1) there is negligible use of pico-hydro, photo-voltaic systems (PVS), and small-generators, certainly too few to analyze these users statistically; (2) grid households reduce their use of kerosene/diesel and car batteries for lighting once they are connected to the grid9; and (3) electrified households continue to use firewood (for cooking). Table B. 4 Percentage of households using each energy source Non-grid Grid All Small generator 0.5 0.0 0.3 Pico-hydro Mini 0.5 0.0 0.3 PVS 0.9 0.0 0.5 nCar battery 37.5 1.6 21.9 Firewood 94.4 80.3 88.3 Kerosene 0.7 0.3 0.5 Diesel 89.8 12.9 56.5 Candle 5.6 7.7 6.5 Grid 0.0 100.0 43.3 No. of households 2,152 1,029 3,181 Note: Sums to over 100% as households use more than one source. Source: Household survey Energy consumption can also be shown in a common metric, such as kg of oil equivalent (koe, see Table B. 5). Electrified households consume more than twice as much energy as non-electrified ones. This is partly an income effect, but also because electrified households pay much less for their energy (Table B. 6) ­ this is the "unequalising effect of electrification" whereby the better off capture the bulk of the benefits. 9. Kerosene is little used; most of the households shown in this row consume diesel. 31 Annex B Table B. 5 Consumption of different fuels, koe/household/month Non-grid Grid All Firewood 56.8 39.4 49.2 Charcoal 5.6 9.6 7.4 Diesel and kerosene 8.1 1.4 5.2 Candles 0.0 0.0 0.0 Car batteries 2.3 0.3 1.3 Small generator 0.0 0.0 0.0 Electricity 0.0 61.4 26.6 o/w lighting 0.0 2.4 1.0 appliances 0.0 59.0 25.6 Total 72.8 173.5 116.3 Source: household survey Once income is taken into account, electrification increases energy consumption by 35 percent as shown by the regressions reported in Table B. 7. There is a small, just significant, difference in the use of firewood and charcoal by grid connected households (Table 6), reflecting the use of electric rice cookers by these households (Table B. 8). Table B. 6 Energy expenditure and the cost of energy Expenditure (kip) Price Non-grid Grid All kip/koe US$/koe Firewood 48,652 33,808 42,225 858 0.08 Charcoal 2,551 9,119 5,395 729 0.07 Diesel and kerosene 27,702 5,486 18,083 3,471 0.31 Candles 8,000 8,000 8,000 562,703 50.21 Car batteries 9,062 412 5,317 4,090 0.36 Small batteries 836 42 492 4,665,319 416.25 Electricity - 9,317 4,034 152 0.01 Total 96,804 66,184 83,546 718 0.06 Table B. 7 Regression of fuel consumption on income (logged) and grid connection Biomass Total Intercept 2.50 *** 1.4 *** (16.92) (6.35) Income 0.11 *** 0.20 *** 9.04 (11.13) Grid -0.04 * 0.35 *** (-1.66) (9.68) Note: OLS with robust standard errors. Source: survey data Uses of energy Appliances The most common appliances are TV and fan, which are owned by almost 80 percent of households connected to the grid and one-third of those using car batteries (Tables B. 8 and B. 9). Of course there is also an income factor in the ownership of these items ­ but a probit regression shows that being 32 Annex B connected to the grid increases the likelihood of owning a TV by 84 percent (Table B. 10).10 The result is that TV usage is, on average, over 2.5 hours a day in grid-connected households compared to less than 15 minutes for unconnected households. Comparing TV owners only, the difference is smaller, but still substantial and significant. On the other hand, grid connection has no significant impact on radio ownership or usage.11 Table B. 8 Ownership of electricity-related consumer goods Non-grid Grid Total TV 12.93 78.15 41.16 *** Radio 19.27 15.25 17.53 Fridge 0.00 52.54 22.74 *** Rice cooker 0.00 26.90 11.64 *** Fan 0.00 72.48 31.37 *** Iron 0.00 52.54 22.74 *** Memo items: Daily TV usage (hours) 0.24 2.56 1.25 *** for owners only 1.86 3.29 3.03 *** Daily radio usage 0.38 0.40 0.38 Note: final column is t-stat on difference in proportions. *** indicates significant at the 1% level. Source: survey data. Table B. 9 TV and radio ownership by source of electricity TV Radio Grid 78.2 15.3 Car battery 35.8 31.8 Alternative energy 23.8 33.3 Small generator 10.5 25.0 All households 41.2 17.4 Source: survey data Table B. 10 Probit regression of TV and radio ownership on wealth index (log) and electricity source TV Radio dF/dX t-stat dF/dX t-stat Wealth (log) 0.31 7.78 *** 0.14 5.4 *** Grid 0.84 35.12 *** 0.00 -0.02 Alternative energy 0.23 1.58 0.12 0.59 Car battery 0.65 12.48 *** 0.18 4.68 *** Children's homework Data are available on how long children spend doing homework each evening. This is given as an average for all children in the household. Hence the relevant sample is households with children of 10. Interpreting the probit may seem odd when non-electrified households cannot have TV. Another way of doing the calculation is as follows. In a non-grid community, one-third of households use car batteries, and one third of these (i.e. one-ninth of the community) own TVs. Once the community is connected to the grid then 80% of households acquire TV, which is an incremental increase of 69%. 11. This finding is consistent with a weak relationship between electrification and radio ownership in cross- country data. 33 Annex B school age who are attending school, so the model needs to correct for sample selection bias. The selection equation is identified with distance to school. Other conditioning variables for homework are household income and parental education, plus the number and age of the children. Lighting source does not have a significant impact on time spent doing homework (see table 11). Table B. 11 Determinants of average homework time each evening Children Adult Propensity to study Study Time Propensity to read Read Time Wealth 0.27* 1.10* 0.78*** -0.14 Education (no. of years) of head 0 0.10*** 0.08*** 0.11*** Age of the head (log) -0.09 0.14 Occupation of head of household Farmer 0.02 0.38 -0.12* -0.2 Home business 0.19** 0.38 0.13 0.14 Children in the household Number 0.71*** -0.19*** Child of more than 10 years age (0/1) 2.79*** Children in school (0/1) 1.07*** Household size 0.04*** Source of Light: Grid 0.14** 1.12*** 0.14* 0.43** Kerosene -0.44** -0.38** Candle -0.22 0.2 Others -0.25 0.68** Regional Dummies Suppressed Constant -0.4 -4.56*** -1.65*** 2.34*** Observations 1992 1992 Wald Chi 407.65*** 95 Source: ESMAP (2000)* * ESMAP (2002) Rural Electrification and Development in the Philippines: Measuring the Social and Economic Benefits, Report 255/02. The analysis is based on data collected from 2000 households in four regions of the island of Luzon. Who has electricity? Determinants of grid connection. There are two aspects to this: (1) being in a community which is connected, and (2) taking the connection. Community level connections The first question can be analyzed using community level data, and is a function of ease of access to the community and its wealth ­ we should expect this to be the case as most companies used a financial return decision rule which accounts for costs (ease of access) and expected benefits (number of connections, which is a function of wealth). There are three aspects of location. First, the grid has been extended from provincial headquarters in stages. Towns and areas close by were electrified under PGI, the lines extended under SPRE and the lines are being extended further still under REP 1. These extensions mostly follow the main roads, 34 Annex B helping to reduce construction costs. In Champasak, for example, many villages are along the Mekong river, whereas the road south is some 10-20 km east of the river. Hence many of these villages were missed under SPRE, though lines were run to them for larger villages, and more are being "filled in" under REP 1. Three factors are important: distance to district headquarters, distance to road, and size of village. A second factor is village income, since utilities connect those communities which can afford the connection (meaning that at least 60 percent of villagers are projected to connect). In practice this factor may not matter, since, as outlined elsewhere, the cost of electricity is less than that of alternative energy sources, so the 60 percent target is likely to be met in all communities, though there will be non-connectors who cannot afford the connection fee. The final factor is locational: it is not economic to connect smaller island communities or those in mountainous regions. These are the areas unlikely to get a grid connection in the next ten years which are targeted by the off-grid program. Unfortunately the questionnaire did not collect information on these variables. However, ethnicity can be a proxy for living in a mountainous region, with the Lao Tueng in particular living in more mountainous regions. The model was estimated using probit (community connected to grid or not) and tobit (proportion of community which is connected). The community size variable is positive and significant, and distance to district capital significantly negative. Distance to road is negative but insignificant; but the variable is correlated with distance to the district capital and when that variable is dropped the distance to the road becomes significant. Table B. 12 Determinants of community connection Probit Tobit Coefficient t-stat Coefficient t-stat Proportion poor 0.65 0.24 0.19 0.48 No. of households 0.00 0.01 *** 0.00 0.03 ** Village layout (reference category: dense) Community along road 0.54 0.15 0.23 0.17 Scattered community -0.05 0.90 -0.06 0.71 Hamlet Hamlet population 0.00 0.80 0.00 0.40 Distance to hamlet 0.01 0.73 0.01 0.56 Ethnicity (reference category Lao Lum) Lao Sung -0.89 0.43 -0.48 0.29 Lao Tueng -1.71 0.00 *** -0.89 0.00 *** Village location Distance to HQ (logged) -0.45 0.00 *** -0.16 0.01 *** Distance to read (logged) -0.12 0.14 -0.04 0.23 Intercept 0.58 0.22 0.28 0.19 No. of observations 148 148 Note: ** significant at 5%, *** significant at 1% Household connection decision Whether a household decides to connect to the grid once it is available in the community is a function of wealth. This is so since, although electricity is available, the poor may not be able to afford the service, high connection charges being a frequent barrier. In Lao PDR it is estimated that 30 percent of the population cannot afford the connection charge of $100. 35 Annex B The implications are that those who can afford to do so connect to the grid once it becomes available, though they may retain off-grid sources as a backup in case of power outages, whilst others may wait many years to connect. Figure B. 1 shows that around 60 percent of households in Lao PDR connect within the first year; the vast majority of households that will connect do so in the first three years of the grid reaching the community. Figure B. 1 A Large Proportion of Households Connect to the Grid Immediately When it Becomes Available - But Some Remain Unconnected after Many Years 90 ... then it takes 7 years for the next 10% to 80 connect Another 10% connect in All households 70 the next two years... 54% connect in first year 60 e Poor households rat noi 50 catifri 40 ectlE 30 20 10 0 0 2 4 6 8 10 12 14 Years since grid connection Source: REP I baseline data 37 Annex C Annex C. Economic Rate of Return Grid extension As discussed in the main report, the ICR calculation relies on the calculation in the PAD for REP I. That analysis used three separate linear demand curves to give a willingness to pay (WTP) of US$72 per household per month. Using instead a single, constant elasticity demand curve, gives a WTP of US$5.60 per household per month. This figure is 11.4 percent of the WTP used in the ICR. Hence the IEG calculation takes the WTP to be 11.4 percent of that given in the ICR, although this is not reduced as consumption expands as is done in the ICR (this in turn results in a small over-estimate for IEG calculations whereas the ICR approach gives an under-estimate). Using this lower estimate of WTP reduces the ERR to 12.6 percent. Table C.1. Grid component economic rate of return (US$) Energy Consumer Energy sales surplus = Investment cost = Total Producer Net (GWh) SalesxWTP Tariff Revenue cost salesx61.24 O&M cost surplus benefit 1998 0.00 0.00 0.00 0 0.00 0.00 0.00 0.00 0.00 1999 0.00 0.00 0.00 0.11 0.00 0.00 0.11 -0.11 -0.11 2000 0.00 0.00 0.00 4.21 0.00 0.09 4.30 -4.30 -4.30 2001 0.00 0.00 0.00 6.98 0.00 0.23 7.21 -7.21 -7.21 2002 0.00 0.00 0.00 7.36 0.00 0.37 7.73 -7.73 -7.73 2003 30.80 1.38 37.80 1.16 5.6 1.89 0.49 7.97 -6.81 -5.43 2004 72.27 3.24 48.28 3.49 0.89 4.43 0.50 5.82 -2.33 0.91 2005 84.08 3.76 49.06 4.13 0 5.15 0.50 5.65 -1.53 2.24 2006 97.87 4.38 49.86 4.88 0 5.99 0.50 6.50 -1.62 2.77 2007 113.92 5.10 50.66 5.77 0 6.98 0.50 7.48 -1.71 3.39 2008 119.62 5.36 51.49 6.16 0 7.33 0.50 7.83 -1.67 3.69 2009 125.60 5.62 52.32 6.57 0 7.69 0.50 8.19 -1.62 4.00 2010 131.88 5.91 53.17 7.01 0 8.08 0.50 8.58 -1.57 4.34 2011 138.47 6.20 54.03 7.48 0 8.48 0.50 8.98 -1.50 4.70 2012 145.39 6.51 54.03 7.86 0 8.90 0.50 9.41 -1.55 4.96 2013 152.66 6.84 54.03 8.25 0 9.35 0.50 9.85 -1.60 5.23 2014 160.30 7.18 54.03 8.66 0 9.82 0.50 10.32 -1.66 5.52 2015 168.31 7.54 54.03 9.09 0 10.31 0.50 10.81 -1.72 5.82 2016 176.73 7.91 54.03 9.55 0 10.82 0.50 11.33 -1.78 6.14 2017 185.56 8.31 54.03 10.03 0 11.36 0.50 11.87 -1.84 6.47 2018 194.84 8.72 54.03 10.53 0 11.93 0.50 12.44 -1.91 6.82 2019 204.58 9.16 54.03 11.05 0 12.53 0.50 13.03 -1.98 7.18 2020 214.81 9.62 54.03 11.61 0 13.16 0.50 13.66 -2.05 7.57 2021 225.55 10.10 54.03 12.19 0 13.81 0.50 14.32 -2.13 7.97 2022 236.83 10.60 54.03 12.80 0 14.50 0.50 15.01 -2.21 8.39 2023 248.67 11.13 54.03 13.44 0 15.23 0.50 15.73 -2.30 8.84 2024 261.11 11.69 54.03 14.11 0 15.99 0.50 16.49 -2.39 9.31 38 Annex C Off-grid component The ICR reported an ERR of 26 percent. The analysis in this report makes two adjustments to the calculation. First, as argued for grid connections, consumer surplus is over-estimated. For off-grid connections this surplus is estimated at US$90.59 a year, equivalent to US$7.55 a month; this is more than the IEG estimate for the consumer surplus from grid connection, which is not plausible. However, insufficient information is available on the method of calculation of the consumer surplus to re-estimate it as was done for grid connections. Two approaches are used here. The first, as an upper-bound, is to use the same consumer surplus as that calculated for grid connections, which was US$5.60 a month, equal to US$67.20 a year. The lower bound is to take the same percentage of the ICR's consumer surplus as that from the grid estimate, which was roughly 10 percent (5.60/584.7=9.5%), i.e. US$9.06 a year. Making this adjustment gives an upper bound of the ERR of 19.9% and a lower bound of -1.7%. A final alternative approach is to assume that a linear demand curve was used in the consumer surplus calculation. Adjusting for a constant elasticity demand curve usually gives a consumer surplus around 50-65 percent less. Taking the value of one half (i.e. US$45.30 per household per year) gives an ERR of 13.5% Table C.2 Off-grid component economic rate of return (US$) No. of Consumer Total Supplier Total Total costs Net households surplus per consumer benefit benefit benefit household surplus 1998 0 1999 210 5.60 0 0 0 0 2000 210 5.60 0 0 0 350104 -350104 2001 210 5.60 1176 0 1176 90583 -89407 2002 686 5.60 3842 14440 18282 522576 -504294 2003 2896 5.60 16218 78520 94738 638203 -543465 2004 4764 5.60 26678 101320 127998 678640 -550642 2005 5888 5.60 32973 123224 156197 49536 106661 2006 5888 2.80 16486 106290 122776 41156 81620 2007 5888 2.80 16486 106290 122776 41156 81620 2008 5888 2.80 16486 106290 122776 41156 81620 2009 5888 2.80 16486 105521 122007 40772 81235 2010 5888 1.68 9892 104442 114334 40232 74102 2011 5888 1.68 9892 104442 114334 40232 74102 2012 5888 1.68 9892 104442 114334 40232 74102 2013 5888 1.68 9892 94411 104303 36888 67415 2014 5888 1.68 9892 39361 49253 16360 32893 2015 5888 1.68 9892 6116 16008 2587 13421 2016 5888 1.68 9892 0 9892 0 9892 2017 5888 1.68 9892 0 9892 0 9892 2018 5888 1.68 9892 0 9892 0 9892 2019 5888 1.68 9892 0 9892 0 9892 2020 5888 1.68 9892 0 9892 0 9892 2021 5888 1.68 9892 0 9892 0 9892 2022 5888 1.68 9892 0 9892 0 9892 2023 5888 1.68 9892 0 9892 0 9892 2024 5888 1.68 9892 0 9892 0 9892 39 Annex C A second adjustment is required to reflect the technical problems which mean that many households get power for between 30 minutes and one hour a day rather than the intended 3- 4 hours. Consumer surplus is not adjusted from 2001-2005, from 2006-2009 it is taken as half the estimate from full power, and from 2010 onwards at 30 percent. This adjustment alone reduces the ERR from 26.0% to 16.0%. Combining with the re-estimates of consumer surplus gives a range of -7.3% to 10.4%, resulting in an average of 1.5%. Using the alternative approach of estimating consumer surplus at half the ICR level gives an ERR of 1.2%. 41 Annex D Annex D. List of people met Philippe Arnou IED, VOPS office Phakdy Dindavong Manager of Finance and Debt Office, Electricite du Laos Gnanhkham Project Manager REP, Electricite du Laos (former Acting Project Douangsavanh Manager SPRE) Jill Engen International Technical Advisor, UNCDF Morten Larsen World Bank, Vientiane Peter Logan Asian Development Bank Anousak Phongsavath Deputy Chief, Rural Electrification Department; Manager PMU REP MEM Component, Ministry of Energy and Mines Keiichi Sato JICA Expert, Power Sector, Ministry of Energy and Mines Andy Schroeter Sunlabob Anjali Shanker IED, VOPS office Hatsady Sisoulath Deputy Director General, Electricite du Laos Seumkham Deputy Chief of Social and Environmental Division, Electricite Thoummavongsa du Laos Vilaphorn Visounnarath Manager, Environmental Office, Electricite du Laos Bounkeua Xayasone Deputy Project Manager REP, Electricite du Laos