Kenya Microfinance for Water Services Project. Project Completion Report TF TF TF TF093392

Kenya Microfinance for Water Services Project Project Completion Report TF057614 TF057615 TF057616 TF093392 June 2015 i

2016 The Global Partnership on Output-Based Aid The World Bank 1818 H Street NW Washington DC 20433 Website: www.gpoba.org E-mail: gpoba@worldbank.org All rights reserved.this report was produced the Global Partnership on Output-Based Aid (GPOBA). The findings, interpretations, and conclusions expressed herein do not necessarily reflect the views of GPOBA or the Board of Executive Directors of the World Bank or the governments they represent. Neither GPOBA nor the World Bank guarantees the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of GPOBA or the World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Note: All currency amounts are in U.S. dollars unless otherwise noted. CURRENCY EQUIVALENTS (Exchange Rate Effective: February 28, 2013) Currency Unit = Kenyan Shilling K Sh 1.00 = US$0.01 US$1.00 = K Sh85.95 US$1.00 = 0.76 US$1.31= 1.00 ii

ACRONYMS AND ABBREVIATIONS CAS Country Assistance Strategy ERR external rate of return EU European Community European currency FGD focus group discussions GPOBA Global Partnership Output-Based Aid Trust Fund Program CR Completion Report IFC International Finance Corporation, part of the World Bank Group IRR internal rate of return ISR Implementation Supervision Report K Sh Kenyan shilling MDG Millennium Development Goal MIGA Multilateral Investment Guarantee Agency, part of the World Bank Group MWI Ministry of Water and Irrigation NEMA National Environmental Management Authority NPV net present value NWCPC National Water Conservation and Pipeline Corporation O&M operations and maintenance OBA Output-Based Aid OVR Output Verification Report PAC Project audit consultant PAD Project Appraisal Document PIC Project Implementation Consultant PPIAF Public-Private Infrastructure Advisory Facility, a World Bank-administered trust fund SNTA PPIAF Sub-National Technical Assistance Program SPA service provision agreement US$ US dollar USAID United States Agency for International Development Water and Sanitation Program (Note: This report does not WSP use WSP to mean Water Service Provider, a common acronym in Kenya) WSP-AF Water and Sanitation Regional Program in Africa WASREB Water Services Regulatory Board WRMA Water Resources Management Authority WSTF Water Services Trust Fund iii

Kenya Microfinance for Water Services Project CONTENTS Data Sheets A. Basic Information B. Key Dates C. Sector and Theme Codes D. Bank Staff E. Results Framework Analysis F. Ratings of Project Performance in ISRs G. Restructuring H. Disbursement Graph Table of Contents 1. Project Context, Development Objectives, and Design... 1 1.1 Context at Appraisal... 1 1.2 Original Project Development Objectives (PDO) and Key Indicators... 3 1.3 Revised PDO and Key Indicators, and Reasons/Justification... 4 1.4 Main Beneficiaries... 5 1.5 Original Components... 5 1.6 Revised Components... 6 1.7 Other Significant Changes... 6 2. Key Factors Affecting Implementation and Outcomes... 7 2.1 Project Preparation, Design, and Quality at Entry... 7 2.2 Implementation... 9 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization... 11 2.4 Safeguards and Fiduciary Compliance... 13 2.5 Post-completion Operation/Next Phase... 13 3. Assessment of Outcomes...14 3.1 Relevance of Objectives, Design, and Implementation... 14 3.2. Achievement of Project Development Objectives... 14 3.3 Overarching Themes, Other Outcomes and Impacts... 16 3.4 Summary of Findings of Beneficiary Survey... 16 4. Assessment of Risk to Development Outcome...16 5. Assessment of Bank and Borrower Performance...17 5.1 Bank Performance... 17 5.2 Borrower Performance... 17 6. Lessons Learned...17 7. Comments on Issues Raised by Grantee/Implementing Agencies/Donors...20 Appendix A. Project Costs and Financing...21 Appendix B. Outputs by Component...22 Appendix C. Economic and Financial Analysis...28 Appendix D. Grant Preparation and Implementation Support/Supervision Processes. Error! Bookmark not defined. iv

Appendix E. Beneficiary Survey Results...37 Appendix F. Beneficiary Survey Results...37 Appendix G. Summary of Grantee's CR and/or Comments on Draft CR...37 Appendix H. Comments of Cofinanciers and Other Partners/Stakeholders...37 Appendix I. List of Supporting Documents...37 Appendix J. Additional Notes on the Selection and Measurement of PDO and Intermediate Outcome Indicators...40 Appendix K. Summary of Project Agreements and Amendments...44 Appendix L. Breakdown of Original and Revised Financing Figures Presented in Data Sheet A: Basic Information...45 v

Data Sheets This is a small-grants project for which an operations portal roadmap is not available and processing is done offline. The Data Sheets have been constructed from various source documents and SAP-based activity management systems. A. Basic Information Country: Kenya Project Name: Project ID: P104075 L/C/TF Numbers: Date: January 28, 2014 Type: Kenya Microfinance for Water Services TF057614 TF057615 TF057616 TF093392 Completion report prepared following Core guidelines, by agreement with TF task manager Lending Instrument: Grant Borrower: K-Rep Bank, Ltd. Original Total Amount*: Revised Amount*: US$1,151,300 Disbursed Amount: US$2,597,119 US$1,151,300 1,400,539 Total equivalent: US$2,988,528 Environmental Category: B Partial assessment Implementing Agency: Co-Financiers and other External Partners:* K-Rep Bank, Ltd. K-Rep Bank, Ltd. International Development Association (IDA) * See Appendix L for additional details. vi

B. Key Dates Process Date Process Original Date Concept Review:* none Effectiveness: March 5, 2007 Appraisal:** April 4, 2006 Restructuring: June 29, 2011 Approval: *** December 6, 2006 Mid-Term Review: none Closing: December 31, 2008 Revised/Actual Date June 30, 2010 June 30, 2011 February 28, 2013 * Scoping study done 2002, published as Mehta and Virjee (2003). Pre-feasibility work conducted 2004-2006. See World Bank 2009 (Appendix A) for a complete timeline. ** Eligibility Note and Proposal for Micro-Financing Project submitted to GPOBA (GPOBA 2006). Panel of Experts review of application completed April 11, 2006. *** Grant agreement signed. C. Sector and Theme Codes C.1 Sector Codes (as percentage of total Bank financing) Sector Original Percentage Revised/Actual Percentage Water supply 100% 100% C.2 Theme Codes (as percentage of total Bank financing) Theme Financial and private sector development Original Percentage Revised/Actual Percentage 100% 100% D. Results Framework Analysis These indicators and targets were compiled from different project source documents, including the project proposal, grant agreement and restructuring paper. Normally, indicators and original target values in the following data sheets are taken from the Results Framework in the Project Appraisal vii

Document (PAD), but this project did not have a PAD. 1 Appendix J explains how the indicators were selected and measured in the absence of a PAD. D.1 Project Development Indicators Indicator Indicator 1: Unit = persons Baseline Value Original Target Value Formally Revised Target Value Actual Value Achieved at Completion Number of people with access to water under the Project Not established* 60,000 100,000 190,119 Dates Achieved -- December 2006 June 2011 February 2013 Comments (incl. % target achieved) 190% target achievement. The actual value at completion represents the estimated number of persons served through new and existing connections (household, kiosk and standpipe, agricultural) on the 35 schemes. * The original baseline value is not available because baseline surveys were completed only for 10 out of the originally envisioned 21 projects during the pilot phase from 2006-2010 (GPOBA 2009, pg. 2, GPOBA 2011, pg. 3). D.2 Intermediate Outcome Indicators Indicator Indicator 1: Baseline Value Original Target Value Formally Revised Target Value Actual Value Achieved at Completion Number of community subprojects accessing finance under the Project Unit = subprojects 0 21 40 35 Dates Achieved Comments (incl. % target achieved) December 2006 87.5% target achieved December 2006 June 2011 February 2013 1 At the time this project was approved, recipient-executed trust fund projects under US$5 million were not required to have PAs. viii

Indicator Baseline Value Original Target Value Formally Revised Target Value Actual Value Achieved at Completion Indicator 2: Number of new water connections Unit = connections 0 Not established 8,000 11,505 Dates Achieved Comments (incl. % target achieved) Indicator 3: Unit = household connections (incl. agricultural connections used for domestic purposes) Dates Achieved Comments Indicator 4: Unit = kiosk and standpipe connections Dates Achieved Comments December December 2006 June 2011 February 2013 2006 134% target achieved: Household, institutional, agricultural, standpipe, and kiosk connections were counted. Institutional connections are not counted in any of the other indicators. No. of new household connections at agreed working conditions and service standards 0 Not established Not established 11,398 December December 2006 June 2011 February 2013 2006 The total includes 231 agricultural connections (built expressly for irrigation), but also used for domestic purposes. Note that the OVRs referred to a number of connections as inactive, and data on the condition and service standards was not available for most schemes at the time of project close. No. of new kiosk facilities at agreed working conditions and service standards 0 Not established December 2006 Not established 62 December 2006 June 2011 February 2013 Note that the OVRs referred to a number of connections as inactive, and data on the condition and service standards was not available for most schemes as of project close. Indicator 5: No. of existing projects with an increase in revenue, measured as the average monthly revenue collected from water sales over two months after project commissioning Unit = Sub-projects 0 Not established Dates Achieved Comments (incl. % target achieved) Not established 12 December 2006 December 2006 June 2011 February 2013 100% achieved: Of the 33 schemes (35 subjects with two schemes having 2 subprojects each), 12 schemes were extant prior to the loan. ix

Total expenditure (USD) E. Ratings of Project Performance in ISRs Not available. This is a data sheet (disbursed amounts by Implementation Supervision Report [ISR] dates and ratings) internally generated by the Operations Portal. The Operations Portal has not been set up to generate CR templates for this project; no ISRs were required. F. Restructuring The project was restructured on June 29, 2011 to extend the closing date for the third time (World Bank 2011, pg. 1). Target values for some indicators were increased. See Appendix J for a discussion of the indicators and target values. See Appendix K for a timeline of the various amendments to the grant agreement, including restructuring. G. Disbursement Profile Normally, the Operations Portal generates a disbursement profile (original disbursement forecast against formally revised forecast against actual disbursement). The Operations Portal is not set up to generate any CR templates for this size project; an CR is not required. The project team produced the following graph depicting the movement of the project grant over the project period. A total of US$2,597,118 was disbursed. 3,000,000 2,500,000 2,000,000 1,500,000 1,000,000 500,000 Grant Movement - Utilization Schedule x

1. Project Context, Development Objectives, and Design 2 1.1 Context at Appraisal Millennium Development Goals By 2006, Kenya required considerable investment to reach the Millennium Development Goal (MDG) to provide 72.5 percent of the population with safe water supply access. At the time, safe water coverage in rural areas was just 52 percent. To meet the MDG required adding infrastructure providing access to an additional 5.6 million rural residents as well as rehabilitation of existing infrastructure. Infrastructure in suburban areas also required expansion or upgrading. Kenya s government budget was unlikely to meet this level of investment (GPOBA 2006). Community-managed Piped Schemes In 2002, community-managed piped schemes provided water to about 30 percent of the eight million rural Kenyans with access to safe water. Communities, self-help groups, the government, non-governmental organizations (NGOs), and external support agencies provided the investment capital for these schemes (Njonjo and Lane 2002, pgs. 5-6). Community organizations with autonomous legal status typically managed these systems (GPOBA 2006). Water Sector Reform The government had been implementing reforms in the water sector when this project began. The 2002 Water Act, which established new institutions to take over much of the responsibility for service provision from the Ministry of Water and Irrigation. The National Water Conservation and Pipeline Corporation would play the key role in these reforms. Key organizations mandated under the 2002 Act include: Water service providers: The principal types of providers are public utility companies, private operators of small piped networks, and community water users associations. Water Service Boards (WSB): WSBs have the legal mandate to ensure that water and sanitation services are delivered within their jurisdictions. WSBs sign service provision agreements (SPAs) with water service providers, and monitor SPA compliance. Water Services Regulatory Board (WASREB): WASREB is the national regulator responsible for tariff setting, quality standards, and licensing for WSBs. Water Services Trust Fund (WSTF): This institution provides grant finance for capital investment in rural water supplies. National Environmental Management Authority (NEMA): This institution provides environmental clearances to water service providers. 2 This section of a Completion and Results Report (CR) is usually completed using information drawn from the Project Appraisal Document (PAD). However, a PAD was not prepared for this project, in line with then World Bank procedures for small trust fund grants. Therefore, this section has been completed instead using information from the GPOBA Eligibility Note and its review (GPOBA 2006, 2006b), the original grant agreements and their amendments (IDA 2006, EU 2007, Bruce 2007), and the Operations Manual (K-Rep 2006). 1

Water Resources Management Authority (WRMA): This government authority protects and conserves water resources. Water service providers must receive a permit from WRMA to abstract a designated volume of water. The above organizations had begun to transfer assets and responsibility for service delivery to autonomous local service providers at the start of this project (GPOBA 2006, pg. 6). Country Assistance Strategy In 2004, the World Bank issued a new Kenya country assistance strategy (CAS), jointly prepared with the International Finance Corporation (IFC) and Multilateral Investment Guarantee Agency (MIGA). The CAS aimed to mobilize private sector participation in the provision and management of infrastructure, including water supply. It also aimed to build capacity and to expand the services of financial institutions. The CAS cited several examples of programs involving Bank and IFC cooperation in promoting private investment in infrastructure building (World Bank 2004, pgs. ii, 2-3, 7, 15, 22-23). Project Antecedent K-Rep Bank Ltd, the grant recipient, isa microfinance institution registered as a commercial bank in Kenya. Prior to this project, the Water and Sanitation Program Regional Program in Africa (WSP-AF) and the Public-Private Infrastructure Advisory Facility (PPIAF) had collaborated with the Ministry of Water and Irrigation, the Nairobi Water Services Board and K-Rep Bank to mobilize market-based funding for small community-managed schemes in Kenya. The work to-date focused on: support in developing credit assessment and appraisal tools for community-managed schemes; assessment of market lending to community-managed schemes in Kenya; assessment of the potential for business development services to assist communitymanaged schemes to develop into small water enterprises; and the development and assessment of three subprojects for K-Rep appraisal and preliminary loan approval (GPOBA 2006, pg. 9). Project Rationale 3 The project aimed to increase access to water supply services by leveraging community equity with commercial financing, and by managing the viability of projects in lower income rural communities with the support of output-based subsidies. To this end, the project sought to introduce a new way of financing community-based water supply by using both commercial financing and output-based aid, instead of relying on government grants to self-help groups. The rationale for taking this approach to providing access was as follows: 3 The following paragraphs are based on the Eligibility Note (GPOBA 2004) and Mehta and Virjee (2007). 2

Traditional financing was limited and not sufficiently focused on sustainability: Donors, NGOs, government allocations, and funds raised by community groups traditionally fund community water supplies. These financing sources have several drawbacks. For example, external funding sources often prefer new investments instead of improving the operations and maintenance of existing schemes. Traditional funders also did not conduct rigorous business planning and appraisal of financial viability, nor did they provide communities access to professional levels of design, implementation, and management services. Microfinance institutions could expand available financing, reduce lag times, and introduce commercial rigor to all phases of the project cycle: Microfinance institutions provide an alternative and immediately available financing option for community water supply projects. This would reduce the time communities spend often over five years searching for financing to improve water supply. Commercial financing would also introduce rigor and professional support to the water supply projects by requiring schemes are sufficiently viable to repay the loans. Output-based Aid (OBA) subsidies could develop the link between microfinance institutions and relatively poor rural and suburban communities: Financial viability studies indicated that consumers could afford tariffs sufficient to cover operations and maintenance (O&M) costs, and some debt service on capital. However, Kenya s relatively poor communities could not afford the high interest rates associated with commercial capital. Such an approach was clearly in line with the CAS and its goals of getting the private sector more involved in infrastructure and service delivery, and building and expanding the capacity of financial institutions to serve the non-government sector. 1.2 Original Project Development Objectives (PDO) and Key Indicators Normally the Project Development Objective (PDO) and key indicators are definitively stated in the Project Appraisal Document (PAD). However, this project did not have a PAD: at the time of project approval, PADs were not required for small recipient-executed trust funds such as this one. The original PDO is clear, despite the absence of a PAD, because the various original project documents state the project objective in nearly identical language. The original 2006 grant agreement states: The objective of the Project is to increase access to and efficiency in water supply services for the poor in rural and suburban areas of Kenya through investments in selected community subprojects (IDA 2006, pg. 6). 4 4 The Eligibility Note (GPOBA 2004, pg. 9), which is the closest equivalent to a PAD available from among original project documents, uses nearly identical language in stating the objective. The objective of the pilot is to increase access to water supply service by the poor in the rural and peri-urban areas of Kenya and to increase the efficiency of water supply services. The Note adds that the purpose of the pilot is to demonstrate that microfinance has an innovative role to play in financing small water infrastructure in Kenya... 3

By contrast, defining the project s key indicators and their original target values in the absence of a PAD is not clear-cut. Various original project documents do not use the same set of indicators, and often do not set any target values. The following indicators were selected from various project documents. Appendix J provides the detailed rationale for their selection. For the purposes of the CR, the first indicator (number of beneficiaries) has been defined as the PDO indicator, and the remaining indicators as intermediate outcome indicators. Table 1.1 Indicators and Original Target Values Indicator Original Target Value Number of people with access to water under the Project 60,000 Number of community subprojects accessing finance under the Project 21 Number of new working connections Number of new individual household connections at agreed working conditions and service standards Number of [new] kiosk facilities at agreed working conditions and service standards Number of existing projects with an increase in gross revenue; measured as the average monthly revenue collected from water sales over two months after project commissioning Source: See Appendix J To be defined during subproject design and appraisal. Target values to equal the sum of values for all approved subprojects 1.3 Revised PDO and Key Indicators, and Reasons/Justification The project was restructured on June 29, 2011. The restructuring paper listed three performance indicators with revised target values, as shown in table 1.2. The target values for these indicators were expanded due to the increased scale of the project as of 2008 (see Section 1.6 for further explanation). 4

Table 1.2 Indicators and Their Target Values, As Given in the 2011 Restructuring Paper Indicator Revised Target Value Number of people with access to water under the project 100,000 Number of community subprojects accessing finance under the project 40 Number of new water connections 8,000 Note: No original target value had been established for the third indicator, number of new water connections. See table 1.1 for further explanation. Source: World Bank 2011. 1.4 Main Beneficiaries Consumers served by the water schemes that received project financing comprised the main beneficiaries. The project originally targeted poor rural areas of the country with an estimated average annual household expenditure of US$150 per month or less (GPOBA 2007, pg. 2). Beneficiaries would receive two types of benefits: Existing consumers would receive better standards of service in terms of quantity and hours of supply, due to improvements in the capacity and management of existing schemes. New consumers would receive access to water services either from new schemes or from the expansion of existing schemes. The original project targeted 60,000 beneficiaries served through existing or new connections from the 21 schemes in the districts and city area under the Athi Water Services Board (formally called the Nairobi Water Services Board (WSB)). The revised project expanded the number of targeted beneficiaries to 100,000 persons served through existing or new connections from 40 schemes located in seven of the eight WSB service areas. Schemes under the Northern WSB were excluded due to logistical difficulties of working in this area of Kenya. 1.5 Original Components The grant agreement specifies two parts: investment financing and project management. Investment Financing The project was intended to provide output-based aid subsidies to community scheme management groups that successfully used commercial loans to rehabilitate and augment existing piped water schemes or to construct new piped schemes. The community group would provide 20 percent of the financing upfront, and receive a loan for K- Rep Bank for the remainder. K-Rep Bank followed its normal due diligence process for the loan 5

appraisal. The community group would also receive a subsidy of up to 40 percent of total capital expenditure after meeting the output targets, which were an agreed number of connections and a certain level of monthly revenues from consumer payments for two months following project completion. This subsidy would allow the group to repay a large portion of the loan immediately. Project Management Community groups were eligible for an additional subsidy to pay Project Implementation Consultants (PICs) for technical services during construction, such as procurement, contractor supervision, reporting, and building the technical, financial, and organizational capacity of the users associations to manage the schemes. This component also financed the Project Auditing Consultant (PAC), which prepared the baseline and Output Verification Reports (OVRs), certifying that outputs had been achieved and a given amount in subsidies could be disbursed. 1.6 Revised Components Project components were not redesigned. However, the scale of the project was expanded from 21 subprojects limited to the area covered by the Athi WSB to a nationwide project involving 40 schemes in areas covered by all but one of Kenya s WSBs. 5 A European Union Water Facility grant that more than doubled the financial size of the original project facilitated this change. 6 1.7 Other Significant Changes Additional Assistance to Prepare Loan Applications: PPIAF awarded a grant of US $523,719 to finance a Project Development Facility at WSTF after project start-up. The final amount disbursed under this facility was US$325,951, with the remaining US$197,768 refunded (World Bank 2009b, pg. 3,6; Advani 2011, pgs. 12-13; personal communication with the author). The facility partly financed the costs of support organizations (NGOs and consultants) to assist communities with feasibility studies and bankable K-Rep loan applications. WSTF signed a contract with Price Waterhouse Coopers (PwC) on May 14, 2009, to act as the Project Development Facility Manager (World Bank 2009b, pg. 3,6; Advani 2011, pgs. 12-13; personal communication with author). Ultimately, 35 project development grants of US$9,000 each were awarded to 33 communities (2 communities had 2 grants each) after the management committees submitted viable expressions of demand and contributed US$2,000 toward the cost of the support organizations to assist in preparing the loan applications (WSP, GPOBA, et al 2011). 5 Schemes under the Northern WSB were excluded due to logistical difficulties in working in that area of Kenya. 6 See Appendix A on changes in the financial size of the project, and Appendix K for a timeline of amendments to the project grant agreement. 6

Partial Credit Guarantee: K-Rep secured a revolving US$5 million guarantee facility from the U.S. Development Credit Authority in 2009. The terms allowed for sharing losses on a 50 percent pari passu basis at a portfolio level (GPOBA 2009; Advani 2011, pg. 12). Project Closing Extension: The project closing was extended three times, as detailed in Appendix K. Allow Municipal Water Service Providers to Participate in the Project: The second supervisory mission agreed that the project could be extended to subprojects that would be owned and managed by WSBs and municipal water utilities, rather than limited to schemes in which community management committees were the designated water service providers. Seven subprojects had this new type of management structure: Mukangu (KIRIWASCO), Nyamasaria (KIWASCO), Kenol-Kabati (MUSWASCO), Kiambi Network (EWASCO), Kamweli (KIRIWASCO), Rabai Power (KIMAWASCO) and Kayole Soweto (NCWSC). In the case of Nyamasaria, KIWASCO delegated the authority to identify and connect new consumers, and carry out meter reading, billing and revenue collection to community contractors, or Master Operators. These operators were identified and trained by KIWASCO (Tertiary Engineering Consultants 2007). 2. Key Factors Affecting Implementation and Outcomes 2.1 Project Preparation, Design, and Quality at Entry The project exceeded most target values for its (imputed) PDO and outcome indicators. First, the project design team correctly identified a niche for providing investment subsidies to water schemes owned and managed by community organizations through a microfinancing institution. As a result of this innovative project design, the borrower (K-Rep Bank) and other banks have become keenly interested in this business line, provided that substantial subsidies are available for the costs of constructing, rehabilitating, or expanding the scheme networks and main civil works. The Dutch WASH Alliance and the United States have developed projects with similar designs, based on the experience with this project. 7 Second, the output-based aid approach created incentives for K-Rep Bank to take on a substantial role in oversight, which in turn led to cost savings and higher construction quality (K-Rep Bank 2013, pgs. 12-13). Significant preparation work was undertaken before the project was approved, notably: 7 K-Rep Bank has signed a memorandum of understanding with the Practical Action to provide loan products to small and medium enterprises for waste management, water, sanitation, and hygiene. The USAID SUWASA Kenya project will partner with several commercial banks to provide loan products for water services to the urban poor. K-Rep has partnered with SUWASA to adapt these loans to financing urban utilities in poor areas. Communication with former K-Rep Special Projects Depart Head, and former project manager, October 17, 2013, Nairobi, and Samual Baiya, WSP-Kenya, email, November 25, 2013. 7

the development of credit assessment and appraisal tools for community managed piped schemes; an industry assessment of the market for financing such schemes in Kenya; an assessment of the potential of consultancy firms to provide business support services to relevant schemes; and the development and assessment of three subprojects for appraisal and preliminary loan approval (GPOBA 2006, pg. 9). Despite this preparation, the project immediately fell behind schedule and faced other challenges, described below. Certain relatively small changes in project preparation and design could have mitigated implementation delays and their consequences, including: An evaluation of the capacity and time needed for WSTF and K-Rep to complete all the steps leading up to loan appraisal. More time allotted to secure ownership of the project processes by the relevant institutions and organizations. Stakeholders could have prepared using project documents such as the ESMF and Project Operations Manual. Stakeholder participation should have been made conditional upon workshops and training sessions attendance. A loan disbursement window that ended well before the project closing date. The above procedures are typical for World Bank projects, particularly when lending to financial intermediaries. It was not clear at the time of project appraisal that these procedures would be necessary given the relatively small recipient-executed trust funds. Readers will recall that for this same reason, no PAD was prepared for this project. Other factors behind the implementation delays indicate some fundamental weaknesses in the original project design, including: Limited private sector capacity: The lack of capable support organizations and project implementation consultants was a major factor in delays. Limited demand for financing from community management committees: Identifying community-managed schemes with management organizations that were able to present adequate loan applications proved difficult. In addition, some communities did not take up the loan offers. The World Bank approved a significant change in project design whereby loans could be extended to schemes owned and managed by WSBs and municipal water utilities, rather than those in which community management committees were the designated water service providers. K-Rep Bank, in its final report on the project, stressed the potential for lending to water utilities (K-Rep Bank 2013, pgs. 13-16). 8

2.2 Implementation The time lag from subproject identification to loan took longer than anticipated during the pilot phase (2006-2010) due to the time required to: obtain community management committees registered as legal entities; 8 sign a Service Provision Agreement with the WSB; 9 fulfill the water abstraction permit requirements from WRMA(World Bank 2009a); obtain approval of project reports by NEMA (World Bank 2009a); carry out the considerable, unfunded work in developing the proposals initially required from the Athi WSB. These responsibilities were later transferred to K-Rep Bank (World Bank 2009a); and resolve title disputes for the land where the scheme infrastructure would be (WSP, GPOBA, et al 2011). PICs initially performed poorly, partly because they lacked experience. Meanwhile, communities lacked the expertise to manage PICs, and the contracts were not sufficiently rigorous in terms of stating outputs and penalties for failing to deliver on time. As a result, there were considerable delays in planned versus actual achievements for the project as originally designed. For example: One subproject was completed by December 31, 2008, the original project closing, versus the planned 21 subprojects (World Bank 2009). One year after the original closing date (December 2009), six of the originally planned 21 subprojects had been completed. Ten of the originally planned 21 subprojects were completed by November 2010, about two years after the original closing date (Advani 2011, pg. 1). 10 8 Most community water scheme management committees were registered simply as self-help groups, which was not sufficient for loan purposes. Legal registration was delayed due to the need to get ministerial signatures (GPOBA 2009a, pg. 5). 9 Getting SPAs signed was delayed because the WSBs had not resolved how to allow for community management committees as providers in areas that had been granted to municipal water companies. The WSBs worked out various arrangements, all of which required approval by the WSRB. The project also encountered situations whether overlapping SPAs had been issued, and where the exclusivity clauses under SPAs had not been enforced (World Bank 2009a). Simplified procedures and contracts needed to be established (World Bank 2009b, pg. 6). 10 By January 2009, one subproject had been completed, and seven were in various stages of construction. A total of 13 subprojects (including the preceding 8) had been identified as eligible for loans (World Bank 2009). By December 2009, six subprojects had been completed, with another six loans approved, of which 3 subprojects were under construction. Meanwhile, the scaling up had begun, with 24 sub-projects under study (World Bank 2009, pg. 4). Note: A February 2010 GPOBA Semi-Annual Report gave conflicting account of outputs, reporting five completed subprojects, and nine subprojects under construction (GPOBA 2010, pg. 2-3). 9

Implementation improved during the second national phase (2011-2013) of the project. The project team deserves credit for taking steps to incorporate these lessons, notably: K-Rep Bank assumed from the communities more contractual responsibilities for the management and supervision of PICs, and set more rigorous standards for PIC prequalification (World Bank 2009a, 2009b, pg. 6). K-Rep Bank branch staff and the Special Operations Department devoted considerable staff resources to supervising the project. The project began extending loans to communities whose schemes were owned by WSBs and managed by municipal water utilities. WSP-AF provided technical assistance to K-Rep Bank. Disbursements indicate that most construction occurred toward the end of the project (see Figure 1 below). Nearly three-quarters of the 25 OVRs for the second phase of the project were conducted less than one month before project closing. Three OVRs were conducted on the last day of the project. (Appendix B, table B.1 presents the dates of the 35 OVRs for project phases 1 and 2.) Figure 1: Cumulative K-Rep Loan Disbursements for Subprojects in US Dollars Source: Prepared by project team, WSP-AF, from data in project monitoring system. Implementation delays led to a several negative consequences, namely: A number of communities failed to receive the planned subsidies because they had not finished the planned number of connections by project closing. Some schemes had little to no period of post-implementation management support and monitoring, as construction was completed very near to the project closing. 11 Revenue projections had to be used in place of actual revenue in order to calculate the subsidies. Key data with which to evaluate the project outcomes were not available on all or most of the schemes. For example, Phase 2 projects had not established whether they could meet their target monthly average revenues. Meanwhile, the PAC reported that most subprojects 11 As will be described in the section on post-implementation support, WSP will try to address this problem through technical assistance starting in late 2013. 10

held their revenues as debt (Tertiary Consulting Engineers 2013, pg. 8). Only eight subprojects were able to report on nonrevenue water, and only 24 subprojects could report on collection efficiency (See Appendix C). Some civil works were apparently not completed by project closing (see Appendix B). 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization (a) M&E Design The borrower s M&E system comprised three components: Pre-loan monitoring: K-Rep Bank tracked the progress of the subprojects toward a loan offer and acceptance. As mentioned above, several of these steps proved time-consuming, such as obtaining legal registration for the scheme management committees; receiving SPAs from the WSBs; obtaining environmental clearances from NEMA and water abstraction permits from WRMA; and securing land (title or access) for the infrastructure. Construction monitoring: The PICs were required to provide monthly diagnostic reports. This monitoring continued for the 24 months during which the contract with the PICS was in effect. OBA output monitoring: The process was designed as follows: The OBA target outputs, derived from the scheme business plans for each subproject were clearly stated in the K-Rep loan. Prior to construction, the Project Auditing Consultant (PAC) visited the subproject to assess the baseline situation relative to specific outputs, for example the number and type of existing connections, and to confirm the eligible project cost, cost per capita, and expected OBA subsidy (i.e. 40 percent of the eligible cost). Upon completion of construction, the water user association submitted an Outputs Report to K-Rep Bank. The PAC then revisited the subproject to verify the outputs that had been achieved relative to the targets, and to recommend payment of the subsidy based on what percentage of the outputs had been achieved. 12 If the outputs were less than 100 percent, the PAC conducted an additional field visit, not more than six months after subproject commissioning. The remaining subsidy could be authorized if all output targets had been achieved (K-Rep 2006, pg. 17). In addition to the above M&E system, the schemes were to be supervised by the sector institutions and NEMA: 12 Each subproject is assigned a score that weights coverage and service levels achieved against established targets. The total score is calculated as = (coverage weights x coverage score + service weights x service score). 11

Water Service Boards: The WSBs were responsible for requiring WUAs to submit annual reports with specific indicators to track performance in sustaining outputs. Recommended indicators included: number of connections revenue total number of households benefiting from increased access average daily consumption per connected household percentage of population in target area with access to water supply collection efficiency (K Sh collected/k Sh billed) average hours of access to water supply per day number of interruptions per given period (K-Rep 2006, pg. 25) Environmental Safeguards: NEMA and WRMA were responsible for monitoring compliance with the environmental clearance and water abstraction permit. The scheme management committees were required to submit an annual audit report to NEMA as part of this process. WSP-AF monitored the major milestones in the loan development and implementation for each subproject. (b) M&E Implementation The M&E system allowed K-Rep Bank and WSP-AF to identify and track problems in the subproject cycle. For example, it spotted weak operations in the Mataraa scheme and management problems in the Gitaru scheme (WSP, GPOBA et al., 2011, pg. 3). However, the M&E system was less effective at collecting data to track the PDO and outcome indicators. Appendix B details the inaccuracies in estimating the number of connections, and Appendix J described other challenges in measuring the PDO and Outcome indicators. A principal problem with the M&E was that it did not track the performance of the schemes beyond the first two months of operations. For a slight majority of schemes, data on the first two months of operations are not even available since construction was completed so close to project closing. There is also no detailed analysis of operations and efficiency for the ten schemes that have been in operation for over two years. The economic and financial analysis had to rely largely on data from the OVRs and focus group discussions in six subprojects. The M&E system was also weak in monitoring the performance of sector institutions. As noted above, the WSBs, WRMA, and NEMA play critical and ongoing roles in the regulation and oversight of community-managed schemes. While the M&E system performed well in tracking whether scheme management applied for SPAs, water abstraction permits, and environmental clearances, it did not track how well WSBs, WRMA, and NEMA evaluated and monitored compliances. 12

2.4 Safeguards and Fiduciary Compliance Safeguards An Environmental and Social Management Framework and a Resettlement Policy Framework were prepared. The initial prescribed steps in these documents were followed, such as obtaining water abstraction permits and environmental clearances for each subproject. It is unclear whether the scheme management organizations submitted the required annual environmental audits, although for most schemes, the first audits were not due until after project closing. Fiduciary Compliance K-Rep Bank submitted quarterly Financial Monitoring Reports on time and with the approval of the World Bank. K-Rep Bank submitted a financial audit report covering the period from inception to June 30, 2010. The audit was conducted in accordance with international accounting standards and the auditor issued an unqualified opinion on the financial statements and special account statements as of June 30, 2010. No major accounting and internal control weaknesses were noted during the audit that would affect the project implementation, and the World Bank accepted the report (WSP et al., 2011). K-Rep Bank submitted a subsequent financial audit report, covering the period between July 1, 2010, and June 30, 2012 (Kimani and Associates 2012). The 2011 Supervision Mission attempted a procurement review, although it did not provide all necessary information. The mission noted the need for K-Rep Bank to (i) improve procurement filing and record keeping; (ii) adhere to the provisions of Bank s Guidelines in carrying out procurement activities; (iii) use Bank s standard bidding and request-for-proposal documents wherever possible; and (iv) prepare and submit to the Bank a procurement plan for the proposed project extension for review and clearance. K-Rep Bank subsequently worked with a World Bank Procurement Specialist to update the project operations manual according to Bank procurement procedures. 2.5 Post-completion Operation/Next Phase WSP-AF recognized that most scheme management organizations lacked the skills and systems to manage the schemes efficiently, and were unlikely to have sufficient revenue to hire private operators. WSP-AF therefore plans to recruit a consulting firm to provide technical assistance to the 33 schemes, and to develop training materials and management systems. WSP-AF has identified four options for a future project building on this one: A repeat project with K-Rep Bank Work with WSTF to use donor funding to provide loans to rural community-managed schemes Work with WSTF to use donor funding to provide loans to rural community-managed schemes using an OBA approach 13

Work with WSTF to borrow funds to provide loans to rural community-managed schemes using an OBA approach 3. Assessment of Outcomes 3.1 Relevance of Objectives, Design, and Implementation Relevance of Objectives: The need to improve access and efficiency in schemes owned and managed by communities remains highly relevant, as this infrastructure will continue to provide water to a significant portion of the rural and suburban populations. Improving access in suburban areas to schemes owned by WSBs and managed by water utility companies also emerged as a relevant project objective. Relevance of Design and Implementation: Project design has a number of elements. A key element in the project design was to provide subsidies through commercial banks for constructing, rehabilitating, or upgrading the network and major infrastructure of the schemes owned and managed by communities. The subsidies indirectly serve to make the water tariffs and interest rates affordable to the consumers and management committees respectively, and to reduce the risk to the bank (K-Rep). 13 Donor institutions in the Netherlands and the United States are developing programs with very similar designs (see Footnote 7). Implementation through a commercial bank worked well. The OBA approach created incentives for the borrower to take on more oversight than would otherwise have been the case, which in turn led to cost savings and higher construction quality (K-Rep Bank 2013, pgs. 12-13). At the same time, there are limits to a bank s role, notably the short duration of bank supervision (24 months in the case of this project), and monitoring environmental and water abstraction compliance. 3.2. Achievement of Project Development Objectives The project reached 190 percent of the PDO s target indicator. As explained in Appendices A and J, this achievement probably underestimates the actual achievement, because new connections (and therefore additional beneficiaries) were likely added after connections had been tabulated for the purpose of paying the subsidies. The project similarly achieved 135 percent of the target value for the number of new connections. Again, this figure underestimates the number of new connections because it does not include institutional connections. 13 The subsidies cover part of the investment costs, thereby reducing the loan amount and limiting the tariff impact of the investment. 14

No target values were established for two additional indicators: the numbers of new household connections and kiosks at agreed working conditions and service standards. The numbers of these new connections represented a successful outcome, given that the overall target for new connections was exceeded. On the other hand, the monitoring system did not clearly establish that all these connections were at agreed working conditions and service standards. One indicator fell short of the target value: the number of subprojects financed totaled 87.5 percent of target. In addition to the PDO and outcome indicators that were selected ex post facto (see Appendix J), the project documents refer to a number of other efficiency indicators: low rates of nonrevenue water; monthly revenue targets for all schemes; revenue collection efficiency; population coverage in the supply areas; etc. Data are not available on these indicators due to an M&E system that largely ended with the OVR, and the short time between construction completion and project closing for most subprojects. 3.3 EfficiencyThe present financial projection indicates the total portfolio s financial internal rate of return is 18 percent (with the subsidy payments taken into account). This is just at the hurdle rate equivalent to K-Rep Bank s lending rate. The portfolio, however, yields a negative net present value of minus US$50,000. The financial performance of the subprojects is summarized in table 3.1, below: Table 3.1 Sub-Project Summary Financial Performance IRR with subsidy IRR without subsidy 15 Net Present Value (US$) Operating Cost Coverage Loan Life Debt Service Coverage Mean 24% 10% -1,400 2.17 1.97 Minimum -9% -19% -166,000 1.19 0.26 Maximum 159 76% 130,000 8.53 4.72 No. meeting 18% hurdle rate 15 13 No. with NPV or OCC > 1 14 35 No. with DSCR > 1.5 18 % of 35 meeting relevant criteria Source: Appendix C, table C.4 43% 37% 40% 100% 51% Despite the portfolio s weak financial performance, the project s outcome confirms that investments in water produce significant positive economic returns. The economic analysis was conducted based on information collected in focus group discussions conducted in six subproject areas. The economic internal rate of return of the six subprojects ranged from 64 percent in Hindi to as high as 392 percent in Nyamasaria. This yielded benefits of between US$2 and US$10 for every U.S. dollar invested in water. The economic rate of return results are summarized in table 3.2 below for the six subprojects: Table 3.2 Summary Results of Economic Analysis Hindi Koibatek Kiptere Nyamasaria Olepolos Gitangu