Document of The World Bank Report No: ICR00001937 IMPLEMENTATION COMPLETION AND RESULTS REPORT (IDA-37950, IDA-37951, TF093909) ON A CREDIT IN THE AMOUNT OF SDR 83.9 MILLION (US$120 MILLION EQUIVALENT) TO THE REPUBLIC OF KENYA FOR AN ARID LANDS RESOURCE MANAGEMENT PROJECT – PHASE TWO June 30, 2012 Agriculture and Rural Development Sustainable Development Department Eastern Africa Country Department 2 Africa Region CURRENCY EQUIVALENTS (Exchange Rate Effective December 31, 2010) Currency Unit = Kenya Shilling (KES) KES 80.3 = US$1.00 US$1.00 = KES 1.0 FISCAL YEAR [January 1 – December 31] ABBREVIATIONS AND ACRONYMS AF Additional Financing AIE Authority to Incur Expenditure ALRMP Arid Lands Management Project ALRMP II Arid Lands Management Project Phase Two ARD Agriculture and Rural Development ASALs Arid and Semi-Arid Lands CAS Country Assistance Strategy CBOs Community-Based Organizations CDC Community Development Committee CDD Community Driven Development DANIDA Danish International Development Agency DCU District Coordinating Unit DfID Department for International Development DHS Demographic Health Survey DMI Drought Management Initiative DO Development Objective DPC District Peace Committee DSG District Steering Group EA Environmental Assessment ECHO European Community Humanitarian Office EDRP Emergency Drought Recovery Project EMF Environmental Management Framework EMP Environmental Management Plan ERR Economic Rate of Return ERS Economic Recovery Strategy ESW Economic and Sector Work EU European Union EWS (Drought) Early Warning System FAO Food and Agriculture Organization FEWSNET Famine Early Warning Systems Network FGM Female Genital Mutilation FM Financial Management ii GAAP Governance and Anti-Corruption Action Plan GEF Global Environment Facility GPS Global Positioning System GoK Government of Kenya HIV/AIDS Human Immunodeficiency Syndrome/Acquired Immunodeficiency Syndrome IAD Internal Audit Department IBRD International Bank for Reconstruction and Development ICB International Competitive Bidding ICR Implementation Completion and Result Report IDA International Development Association IE Impact Evaluation IEG Independent Evaluation Group IFMIS Integrated Financial Management Information System IFR Interim Financial Report IGA Income Generating Activity IL-ICR Intensive Learning Implementation Completion Report ILRI International Livestock Research Institute INT Integrity Vice Presidency IOI Intermediate Output/Outcome Indicator IP Implementation Progress IPCC Intergovernmental Panel on Climate Change ISR Implementation Status Report KACCAL Kenya Adaptation to Climate Change in Arid and Semi-Arid Lands (GEF Project) KENAO Kenya National Audit Office KES/KSH Kenya Shilling KESSP Kenya Education Sector Support Project KFSM Kenya Food Security Meeting KFSSG Kenya Food Security Steering Group KPI Key Performance Indicator KRDS Kenya Rural Development Strategy LCB Local Competitive Bidding M&E Monitoring and Evaluation MET Mobile Extension Team MIC Multiple Indicator Cluster MIS Management Information System MOF Minister of Finance MTR Mid-Term Review MUAC Mid-Upper Arm Circumference NCB National Competitive Bidding NDMA National Drought Management Authority NDVI Normalized Difference Vegetation Index NEMA National Environmental Management Authority NGO Non-Governmental Organization NPV Net Present Value iii NRM Natural Resource Management O&M Operations and Maintenance OCHA Office for the Coordination of Humanitarian Affairs OED Operations Evaluation Department OP Operational Policy OXFAM Oxford Committee for Famine Relief PAD Project Appraisal Document PCU Project Coordination Unit PDO Project Development Objective PICD Participatory Integrated Community Development PIP Project Implementation Plan PMP Pest Management Plan PP Project Paper PPAR Project Performance Audit Report PPF Project Preparation Facility PPRs Post Procurement Reviews PRSP Poverty Reduction Strategy Paper SA Social Assessment SCCF Special Climate Change Fund SOE Statement of Expenditure SWAp Sector-Wide Approach UNDP United Nations Development Program UNICEF United Nations Children’s Fund USAID United States Agency for International Development USD/US$ United States Dollar WFP World Food Program WKCDD Western Kenya Community-Driven Development Vice President: Makhtar Diop Country Director: Johannes C.M. Zutt Acting Sector Manager: Martien Van Nieuwkoop Project Team Leader: Johannes Woelcke ICR Team Leader: Luis O. Coirolo iv REPUBLIC OF KENYA Arid Lands Resource Management Project – Phase Two CONTENTS Data Sheet A. Basic Information B. Key Dates C. Ratings Summary D. Sector and Theme Codes E. Bank Staff F. Results Framework Analysis G. Ratings of Project Performance in ISRs H. Restructuring I. Disbursement Graph 1. Project Context, Development and Global Environment Objectives Design ...... 1 2. Key Factors Affecting Implementation and Outcomes ........................................ 6 3. Assessment of Outcomes ................................................................................... 19 4. Assessment of Risk to Development Outcome: Moderate................................ 33 5. Assessment of Bank and Borrower Performance ............................................... 35 6. Lessons Learned ................................................................................................. 41 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partner ..... 45 Annex 1. Project Costs and Financing .................................................................... 47 Annex 2. Outputs by Component............................................................................ 48 Annex 3. Economic and Financial Analysis ........................................................... 61 Annex 4. Bank Lending and Implementation Support/Supervision Processes....... 72 Annex 5. Stakeholder Workshop Report and Results ............................................. 75 Annex 6. Borrower's Comments on Draft IL-ICR .................................................. 79 Annex 7. Comments of Co-financier ...................................................................... 87 Annex 8. List of Supporting Documents ................................................................ 88 MAP .............................................................................................................................. 90 v A. Basic Information Arid Lands Resource Country: Kenya Project Name: Management Project - Phase Two IDA-37950,IDA- Project ID: P078058 L/C/TF Number(s): 37951,TF-93909 ICR Date: 06/30/2012 ICR Type: Intensive Learning ICR MINISTRY OF Lending Instrument: SIL Borrower: FINANCE AND PLANNING Original Total XDR 43.60M Disbursed Amount: XDR 76.89M Commitment: Revised Amount: XDR 77.14M Environmental Category: B Implementing Agencies: Office of the President Ministry of State for the Development of Northern Kenya and Other Arid Lands Cofinanciers and Other External Partners: European Commission (EC) B. Key Dates Revised / Actual Process Date Process Original Date Date(s) Concept Review: 02/18/2003 Effectiveness: 09/08/2003 09/08/2003 Appraisal: 04/23/2003 Restructuring(s): 08/03/2006 Approval: 06/19/2003 Mid-term Review: 11/06/2006 11/06/2006 Closing: 06/30/2009 12/31/2010 C. Ratings Summary C.1 Performance Rating by ICR Outcomes: Moderately Satisfactory Risk to Development Outcome: Moderate Bank Performance: Moderately Unsatisfactory Borrower Performance: Moderately Satisfactory C.2 Detailed Ratings of Bank and Borrower Performance (by ICR) Bank Ratings Borrower Ratings Quality at Entry: Moderately Satisfactory Government: Moderately Satisfactory Moderately Implementing Quality of Supervision: Moderately Satisfactory Unsatisfactory Agency/Agencies: vi Overall Bank Moderately Overall Borrower Moderately Satisfactory Performance: Unsatisfactory Performance: C.3 Quality at Entry and Implementation Performance Indicators Implementation QAG Assessments Indicators Rating Performance (if any) Potential Problem Project Quality at Entry No None at any time (Yes/No): (QEA): Problem Project at any Quality of Yes None time (Yes/No): Supervision (QSA): DO rating before Satisfactory Closing/Inactive status: D. Sector and Theme Codes Original Actual Sector Code (as % of total Bank financing) Animal production 30 30 General agriculture, fishing and forestry sector 40 50 Other social services 30 20 Theme Code (as % of total Bank financing) Natural disaster management 40 55 Other environment and natural resources management 20 15 Other rural development 20 18 Participation and civic engagement 20 12 E. Bank Staff Positions At ICR At Approval Vice President: Makhtar Diop Callisto E. Madavo Country Director: Johannes C.M. Zutt Makhtar Diop Sector Manager: Martien Van Nieuwkoop Karen Mcconnell Brooks Project Team Leader: Johannes Woelcke Christine E. Cornelius ICR Team Leader: Luis O. Coirolo ICR Primary Author: Luis O. Coirolo vii F. Results Framework Analysis Project Development Objectives (from Project Appraisal Document) The objective of the second phase of ALRMP is to enhance food security and reduce livelihood vulnerability in drought-prone and marginalized communities in 22 ASAL districts. Revised Project Development Objectives (as approved by original approving authority) Reduced livelihood vulnerability, enhanced food security, and improved access to basic services in 28 drought prone arid and semi-arid districts in Kenya (restructured through additional financing amendment). (a) PDO Indicator(s) Original Target Formally Actual Value Values (from Revised Achieved at Indicator Baseline Value approval Target Completion or documents) Values Target Years Decreasing proportion of people assessed as needing free food aid in each arid Indicator 1 : and semi-arid district affected by severity of drought Food aid needs during Food aid needs Value Reduced food aid Reduced food 2004/2005 drought. See reduced in 28 quantitative or needs in 22 ASAL aid needs in 28 section 3for discussion districts in which Qualitative) districts ASAL districts and quantitative data. ALRMP II operated Date achieved 07/01/2005 06/30/2009 06/30/2010 12/31/2010 Achieved. Cumulative project expenditures negatively correlated with the share Comments of people needing food aid. Proportion of people needing food aid was 19.7% (incl. % lower in the project areas when compared to control areas between 2005 and achievement) 2009. See Section 3 for details. Indicator 2 : Reducing the time lapse between reported drought stress and response 3.5 weeks across Value 7.6 weeks in 11 original Reduced time Not a PAD ASAL districts in quantitative or ALRMP arid districts in lapse in all indicator 2008/09 drought Qualitative) 2000/01 drought year ASAL districts year. Date achieved 06/30/2001 06/30/2009 06/30/2010 12/31/2010 Achieved. ALRMP Bulletin became main source of drought information for Comments ASALs. Bulletin use is significantly and negatively associated with response (incl. % time: mean response time of Bulletin users decreased from 7.6 weeks in 2000/01 achievement) to 3.5 weeks in 2008/09. Improved nutritional status of children below 5 years of age affected by severity Indicator 3 : of drought over time Improved Positive impacts on Improved child Value Nutritional status of child child nutrition in nutritional status quantitative or children in rural ASALs nutritional project treatment in 22 ASAL Qualitative) extremely low status in 28 areas, especially for districts ASAL districts worst off children Date achieved 06/30/2005 06/30/2009 06/30/2010 12/31/2010 Comments Achieved. For the 10th percentile of Middle-Upper Arm Circumference (MUAC) viii (incl. % scores (poorer nutritional status), 15% of intervention locations showed achievement) deteriorating nutritional status, while 25% of control locations experienced deterioration in nutritional status. Increased number of people with access to basic services (water, human and Indicator 4 : animal health services and education) Access to many Increased basic services Value Access to basic services Not a PAD access to basic improved in areas quantitative or generally poor in rural indicator services in 28 in 28 districts Qualitative) areas of ASAL districts ASAL districts which ALRMP II operated Date achieved 06/30/2006 06/30/2009 06/30/2010 12/31/2010 Achieved. ALRMP II treatment communities had statistically significant Comments improvements compared to control groups in: the quality of water sources; (incl. % spending on domestic water; and access to human health services, animal health achievement) services, and education. Increase people's participation in the project districts in local and national Indicator 5 : development, as demonstrated by the reflection of arid land concerns in the Economic Recovery Strategy and other relevant national policies. Increased empowerment Increased people’s Increased evident across participation and peoples’ districts where empowerment had begun Value participation, ALRMP II in original 11 ALRMP Not a PAD quantitative or and greater operated, and arid districts. First phase indicator Qualitative) influence on ASAL concerns project had also begun to national policy reflected in ERS, draw attention to ASAL formulation Vision 2030 and development issues several important draft policies. Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Partially achieved. ALRMP II had a measurable impact on empowerment. It Comments influenced relevant policy formulation through advocacy, knowledge and (incl. % advisory support. However, some policies had not yet been formally adopted. achievement) See Section 3 for details. (b) Intermediate Outcome Indicator(s) Original Target Actual Value Formally Values (from Achieved at Indicator Baseline Value Revised approval Completion or Target Values documents) Target Years District staff and a minimum of 600 communities trained in participatory and Indicator 1 : natural resource management Train 600 1,990 training Value Some training activities communities activities (including Not a PAD (quantitative undertaken in original 11 and staff in 28 614 communities) indicator or Qualitative) ALRMP arid districts ALRMP II in the 28 districts arid and semi- on participatory ix arid districts NRM Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved. Source: Borrower ICR; ICR Team achievement) Key potential conflict flashpoints identified in each district and conflict Indicator 2 : management initiatives strengthened Conflict flashpoints identified and Conflict Identify District Peace management flashpoints Committees Some conflict Value mechanisms for and implement established in all 28 management activities (quantitative NRM established initiatives in districts; 214 implemented in original or Qualitative) at potential 28 ALRMP II conflict 11 ALRMP arid districts flashpoints in 22 arid and semi- management districts arid districts meetings and 219 conflict resolution events held Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved. Source: Borrower ICR achievement) Kenya Food Security Meeting (KFSM) and 28 District Steering Groups Indicator 3 : institutionalized KFSM and 28 KFSM and 28 Value KFSM and DSGs KFSM and 22 DSGs functioning DSGs (quantitative functioning in original 11 DSGs actively, but not yet institutionalize or Qualitative) ALRMP arid districts institutionalized legally d institutionalized Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments Partially Achieved. Arrangements functioning, but organizational changes (incl. % (shifts in ministerial responsibility for ALRMP II) and constitutional reforms achievement) have delayed full institutionalization. Drought Early Warning System (EWS) monthly bulletins produced and Indicator 4 : disseminated in the project area and at national level on a regular timely basis Bulletins produced Produce and for 28 districts. EWS bulletins produced disseminate Value Distribution within and distributed nationally Not a PAD EWS bulletins (quantitative one month at and in original 11 indicator for expanded or Qualitative) national and district ALRMP arid districts ALRMP II level and on project area ALRMP website Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved. Source: Borrower ICR; ICR Team achievement) District Long Term Drought Preparedness and Annual Work Plans developed in Indicator 5 : each district of the project area Value District Plans produced in Not a PAD Develop plans Plans developed in x (quantitative original 11 ALRMP arid indicator in 28 ALRMP 28 districts with or Qualitative) districts II arid and participation of semi-arid communities and districts other stakeholders Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved. Source: Borrower ICR; ICR Team achievement) District Contingency Plans developed in each district in the project area and Indicator 6 : disseminated to stakeholders Plans developed in Develop plans Contingency plans 28 districts with Value in 28 ALRMP prepared and Not a PAD community (quantitative II arid and disseminated in original indicator participation, and or Qualitative) semi-arid 11 ALRMP arid districts disseminated to districts stakeholders Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved. Source: Borrower ICR; ICR Team achievement) Indicator 7 : Drought Contingency Fund and associated funding mechanism established Drought Establish Contingency Fund Drought Contingency Drought functioning as a Value Fund operating as a Not a PAD Contingency project instrument (quantitative project instrument for indicator Fund as now covering 28 or Qualitative) original 11 ALRMP arid permanent districts, but not yet districts institution legally institutionalized Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments Partially Achieved. Source: Borrower ICR; ICR Team. Legal documents are (incl. % finalized and Presidential Decree is pending. achievement) Mobile Extension Teams and line ministries core staff in specified arid lands Indicator 8 : districts trained to conduct modified Participatory Rural Appraisal (PRA) and provide backstopping to communities in CDD Enhanced CDD training training in 11 Some training in delivered in 11 arid Value arid districts participatory techniques Not a PAD districts to 148,178 (quantitative for greater provided in original 11 indicator people (of which or Qualitative) delegation of ALRMP arid districts 52% women) in CDD authority 2,477 events to local level Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments Achieved. Training covered Participatory Integrated Community Development, (incl. % Finance and Community Procurement, Cross-cutting Issues and Thematic areas. achievement) Source: Borrower ICR; ICR Team. Effective screening by specified DSGs routinely undertaken for community plans Indicator 9 : and project proposals for technical, environmental and financial soundness xi DSG screening capacity has been DSGs screened Better capacity strengthened in 11 Value community proposals in Not a PAD to screen CDD arid districts. (quantitative original 11 ALRMP arid indicator proposals in Evidence on micro- or Qualitative) districts the 11 districts project results suggests generally good screening Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Achieved. Numerous capacity building activities are detailed in Borrower ICR. Comments Technical audit in 2010 of a sample of micro-projects found high levels of (incl. % community ownership (93%) and satisfactory construction (87.5% with achievement) relatively few design issues). 600 Community Action Plans prepared and micro-projects implemented by Indicator 10 : communities 2,456 micro- Value projects (quantitative 0 600 implemented in 614 or Qualitative) communities Date achieved 06/30/2003 06/30/2009 12/31/2010 Achieved and target exceeded. Although Project Document had a figure of 800 Comments community action plans, this appears to be a typo since all Borrower and IDA (incl. % reports consistently refer to assisting 600 communities through the CDD achievement) component. Source: Borrower ICR; ISRs. At least 1,000 groups assisted in developing savings capacity through top-up Indicator 11 : saving grants 1,588 transactions completed, Value benefitting 57,095 Not a PAD (quantitative 0 1000 males and 83,790 indicator or Qualitative) females through income generating activities (IGAs). Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments Achieved. Community groups saved and contributed approximately 35% of total (incl. % costs for these IGAs. Source: Borrower's comments on draft IL-ICR; project achievement) files. Indicator 12 : National Pastoral Policy formulated through participatory process Value Policy formulated, Not a PAD Pastoral Policy (quantitative No pastoral policy but not yet Indicator formulated or Qualitative) officially adopted Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved. Source: PCU documents achievement) Indicator 13 : Mobile schools and nomadic education centers established in 3 project districts Value Mobile school concept Establish 30 mobile schools Not a PAD (quantitative was piloted under first mobile schools and nomadic Indicator or Qualitative) ALRMP project in 3 project education centers xii districts established in 6 project districts Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved and target exceeded. Source: Borrower ICR achievement) At least 600 communities reached with awareness raising program on human Indicator 14 : health and HIV/Aids Value Not a PAD 600 614 (quantitative 0 indicator communities communities or Qualitative) Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved and target exceeded. Source: Borrower ICR achievement) Guidelines for emergency livestock off-take prepared and articulated in all Indicator 15 : Project districts ALRMP supported Prepare and Value emergency livestock off- disseminate Guidelines prepared Not a PAD (quantitative take in 11 arid districts guidelines in l and disseminated in indicator or Qualitative) but did not have clear 28 ALRMP II 28 districts guidelines districts Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Comments (incl. % Achieved. Source: PCU documents achievement) Options for improved financial services delivery considered and implemented Indicator 16 : where possible within the project area Value Access to financial Study options Only some training Not a PAD (quantitative services very limited in and implement and study tours indicator or Qualitative) ASAL districts where possible implemented Date achieved 06/30/2003 06/30/2009 06/30/2010 12/31/2010 Not achieved. Only some training and study tours were implemented. The Comments wording of the indicator suggests that even at design this objective was (incl. % considered tentative, and during implementation it did not receive priority achievement) attention. Source: ICR mission. xiii G. Ratings of Project Performance in ISRs Actual Date ISR No. DO IP Disbursements Archived (USD millions) 1 11/26/2003 Satisfactory Satisfactory 5.07 2 06/06/2004 Satisfactory Satisfactory 5.07 3 06/18/2004 Satisfactory Satisfactory 5.07 4 12/14/2004 Satisfactory Satisfactory 10.27 5 06/15/2005 Satisfactory Satisfactory 18.53 6 12/21/2005 Satisfactory Satisfactory 22.84 7 05/08/2006 Satisfactory Satisfactory 28.26 8 12/21/2006 Satisfactory Satisfactory 36.11 9 06/23/2007 Satisfactory Satisfactory 43.69 10 12/21/2007 Satisfactory Satisfactory 71.90 11 05/31/2008 Satisfactory Satisfactory 76.31 12 12/22/2008 Satisfactory Satisfactory 88.20 13 06/26/2009 Satisfactory Satisfactory 97.28 14 12/15/2009 Satisfactory Satisfactory 112.28 15 06/29/2010 Satisfactory Satisfactory 118.39 Moderately 16 04/12/2011 Unsatisfactory 118.39 Unsatisfactory xiv H. Restructuring (if any) Restructuring Board ISR Ratings at Restructuring Amount Reason for Date(s) Approved Development Implementation Disbursed at Restructuring and PDO Objective Progress Restructuring Key Changes Made Change (US$ m) 08/03/2006 Yes Satisfactory Satisfactory 36.11 Additional Financing (US$60 m); minor change to PDO; improved Key Performance Indicators (KPIs) Summary Ratings of Original and Revised PDO Targets Outcome Ratings Against Original PDO/Targets: Moderately Satisfactory* Against Formally Revised PDO/Targets: Moderately Satisfactory Overall (Weighted) Rating: Moderately Satisfactory** * The PDO changed only slightly in the 2006 Additional Financing amendment, to add reference to improved access to services and the expansion from 22 to 28 project districts. However, the core of the PDO -- reducing livelihoods vulnerability to drought -- remained the same. IDA rated DO and IP as Satisfactory at the time that the AF was approved by the Board. The KPIs to achieve the PDO were reformulated in the AF amendment. As the task team and counterparts had been working on refining the results framework from early on in project implementation, little emphasis was placed on accumulating data for some of the original indicators. The final IE and the Borrower’s ICR used the revised KPIs to assess outcomes. Given that 3 of the 6 original PDO indicators were similar to revised KPI 1 on food aid, and a fourth was virtually the same as new KPI 3 on the nutritional status of children under 5 years of age, the IL-ICR agrees with IDA ratings for these four original indicators, and assesses achievements to be in the Achieved/Satisfactory range. The other two original indicators called for increases in the volume and value of various livestock and other productive assets. ALRMP II made successful investments in both, therefore the outcomes would also probably have been assessed to be either Partially Achieved or Achieved. However, since they were dropped, and there were no similar indicators after the AF, the IL-ICR team did not have sufficient supporting information on them and has no choice but to assess achievement of these two original indicators to be Moderately Unsatisfactory. With four original indicators rated Satisfactory and two Moderately Unsatisfactory, the IL-ICR team assesses overall achievements against original project targets to be Moderately Satisfactory. ** Approximately 30.5% of total project resources were disbursed before the minor reformulation of the PDO and revision of KPIs. See first part of Section H above. The overall rating of Moderately Satisfactory at completion is based on a weighted average of the outcome ratings for the original and revised KPIs. xv I. Disbursement Profile P078058 xvi 1. Project Context, Development and Global Environment Objectives Design 1.1 Context at Appraisal 1. The Arid Lands Resource Management Project Phase Two (ALRMP II) was the latest stage in a nearly two-decades long partnership between International Development Association (IDA) and the Government of Kenya (GoK) to strengthen drought management and reduce vulnerability of pastoralists and small farmers in Kenya’s very poor and fragile arid and semi-arid lands (ASALs). The partnership evolved from: (i) emergency support after a severe drought in the early 1990s (Emergency Drought Recovery Project, EDRP1); to (ii) the development under the first phase of ALRMP2 of a community-based drought early warning system (EWS) for the arid districts; contingency planning at the district level; institutional and financial arrangements to reduce response times and improve drought management; small investments to reduce livelihoods vulnerability; and finally (iii) to ALRMP II, which aimed to continue working in the arid districts while extending parts of the program to the semi-arid districts, enhance and substantially scale up the EWS and other institutional, financial and inter-agency coordination arrangements, expand investments to improve livelihoods, and intensify advocacy for a much stronger policy focus on the ASALs. 2. Although not characterized as climate change adaptation projects, both the first and second phases of ALRMP aimed to strengthen the coping, resilience and adaptive responses of the rural poor in the ASALs to worsening drought conditions that are now recognized to be consequences of long-term climate change and climate variability in the Horn of Africa. 3. The Implementation Completion and Result Report (ICR) by the Africa Region and subsequent Project Performance Audit Report (PPAR) by IDA’s Operations Evaluation Department (OED, now the Independent Evaluation Group--IEG) assessed that the first phase of ALRMP had high strategic relevance, that outcomes were satisfactory with some drought management activities highly satisfactory, and that institutional impact was also high. 4. At the time of ALRMP II preparation, the IDA Country Assistance Strategy FY04-07, Kenya’s draft 2001-04 PRSP (Economic Recovery Strategy for Wealth and Employment Creation--ERS), and its Rural Development Strategy (KRDS, 2002) emphasized poverty reduction, greater popular participation, and increasing decentralization of decision making over resource allocation and service delivery. This 1 Cr. 2460-KE (1993); US$20 million equivalent. The project was also financed by amendments to three existing credits: Cr. 1758-KE (Animal Health Services – US$2.5 million), Cr. 1974-KE (Rural Services Design – US$2.75 million) and Cr. 2199-KE (Second National Agricultural Extension – US$2 million equivalent). 2 Cr. 2797-KE (1995); US$22 million equivalent. 1 policy orientation reflected dramatic changes in the Kenyan political environment, as the presidential elections of December 2002 brought about a peaceful transition from a government that had been in place for 24 years to a new, participatory, and forward- looking administration. 5. The change in government and trend towards decentralization offered opportunities for ALRMP II to scale up the agenda that had been initiated under the first project, i.e., building capacity and empowering pastoralist and small farmer communities in the ASALs to participate more actively in identifying and advocating for their own development priorities – centered very much around reducing risks and vulnerability to droughts. IDA also considered that the time was propitious for a stronger emphasis on governance issues in Kenya, and saw community-driven development (CDD) as one instrument for improving transparency, citizen capacity to engage upward, and downward accountability by national and local governments. The first phase of ALRMP had been in the forefront in piloting CDD in Kenya, and ALRMP II continued this effort. 6. ALRMP II, the subject of this Intensive Learning Implementation Completion Report (IL-ICR), was financed by IDA Credit 37950 of US$60 million equivalent, approved on June 19, 2003, and by Additional Financing (AF) under IDA Credit 37951 of US$60 million approved on August 8, 2006, for a total of US$120 million equivalent. The total value increased over time to US$124.1 million, as a result of exchange rate changes.3 1.2 Original Project Development Objectives (PDO) and Key Indicators (as approved) 7. The original PDO of ALRMP II was to enhance food security and reduce livelihood vulnerability in drought-prone and marginalized communities in 22 ASAL districts. This would be achieved by: (i) strengthening and institutionalizing natural resources and drought management; (ii) empowering communities so that they could successfully identify, implement and sustain their development priorities through CDD; and (iii) fostering a conducive enabling environment for ASAL development through policy support, advocacy and improvement of essential services, complementing existing sector programs. 8. PDO results were to be measured by the following original key performance indicators (PAD A.2 and Annex 1): (i) Decreasing annual trend in the number and percentage of people of targeted communities in each district seeking/receiving free food aid; 3 In November 2006, the GEF Council allocated US$6.5 million from the Special Climate Change Fund (SCCF) for a Kenya Adaptation to Climate Change in Arid and Semi-arid Lands (KACCAL) project, of which IDA would develop US$5.5 million and the UNDP US$1 million. IDA approved the trust fund in June 2010, but it is not yet effective. This IL-ICR does not cover KACCAL, although it does refer to some preparatory studies that provided useful information for evaluation of ALRMP II outcomes. 2 (ii) Increasing annual trend in the percentage of communities with food consumption above national food poverty line at historically driest month (pastoralists) and before harvest (farmers); (iii) Decreased variability in food consumption at household level within and between years; (iv) Increased value and diversity of livestock, productive and household assets, with focus on female-headed households and marginal groups; (v) Increased diversity and amount of household income from sources that are drought-tolerant and environmentally non-degrading for targeted communities; (vi) Improved child nutrition and health indicators for children below 5 (weight for height/weight for age as an anthropometric proxy). 1.3 Revised PDO (as approved by original approving authority) and Key Indicators, and reasons/justification 9. The PDO was modified very slightly, and the key indicators were revised to (i) reflect changes associated with the AF; (ii) consolidate and improve the measurability of several of indicators; and (iii) better align the indicators with the evolving focus of ALRMP II and expected causality between project interventions and outcomes. 10. Revised PDO: To reduce livelihood vulnerability, enhance food security, and improve access to basic services in 28 drought prone arid and semi-arid districts in Kenya. 11. The minor changes to the PDO were the reference to access to basic services, which had always been part of the project but were not explicitly captured in the original PDO formulation; and the increased number of target districts from 22 to 28 made possible by the AF in 2006. Revised PDO Indicators: (i) Decreasing proportion of people assessed as needing free food aid in each arid and semi-arid district affected by severity of drought; (ii) Reducing the time lapse between reported drought stress and response; (iii) Improved nutritional status of children below 5 years of age affected by severity of drought over time; (iv) Increased number of people with access to basic services (water, human, and animal health services and education); (v) Increased people’s participation in the project districts in local and national development as demonstrated by the reflection of arid lands concerns in the economic recovery strategy and other relevant national policies. 12. The revisions to the PDO indicators partly reflected changes associated with the AF (the increased emphasis on drought management and recovery). However, they were mainly the result of work over the first few project years to strengthen the Monitoring & Evaluation (M&E) framework. Revised Key Performance Indicator (KPI) (i) on the share 3 of people needing food aid consolidated and replaced original PDO indicators (i), (ii), and (iii), which had also dealt with food needs, and it made explicit the notion of normalizing for drought severity to ensure proper inter-annual comparisons. Revised KPI (iii) on child nutrition replaced original indicator (vi), making more explicit the focus on drought impacts (amongst the many other factors that can influence child nutrition). Revised KPIs (ii), (iv), and (v) were new; they provided monitorable measures of outcomes that had always been considered important causal factors for achieving the PDO, but for which there had not been original indicators. Original indicators (iv) and (v) on assets and incomes were dropped, as the project emphasis was shifting towards drought management and recovery. The IL-ICR team’s views on these changes are included in Section 2.3. 1.4 Main Beneficiaries 13. The main intended beneficiaries of ALRMP II were (i) those households in the ASALs whose vulnerability to droughts was reduced and access to services improved as a result of project interventions, as well as (ii) those for whom the consequences of droughts were lessened because of the impact of the project EWS, contingency fund and coordination mechanisms on the timing of drought response efforts by GoK and other actors. ALRMP II planned to focus on the most vulnerable people within the ASAL districts, both from the standpoint of poverty4 and also with reference to gender (women), age (children) and marginalized groups. 1.5 Original Components (as approved) 14. ALRMP II originally had three components, which are described below. All three were to be implemented in the 11 arid districts; only Component 1 would be implemented in 11 semi-arid districts. The semi-arid “extension� districts as they are referred to in project documents were selected based on rainfall and vegetation types of agro-ecological zones 5-7, and levels of food insecurity measured by qualification for World Food Program (WFP) emergency food in the 2001 drought assessment. 4 The 28 ASAL districts include 80% of Kenya’s total land area (58 million ha) and 25% of its population. These are the poorest areas in Kenya – 75% of the inhabitants of the North-Eastern Province and over half those in the Eastern Province live below the national poverty line (Kenya Poverty and Inequality Assessment, 2009). The seven districts with the highest poverty rates are all in the ASALs. A large proportion of the North-Eastern population is engaged in pastoral or agro-pastoral livelihoods, just 4% has access to electricity, 88% of adults have not completed primary education, and 50% of children are not enrolled in primary school. Due inter alia to lack of public infrastructure, only 8.7% of the population has access to formal finance, significantly lower than the national average (18.5%). In 2003, the mean household time to a water source was 85 minutes in the rural North-Eastern province -- more than 3 times the national average (DHS 2003). The area is prone to increasingly unpredictable and prolonged periods of severe drought. 4 15. Component 1: Natural Resource Management and Drought Management (US$99.7 million total cost, including IDA and GoK) had two sub-components: (i) Natural Resource Management; and (ii) Drought Preparedness and Management. The first sub-component aimed to improve knowledge, planning and implementation of activities to make the most beneficial and sustainable long-term use of the natural assets of the ASALs. The second sub-component supported enhanced work on drought preparedness and response, to create a more effective drought cycle management system at the national and the district level. 16. Component 2: Community Driven Development (US$28.6 million total cost, including IDA, GoK and communities) aimed to empower communities to take greater charge of their own development agenda, through three sub-components: (i) Support to CDD implementation; (ii) Community Capacity Building; and (iii) Capacity Building for Backstopping Services. The first sub-component financed community-prioritized micro- projects, to which communities would contribute 30 percent. The second sub-component supported the training of communities and their management committees to build capacity to identify priority investments, keeping in mind the special needs of marginalized groups, and to implement the projects and account for the funding received. The third sub-component financed communities’ support and backstopping from the district level project staff and Mobile Extension Teams (METs). 17. Component 3: Support to Local Development (US$16.6 million, including IDA and GoK) had three sub-components: (i) Policy, Advocacy and Research; (ii) Local Services Development; and (iii) Piloting Financial Services. The first sub-component sought to ensure that the needs of the arid districts were adequately reflected in national and sector policy initiatives. The second sub-component supported the strengthening of local services in sectors that are essential to the sustainable development of arid lands communities by complementing existing government and partner programs. The third sub-component supported the piloting of low-cost delivery of financial services, to which the ASAL population had very limited access. 1.6 Revised Components 18. In 2006, to help GoK scale up selected project activities to deal with severe drought consequences of 2.5 years of poor rains, IDA doubled its support for ALRMP II by providing AF of US$60 million equivalent. The AF maintained the same components, but changed the relative emphasis among them. Resources for Natural Resource Management (NRM) and Drought Management (Component 1) were substantially increased, although there were also small increases for Community Development (Component 2) and Local Development (Component 3). The possibility of such a shift in emphasis had been foreseen in the Project Appraisal Document (PAD), which stated that if severe drought persisted, relatively more of ALRMP II project effort would necessarily shift into drought mitigation, response and recovery (PAD Critical Risks template, p. 43). 5 1.7 Other significant changes 19. Expansion of Geographic Coverage. When the AF was approved, six more semi- arid districts were added that had initially been excluded from ALRMP II due to resource constraints. 20. Shift of GoK Ministerial Responsibility for ALRMP II. In 2003, the PCU moved from the Ministry of Provincial Administration to the Ministry of State for Special Programs, both in the Office of the President. In 2008, it moved again to a newly created Ministry of State for Development of Northern Kenya & Other Arid Lands, coordinated by the Prime Minister. 21. Extension of IDA Credit Closing Date. In 2006, the Credit closing date was extended from June 30, 2009 to June 30, 2010 to accommodate activities under the AF. On June 30, 2010, it was extended to December 31, 2010, i.e., a total of 18 months beyond the original closing date, to facilitate completion activities and transition to an expected follow-on operation. 22. Informal Suspension of IDA Credit. Based on an understanding of preliminary findings of an Institutional Integrity (INT) forensic audit, in July 2010 management informally suspended IDA financing for ALRMP II, and the suspension remained in place through closing on December 31, 2010 (Sections 2.1, 2.2, 2.4. 5.1(b)). 2. Key Factors Affecting Implementation and Outcomes 2.1 Project Preparation, Design and Quality at Entry 23. The task team drew on a wide range of inputs that collectively provided a sound basis for the design of ALRMP II: (i) lessons learned during implementation of the first phase of ALRMP and the predecessor Emergency Drought Project; (ii) a Beneficiary Assessment Study of the first phase project; (iii) the ICR for the first phase, which was completed in time to inform the appraisal of ALRMP II; (iv) a Social Assessment of conditions affecting pastoralists and small farmers in the ASALs; and (v) work of the pastoral thematic team that had participated in drafting Kenya’s 2002 PRSP, including extensive consultations in the ASALs.5 As CDD activities had been piloted under the first phase and were expected to be scaled up in the arid districts under ALRMP II, the team also reviewed CDD experiences in the Africa Region.6 5 (i) Appropriate Development Consultants Limited (2002). Beneficiary Assessment Study of the Arid Lands Resource Management Project. Nairobi, Kenya; (ii) Support for Tropical Initiatives in Poverty Alleviation (2003). Social Assessment Report: Arid Lands Resource Management Project. Nairobi, Kenya; (iii) Kenya (2001). Poverty Reduction Strategy Paper for the Period 2001-04. 6 World Bank (2002). Kenya Community Driven Development: Challenges and Opportunities. Washington, D.C. 6 24. Lessons cited from the first phase project ICR included: (i) the importance of deepening community participation in preparation of the drought risk management plans; (ii) the need to strengthen ALRMP’s M&E systems; (iii) the need for a clear framework for management, protection and sustainable development of the arid lands natural resource base; and (iv) the importance of requiring community contributions to CDD micro-projects to avoid dependency and build ownership. The Beneficiary Assessment came to some of the same conclusions and also emphasized the importance of an enabling policy environment and institutions for the long-term sustainability of ALRMP impacts. The CDD review focused on: (i) models of community participation and indigenous management of water systems; (ii) the need to link CDD actions with local government structures, and empower and strengthen local authorities; and (iii) potential CDD contributions to improving local governance and accountability. 25. The design team had a good range of skills and drew on appropriate peer guidance (Annex 4), with a few exceptions. Although the CDD component of ALRMP II reflects lessons from experience in the Africa Region, the team could also have drawn on best practice knowledge available at the time from other regions (e.g., Indonesia, India, Bangladesh, and Brazil). Inter alia, these would have included: (i) the importance of having much clearer ‘rules of the game’ defined up front; (ii) greater clarity as to how communities, beneficiary groups and micro-projects would be selected; (iii) how decision-making would take place at all stages; (iv) the role of communities and other actors in implementation, including how financial management (FM) and procurement activities would be carried out; and (v) how O&M responsibilities would be defined and discharged. The CDD manual was prepared during implementation, rather than up-front. Indicators to measure CDD outcomes/outputs were weak, and some of these were dropped at the time of the AF, which complicated the task of assessing empowerment achievements at project completion. The M&E framework also had several shortcomings (discussed below). The financial services activity responded to a perceived need to strengthen access to finance by the ASAL population, but was not well grounded in analysis of the Kenyan micro-finance sector. Objectives were vague and implementation arrangements not commensurate with the challenge that the activity proposed to address. 26. There were several areas of excellence in design, including: (i) innovative work linking natural resource management and conflict resolution; (ii) multi-sectoral and inter- agency coordination on drought management at national, district and local levels; and (iii) the community-based drought Early Warning System (EWS) in which Kenya has become a recognized regional leader. 27. Project content and organizational structure were straightforward, and tightly linked to the results chain to achieve the PDO. For the most part, the team avoided the risk of including activities, however worthwhile, that were not critical to the central objective of enhancing food security and reducing livelihood vulnerability in drought prone and marginalized ASAL communities. One exception was the financial services activity. The inferred objectives may well be important for long-term adaptation to 7 drought conditions in the ASALs, but the link to the PDO was not clear, sub-component activities were vague and little was accomplished. 28. Several critical risks and mitigation measures were well defined. Some significant risks, however, were underestimated, and the task team did not foresee fiduciary and accountability risks that were noted in the course of implementation. (i) The team’s assessment as ‘negligible’ of the risk that GoK would not institutionalize the drought management system into government structures proved optimistic, as this has been partially achieved to date. The system is functioning well on the ground, and GoK established a new National Drought Management Authority (NDMA) in November 2011 after ALRMP II completion. However, other important institutional arrangements have not yet been formalized, although the failure to do so thus far has been due more to delays caused by governmental organizational changes and the constitutional reform process than to lack of GoK commitment. (ii) Fiduciary risks were detailed in the PAD Annexes. Since some procurement and financial management sub-risks were assessed to be Substantial or High, they should also have been reflected in the main PAD Critical Risks template. Risks related to accountability of expenditures at local levels and capacity for record-keeping were underestimated. (iii) The design team did not foresee the risk of suspension of financing, linked to the fiduciary risks noted above. The informal suspension undermined ALRMP II’s impact in the final project period and elevated risks to sustainability of project achievements. Subsequent to project closing, however, and with better understanding of the limitations of the findings of INT, both IDA and other partners have renewed commitment to future support for the agenda supported under ALRMP II. The risk to sustainability of the core interventions is thus rated Moderate. 2.2 Implementation 29. Factors that affected implementation: ALRMP II implementation was affected by several positive and adverse factors. 30. Positive factors: A very seasoned and highly motivated Project Coordination Unit (PCU), with good continuity in staffing and access to high-level decision-makers. The PCU had gained considerable experience under the predecessor ALRMP (and EDRP) operations. Although reporting responsibilities were changed twice during the project period (see Section 1.7), key staff continued in place and the project enjoyed consistently high priority, good access to senior officials, and timely and adequate counterpart financing. 31. Support by a wide range of stakeholders. Information generated by the project (through the EWS) and its institutional structure (particularly the District Steering 8 Groups--DSGs) came to be used increasingly by government, all major donors and other stakeholders, and this in turn strengthened the project’s influence and effectiveness. 32. Increasing attention to global climate change adaptation. When the 2006 Stern Report and 2007 findings of the Fourth Intergovernmental Panel on Climate Change (IPCC) were published, ALRMP II – which was focused precisely on strengthening coping skills and supporting adaptive responses to accelerating climate change – was under implementation and producing results. As many countries and donors, including the World Bank Group, began considering how to design such programs, ALRMP II was already generating valuable lessons that were helpful to GoK in formulating its own national position on climate change adaptation. This also enabled ALRMP II to attract additional support, such as the Global Environment Facility—Kenya Adaptation to Climate Change in Arid and Semi-Arid Lands Project (GEF-KACCAL) European Union Drought Management Initiative Trust Fund (EU-DMI) and the proposed follow-up Sector-wide Approach (SWAp) operation. 33. Growing recognition of ALRMP II’s effectiveness as a social safety net for the rural poor in the ASALs. When the global food security crisis emerged in 2008, although the community-based EWS and organizational structure at district level in the ASALs had been developed for other purposes, they served as another reminder of the effectiveness of ALRMP II as a safety net able to reach the grass roots level quickly in response to crises, and this added to the project’s legitimacy and influence. 34. Economic growth. Much of ALRMP II implementation took place in a context of strong economic growth (3.7 percent p.a. 2002-07) that contributed to a stable fiscal situation and helped to ensure steady counterpart funding. A combination of shocks (on the domestic front: drought and electoral crisis; on the global front: food security and financial crises) made the situation in 2008-10 more difficult, but did not affect domestic financing for ALRMP II. 35. Adverse developments: Political instability. Highly contested elections in 2007 led to outbreaks of violence that were unprecedented in Kenya in recent years, and these affected implementation of some project activities. They also resulted in a number of organizational changes within government, which delayed consideration of several draft policies that ALRMP II had helped to formulate, and the incorporation of some aspects of ALRMP’s institutional arrangements into government structures (DSGs, Peace Committees, and Drought Contingency Fund). Establishment of a long proposed National Drought Management Authority took place in November 2011, after project completion. The ongoing constitutional reform process will result in a new county structure at the sub-national level, in place of districts, to which the institutional arrangements used under ALRMP II will need to adjust. On the other hand, the new Constitution provides for greater decentralization of government roles and responsibilities, consistent with ALRMP II themes of greater local participation and voice. 36. Increasing frequency, severity and duration of drought. The implementation period of ALRMP II has coincided with several years of severe, prolonged droughts, 9 together with more difficult ‘regular’ drought periods. ALRMP II responded well and produced some very strong results in terms of assisting the ASAL population in coping. However, these crises did cause the project to focus heavily on drought management and recovery, and reduced somewhat the scope for concentrating on community development and longer-term adaptation investments, as many households struggled during certain periods with the challenges of day-to-day survival. 37. Pace of implementation and disbursements proceeded largely as planned. The Credit was signed and made effective as scheduled. Disbursement lags through 2006 when the AF was approved were marginal and usually related to timing of withdrawal applications rather than real delays in activities. From approval of the AF through FY09, activities stayed on track although absorbing the full incremental financing was challenging and GoK therefore requested a six-month extension of the closing date to December 31, 2010. IDA management approved the extension on June 30, 2010, but in July it informally suspended disbursements. The final six months of implementation were therefore disrupted, and project activities on the ground slowed down. As INT did not provide its preliminary report to the government until April 2011, it was not possible to process withdrawal applications and IDA therefore extended the normal four-month period for final reconciliation of accounts to six months, to June 30, 2011. 38. Implementation Status Report (ISR) ratings. Project ISRs from Effectiveness through the Closing Date rated overall progress toward achievement of the PDO and implementation progress (DO and IP, respectively) as Satisfactory. The only sub-rating that slipped into the Unsatisfactory range on one occasion during this period concerned financial management, for the reasons discussed in the fiduciary section below. For the most part the IL-ICR team found ratings well supported by information in the ISRs, aide- memoires and/or project files. However, the ICR team would have lowered the sub- assessment of M&E in the early years of the project. An additional ISR, entered into the Operations Portal four months after Closing, downgraded the DO to Moderately Unsatisfactory and the IP to Unsatisfactory, mainly to reflect preliminary INT findings (Section 2.4). The post-Closing Date ISR is listed in the Data Sheet. Pursuant to IEG guidelines, the IL-ICR uses the final ISR before Closing for purposes of identifying disconnects between completion and supervision ratings, and on this basis finds that there is no disconnect. Furthermore, the INT findings have been revised several times since the filing of the post-closing ISR, a development that was not foreseen at the time of the final ISR. 2.3 Monitoring and Evaluation (M&E) Design, Implementation and Utilization Design 39. The ICR on the first phase of ALRMP concluded that: “The main faults in the project design were in not providing a sufficiently clear and detailed design for monitoring and evaluation…� GoK and the task team attempted to address this in the design of ALRMP II. In the Project Performance Audit Report (PPAR), IEG advised that 14 of the 19 key indicators in the staff appraisal report for the first phase project related to implementation rather than impact. While the other 5 indicators did seek to measure 10 impact, none focused on qualitative aspects, e.g., responsiveness of district level agencies. As the PPAR was issued in 2005, the recommendations were too late to influence the design of ALRMP II’s original results framework, but they were helpful to the task team in revising the KPIs at the time of the AF amendment in 2006. 40. Despite the avowed intent to strengthen M&E, the ALRMP II PAD discussion on the subject is somewhat diffuse. It refers to a baseline to be initiated before credit effectiveness, a program of ongoing and periodic evaluations (including mid-term and final termination evaluation), and a system of routine records and periodic monitoring reports at community, district and project levels. The statement that “clear responsibilities and procedures for M&E, including frequency, content and audience of project reports, will be established early on in the project life� (PAD, p. 19) suggests inadequate consideration of the project’s M&E needs during preparation. The KPIs were an improvement over those of the first phase project, although there was some overlap among the first 3 PDO indicators (on food aid) and one may have been difficult to measure (i.e., consumption levels in respect of the national food poverty line in particular months each year, for pastoralists and farmers). These were complemented by an unwieldy set of 25 intermediate outcome/output indicators. The PAD had a logical framework matrix and stated that a more detailed log frame was included in the Project Implementation Plan (PIP), but the IL-ICR team was unable to find this. Baseline data were not included in the log frame, although in many cases the team could have done so based on its knowledge of the first phase project. Very few indicators had any quantitative targets. 41. Implementation and utilization of M&E arrangements have been mixed. On the one hand, ALRMP II has a very elaborate Early Warning System (EWS) database drawn from some 10,000 household questionnaires administered monthly, processed and with results available on the web and distributed to district level monthly. The information is widely used by the Kenya Food Security Steering Group (KFSSG), national and international partners, and has shortened the response time and helped to improve the transparency and governance of food relief efforts. ALRMP II also has: (i) participatory needs assessments and plans for all 28 districts; (ii) assessments of potential conflict flashpoints for the districts; (iii) annual project reports for each district; and (iv) environmental audits, which were conducted almost every year. In other words, there is an extremely rich amount of data that could support various research efforts. The absence of a clear analytical structure to guide data collection, entry and analysis means, however, that much of the information on hand may be difficult to reconcile/compare in a statistically robust manner. These weaknesses also made it difficult for the project to respond quickly and with comprehensive information to public criticisms when they arose. 42. Successive ISRs indicate a constant preoccupation by the task team to strengthen M&E. The team developed a tighter set of PDO and output indicators, which were adopted as part of the AF amendment in 2006 (see Section 1.3). Most of the changes improve the log frame, although the IL-ICR has several reservations: (i) the KPI on changes in the share of the population assessed as needing food aid refers to drought 11 normalization (to be able to compare needs across drought years), but does not clarify the methodology. There are different views on the subject and this has complicated end-of- project evaluation; (ii) KPIs on asset/income accumulation and diversification were dropped, although ALRMP II continued to invest in income-generating activities (IGAs); (iii) there are no intermediate indicators to measure effectiveness of participatory arrangements, and some indicators on gender and marginalized groups were dropped; (iv) there is no effort to track specifically which other stakeholders are using the EWS, DSG plans and coordination structures or the amount of additional resources being channeled to the ASALs as a result of the project; (v) output indicators on financial services were retained, although very little implementation was taking place/planned; and (vi) the revised log frame still lacked baselines and targets for many indicators (although in a few cases ISRs had more detail). 43. A statistical baseline survey was carried out by the Central Bureau of Statistics on behalf of the project in 21 of the 28 ASAL districts in 2004/05, covering 4,000 randomly selected households. Later, as ALRMP II completion approached, the International Livestock Research Institute (ILRI) was commissioned to conduct the final Impact Evaluation (IE); a team including some very experienced researchers prepared a useful report. Between June-August 2009, they re-surveyed a subset of the 4,000 households using a similar questionnaire. Sample intervention and control communities were identified and econometric techniques used to control for any underlying differences. A survey was conducted of participants in the DSGs, and anthropometric data measuring nutritional status of over 600,000 children was obtained for the period 2005-09. The nutritional data (middle-upper arm circumference, MUAC) was then used to evaluate welfare changes over time in a stochastic dominance framework, the first time such techniques had been applied for that purpose, and the first to use stochastic dominance analysis for MUAC data in particular. This analysis made it possible to detect project impacts on the nutritional status of the worst off children. Qualitative information was also gathered through focus group discussions in 21 of 24 treatment communities; and the IE used this to prepare a social network analysis of changes in empowerment and ability to access services. 44. This IL-ICR is based on: (i) findings of the IE; (ii) several studies carried out in preparation for the KACCAL GEF activity associated with ALRMP II, and others financed by Danish International Development Agency (DANIDA) and the EU in preparation for a new follow-on project; (iii) ALRMP II project files; (iv) the Borrower’s ICR, Borrower comments on IDA’s draft IL-ICR and other PCU contributions; and (v) IL-ICR team analyses and consultations with a wide range of stakeholders in individual and focus group meetings and three stakeholder workshops. 2.4 Safeguard and Fiduciary Compliance 45. Safeguards. The ALRMP II Environmental Category was B (Partial Assessment). The project triggered IDA’s Environmental Assessment, Natural Habitats, Forestry and Pest Management safeguard policies (OPs 4.01, and 4.09, respectively). The project did not trigger any social safeguard policies and none arose 12 during implementation. An Environmental Assessment (EA) and a Social Assessment (SA) were carried out and an Environmental Management Framework (EMF) was prepared and disclosed in the Bank Country Office and InfoShop. Environmental considerations were integral to ALRMP II, since the project aimed to reduce vulnerability in an environmentally fragile area subject to accelerating and negative climate change impacts. Preparatory work on environmental issues was thorough and thoughtful, both as part of the EA and EMF, and in the design of the three sub-components. The main risks identified concerned potential cumulative impacts of small-scale investments that in themselves were unlikely to be harmful, but in the aggregate could pressure the fragile resource base (e.g., through over-extraction of groundwater, saline intrusion of aquifers due to borehole investments). These aggregate effects were to be addressed both by training and strengthening planning at the overall district level, and environmental screening of individual sub-projects. ALRMP II design also provided for annual environmental audits by independent consultants, in close consultation with Kenya’s National Environment Management Authority (NEMA).7 46. The IL-ICR team found Environmental Audits were done regularly during 2005- 08 and in 2010. It reviewed two of the audits and found them technically sound and thorough. NEMA evaluated results, and comments were transmitted to ALRMP (to component heads and partners) for follow-up. The reports were posted on the website and the Drought Management Officers notice boards. Key findings were included in flyers and brochures disseminated in the project area. ALRMP organized a public awareness week, and used some District Agriculture Shows (e.g., in Garissa, Kilifi, Malindi) to further disseminate environmental messages.8 A range of activities had taken place as planned for DSGs, District Environment Officers, METs, Community Workers and stakeholders. Prior to project implementation, ALRMP had profiled potential impacts of the most common sub-projects, which helped alert beneficiaries and the DSG to the need for mitigation measures. Environmental screening of individual investments included public consultations. ALRMP II had at least two supervision missions per year. These included frequent participation by environmental specialists; and aide-memoires routinely addressed environmental issues, actions and follow-up. 7 NEMA was under establishment during the first phase of ALRMP and became fully operational shortly before the start-up of ALRMP II. Cooperation with IDA through ALRMP II and several IDA-GEF activities has helped to strengthen NEMA capacity; in October 2011, NEMA achieved ISO 9001:2008 certification (www.nema.gov.ke). 8 In field visits the IL-ICR mission was able to confirm dissemination of important messages. For example, in a visit to the Kawana sand dam sub-project under the drought management component (aimed at creating youth employment opportunities and improving livelihoods through enhanced irrigation activities and sale of produce such as tomatoes, spinach, onions, capsicum and kale), the mission observed that youth group had met often with the Mobile Extension Team (MET) on environmental safeguards, development of an environmental management plan (EMP) and a Pest Management Plan (PMP). The latter recommended a mix of farmer-driven, ecologically based pest control practices to reduce reliance on synthetic chemical pesticides; the youths were fully aware of the non-eligibility of pesticides that include diazinon. 13 Fiduciary 47. Procurement. GoK and the task team faced several challenges in designing the ALRMP II procurement arrangements. The predecessor ALRMP had been the first IDA CDD project in Kenya, at a time when national operational policies and procedures did not envisage public funding of such decentralized community-led investments. Moreover, the project was to be implemented in an immense land area, with the least developed road network in the country, inhabited by dispersed pastoralist communities with very high illiteracy. Procurement supervisions were done twice a year. Each mission covered a few districts on an alternating basis due to the widespread nature of the project area, i.e., 28 districts. Post procurement reviews (PPRs) were conducted once a year for the districts covered during supervision missions. Although there was an expectation of a gradual improvement in governance following introduction of the multiparty system in 2002, the pace of procurement reforms initiated in 1998 was sluggish and project procurement risk was rated High (and has remained so throughout the implementation of ALRMP II). 48. In total, US$118.3 million was disbursed out of a total IDA credit of US$124.1 million (original Credit of US$60 million equivalent plus AF of US$60 million and gains from shifts in the USD : SDR exchange rate). Of this, nearly US$80 million was spent on works, goods and consultancy contracts. Expenditures on CDD activities were 16 percent of the total (of which 11 percent for micro-projects and 5 percent for training and other support activities); contracts executed by the PCU represented 64 percent; and the remaining 20 percent or less related to procurement transactions carried out by the Project District Coordination Units (DCUs). 49. GoK prepared a Project Implementation Plan (PIP) that outlined the procurement procedures for ALRMP II. This was supported by a Community Procurement Manual and Community Accounting and Financial Reporting Manual that guided the Community Development Committee (CDC) in carrying out day-to-day procurement activities. The PIP prescribed a three-tier procurement implementation arrangement: (i) The PCU was charged with the responsibility of procuring all International Competitive Bidding (ICB) and a large proportion of National Competitive Bidding (NCB) goods and works contracts, and all consultancy services contracts; (ii) DCUs were responsible for procurement of goods and works required for common community facilities, and procurement of goods for Components 1 and 3. Also, DCUs conducted procurement training for CDC members, and provided technical support and oversight to microprojects; and (iii) A procurement sub-committee of the Community Development Committees (CDCs) was trained and designated to carry out procurement of goods and works for Component 2. 50. IDA supervised procurement implementation in two ways. ICB contracts and consultancy contracts costing US$200,000 or more were subjected to prior review. IDA also conducted annual PPRs on the procurement documentation for non-prior review contracts at the PCU, DCU and CDC levels. The findings of all PPRs rated the 14 procurement performance at the three levels as satisfactory, and IDA did not identify any incidents of mis-procurement. PPRs at the district and community levels were carried out during the supervision missions, and physical verification of goods and works procurement constituted an integral part of the exercise. 51. Financial Management. The IL-ICR team reviewed the design and implementation of FM arrangements. 52. Flow of funds. Delays in disbursement were to be addressed to ensure timely release of funds and authority to incur expenditures (AIEs) to districts. This action was fully met and the disbursement rate of the project was one of the highest in the Kenya portfolio. The PCU ensured that funds to districts were released together with the AIEs. The disbursement rate on the credit was high in spite of the fact that ALRMP II financing was suspended during the final six months before the Closing Date. There were some challenges in raising community contributions and/or valuation of contributions in kind. Disbursements were made to some community groups without the requisite community contribution being made, as specified in the PAD.9 The reasons advanced by the PCU for this were that some communities were too poor to make the contribution, or that it had been made in kind but could not be quantified/valued and recorded. There were challenges with in-country funds flow arrangement of moving funds from the Special Account to the project account and eventually to the CDD accounts. 53. Budgeting. A budget monitoring committee was established, but the line budget approach of GoK differs from the IDA’s eligible expenditure format and this created challenges in monitoring budget implementation. Therefore, there is a need to harmonize IDA reporting with the GoK chart of accounts and budgets. GoK has now adopted a new chart of accounts on the Integrated Financial Management Information System (IFMIS) platform that enables reporting by both activity and line item up to district level. 54. Accounting and Financial Reporting. The accounting capacity throughout project implementation was assessed as adequate with 3 accountants at the PCU and project district account in each of the 28 districts. The project jointly with IDA’s FM team conducted regular capacity building training at Headquarters and districts. Delays in rollout of IFMIS both at national and district level adversely affected the accounting system with various material issues noted by auditors between project accounts and IFMIS ledgers. This created material variances in the financial statements prepared by the project and the corresponding project ledger accounts at the ministry and Minister of Finance (MOF) level, and resulted in additional work of reconciling year-end accounts for audit. This issue was the basis for audit qualifications by the Kenya National Audit Office (KENAO) during the year-end audit review. The IDA FM team provided training for the GoK Internal Audit Department (IAD). This allowed the IAD to conduct a 9 However, in the aggregate community contributions of 30.9% slightly exceeded the share defined in the project documents (30% - see Annex 2). 15 comprehensive risk-based fiduciary review of the project in January-March 2009. The review consisted of detailed transaction review and on-site field audit at district and community level. The review identified areas of internal control weaknesses that were brought to the attention of the implementing agency for remedial measures. 10 On the basis of the fiduciary review, the IAD included the project in the list of those that were eligible for in-depth audit review that commenced in July 2009. Because INT had already decided to undertake a separate and independent forensic audit, IAD, with agreement of the World Bank, dropped the project from the in-depth audit. This decision was taken to avoid duplication of effort, but ultimately put the project at a disadvantage due to delayed completion of the forensic audit by INT and ambiguity regarding interpretation of the findings. While the IDA Team detected control weaknesses, it is questioned whether the severity and scope of potential issues was realized and whether the recommended action was sufficient and followed-up systematically. 55. The project was able to prepare and submit Interim Financial Reports (IFRs) to the Bank within the stipulated time lines. However, the accounting weaknesses at district and community level as well as the variances in project and ministry level accounting records negatively affected the quality of the reports. The annual financial statements were similarly affected as noted from the audit qualifications. In addition, the IFR was prepared largely to meet the Bank’s financial covenants, but was not utilized fully by the implementing ministry as a management tool. As part of the IDA FM team capacity building measures, the project was able to submit analysis of funds disbursed to communities per district as part of the quarterly IFR. However, the IFR did not indicate the status of accountability of these funds once disbursed. This was in line with the agreed FM arrangements of the CDD components, whereby disbursement via the IFR was made on the basis of advances to communities and not final accountability by the same. This was a portfolio level limitation for all Kenya CDD projects. The FM arrangements at district and community level at the time would not have been able to support such accountability within the deadline required for submission of the IFR. As the IFMIS is rolled out to the district/county level, it will be possible in future Kenya CDD projects to provide accountability of community advances before reimbursement is made via IFR/Statement of Expenditure (SOE). The IFR is also an unaudited financial report and supporting documents are not attached at time of submission and clearance. However, the Bank FM supervision, and audit reports (KENAO and IAD) provide for regular detailed transaction reviews to reconcile the IFR expenditures to the cashbooks, payment vouchers and other underlying supporting documents. 56. Internal Controls. The project developed a community grant manual for the CDD component, and relied on GoK financial guidelines and regulations for Headquarters and district levels. There was good compliance on both the grant manual and the GoK 10 IDA had downgraded the FM rating to MS in the ISRs of December 2007, May 2008 and December 2008. On the basis of actions taken to correct weaknesses noted, the rating was upgraded to S in June 2009, then lowered to MU in December 2009 and upgraded again to MS in June 2010, the final ISR before the closing date. 16 financial regulations. However, internal control lapses occurred at Headquarters, district and community level. This was partially due to the fact that GoK regulations on CDD operations are not comprehensive. This is a country level FM capacity issue that was flagged in FM supervision reports and GoK internal and external audits.11 In addition, material internal control weaknesses were noted at the PCU, districts and community levels. There was weak accounting capacity at community level that adversely affected the project internal control environment. The FM reviews revealed instances whereby project funds were not properly accounted for, or the payments were otherwise not properly supported (e.g., weaknesses with fixed assets control, vehicle usage and costs, workshops, staff advances/imprest, staff allowances, micro-project control, value-for- money). These weaknesses were not detected before 2007, when a complaint was lodged and an internal review undertaken. Adjustments in internal controls were introduced at the time in response to these observations, but the effectiveness of these adjustments is not known due to loss of records during the forensic audit and the absence of a transaction-level audit subsequent to the introduction of changes. Some cases of suspected fraud and corruption were noted in the audit reports, which, together with additional outside reports of irregularity resulted in the subsequent INT forensic audit. Follow-up on suspected fraud and corruption is still under way and not yet conclusive. The questions about internal controls and difficulties reaching definitive answers confirm that the risks were high. The fact that questions raised proved difficult to answer confirms the need for better record keeping and faster access to primary records in response to queries. 57. Auditing. Audit reports were submitted on time and the PCU followed up on findings. The audits were conducted in line with International Standards of Accounting. The FM team conducted FM capacity building training of the KENAO which enabled the auditors to conduct on-site field audit reviews at district and community level. The audit reports were qualified on the basis of variances between the line ministry ledger accounts of the project and financial statements prepared by the PCU, as discussed below. 58. Social accountability. Although the original project design did not specifically provide for social accountability mechanisms, agreements reached at the time of the AF strengthened requirements for public disclosure and access to information, and scaled up community, district, and national monitoring. The grant manual was revised over time accordingly, particularly with respect to public reporting and complaints handling. There were requirements that cheques to the communities be issued in public forum, project budgets be disclosed at district level, and that district level financial reports be accessible to community members. Finally, information related to the micro-projects was to be disclosed at the project office in each district. Implementation of some requirements was fairly consistent (e.g., public disclosure of micro-project information at district offices), but less so in other cases (e.g., branding of micro-projects and other local investments), 11 IDA is now working with an MOF-led effort to develop a new policy to ensure effective governance structures and good financial management and procurement systems and guidelines for devolved funds at national, county, facility and community levels (see Annex 2). 17 and supervision by IDA of social accountability dimensions was not systematic. Overall, more could have been done to enhance social accountability mechanisms earlier, as was already the case in some CDD projects in other countries. 59. INT Review. In 2009, INT began a review of possible fraud and corruption in ALRMP II, in response to information from outside the Bank and from a report by the Africa Region conveyed in May of 2007. At the time of the 2007 report, INT reviewed the response of the GoK and determined that further action by INT was not then required. The issue was reopened in early 2009 in response, as noted above, to an in-depth review of several projects by the FM team and to a renewed external report of irregularities. In July 2010, INT provided a verbal report to Management, indicating that a high share of expenditures incurred in financial years 2007 and 2008 in seven districts reviewed by INT were questionable and suspected fraudulent. Upon receipt of the verbal report, Management informally suspended IDA financing for ALRMP II. A preliminary written report of findings was given to GoK nine months later, in April 2011, with further annexed details provided in June 2011. The written report had a share of questionable expenditures considerably lower than that contained in the earlier verbal report to management. INT’s preliminary written report was published on the INT web site on July 15, 2011. Kenya’s IAD reported back its assessment of the INT findings on August 16, 2011, questioning the methodology used by INT and the findings on the extent of misuse of funds and/or corruption. The IDA Board of Directors was briefed on the status of the review at several points, and requested a discussion that was held in October 2011. Subsequent to the Board discussion, IAD and INT undertook an effort to resolve the differences in approach and findings. In November 2011 the GoK agreed to repay KES 340 million (US$3.8 million at exchange rate of November 30, 2011), although agreement has not yet been reached on the extent of and reason for ineligibility of expenditures due to continued questions on methodology and procedure of the forensic audit. This IL-ICR comments on the process of the review to the extent that it had a bearing on supervision (Section 5.1(b)), on the impacts on implementation of the ALRMP II program (Section 2.2), and the risks to sustainability of its achievements (Section 4). The findings of the review and their interpretation are still under discussion and beyond the technical scope of the ICR team. 2.5 Post-completion Operation/Next Phase 60. Until the final stages of ALRMP II implementation, GoK, IDA and other partners had expected that there would be a follow-up operation. Substantial preparatory work was completed. As a result of issues related to the forensic audit, however, Bank management stopped preparation work in late FY11, and no decision has been made on a next round of support. IDA has been the lead partner for ALRMP. Other donors that had planned to co-finance the next phase are therefore waiting for a decision by IDA. 18 3. Assessment of Outcomes 3.1 Relevance of Objectives, Design and Implementation 61. The objectives of ALRMP II had high strategic importance for Kenya when it was designed, and this continued throughout implementation and remains the case today. Development challenges of the ASALs figured prominently in Kenya’s PRSP (Economic Recovery Strategy – being drafted at the time the project was under preparation and finalized in 2004), and were reiterated in the letter of Sectoral Policy submitted to IDA at the time of credit negotiations. They were also well reflected in IDA’s Country Assistance Strategy at the time. The PRSP states that “to address poverty reduction targets, the ERS identified some priority programs that target poor communities, the poor in arid and semi-arid areas, the urban poor and the marginalized poor.� The Country Assistance Strategy (CAS) from 2004 highlights that targeted programs under the pillar 2 poverty reduction and equity include community-driven development in the arid and semiarid areas of Kenya The challenges of the ASALs figure in the more recent Vision 2030, and ALRMP II was expected to play an important role in helping to achieve several outcomes of IDA’s current Country Partnership Strategy for Kenya, in relation to poverty, decentralization and improved local governance, environment, natural resource management and climate change adaptation. As global attention to climate change has increased, Kenya has been identified as one of the countries likely to be severely affected – some, though not all, of these challenges relate to worsening conditions in the ASALs. Though not described as such, because the terminology was not in common use at the time, ALRMP II is very much a climate change adaptation project, and its successful implementation has yielded a number of important lessons that need to guide future efforts and outcomes that should be sustained. 62. The initial design of some ALRMP II activities, in particular the CDD component and M&E arrangements could have been improved, but actions taken during implementation overcame several shortcomings. Annex 2 offers suggestions for further strengthening indicators for related interventions in the future. The project had several innovative features that have attracted attention, in particular the community-based, multi-sectoral EWS that incorporates indicators relevant to pastoralist communities and the inclusion of conflict resolution activities as an integral part of the natural resource management component. Impact evaluation also involved application of new research techniques to assess impact of droughts, normalized for severity, on children at different nutritional levels. This approach has the potential for much broader application. 63. The project design and implementation were relevant and appropriate for the achievement of the PDO. Shortcomings in design and implementation did not compromise the achievement of the PDO, as confirmed through assessment of the agreed indicators at project close. 3.2 Achievement of Project Development Objectives 64. The project achieved the PDO in its three core elements, i.e., reduced vulnerability of livelihoods, improved food security, and better access to basic services. 19 Focusing on the direct impacts, the IL-ICR team considers that the PDO has been achieved and that the results can be attributed to ALRMP II. The three project components were well justified from the standpoint of causality (the contribution that they were expected to make, individually and collectively, towards meeting the PDO is plausible); the revised PDO indicators were appropriate to assess the PDO. It would have been useful to capture the donor and interagency coordination aspects of ALRMP II either as a sixth PDO indicator or as an Intermediate Output/Outcome Indicator, because the lessons may be very important for other programs. 65. Drought-related vulnerability of beneficiaries in the ASALs declined compared with the population not assisted by the project. Drought management, natural resource management and access to services in the ASALS improved through activities that engaged affected communities at the local level and increased attention to ASAL development issues at the national sphere. Reduced vulnerability was measured mainly by changes in beneficiary households’ needs for free food aid (KPI 1). For those households that still needed food aid, ALRMP II facilitated inter-agency coordination and shortened the time interval before they received assistance (KPI 2), thereby reducing the extent of negative welfare impacts, for example the nutritional status of children under 5 years (KPI 3). 66. The PDO is clear and the project structure and revised results framework provided a good basis for implementing, and assessing progress towards achieving the expected outcomes. Table F summarizes the findings with regard to the main PDO and intermediate outcome/output indicators. Further details are provided below and in Annex 2, including linkages between component activities and indicators. Findings are based on an assessment of available quantitative and qualitative data, including extensive consultations with partner organizations, Non-Governmental Organizations (NGOs), line agencies and beneficiaries. 67. Decreasing the proportion of people assessed as needing free food aid in each arid and semi-arid district affected by severity of drought (KPI 1: Achieved). KPI 1 addresses the food security and livelihood vulnerability dimensions of the PDO. The Kenya Food Security Group’s Long Rains Needs Assessments estimate a decrease in the share of people needing food aid in the ASALs districts, from 51 percent in 2000/01 to 28 percent in 2008/09. However, there are some differences in geographic coverage of the two years, drought severity is not normalized, and it is not possible to determine attribution. 68. To address this, two further analyses were done, by ILRI for the IE and by the IL- ICR team, to assess the impact of ALRMP II on the share of people assessed as needing food aid in the project area. ILRI and the IL-ICR team also sought qualitative feedback from stakeholders.  The IE analyzed data on changes in food aid needs and on the relationship with relevant ALRMP II expenditures in 10 randomly selected project districts for the period 2004/05 to 2008/09. It concluded that: (i) there was a decrease 20 in food aid needs in most districts, compared with the reference district; and (ii) ALRMP II cumulative expenditures were negatively correlated with the percentage of people needing food aid. The relationship was statistically significant in the arid districts (where all three project components had operated), but not in the semi-arid districts (where only Component 1 operated).  The IL-ICR team analyzed sub-district level data from the ALRMP II household-based EWS, satellite-based rainfall estimates and ALRMP II micro-project investment data for the 2005-09 periods. Treatment and control divisions (sub-district administrative units) were identified based on cumulative ALRMP II expenditures on water, livestock and agriculture (KES 100 per capita cut-off). The analysis covered 51 treatment and 46 control divisions, approximately equally divided between arid (52) and semi-arid (45). To adjust for drought severity, a cumulative rainfall index was developed based on a three-year weighted average of rainfall.12 The results indicate that treatment divisions did much better than control divisions in terms of the percentage change in the proportion of people assessed as needing food aid, and that the difference (-19.7 percent) was statistically significant at 5 percent. Also, at the division level where the IL-ICR team focused its analysis, for the population considered part of the treatment group the average share of people needing food aid fell from 31.9 percent in 2005 to 30.2 percent in 2009; the weighted overall average for the ASALs fell from 27.2 percent to 23.2 percent. The fact that the share originally needing food aid was higher in the treatment divisions suggests that ALRMP targeted the worst off areas in the ASALs.  In terms of qualitative assessment, these trends and differences between project-assisted and control areas coincide with the perceptions expressed to the IL-ICR team by development partners, government line agencies, NGOs and beneficiaries during stakeholder consultations (Annex 6). The qualitative information gathered by ILRI in 21 focus groups also found a decrease in vulnerability to drought, particularly in the arid districts where all project components had operated. 69. Taken together the above results indicate that ALRMP II was able to reduce food aid needs in those areas in which it operated. 70. Reducing the time lapse between reported drought stress and response (KPI 2: Achieved) KPI 2 captures the food security dimension of the PDO. Although some studies (e.g., by Tufts University) have reported that overall response time for the 2000/2001 drought was six months (and some of this information was included in ISRs), 12 This drought normalization methodology differs from the Normalized Difference Vegetation Index (NDVI) used by ILRI in the IE; it is based on specialist advice from NASA. 21 the ILRI-IE found that users of the ALRMP Bulletin had a mean response time of 7.6 weeks in 2000, decreasing to 3.5 weeks in the drought of 2009. The analysis is based on a survey by ILRI of DSG members in 10 study districts to find mean response times for users and non-users. The DSG comprises all relevant actors (international organizations, line Ministries, NGOs, community-based organizations and donors, with ALRMP II acting as secretariat) that influence or are directly involved with emergencies at district level. The ALRMP II EWS Bulletin has become, by far, the most common information source on droughts: Seventy-eight percent of all relevant organizations used the Bulletin in 2009, up from 32 percent in 2000. In sum, controlling for other factors affecting change in response time, ILRI found that the use of the ALRMP II Bulletin was significantly and negatively associated with the change in response time over the project period. The IL-ICR team estimates that if only the activities of the Drought Contingency Fund are considered, the response time would be even lower at 2.5 weeks. 71. In terms of qualitative feedback, in the three stakeholder workshops conducted by the IL-ICR team, participants (communities, line agencies, partner organizations operating at field level) uniformly reported that they felt the drought response time had decreased compared with the pre-project situation, and attributed this to the EWS, the project contingency fund and the role ALRMP II played in promoting coordinated action by all parties in using the EWS information. They felt this had significantly improved knowledge about impending drought developments and their ability to take mitigating actions. 72. Improved nutritional status of children under 5 years of age affected by severity of drought over time (KPI 3: Achieved). KPI 3 covers the food security and vulnerability dimensions of the PDO. The IE used MUAC as an indicator of child nutrition status. A two period (2005 and 2009) panel was constructed, summarizing MUAC measurements of 602,000 children in 10 districts. Stochastic dominance analysis across the distribution of MUAC scores indicates improvements in nutritional status over time. The analysis made it possible to capture changes at the lower percentiles rather than only those at the mean or median. For the 10th percentile, 15 percent of intervention sub-locations had a negative change in MUAC scores while 25 percent of the control sub- locations experienced a negative change. Put differently, 85 percent of intervention sub- locations did not have a negative change (stayed the same, or improved) versus only 75 percent of control sub-locations. The fact that MUAC levels in treatment sub-locations were initially worse than in control areas suggests that ALRMP II targeted the worst off sub-locations. The IE concluded that the ALRMP II project effectively functioned as a nutritional safety net for children under 5 years of age. 73. Increased number of people with access to basic services (water, human and animal health services and education) (KPI-4: Achieved). KPI 4 addresses the access to basic services elements of the PDO. The IE compared changes in access to services in treatment and control locations. Treatment communities experienced statistically significant improvements in access to quality water sources; primary, secondary and adult education; and to veterinary medicines than at the start of ALRMP II. The percentage of households consulting medical professionals, using bed nets, and seeking 22 veterinary extension services also increased, and negative health-related outcomes (prevalence of child diarrhea and livestock mortality) decreased significantly. Animal deaths declined by a quarter and deaths from diseases by nearly half. Changes from 2004 to 2009 were significant within intervention sites for most services (ALRMP II had a significant impact on access to services in the areas where it made investments), however differences between treatment and control areas were not statistically significant, suggesting that investments by others also took place in control areas. This may be a by- product of the good inter-agency coordination through the DSG that stakeholders emphasized during IL-ICR consultations in Nairobi and the field (Section 3.6). ALRMP II advocacy on behalf of the ASALs also probably helped to improve service delivery in both intervention and control communities. 74. The IL-ICR team also paid particular attention to water issues. The IE found that time to water had increased in the treatment and control locations. The IL-ICR team found that between 2003-09, the mean household time to water fell from 85 minutes to 42.6 minutes in the rural North-Eastern province, while the national average did not change significantly (IL-ICR calculations based on Demographic Health Survey (DHS) 2003 and 2008/09). On distance to water, the EWS shows that between 2005-10 mean distances decreased across the project area: (i) to boreholes, from 6.4 to 4.3 km; (ii) to shallow wells, from 5.2 to 2.9 km; and (iii) to pans and dams, from 3.8 to 3.0 km. Although it was not possible to determine attribution, ALRMP II did make a significant number of water investments between 2004 and 2008. 75. Increase people’s participation in the project districts in local and national development as demonstrated by the reflection of arid lands concerns in the economic recovery strategy and other relevant national policies (KPI-5: Partially Achieved). KPI 5 addresses the vulnerability dimensions of the PDO. Participation/community empowerment: using social network measures, the ILRI IE found measurable improvements in empowerment, with general increases greater in arid compared to semi-arid districts (possibly because arid districts had been part of both ALRMP and ALRMP II, and the CDD component did not operate in the semi-arid districts). Influence on national policies: ALRMP II also had a significant impact on formulation of relevant national policies, both through advocacy activities and as a source of reliable data and advisory support. The project was involved in the formulation of nine (9) draft policies of direct relevance to the ASALs. The Economic Recovery Strategy for Wealth and Employment Creation (Kenya’s PRSP, 2004) includes a full chapter influenced by the ALRMP II PCU and project beneficiary communities at the national and local levels. Among other policies and sessional papers to which ALRMP II provided important inputs are the National Policy for Sustainable Development of Arid and Semiarid Lands of Kenya, National Land Policy, National Disaster Policy, Policy Framework for Nomadic Education in Kenya, National Policy on Land Reclamation, National Policy on Peace Building and Conflict Management, National Food Security and Nutrition Policy, National Livestock Policy, and National Irrigation and Drainage Policy. However, some of these policies were still in draft (had not yet been formally adopted) at the time of the IL-ICR team’s visit. Kenya’s Vision 2030 (successor to the Economic Recovery Strategy) also reflects an emphasis on the ASALs. ALRMP II 23 advocacy also led to the creation of a new Ministry of State for the Development of Northern Kenya and Other Arid Lands (April 2008), and more recently to the establishment of a National Drought Management Authority (November 2011). Finally, the experience and knowledge generated by ALRMP II contributed to the formulation of Kenya’s official thinking on climate change adaptation. 76. As described above, 4 KPIs have been fully achieved and 1 KPI partially. Each KPI addresses at least one of the three dimensions of the PDO, i.e., reduced vulnerability, improved food security, and better access to basic services. Thirteen out of 16 Intermediate Outcome Indicators (IOIs) have been achieved. Annex 2 describes in details how the IOIs and KPIs are linked to the achievements of the three different components. While these achievements would usually justify a satisfactory rating for efficacy, a conservative approach is taken and the rating is downgraded to moderately satisfactory, mainly since uncertainties remain with respect to the extent of possible fiduciary issues. 3.3 Efficiency 77. Micro-project investments. The PAD did not provide an ex-ante overall Net Present Value of Economic Rate of Return for ALRMP II, on the grounds that the micro- projects and similar small investments under the three components have heterogeneous impacts that do not lend themselves to such analysis. Instead, the PAD assessed the viability of several individual activities (irrigation, rain-fed agriculture, petty trade, livestock trade and bee keeping) and found them all to be financially viable. 78. Since the ALRMP II results framework as revised in the AF amendment does not include indicators on returns to micro-project investments, the ILRI-IE did not cover this subject. The IL-ICR team therefore conducted a Cost-Benefit Analysis, which quantifies the benefits that result from two sets of mechanisms that can be attributed to the project. First, the project financed various types of micro-project investments. Secondly, the project invested resources in the water and livestock sectors (e.g., livestock restocking, veterinary services and improved access to animal water sources). These investments resulted in (counterfactual) gains in (indigenous) livestock wealth. With respect to the micro-projects, four were randomly selected that were representative of the most common types of project investments (see Annex 3). The micro-projects were selected from a comprehensive roster of geo-referenced project investments, and the analysis was carried out during field visits in connection with the stakeholder workshops. They include investments in sand dams, irrigation, goat dairy breeding and fishponds. In the aggregate, these activities account for 65.5 percent of the total number of micro-projects supported by ALRMP II and 43.4 percent of the total value of investment in all micro-projects. The benefit-cost ratios ranged from 1.1 to 2.1, and the ERRs from 24 percent to 78.5 percent. Weighting these returns in proportion to overall project expenditures yields an overall ERR on micro-project investments of 37 percent. In addition, the IL-ICR team calculated decreases in livestock after each of three previous droughts and estimated counterfactual changes in livestock without project interventions. The gains from investments in the water and livestock sectors can, in part, be quantified by the (counterfactual) increases in livestock wealth and these gains were estimated to be consistent with an ERR of 28.2 percent (see Annex 3 for details). Further, because of the still open issue of questionable 24 expenditures from INT’s findings, sensitivity analysis was conducted to assess the impact of potential misuse of funds on the project’s economic viability. Assuming no funds were misused results in an ERR of 40.4 percent. Under an assumption that 20 percent of funds were misused, the project would yield an ERR of 31.1 percent. Misuse at the level of 50 percent (considerably above INT’s finding of 29 percent questionable expenditures and no confirmed misuse) would result in an ERR of 15.9 percent. These results lead to the conclusion that under a range of assumptions about the extent of possible diversion of funds from 0 percent to 50 percent the project was economically viable and the costs incurred were justified. 79. Cost-effectiveness and efficacy. Regarding cost effectiveness relative to normal GoK investments for similar activities (e.g., schools, health dispensaries), ALRMP II did not capture comparative data. However, the IL-ICR team does not anticipate that there would be major differences because line agencies were so heavily engaged through the DSGs and METs that standards probably did not differ greatly (the 30 percent community contribution under the CDD component would have made those micro- projects less expensive for the government, but this would not alter overall cost- effectiveness, taking into account all sources of finance). Regarding efficacy, the fact that the micro-projects resulted from a consultative planning process in which beneficiary communities and the range of DSG technical members were able to participate suggests that they responded to perceived needs from the standpoint of these various actors. In 2010, the firm Batch Associates Ltd. conducted a technical audit of ALRMP II community-level infrastructure investments covering 27 randomly selected micro- projects in 4 districts (out of a total universe of 2400 microprojects). The audit found 82 percent of communities ‘very satisfied’ and 11 percent ‘satisfied’ with the micro-projects; 7 percent (2 communities) had issues. The same review rated the technical quality of construction for most micro-projects visited where this was relevant (in a few cases, such as purchase of pumps, construction quality was not relevant). The study team assessed construction quality as ‘satisfactory’ for 87.5 percent of micro-projects; 12.5 percent (3 micro-projects) had issues. They also noted that for a number of micro-projects, operations and maintenance (O&M) arrangements were not clear or needed improvement. Rigorous data comparing costs of micro-projects under the ALRMP II and which similar micro-projects financed by other donors or NGOs were not available to the ICR team. 80. Other efficiency considerations: The IL-ICR team noted that the PCU structure (Nairobi and field) is lean, with no obvious excess staffing (indeed, perhaps too thin in a few areas; see Section 5.2(b)). 81. Finally, the team was unable to judge the cost-effectiveness of the ALRMP II EWS vis-à-vis those of other countries, because each is quite different in terms of needs, coverage, technological features, and other factors. However, to the extent that this information is being widely used (the IE found that 78 percent of survey respondents use the EWS Bulletin as their primary source of early warning data on droughts, up from 32 percent at the start of ALRMP II), and that ALRMP II institutional arrangements appear to have greatly increased inter-agency cooperation and specialization of function, the 25 project has very likely contributed to greater efficiency in the delivery of relief activities, investments and services, by a wide array of actors. 82. Overall, the efficiency is rated moderately satisfactory. In general, the analyses conducted indicate high efficiency of project investments and would justify a satisfactory rating. However, due to FM issues identified, remaining uncertainties of the INT forensic audit, and some limitations with respect to the ERR calculations and data availability, the ICR team rates the efficiency conservatively as moderately satisfactory. 3.4 Justification of Overall Outcome Rating: Moderately Satisfactory 83. ALRMP II has fully achieved 4 of the 5 PDO outcome indicators and partially achieved the fifth indicator. It also achieved 13 of 16 intermediate outcome indicators, and partially achieved two. In light of the rating of satisfactory or substantial for relevance, moderately satisfactory for efficacy, moderately satisfactory for efficiency, and the remaining uncertainties with respect to the fiduciary issues, the overall outcome is rated moderately satisfactory. 84. ALRMP II is also a good practice example of real cooperation amongst all actors – international partners, NGOs, government officials from different line ministries, especially at district level. The high degree of coordination complicated comparisons between treatment and control areas for some indicators. The IL-ICR has taken a fairly conservative approach to direct project outcomes, confining the analysis to those that could be linked to relevant ALRMP II investments. However, the project’s advocacy, policy and donor coordination activities at the national level also benefited control areas, by drawing greater GoK and development partner attention and resources to the ASALs. Because of the very successful inter-agency coordination, other actors report having prioritized control areas due to the presence of ALRMP II in treatment areas. The IL-ICR does not consider these indirect impacts and, in that regard may well understate the full benefits of ALRMP II. 3.5 Overarching Themes, Other Outcomes and Impacts (a) Conflict Resolution, Gender, Inclusion and Empowerment 85. Project design drew on both the Social Assessment done as part of preparation for ALRMP II, and on the Beneficiary Assessment that had just been completed to assess the impact of the first phase of ALRMP. The project placed strong emphasis on community participation, conflict resolution, inclusion of marginalized groups, gender mainstreaming, and enhancing the voice of pastoralists and poor farm households from the ASALs at district and national levels. 86. Evaluations and focus group discussions facilitated by the project, IDA, the IL- ICR team and external evaluators repeatedly document community and donor perceptions that ALRMP II made substantial progress in empowering communities, putting in place conflict resolution mechanisms that worked, enhancing awareness of inequalities and opportunities for women and marginalized groups to participate in development activities. 26 At the same time, social development objectives articulated in project documents were not as fully implemented or consistently supervised as initially planned, and weaknesses in monitoring indicators make it difficult to rigorously determine some of these outcomes. 87. Conflict resolution. The Beneficiary Assessment for the first phase of ALRMP noted that although not part of the original work plan, conflict resolution had been so important on the ground in the arid districts (e.g., unresolved land tenure issues among neighboring pastoralist communities and between pastoralist and farming groups; the general state of insecurity and influx of small arms from neighboring countries) that it had become a de facto fourth project component.13 The design of ALRMP II therefore highlighted the linkages between conflicts and natural resource management, and built on a combination of customary and formal institutions. Peace building and conflict management was one of six activities funded under the Component 1 of the project, which included a wide variety of NRM training activities and investments that responded well to local priorities and need. Conflict management was also mainstreamed into the CDD manual, which includes multiple exercises devoted to identifying past, current and potential sources of conflict, and to developing conflict-sensitive, and cross-ethnic/clan activities in the participatory integrated community development (PICD) process. 88. Drawing on lessons from the ALRMP II mid-term review, the project increased the emphasis mechanisms that would be proactive in identifying sources of conflict and conflict resolution strategies, rather than simply reacting to conflicts that had already broken out. The results framework in the PAD had as an output indicator that “conflict management mechanisms for NRM established at potential flashpoints�. The revised results framework adopted in connection with the AF in 2006 strengthened the indicator: “Key potential conflict flashpoints identified in each district and conflict management initiatives strengthened.� Following the post-election violence in 2008, the PCU hired a conflict officer who focused on capacity building and on establishing and empowering District Peace Committees (DPCs). The DPCs sought to link traditional and formal mechanisms for conflict prevention, resolution and management by bringing together stakeholders who work on peace and security issues in their districts. Unlike traditional conflict management that typically targets higher-level authorities, ALRMP II focused assistance on local level conflict management, including using the CDC to provide a link between local community leaders and district level authorities. 89. At project closure, an assessment of potential conflict flashpoints had been carried out for all 28 districts, and DPCs had been established in many locations in the ASALs. The Borrower’s ICR states that 423 conflict management activities had been carried out at normal/alert and drought situations. The IL-ICR team received consistently positive 13 “ALRMP upon realizing that development could not proceed until some of the local emergency problems were resolved, undertook the lead in conflict resolution and assisting victims with emergency measures these issues became an important entry point for ALRMP, meeting with unprecedented success. The credible manner in which ALRMP handled these emergency situations and the transparent response determined the future relationship between ALRMP and the communities.� ADCL, 2002. 27 feedback from donors, beneficiaries and other stakeholders on the effectiveness of ALRMP II’s conflict resolution efforts. The mission does not infer from this that there has been a net decrease in the level of tension and violence in the ASALs, because there is not sufficient long-term data to make such an assessment. However, it is comfortable in concluding that ALRMP II has had a significant impact on addressing the incidence of natural resource-related conflict in the project area. 90. Gender mainstreaming and inclusion. The project also successfully strengthened its emphasis on gender mainstreaming. The Borrower’s ICR, other evaluations and feedback from beneficiaries and stakeholders note a substantial positive impact on women and girls in the arid and semi-arid districts where ALRMP II operates. Increased access to water, education, and health care is credited with having benefited girls and women disproportionately, as they are traditionally disadvantaged in access to these services, and bear the burden of providing water for the family. The participatory methodology for the development of community priorities for the CDD component mandated women’s participation, as did the formation of CDCs. Gender focal points comprising eight members were created in every district to ensure that gender issues and needs were incorporated in all development projects at the district level. Unfortunately, the revised project results framework did not include gender indicators, however the Borrower’s ICR reports that in most districts over 40 percent of the CDC members are women (i.e., exceeding the one-third requirement) and almost the same number are members of the procurement sub-committee, and women accounted for slightly more than half (52 percent) of all CDD trainees. During the IL-ICR team visits to field sites and in stakeholder workshops, it was noticeable that women were active participants in the meetings, were part of the CDCs, and themselves pointed out that in the past they would have not thought of standing up and talking in such meetings in the presence of male community members. Development partners reported similar findings to the team. 91. ALRMP II was also actively involved (through the Support to Local Development component) in policy formulation and advocacy work, and it used this platform to promote the incorporation of gender concerns into a range of relevant policy papers (e.g., draft National Land Policy, draft Peace Building and Conflict Policy, and draft Disaster Management Policy). 92. Inclusion of vulnerable and marginalized groups and people. The Borrower’s ICR notes that all districts provided some evidence of communities making special provisions for the neediest groups in society, and describes specific examples of these initiatives. The PICD process placed special attention on integrating women and marginalized groups into the planning process in ALRMP II. Communities used a wealth-ranking tool to identify the neediest people amongst themselves. The ICR also notes that strategies to integrate the needs of marginalized groups varied from one community to another, e.g., in some cases, efforts were made to incorporate representatives of marginalized groups into existing CDC structures while in others, attempts were made to establish separate CDCs where the marginalized were in the majority. Safety net micro-projects were aimed at vulnerable people and the Borrower’s ICR documents multiple investments aimed at supporting the most vulnerable, primarily 28 through restocking, shelter improvement, and support to income generation activities for vulnerable groups. These interventions are credited with giving vulnerable groups greater voice, priority in restocking programs and, for some, enhancing livelihoods, nutritional status, and enabling education of children. Unfortunately, the lack of relevant monitoring indicators and reporting during supervision limits the degree to which these impacts can be quantified. 93. Empowerment. Although the theme of increasing voice and empowerment of the ASAL population was central to ALRMP II, the results framework did not have specific indicators to monitor change over time. Nonetheless, the ILRI Impact Evaluation employed social network analysis techniques to measure changes in community empowerment in two ways. The first was the extent to which communities sought services from providers rather than simply waiting for the services to be received. The second was a measure of ‘node betweenness centrality’, which indicates the extent to which an actor (in this case the ALRMP II) is an intermediary between other pairs of actors in the network who are not directly linked. On the first measure, the IE found an increase of 54.2 percent across all sites between 2004 and 2009. On the second measure, the community betweenness centrality value increased from 50 before the project to 160 afterwards, which indicates that ALRMP II has played an important linking function. In both cases, changes were greater in arid versus semi-arid communities, possibly having to do with the fact that all project components operated in the arid districts, whereas only Component 1 operated in the semi-arid communities (and some only after the AF). In a survey conducted by ILRI for the IE, DSG members were asked to identify the most significant changes that they had observed at community level as a result of ALRMP II community development and policy advocacy activities: 35 of 40 respondents cited empowerment as the most significant impact of community development, and 12 of 18 respondents cited it as the most significant change resulting from policy advocacy. The IL-ICR team received similar feedback in its consultations with stakeholders and beneficiaries. (b) Institutional Change/Strengthening 94. ALRMP II had a positive direct impact on the well-being of beneficiaries due largely to the effectiveness of the institutional arrangements that it put in place, or strengthened. These arrangements also helped other actors to be more effective in the delivery of their drought management and vulnerability reduction efforts in the ASALs. 95. The EWS and its relationship with the Kenya Food Security Meeting (KFSM) and KFSSG, the DSGs and Drought Contingency Fund were in place at the close of the first phase project, but were strengthened considerably and their geographic coverage was extended from 11 to 28 districts under ALRMP II. DPCs were created for the first time under ALRMP II.  Regarding the Drought Early Warning System, reliance of all parties (including participants in the KFSM and KFSSG and others active in the ASALs) increased significantly during implementation of ALRMP II (Section 29 3.2 (iii)). Technical improvements under ALRMP II included the following: (i) while collection had been ongoing since the first phase, poor data storage protocols and unclear data collection methodology had resulted in the waste and, in some cases, loss of data. Under ALRMP II, the survey instruments were streamlined, focusing the questionnaire on the most pertinent questions and re-designing it to be equally relevant to the semi-arid districts that became part of the project at the start of ALRMP II (and the others that joined after the AF); (ii) new data storage and analysis software (REWAS III) was adopted; (iii) this in turn enabled more systematic analysis and inclusion of a broader set of non-food indicators that offer a more holistic view of the situation on the ground; (iv) the data are now stored electronically and updated monthly as information comes in from each site; and (v) EWS bulletins are disseminated immediately on the project and KFSSG websites.  New District Steering Groups were created for the17 semi-arid Districts that came into ALRMP II, and considerable capacity building investments were made to upgrade planning and investment screening capabilities. In every single individual and focus group meeting that the IL-ICR team held, participants (other donors, GoK officials at Nairobi and field level, key NGOs active in the ASALs) pointed to the DSGs as having enabled them to be more effective. The DSGs are well regarded because of their inclusiveness, efficiency in ensuring good coordination among participants, transparency and knowledge sharing, and technical skills. The plans developed by the DSGs are used as reference points for all actors and have reportedly helped to avoid wastage and duplication of efforts (not to mention reducing the burden on communities of having to explain their priorities to multiple parties).  Procedures and operational guidelines for the Drought Contingency Fund were strengthened under ALRMP II, and the EU also began channeling part of its Disaster Management Initiative financing through the Fund.  District Peace Committees. Although the first phase project had undertaken many conflict resolution activities, there was no structured arrangement for handling these and most interventions were reactive. ALRMP II created District Peace Committees in most ASAL districts, and initiated proactive work to pre-identify conflict flash points with a view to mitigating risks rather than only responding to conflicts that had already broken out. 96. The areas where ALRMP II was less successful included: (i) legal formalization of the DSG, Contingency Fund and DPC structures (the first two were legal covenants under the Credit Agreement that had not been fulfilled by project completion); and (ii) establishment of NDMA before project completion (although this has since taken place in November 2011). Delays appear to be largely the result of ministerial reorganization (the shift of the PCU from a Ministry in the Office of the President to a newly created Ministry coordinated by the Prime Minister), on the one hand, and the constitutional reform process on the other, rather than lack of GoK commitment to the objectives. 30 Formalizing district-based institutions in particular would have been difficult since constitutional reforms were moving towards replacement of districts by counties. The constitution was ratified in August 2010, and details on county arrangements are still being worked out. Meanwhile the IL-ICR team observed that the DSGs and DPCs are still functioning actively. 97. Failure to institutionalize the Drought Contingency Fund was very unfortunate because it meant that donors, who would otherwise have been willing to add resources during the 2010 drought year, could not do so because IDA had informally suspended the project and the DCF did not have a legal identity apart from the project. However, the Borrower has advised in its comments on the draft IL-ICR that since the contingency fund is expected to be one of the main operational instruments of the recently established NDMA, it anticipates Cabinet approval in the near future. Finally, as GoK is considering further changes in central government structures, including streamlining of ministries, it will be important to ensure that (i) units dealing with food aid assessments and delivery, currently in separate ministries reporting to the Offices of the President and Prime Minister, are brought together, as was the case when ALRMP II was designed; and (ii) existing and proposed institutions for drought management, disaster management and climate change adaptation are clearly delineated so that the resulting arrangements are synergistic and avoid overlapping mandates and fragmentation of effort. (c) Other Unintended Outcomes and Impacts 98. Early Warning System Model. Kenya’s EWS has attracted attention and is helping to inform thinking on early warning systems in other countries. In recent years, ALRMP II has been visited by, inter alia, Ethiopia, South Sudan and Syria, and is presently helping Uganda to establish a drought EWS in the Karamoja region. Features of the ALRMP II EWS that have been of particular interest include: (i) the incorporation of data and indicators relevant to pastoralist communities; (ii) its community-based nature with communities providing real-time information that can be acted on; (iii) its multi-sectoral nature, since the EWS looks at a broad set of inter-related food security indicators, rather than narrowly at, for example, precipitation or vegetation (coverage includes water, agriculture, livestock, changes in household nutrition and other coping strategies); and (iv) the EWS has a clear link to decision-making so that actions needed at each drought stage can be taken to prevent further deterioration. 99. Kenya’s Dialogue and Positioning on Climate Change Adaptation. Both the first phase project and ALRMP II have generated considerable knowledge about the impact of accelerating drought conditions on pastoralist and small farmer communities in the ASALs, including their coping strategies, resilience, willingness and capacity to adapt to climate change. This is quite important, considering that the ASALs include some 75 percent of Kenya’s total land area. Because ALRMP II focused not only on development work on the ground, but also on improving the policy and legal environment at national level, it has had an impact on understanding and policy formulation in a number of areas (Section 3.2(v)). In the aftermath of the 2006 Stern report on the Economics of Climate Change and the 2007 Fourth Report of the IPCC, the international community has 31 stepped up attention to adaptive needs of developing countries. ALRMP II made very useful contributions to GoK’s understanding, global dialogue and strategic positioning on the issues. Preparation of the KACCAL GEF project (Sections 2.2 and 4) by the ALRMP II team was also useful, and its implementation would contribute to the knowledge base on climate change adaptation in the ASALs. 3.6 Summary of Findings of Beneficiary Survey and/or Stakeholder Workshops 100. The IL-ICR team conducted two separate rounds of stakeholder consultations to obtain feedback and validate the team’s preliminary findings. The first round involved a wide range of development partners at the national level: (i) UN agencies (WFP, UNDP, UNICEF, FAO, OCHA); (ii) bi-lateral donors and NGOs (EU, ECHO, DANIDA, USAID/FEWSNET, OXFAM); and (iii) officials, and authorities in the Ministry for the Development of Northern Kenya & Other Arid Lands, and the Ministry of State for Special Programs (Office of the President). The second round involved stakeholder workshops in the field in three project districts with 148 participants, of which slightly more than half were from beneficiary communities, and the balance from local and international agencies working with food aid and drought issues, local officials and line/sector agencies, and local ALRMP II officials. The team also met with the beneficiaries of several randomly selected micro-projects on site in their communities. The key feedback obtained is summarized below and detailed in Annex 6. (i) UN agencies, international donors and NGOs - Nairobi. There was unanimity on the relevance of ALRMP II, and the high level of confidence in the project EWS and institutional set-up. Before the creation of DSGs, each agency had conducted drought management activities in an uncoordinated manner, often with duplication. Instead, the DSGs have provided a forum for joint discussion of local needs and harmonized and coordinated action on the ground. The Drought Contingency Fund has facilitated quicker response to emergencies. The project was considered a best practice case that should be replicated. The main concerns related to: (i) IDA’s suspension of disbursements in the midst of a serious drought, which also affected the work of others; (ii) uncertainty about a possible follow-up project to consolidate and ensure sustainability of ALRMP II gains; and (iii) the need to create a National Drought Management Authority14 and to institutionalize National Drought Contingency Fund. (ii) ALRMP II Staff and Management – Nairobi and Field. There was a high level of frustration among project staff and Ministry officials about the protracted forensic audit, IDA’s suspension of disbursements in the midst of a severe drought, and its abrupt halting of work on a new project without ever formally communicating this to counterparts. These factors have 14 NDMA was subsequently established after the IL-ICR mission and Stakeholder Workshops. See Section 3.5 above, on Institutional Change and Strengthening. 32 negatively affected GoK’s ability to deal with the 2010/11 drought. GoK has maintained key structures and staff, but activities are more limited, and this is affecting people and livestock. (iii) Communities and Other Civil Society Organizations - Field. There was consistent feedback that project structures (especially the DSGs) were beneficial, performed proactively in response to drought emergencies, and provided access to other organizations for assistance and investments that had previously been lacking. There was better focus and transparency in food aid. Training, capacity building and project investments have improved access to services, incomes, resilience and drought preparedness. Impacts on empowerment (and participation of women) and conflict resolution have been very important. Concerns included: (i) the increasing frequency of droughts; (ii) inadequate resources during the 2010/11 drought; (iii) the need for IDA to continue its support; and (iv) the need for additional focus on IGA activities, health, education and women’s groups. (iv) Line/sector Agencies and Partner Organizations - Field. These agencies echoed views on the importance of ALRMP II institutional structures. By including all key ministries, NGOs and international agencies, the DSGs have improved coordination and access to communities, with clearer identification of investment priorities. Capacity building; investments in water, agriculture (better farming practices), livestock; women’s empowerment; the Drought Contingency Fund and rapid response to emergencies were the most commonly cited interventions of importance. Creation of DPCs and cross border peace committees, training and facilitated meetings among communities had reduced conflicts over access to water and pastures. Concerns included: (i) limited resources; (ii) the need to sustain project institutional arrangements; (iii) deploy more line agency staff to supervise projects on the ground; and (iv) expand training to ensure maintenance and sustainability of micro-projects. 4. Assessment of Risk to Development Outcome: Moderate 101. The IL-ICR team considered risks to sustainability of ALRMP II achievements in three categories: (i) the EWS, institutional arrangements and inter-agency cooperation; (ii) reduction in vulnerability to drought in the ASALs; and (iii) empowerment and other social development gains. The severe drought of 2011/12 demonstrated the importance of the agenda supported by the project, and the institutions and activities established continued without IDA financing, albeit at lower levels of funding. The institutions established and supported under the project performed well for the resident population, although they were not sufficient to handle the influx of refugees from across the Somali border. IDA and other development partners remain engaged in the arid and semi-arid areas and have pledged to increase their commitment to the challenges that pastoral communities face, including the transboundary ones. The Regional Pastoral Livelihoods Recovery and Resilience Project and the Kenya Drought Recovery Project will receive 33 IDA resources. The Hunger Safety Net Program, supported by Department for International Development (DfID) and IDA, is exploring ways to channel resources for cash transfer through the NDCF. The EU and IDA have agreed to disburse EU resources through active Bank projects for the NDCF. Finally, KACCAL, originally designed as a companion to the ALRMP II project, will be restructured to link to the Kenya Agriculture Productivity and Agribusiness Project. Although the modalities will change with the end of support through ALRMP II and the changes effected under the new constitution, the commitment of support by the GOK and development partners remains high. The risk to sustainability is thus rated moderate. 102. EWS, institutional arrangements and inter-agency cooperation. Kenya’s community-based EWS, with its unique adaptation to the needs of pastoralist communities, is likely to continue. Despite the fact that ALRMP II had closed in December 2010, the IL-ICR team observed that the EWS system of data collection and preparation of the regular monthly bulletins was still taking place and playing a critically useful role in supporting the response by the KFSSG and partners to the severe drought in 2010/11. Systems depend on people, in this case the dedicated ALRMP staff at national and district levels who have gained experience in applying the 10,000 or so monthly household questionnaires, and on the confidence of communities who provide the information. The GoK retained staff during 2011, although there were several months with no salary payments. Similarly, the IL-ICR observed continuing close inter-agency cooperation during the 2010/11 drought, the result of several years of ‘practice’ during ALRMP II implementation. The GoK is still in the process of mainstreaming some critical institutional features (the Drought Contingency Fund, and others may require modification to accommodate the new constitution. The lack of resolution of questions raised during the forensic audit has had reputational implications for the personnel and institutions involved in implementation of ALRMP II, and initially raised questions about sustainability. With better understanding of the limitations of the INT findings, however, and desire of the GoK to internalize design features to reduce identified risks, the risks to sustainability that appeared high at the project’s closing subsequently attenuated. 103. Reduction in livelihoods vulnerability to droughts in the ASALs. Achievements in vulnerability reduction have been impressive, considering the frequency, intensity and duration of droughts in the ASALs since the start of ALRMP II. The share of people needing food aid in the arid districts, where the full set of project components has been operating, has been reduced and the change is statistically significant. Although gains are also noted in the semi-arid districts, where only one component has operated (and, for six districts, only since late 2006/early 2007), statistical significance could not be detected yet. But these gains are tenuous and will require continued, scaled up investment support to ensure their sustainability. A number of the CDD micro-projects also require immediate attention to define O&M responsibilities and funding. Similarly, while some income generating activity (IGA) micro-projects are sufficiently profitable to self-finance future operating costs and investment needs, others will likely be constrained by the continuing lack of micro-finance services in the ASALs. Finally, as climate change and climate variability processes make life increasingly difficult for inhabitants of Kenya’s ASALs and similar areas in neighboring countries, there is potential for increasing 34 conflict related to natural resource scarcity. ALRMP II has been effective in averting and resolving a number of conflicts by bringing together key actors at the local level and serving as a bridge to other GoK structures. As mentioned above, the DPCs need to be institutionalized, and their activities scaled up, in order to address proactively conflict risks on the scale likely to develop. Design work on all four fronts (scaled up investments, O&M, financial services, and conflict resolution) has been undertaken and is available for incorporation into new operations financed by IDA and/or other partners. 104. Empowerment and other social development gains. Some ALRMP II achievements noted in the ILRI-IE’s social networking analysis, and reported in numerous documents and during the IL-ICR team’s own consultations, are likely to be enduring. Women who are now standing up to speak in local meetings, especially those whose own income opportunities have been improved through micro-project investments, are not likely to sit down again. Communities that have developed the confidence and skills to seek out service providers rather than wait to be attended, if at all, will not lose these capacities. However, as in the case of vulnerability reduction to droughts, the gains are measurably stronger in the arid districts that have benefited from all three project components, although they do not yet reach all areas in these districts. The semi-arid communities are newer to the program, in some cases with only 2-3 years of experience. Sustaining and building on achievements to date will require not only scaled up implementation of ALRMP II’s participatory investment activities, but – equally important -- a continuing strong voice at the national policy-making level on behalf of the ASALs, supported by a sound knowledge base drawn from real-time work on the ground and informed by active local participation, increasingly robust data and good analytical capacity. 5. Assessment of Bank and Borrower Performance 5.1 Bank Performance (a) Bank Performance in Ensuring Quality at Entry: Moderately Satisfactory 105. IDA’s performance at design stage had many strong aspects: (i) the objectives had high strategic relevance for Kenya; (ii) a good effort was made to learn from the experience of the first phase project (a Beneficiary Assessment and the ICR were prepared in time to feed into the process) and from experience in the Africa Region on CDD issues; (iii) the range of skills that the design team mobilized (from IDA and FAO, and other consultants) was impressive overall; (iv) the peer reviewers were suitable; and (v) technical, safeguards and fiduciary aspects were well prepared. Country and Sector Management paid adequate attention, bearing in mind that all concerned considered that ALRMP II was building on a successful first phase operation, and therefore the task at hand was more one of refining particular features of the program, than of starting anew. 35 106. Preparation time and costs were reasonable, although the extensive FAO inputs were not captured in SAP and therefore figures in the system under-state the intensity of effort that went into the preparation and appraisal process. 107. The shortcomings of project design were: (i) the weaknesses in the original project results framework (the PDO made sense, but some PDO indicators would be difficult to measure or duplicative; 25 Intermediate Outcome Indicators were excessive; and it is perplexing why no preliminary baseline estimates were included given the knowledge the team had after implementing the first phase project); (ii) the missed opportunity to draw on relevant best practice CDD experience from outside the Africa Region which could have strengthened that component, in particular social accountability arrangements; (iii) the underestimated risk of accountability of local expenditures; and (iv) the poorly prepared financial services sub-component. (b) Quality of Supervision: Moderately Unsatisfactory 108. In assessing Quality of Supervision by IDA, the IL-ICR team took into account not only performance by the task team, but also IDA management, and other parts of the Bank Group that had an impact on IDA’s overall oversight of ALRMP II. 109. Supervision by the task team was satisfactory in many respects. (i) The skill range of the team was quite strong, considering not only the core members, but also IDA consultants and FAO staff; (ii) There was a good mix of Headquarters and Country Office-based staff, and considerable continuity of team members; (iii) Supervision missions took place regularly at roughly six-month intervals. ISRs and aide memoires were timely and informative; (iv) Environmental audits and fiduciary post-reviews took place on schedule and were followed up; (v) FM supervision identified weaknesses in the project FM arrangements and lowered the FM rating to Moderately Satisfactory in 2007 and 2008. FM also leveraged the Borrower’s country system by partnering with the Internal Audit Department (IAD) to undertake a risk-based portfolio review in January- March 2009, which included physical verification of outputs of ALRMP II on a sample basis. Internal control weaknesses were identified by this joint World Bank-IAD review, as a result of which the project was selected for further in-depth review. Due to the subsequent start of the INT forensic audit, however, this particular in-depth review was not pursued; (vi) The baseline survey was completed within one year of project approval, and the mid-term review took place almost to the day it had been scheduled in the PAD; (vii) A thorough Impact Evaluation was commissioned sufficiently well in advance of closing to permit a good quality product with lessons that could feed into a follow-up operation; (viii) Working relations with other donors and NGOs active in the ASALs were excellent; (ix) The team worked hard to strengthen M&E arrangements and used the occasion of the AF amendment to introduce a somewhat stronger set of KPIs; (x) Country and Sector management commented on most ISRs; and (xi) Most DO and IP ratings seemed appropriate to the IL-ICR team, considering the knowledge the task team and management had at the time. 36 110. There were also weaknesses. (i) ALRMP II included a very important set of activities aimed at improving the national policy and legislative framework relevant to the ASALs, but this work did not appear to receive much support through other instruments in IDA’s country program (e.g., country economic work, other ESW, policy- based lending). Although IDA did not extend policy-based lending to Kenya during the early years of the project, there were plans for an adjustment credit in support of the ERS, which was dropped, but a review of project files indicates that it would not have reflected any of the ASAL-related policy agenda; (ii) The task team and IDA management should have pressed more strongly for formalization of ALRMP II’s institutional features, in particular the Drought Management Authority (which was not formally established until November 2011, almost a year after the project closed), Drought Contingency Fund, DSGs and DPCs. The IL-ICR team understands that this might not have been successful given the turmoil surrounding the 2007 elections and the constitutional reform process, but the record does not establish that these issues received the attention that they required; (iii) The team should have provided closer supervision of the CDD component, in particular social accountability aspects; (iv) The team and management appear to have under-estimated the supervision challenges of extending ALRMP II into the semi-arid districts, nearly doubling the geographic coverage of the project area and, a few years later, also doubling the volume of financing and further expanding the project area (to some 75 percent of Kenya’s total land area). Instead of largely continuing the supervision arrangements that had worked for the first phase project, the team should have built into the design of ALRMP II provision for contracting regular technical field audits to support their own and the PCU’s oversight responsibilities; and (v) While the Bank Team identified some weaknesses in project FM arrangements in 2007, 2008, and 2009, the team should have identified these weaknesses earlier, particularly with respect to internal control systems and record keeping, and should have followed-up more systematically. No conclusions can be drawn as to whether these shortcomings materially affected the efficiency of the Bank’s operation or its ability to deliver results for beneficiaries. As discussed below in more detail, it is also not clear whether these weaknesses allowed fraud and corruption to occur, since the INT report does not prove whether and to what extent there was fraud and corruption under the project. 111. Managing the project during the INT field work as well as following INT’s initial verbal and written reports of its findings was challenging, given the long period over which INT conducted its work (January 2009 through June 2011), the seriousness of INT’s initially reported findings, and the Bank’s institutional difficulties in taking action on the basis of informal (e.g., non-public, unofficial and unwritten) information. Based on an INT verbal briefing in July 2010 that as much as 70 percent of the expenditures reviewed in its ongoing audit work were ineligible, Bank management decided immediately to suspend IDA disbursements to ALRMP II informally (as it did not have a formal basis for suspension), effectively bringing project activities financed by the World Bank to a halt. Because the suspension was informal, it was not processed through the Legal Department as per OP/BP 13.40, although the Board received a verbal briefing of the suspension the day after it was put into effect (in an informal briefing on Kenya that had previously been scheduled). The project remained informally suspended through the final six months of its life, and closed as scheduled in December 2010 in that status. 37 Findings of the INT forensic audit were disclosed to the GoK for the first time in April 2011, about 10 months after INT had provided its preliminary verbal report to Bank management and after management had taken its precautionary action of informal suspension. By that time, a drought-related emergency was fully under way in the Horn of Africa. Although ALRMP II was the single most important program directly related to the drought management agenda in the ASALs, the findings from INT’s forensic audit had not yet been confirmed by IAD or acted upon by the project, and Bank management therefore considered that it was not yet able to resume financing of ALRMP II institutions to respond to the emergency, as it could not provide adequate assurance that project financing would be used for the purposes intended. While the GoK attempted to shift some of its own resources to fill the ensuing financing gap, especially for the Drought Contingency Fund, this involved months of delay. The Bank’s informal suspension in effect reduced IDA financing to the project by about US$5.7 million at a critical time in the project’s history, and it also reduced funding from other sources. The informal suspension also precluded use of the ALRMP II as a vehicle for quick conveyance of additional financing to the affected area during the peak of the drought. In particular US$5.9 million of EU drought financing could not be transferred to Kenya as the EU had been using the ALRMP II Drought Contingency Fund for disbursement15 and was constrained by IDA’s decision to suspend ALRMP II informally; and a US$5.5 million IDA-managed GEF grant, which would have relied on ALRMP II institutions for its implementation, was not made effective. Later in FY11, IDA management also suspended work on a successor project (expected to be a SWAp) that the GoK and its development partners had been assuming would be prepared, pending a final resolution of the issues that the INT forensic audit had brought to light. Despite these developments, a few other donors, such as WFP, remained engaged with ALRMP II institutions, in recognition of the then ongoing drought crisis and they continue to express confidence in the ALRMP institutional arrangements (see Section 3.6 on stakeholder feedback to the IL-ICR team). 112. INT’s conduct in the course of the ALRMP II forensic audit involved shortcomings that affected project implementation. In particular: (i) INT’s initial decision, in early 2009, to make the audit an INT-led exercise, rather than a collaborative exercise involving the GoK’s IAD, meant that INT findings when ultimately provided to the GoK could not be immediately addressed by the GoK, as the GoK (through IAD) needed to validate the findings, thus delaying the GoK’s response; (ii) INT’s report, while strongly suggesting that fraud and corruption had in fact occurred, did not prove their occurrence 15 The objective of the EU Drought Management Initiative Trust Fund (TF093909) was to contribute to the improved effectiveness and efficiency of the drought management system in Kenya by strengthening the capacity of all actors to intervene in an appropriate, effective and timely fashion during the drought cycle. The activities supported risk prevention and livelihood improvement through drought mitigation, response and recovery activities based on drought contingency plans and triggered by early warning systems. The resources were disbursed through the ALRMP Drought Contingency Fund. The TF effectiveness date was September 29, 2007 and it closed on December 31, 2010 when the IDA credit closed. The total amount of the TF was US$11.56 million (Euros 8.5 million), of which US$5.68 million (Euros 4.16 million) was disbursed. Disbursements were stopped in July 2010 when the IDA credit was informally suspended. The EU had no alternative way to disburse the remaining Euros 4.84 million (US$5.9 million equivalent). 38 (it found indicators of possible fraud or corruption that needed to be verified through an investigation); and (iii) INT’s initial verbal estimate of the extent of ineligible and questionable expenditures, on which the decision to suspend was based, was approximately twice the amount contained in the subsequent written report, which was further revised thereafter, raising questions about the rigor of the methodology employed. 113. The INT forensic audit also involved another failure on the part of the Bank. OP 13.40 was not observed when Bank management informally suspended. Compliance with OP 13.40 would have required review and clearance by the Legal Department that a case for suspension was justified under the terms of the Credit Agreement; formal notification to the Board; formal consultation and notice to any affected Cofinanciers (in this case, the EU); and formal notification to the Borrower with opportunity to respond. Because INT in July 2010 had provided a verbal rather than a written preview of its findings, management was unable to provide the Legal Department with a documented basis for formal suspension. But because management also felt that the INT findings were alarming and that it could not fail to act on them, however preliminary, it decided to implement an informal suspension, pending the availability of documentation to provide the basis for a formal suspension. In addition, to ensure that OP 13.40 was observed in spirit to the greatest extent possible, the Board was verbally notified of the informal suspension one day after it was decided; and the GoK and affected Cofinanciers were verbally informed of INT’s findings and the management decision to suspend the project within a few weeks of the informal suspension. The lack of a documentary basis for the decisions taken impeded the ability of management, the GoK and the Cofinanciers to discuss the issues raised or indeed to respond to any of the INT findings, putting a serious strain on all project-related discussions, decisions and activities. 114. Given the severity of the drought, the centrality of the project in addressing it, and the paucity of other instruments, redoubled effort to complete the forensic audit and clear up methodological questions would have been in order. (c) Justification of Rating for Overall Bank Performance: Moderately Unsatisfactory 115. IDA performance has been strong in many respects, and this has contributed to the successful achievement of most outcome indicators of ALRMP II. However, there were some deficiencies in design, supervision and important shortcomings in the Bank Group’s handling of the final phase of ALRMP II, including the forensic audit. 5.2 Borrower Performance (a) Government Performance: Moderately Satisfactory 116. Borrower commitment to ALRMP II was strong during project design and has remained so throughout implementation. Counterpart funding has generally been adequate and timely. There was good continuity of key PCU managers and staff, including staff in the field who were well respected for their technical competence (although turnover of METs due to salary and other issues has been more problematic). 39 Inter-Ministerial coordination at district level has been very constructive. Despite the unexpected suspension of IDA financing for ALRMP II and of work on a follow-up project, GoK has attempted to keep most staff in place and to continue project activities, especially those related to drought management, albeit on a lesser scale. The main weakness in GoK performance was its failure to establish the proposed NDMA during the life of the project (although this has since been achieved in November 2011), and to formalize other important ALRMP II institutional structures (DSGs, DPCs and the Drought Contingency Fund). Also, decisions made in the course of ministerial reorganizations have resulted in a questionable separation of the units responsible for estimating food aid needs from those carrying out food aid and drought management activities (between Ministries assigned in the Office of the President and the Office of the Prime Minister). This situation is working better than might otherwise be expected, because of the good personal relationships and competence of the staff and managers of both units, but it is not ideal. (b) Implementing Agency or Agencies Performance: Moderately Satisfactory 117. The PCU has a thin and efficient structure (Headquarters and field combined). It has had good continuity in management and core staffing, and it has managed to attract and retain some excellent staff at field level. The working culture is proactive and problem-solving, and the PCU’s ability to encourage individuals from across agencies to work together collaboratively has been impressive. When alerted by the World Bank in early 2007 to reports of fraud and corruption in one of the districts, the PCU responded quickly and appropriately to that specific instance, and made changes in handling of the transactions found to be of highest risk. Staff implicated in wrong-doing were dismissed. Thus the PCU established a record of rapid response to problems when information was conveyed to them. Throughout the period of the forensic audit, the project had no concrete information to which to respond, and the existence of the forensic audit thus does not represent evidence of lack of responsiveness or proactivity. Despite the difficulties related to suspension of IDA financing, the PCU prepared a thorough Borrower’s ICR and cooperated with IDA’s efforts to prepare its own ICR. Areas where PCU performance could have been stronger are similar to some of the points raised in relation to IDA (Section 5.2(b)). In particular, the sizeable expansion of project area at the start of ALRMP II, and again at the time of the AF, should have been occasions for the PCU to rethink some organizational issues in order to oversee effectively the much scaled up set of responsibilities. Apart from the possibility of contracting more frequent technical field site reviews to obtain feedback on the many dispersed activities that ALRMP II supports, the PCU might also have reinforced its M&E related staffing. Also, while the PCU’s commitment to empowering communities is deep and this theme cuts across all ALRMP II activities, not just the CDD component, the PCU could have detected earlier the need to clarify decision-making responsibilities and accountabilities at community and district level, and ensure adequate O&M arrangements. The Project Paper of the AF for ALRMP II and a letter from the GoK dated June 26, 2006, list specific disclosure requirement for the project. While the majority of the required documents have been disclosed, some of the District Annual Progress Reports, District Annual Work Plans, and Printed Estimate (at Headquarters Level) have not been disclosed on the project’s website and through other media. 40 (c) Justification of Rating for Overall Borrower Performance: Moderately Satisfactory 118. The successful outcome of ALRMP II is largely a credit to the strong support the project has received from GoK, and the dedication and technical competence of national and district PCU management and staffing. Sustaining those gains will require, inter alia, actions to address some issues that were not resolved during implementation, in particular formalization of some core ALRMP II institutional structures. 6. Lessons Learned 119. Adaptation to longer-term climate change can be pursued jointly with management of short-term emergencies, but requires explicit attention. ALRMP II experienced the same pattern of needing to refocus efforts on drought management, recovery and protection of livelihoods versus longer-term adaptation, as had occurred under the first phase project. In both cases this occurred because of prolonged, severe drought in the Horn of Africa. Given advances in scientific knowledge on climate change and climate variability, it appears that more frequent – though still unpredictable – and severe droughts are likely to be a constant feature of life in that region. This suggests that future projects in the ASALs need to include, from the start, very strong drought early warning, management and recovery components. They need to be large enough and with flexible design features to be able to respond quickly, and on a scale commensurate with drought severity. However, the projects also need to include robust components that facilitate sustainable asset creation/income improvements and long-term out-migration from fragile areas, and these components should not become victim to ‘unanticipated’ severe droughts. That is, the pattern of shifting efforts from strengthening the long-term adaptive response capabilities of the ASAL population, towards (under- dimensioned) drought management/short-term coping/recovery activities, should be avoided. 120. Conflict resolution should be integrated into natural resource management projects in fragile areas experiencing climate change stress. The first phase of ALRMP did not include conflict resolution, but it very quickly became necessary and the project responded creatively. With this experience, ALRMP II designers planned to integrate conflict management into the natural resource management component. Beneficiaries and other stakeholders point to the resulting activities as having been among the project’s most important contributions. Addressing conflicts over natural resources needs to be an integral part of projects in fragile locations that are experiencing in-migration and/or are over-populated (in relation to the resource base) and subject to increasingly adverse climate change impacts. 121. Putting in place flexible institutional arrangements that development partners can use to channel support quickly, before climate emergencies occur, can help to improve coordination and efficiency. Climate-related emergencies require good early warning systems, contingency planning in which the affected population has 41 had a say, and contingency resources that can be accessed quickly and in a well coordinated manner by stakeholders who can provide assistance. A mechanism like the ALRMP II Drought Contingency Fund can be very useful and IDA should consider including such arrangements in projects for countries that are highly vulnerable to climate change and climate variability risks. In structuring the funds, advance thought should be given to legal, financial and reporting features that will make it easy for other donors to come in very quickly if needed, and avoid losing scarce time in bureaucratic negotiations. 122. Lessons from Kenya’s Drought Early Warning System may be useful for other countries. The effectiveness of the ALRMP II Early Warning System provides lessons for countries/regions, particularly in the Africa region, in designing or improving their own EWS. These include: (a) its incorporation of data and indicators relevant to pastoralist communities; (b) its community-based nature with communities providing real-time information that can be acted on; (c) its multi-sectoral nature - looking at a broad set of inter-related food security indicators; and (d) the institutional set up, particularly at the field level, that integrates EWS and other relevant information available and involves and coordinates with all relevant stakeholders engaged in preparedness, emergency relief/management and recovery. 123. Climate change adaptation requires time, putting communities in the drivers’ seat, and sound risk minimization arrangements. Development programs in areas subject to frequent droughts and that target highly vulnerable populations, as in the ASALs, require long enough implementation periods (through one, or a sequence of, projects) to have a meaningful impact in strengthening coping skills, creating resilience and undertaking sustainable adaptive responses. Involvement of communities in the process is essential. This implies decentralization of responsibilities for implementation, recognizing the reality that the project will be operating in a fragile environment and therefore creating systems conducive to reducing risks, with good management information systems that early detection of problems and corrective measures. 124. CDD projects need strong social accountability arrangements to minimize corruption risks and strengthen local governance. Community-driven development projects, or components, through the emphasis on beneficiary participation and empowerment, can be instruments for reducing the risks of corruption and poor governance, especially at the local level, provided that social accountability arrangements are properly designed and implemented by the Borrower and closely supervised. There is a particular need to design workable disclosure and complaints handling mechanisms, assign responsibilities for implementing them, and ensure that this takes place. Selected measures (e.g., timely posting of updated micro-project implementation and financing on signboards and on the web, report formats that enable easy comparisons of sub-projects across communities and districts, etc.) need to be included in the MIS and then specific project and community members tasked with implementation and monitoring. The functioning of these measures should also be a key focus of the CDD training, third-party monitoring, financial audits, and of IDA supervision. Finally, it may also be possible to consider more holistic, portfolio-level social accountability mechanisms supported by 42 reputable civil society organizations. Such an approach can help to enhance public reporting and complaints handling of financial information. 125. The approach to fiduciary supervision should be strategic; technology can play a role. On the basis of the Arid Lands experience, there is need to work strategically, both at the country and project levels. This includes further strengthening of country fiduciary agencies, particularly IAD Treasury and KENAO for audit of CDD and decentralized projects. Consideration should also be made of utilizing fiduciary agents to monitor CDD projects where the risk is deemed to be high. There is also scope for integrating modern technology into oversight of CDD projects. Poor rural communities will often not be able to maintain sophisticated documentation on micro- projects they carry out, however a combination of simple records (i.e., receipts, notes on decisions made on community meetings, quotations for purchases) and proper use of the technology by the PCU (e.g., use of low cost portable scanners and cameras with Global Positioning System (GPS) to record information in situ and upload to the project management information system (MIS) during field visits) could go a long way towards strengthening oversight without imposing excessive controls from above. 126. Implementation arrangements need to allow climate-stressed communities ‘time out’ to cope with emergencies. For projects targeting highly vulnerable, climate- stressed populations, implementation plans need to be flexible and take account of the fact that during peak emergency periods (e.g., severe drought) many households will temporarily be unable to focus on community development, while they are dealing with the day-to-day challenges of survival. This will require special consideration of how to handle community financing, procurement and other implementation arrangements in a manner that allows beneficiaries ‘time out’ to deal with emergencies for which there may be little advance notice, not to lose work already started, and to be able to return to the longer-term adaptation agenda (e.g., micro-projects) when crises subside. 127. Task teams need stronger M&E specialist support. Task teams and counterparts often struggle with assessment of outcomes at completion because baselines, results frameworks and/or impact evaluations were not well designed. In the case of ALRMP II, a very large amount of data was recorded (more than 10,000 household surveys/month for the EWS system; nutritional records for over 600,000 children). The task team made several efforts to strengthen evaluation; and it has been possible for the IL-ICR to make a final assessment of results that meets standards of significance and attribution. However, it should have been less complicated, and it should also have been possible to generate more knowledge. Given the importance of proper evaluation, the region/Bank should consider developing a small group of M&E specialists with proven practical experience in the design of simple, measurable results and risk frameworks and indicators, who can advise teams at design, mid-term and impact evaluation stages. With regard to monitoring, a strong MIS (i.e., one that links sub-project performance and finances, provides meaningful reports on project investments across communities and districts, and enables regular disclosure of project activities at community, district/county, and national levels in user-friendly formats) is critical for effective project management, and also for ensuring transparency and accountability downward to communities and 43 citizens. The need for accessible information on all aspects of ALRMP II was demonstrated during implementation when the project came under questioning, and struggled to address the critics. 128. A methodological note on evaluating results where climate events need to be normalized across years would be helpful. Monitoring trends in drought vulnerability requires a clear methodology for normalizing the severity of these climate events across years. The IL-ICR team found this a challenge, as the Impact Evaluation had used a vegetation-based methodology (NDVI), while other specialists recommended precipitation-based approaches. It may be that alternative methods are appropriate, within the same project, for different kinds of interventions. Given the likelihood that many task teams will face similar issues in the context of accelerating climate change, on the one hand, and the importance of being able to evaluate results properly over time within and across projects, on the other, it would be important for the Agriculture and Rural Development Department (ARD) to develop clear guidance for all teams to use on appropriate methods for normalizing drought severity for the most common interventions supported by Bank projects. The same may well be true for other climate change phenomena (e.g., floods, erosion). 129. Fraud and corruption investigations need to be conducted in a manner that promotes Borrower ownership. When fraud and corruption is suspected or alleged, there is no question that the Bank and Borrower need to act quickly and decisively to investigate and take appropriate measures. The Bank has learned from long experience in other areas (e.g., economic and sector work (ESW), program and project design) that ownership is improved when clients participate. The experience of ALRMP II shows that a closer working relationship between INT and Kenya’s IAD from the start – a model that was followed in the forensic audits initiated in early 2009 for the Western Kenya Community-Driven Development project (WKCDD) as well as the Kenya Education Sector Support project (KESSP) – might have led to an agreement on approach as well as much earlier findings and recommendations (the WKCDD and KESSP audits were completed in 2009, and because IAD was involved in the process from the beginning, the government was able to respond to the findings and recommendations immediately, including by suspending staff involved in fraud and corruption). When the Bank decides to undertake an investigation without involving the client, it needs to do so as expeditiously as possible – particularly in the case of projects that are designed to address emergencies, and when such emergencies are already occurring and human lives are at stake. A protracted process in the case of a project that is designed to address emergency conditions not only limits the ability of the Bank and other donors to provide timely support; it also may undermine the credibility of those in the country who are working on the program and reduce their ability to be effective. Finally, if the process is not perceived as fair or transparent, and in accordance with the Bank’s own fiduciary guidelines, accepted international standards and government legislation and regulations, the result is unlikely to achieve its objectives of improving governance. 130. World Bank Group operational policy should be followed when invoking suspension remedies, but some additional guidance may be needed. When cause for 44 suspension is clear and documented, it is important that procedures be followed, i.e., OP 13.40. However, cases may arise in which continuation of disbursement is considered to be unacceptably risky, and yet the documentary information required to adhere to OP 13.40 is not available. The experience of ALRMP II shows that additional guidance is required to manage such situations. Special effort should be undertaken quickly to gather the documentary information needed to allow either formal suspension under OP 13.40 or full resumption of disbursement. In an emergency situation, limited continued disbursement with special fiduciary oversight may be warranted. All effort should be made by INT and the Regions to assure that projects do not go into an indeterminate status of informal suspension for a prolonged period of time. Informal suspensions do not allow for careful review by the Legal Department of evidence justifying the remedy; there is no formal communication to a Borrower detailing the reasons for suspension and what, if any, actions can be taken to correct the situation; the Board and co-financiers are not formally advised; the Bank does not hold itself to any clear timetable for follow-up; and the process generally lacks the transparency that should accompany such serious actions. 7. Comments on Issues Raised by Borrower/Implementing Agencies/Partner (a) Borrower/implementing agencies 131. The Borrower’s Ministry of Finance submitted comments on the draft IL-ICR to IDA on December 20, 2011 (see Annex 6). These largely concurred with the IL-ICR findings, particularly those concerning achievement of ALRMP II development objectives and most KPIs, difficulties posed by M&E issues, and the impact of the forensic audit process in the final stages of ALRMP II and on the Government’s relationship with IDA. GoK also identified a few differences of opinion on some of the IL-ICR fiduciary findings, summarized steps it has taken to strengthen risk assessment and governance arrangements over the life of ALRMP II, and provided an update on its recent establishment of the long-proposed NDMA in November 2011 after the IL-ICR team had visited Kenya. IDA has corrected factual points, one of which resulted in upgrading of achievements for Intermediate Outcome Indicator (IOI) 11 and reflected the status of the new NDMA in the final IL-ICR. GoK did not question other ratings, and the change in IOI 11 does not affect the overall assessment of project results, risks, IDA and Borrower performance. (b) Co-financier 132. The Delegation of the European Union to Kenya submitted comments on the draft IL-ICR to IDA on December 15, 2011 (see Annex 7). EU comments largely concurred with the draft IL-ICR findings. They emphasized the impacts of IDA’s suspension on ALRMP II funding and interruption of support to sustain and scale up results jointly with other development partners, at a time when Kenya faced a very severe drought. The EU Delegation also updated information on its own continuing institutional support to the Ministry for Development of Northern Kenya and Other Arid Lands and on the establishment of the NDMA in November 2011, and indicated the Delegation’s openness 45 to consider a renewed partnership with IDA on drought management activities in Kenya, taking into account the final conclusions of the forensic audit. 46 Annex 1. Project Costs and Financing (a) Project Cost by Component (in US$ million equivalent) Arid Lands Resource Management Project Phase Two - P078058 and P091979 Actual/Latest Appraisal Estimate Percentage of Components Estimate (US$ millions) Appraisal (US$ millions) Natural Resources and Drought 99.70 115.99 116 Management Community-Driven 28.60 24.20 87 Development Support to Local Development 16.60 11.81 70 Total Project Cost 144.90 152.00 105 (Physical Contingencies) (1.20) (Price Contingencies) (6.60) (b) Financing Arid Lands Resource Management Project Phase Two - P078058 and P091979 Appraisal Actual/Latest Type of Percentage of Source of Funds Estimate Estimate Financing Appraisal (US$ millions) (US$ millions) Borrower Budget allocations 19.85 28.35 142 Cash, materials & Local Communities 5.05 5.96 118 labor IDA Credit 120.00 118.39 99 Total 144.90 152.70 105 47 Annex 2. Outputs by Component 1. Unless otherwise indicated, the information below draws on data from the ILRI- IE of 2010, the Borrower’s ICR of September 2011, Project Files and IL-ICR team findings. Component 1: Natural Resource Management and Drought Management Link to Output/Intermediate Outcome Indicators (IOIs). The natural resource and drought management component was linked to seven IOIs. (i) District staff and a minimum of 600 communities trained in participatory natural resource management (IOI-1: Achieved); (ii) Key potential conflict flashpoints identified in each district and conflict management initiatives strengthened (IOI-2: Achieved); (iii) Kenya Food Security Meeting (KFSM) and 28 District Steering Groups institutionalized (although the structure has been established and was functioning very well before the end of the project, because it has not been institutionalized and this poses a threat to sustainability, the IL-ICR team considers this IOI-3: Partially Achieved); (iv) Drought Early Warning System monthly bulletins produced and disseminated in the project area and at national level on a regular and timely basis (IOI-4: Achieved); (v) District Long Term Drought Preparedness and Annual Work Plans developed in each district in the project area (IOI-5: Achieved); (vi) District Drought Contingency Plans developed in each district in the project area and disseminated to stakeholders (IOI 6: Achieved); and (vii) Drought Contingency Fund and associated funding mechanism established (IOI 7: Partially Achieved). Link to PDO Indicators. Component 1 activities were linked to achievement of three PDO indicators: (i) decreasing the proportion of people in the ASAL districts assessed as needing free food aid, normalized by severity of drought (KPI-1: Achieved); (ii) reducing time lapse between reported drought stress and response (KPI-2: Achieved); and (iii) improving nutritional status of children below 5 years of age over time normalized by severity of drought (KPI 3: Achieved). The natural resource management sub-component activities also contributed to increasing the number of people with access to basic services, including water, human and animal health services, and education (KPI-4: Achieved). 2. This component had two sub-components: (i) Natural Resource Management; and (ii) Drought Preparedness and Management. It operated in both arid and semi-arid districts, and was the largest ALRMP II component in terms of and resources allocated and spent. Actual component expenditures were US$115.99 million. Expected Outputs/Results 3. Sub-component (i): The objective of the Natural Resource Management sub- component was to establish and strengthen initiatives to reduce the livelihoods 48 vulnerability of ASAL communities to the effects of drought through the sustainable use and management of natural resources. The strategy was to promote a better understanding of the natural resource base, thereby leading to improved planning and implementation of coordinated activities that would make the most beneficial and sustainable long term use of the natural assets found in the ASALs, in the context of low and erratic rainfall compounded by other external uncertainties. 4. Expected outputs included: (i) training of district staff and communities in participatory NRM and Drought Preparedness; (ii) development of a Vision and Strategy for ASAL NRM; (iii) generation and dissemination of better information and awareness of NRM; (iv) preparation of discrete development area/livelihood zone NRM plans with financial allocations for drought preparedness prioritized; (v) preparation of community level NRM plans through a Participatory Rural Appraisal/Participatory Learning in Action process; and (vi) peace building and conflict management activities. 5. Actual outputs included: (i) 717 NRM/environmental awareness training sessions at district and community levels and including all 28 district field offices and development partners; (ii) Natural Resource Management Plans were prepared for all 28 districts and were used for identification, prioritization and implementation of NRM micro-projects. The vision and strategies emerging from these plans provided the basic framework for design of the ASAL Development Policy and the District and National Environmental Action Plans; (iii) ALRMP II invested in 1,341 drought preparedness micro-projects; (iv) Community level NRM plans constituted the major part of the 614 community development plans for the CDD component; and (v) 214 conflict management meetings and workshops were held. This sub-component also financed independent annual environmental audits with the recommendations shared among the different project actors, and developed environmental screening tools for micro-projects and inter-community projects. 6. Improving conflict the management of conflicts that have their origin in the increasingly fragile ASAL natural resource base (especially water, pasture and cattle) is crucial for any development effort. ALRMP II therefore sought to integrate conflict management into NRM, with some significant achievements. District Peace Committees have been established and strengthened in all districts. Unlike other peace initiatives that have been driven by higher-level actors, ALRMP II targeted assistance at the local level, involving key community leaders, using a combination of customary and formal institutions and through the CDCs provided a link with district level authorities. Inter- community NRM planning has also helped to reduce conflict. The project brought together communities to agree on how to use pasture and water resources, especially in the case of drought. These created ties and relationships among the parties that make cattle rustling less likely and created an environment in which reconciliation of former hostile communities became possible. 7. Nonetheless, while ALRMP II has piloted a number of catchment conservation, rangeland management and other NRM/environmental conservation activities, achieving an overall impact in the ASAL districts will require considerable scaling up. The 49 National and District NRM visions and strategies have contributed to a more holistic understanding of needs, but implementation is a challenge, and more focused and action- oriented planning tools are needed. Similarly, ALRMP II has contributed greatly to the understanding of issues related to climate change, and it has already enhanced the capacity of major stakeholders in the ASALs to cope with climate variability and change. However, a more comprehensive and explicit approach to climate risk management, and adequate financing to apply the knowledge learned towards increasing climate resilience and implementing climate-smart investments is now needed. 8. Sub-component (ii): The objective of this sub-component was to create a more effective drought management system that would minimize the need for emergency operations, on the one hand, and enhance response mechanisms during acute drought emergencies on the other. This required: (i) building institutional and technical capability at district, regional and national level to respond effectively to drought; (ii) developing and maintaining working relationships with target communities and district-based agencies through the implementation of drought related activities during “normal� periods, in order to create a readiness for quick response at the onset of emergencies; and (iii) preventing drought from degenerating into famine by implementing appropriate drought management interventions and initiating a rapid post-drought recovery and a return to “normal life� as soon as possible. The main interventions supported by ALRMP II towards this end are detailed below. 9. The EWS that had been developed under the first phase of ALRMP was enhanced (in terms of the data collected and analytical processes) and expanded (from 11 to 28 districts), and it played a central role in informing decisions during the main drought operations of 2004/05 and 2008/09 (and intervening less severe years), including forecasting drought severity, needs identification and targeting of relief interventions. The IL-ICR team received feedback from all major national and international development partners involved in drought preparedness and drought relief on the reliability of the reports and how that has improved during ALRMP II implementation. There were important increases in the frequency, timeliness and scope of the Drought Monitoring Bulletins from all districts. The backbone of the ALRMP community based EWS is the monthly collection of field data from local communities at specific sentinel sites, which are sample locations within administrative sub-locations, representative of all main types of livelihoods in the ASAL area. Each of the 28 ASAL districts obtains EWS data from 360 to 450 households each month (about 10,000 to 12,600 households surveyed monthly); the data is analyzed at district level and used at district and national levels. 10. Early warning information was provided to all stakeholders to ensure timely response to drought stress. A total of 2,965 activities were undertaken, including production of 2,336 EWS bulletins, as well as rapid food security assessments and community feedback meetings in all 28 districts. The provision of timely and reliable information enabled stakeholders to make appropriate responses to prevent or mitigate the effects of drought and other disasters. According to ILRI-IE, about 78 percent of national and international institutions involved with drought preparedness and drought 50 relief in 2009 used the project EWS as their main source of information to plan their activities/interventions in the area (up from 32 percent in 2001). 11. The Kenyan Food Security Coordination Structure at the National and District levels includes: (i) the KFSM as the main coordinating body that brings together food security actors in a forum where information is exchanged, options debated and decisions on activities formulated for referral to the Government of Kenya and donors; (ii) the KFSSG, which as the technical “think tank� and advisory body to all relevant stakeholders on issues of drought management and food security; holds monthly meetings to discuss emerging issues, and produces monthly updates that summarize pertinent issues from a series of sources (on weather, rainfall, crop and livestock production, and food aid. The updates rely heavily on the 28 ALRMP bulletins. KFSSG also produces bi-annual assessments on food security and drought status; and (iii) the DSGs that are the equivalent of KFSM at the district level. 12. Each of the 28 ALRMP districts has a DSG that performs food security and drought management coordination work, among other activities. DSGs provide the mechanism to channel information and decisions upwards to the appropriate government bodies and donors, but are also critical for coordinating action at the district level. DSGs play a central role in coordinating needs assessments, particularly during the biannual rainfall assessments and other occasional food security assessments. Members of DSGs include the relevant food security government ministries, local government representatives, as well as UN, NGOs, Community-Based Organizations (CBOs), faith based organizations and local community leaders of each district. The technical arms of the DSGs are made up of relevant line ministries and relevant agencies. The DSGs are chaired by the District Commissioners, with the ALRMP Drought Management Officers acting as secretariat to the DSG. While the project has provided support to the whole Kenyan Food Security, it is at the field level (at the level of the DSGs) created by ALRMP, that the project had its greatest impact. The ILRI-IE conducted a survey among national and international agencies involved with drought preparedness, management and relief, and concluded that it was a great appreciation for the role of the role of the DSGs and the ALMRP II project in terms of coordination and helping to increase the efficiency of the work of those agencies in the field. The IL-ICR team received similar feedback from UN agencies, bi-lateral donors, NGOs and line agencies both during interviews in Nairobi and in the three workshops it held in the field. This feedback singles out the role of the DSGs and ALRMP II as the most important component for coordination of all donors, NGOs and line agencies activities in the field resulting in avoiding duplications, knowledge of where to direct their efforts once they got to the field in a way that best fitted their own comparative advantages, and an improvement in overall response and effectives to droughts. 13. The ALRMP project financed the preparation of District long-term Drought Preparedness and annual Work Plans, as well as District Contingency Plans for the 28 project districts. Numerous training activities were also undertaken to build the capacities of all stakeholders to undertake drought preparedness and emergency interventions. The 51 ALRMP II Contingency Fund was often the first respondent once drought emergencies were declared. 14. Drought preparedness investments. According to the Borrower’s ICR, 2,934 drought preparedness investments were made to improve water availability and access for human and livestock use, including: excavation of water pans; pipeline development; drilling, equipping and rehabilitation of boreholes; shallow well development and equipping with hand and rotary pumps; provision of water tanks; construction of rock catchments, masonry tanks, underground water tanks and ferro-cement tanks, construction of pump houses, water kiosks and cattle troughs; water treatment works; construction of sand dams, as well as rehabilitation of existing water structures, hydrological surveys and solar panel kits. In agriculture, activities included mainly irrigation; on farm storage structures; crop seed/planting materials; seed banks and seed multiplication; and water and soil conservation structures. In the livestock sector, ALRMP II financed some 4,000 activities to improve the availability and accessibility of livestock feeds and water, livestock disease prevention, surveillance and control, improvement of livestock breeds and livestock marketing. To improve access to health care, four investments were made in the district of Tharaka (which lacks adequate health facilities); three other projects investments supported nutritional surveys and information system in Mandera to provide people with reliable health and nutrition information. Main livestock activities included pasture production and fodder establishment helping to maintain body condition of livestock left behind and making milk available to families during drought periods; purchase of improved breeds of cattle, goats and chicken, to improve genetic potential of pastoralist herd, improve productivity, maturity rate, milk production and income from livestock sales; purchase of camels as drought tolerant species, particularly to improve milk availability which is in low supply during droughts; sale yards to facilitate marketing of livestock and hence increased income for pastoralist; other livestock activities included vaccination campaigns; water trucking; purchase of incubators. In order to introduce alternative income, modern bee-keeping micro-projects were undertaken and honey processing equipment purchased, benefiting 100,768 persons. Replacement of traditional hives increased honey production from 20 kgs per hive to over 80 kgs. Other alternative source of income introduced by the project included fish farming in the districts of Kitui, Kajiado and Nyeri. The IL-ICR team visited one of these fish-farming projects in Kitui, which was obtaining very good results. 15. Drought emergency activities related to water, livestock, health and nutrition, agriculture, and coordination. In order to improve water availability and access to clean water during drought periods a total 2,069 water activities were undertaken in all districts, except Nyeri, Tharaka and Transmara. This benefited a large number of households by reducing distances to water points, and providing a constant supply of water to humans and livestock during drought. Water related activities included support to rapid response teams to sustain continuous functioning of boreholes during long operation hours ensuring constant supply of clean water to communities during drought thus preventing morbidity and mortality of humans and livestock; water trucking; drilling and rehabilitation of boreholes; purchase of plastic and collapsible water tanks; purchase of submersible pumps and hand pumps; provision of fuel subsidies, and rehabilitation of 52 other water supply structures among others. In the livestock sector emergency livestock response activities were undertaken during acute drought periods. A total of 7,454 emergency response interventions were undertaken, including emergency livestock off- take which provided pastoralist with an avenue to sell off emaciated stock hence reducing potential loss, selling at markup price higher that offered at local market level; livestock disease control and surveillance and treatment, specially with vaccination campaigns for notifiable diseases reducing morbidity and mortality and allow markets to operate without closures and households to maintain flow of income during most needed periods; livestock supplementary feed provision to maintain livestock body condition and reduce livestock death, especially of the few milking stock left behind while main stock migrates to areas with better pastures. A total of 223 health and nutrition emergency responses included rapid response to disease outbreak such a malaria, cholera, Kalazaar, diarrhea, measles and aflatoxicosis. Nutritional activities included purchase of supplementary feeds for vulnerable groups (under five children and pregnant and lactating women), and nutritional surveys/assessment support to Integrated Management of Acute Malnutrition. Other activities in this sector including purchases of vaccines, material for laboratories, hospital equipment, etc. Emergency health and nutrition activities were funded using the Drought Contingency Fund. Some 107 agriculture related emergency activities were undertaken in order to enhance food security and increase food production. In terms of coordination, they entail the monthly meetings of the District Steering Groups, but also monitoring and evaluation of project activities, DSG tours and visit to the project areas, and activities allowing for effective project planning and implementation activities through collaboration. 16. Areas for improvement in the results framework. The framework relies on changes in assessed food aid needs as a proxy for changes in livelihoods vulnerability. The assumed relationship is reasonable, although other factors could influence food needs, such as changes in remittances or other safety net support to beneficiaries. The IL-ICR team is persuaded that ALRMP II was far and away the main program in the ASALs; and there was little, if any, safety net support from other sources – i.e., such factors could not have caused changes in food aid needs, which are more likely the result of changes in vulnerability. However, the framework would be stronger if it attempted to measure changes in economic activities and NRM practices on the ground (e.g., extent of beneficiary adoption of more drought resistant crops/cultivation/animal management techniques) as intermediate outcome indicators. 53 Component 2: Community Driven Development Link to Output/IOIs. Component 2 was most directly related to four IOIs: (i) Mobile Extension Teams and line ministries core staff in specified arid land districts trained to conduct modified Participatory Rural Appraisal (PRA) and to provide backstopping to communities to participate in CDD (IOI-8: Achieved); (ii) effective screening by specific DSGs routinely undertaken for community plans and project proposals for technical, environmental and financial soundness (IOI-9: Achieved); (iii) 600 Community Action Plans prepared and micro-projects implemented by communities (IOI-10: Achieved); and (iv) at least 1,000 groups assisted in developing savings capacity through top-up savings grants (IOI-11: Achieved). Link to PDO Indicators: The CDD component has contributed most directly to one PDO indicator: increased number of people with access to basic services – PDO4: Achieved). However, it has also played an important role in helping to achieve three other PDO indicators: (i) reduction in the number of people assessed as needing food aid (PDO 1: Achieved); (ii) improved nutritional status of children under 5 years of age (PDO 3: Achieved); and (iii) increased people’s participation in the project districts (PDO 5: Partially Achieved). ILRI-IE findings on the statistical significance of outcomes in the arid districts (where the CDD component operated) and the semi-arid districts (where it did not) underscore the likely importance to PDOs 1, 4 and 5 of employing a participatory CDD strategy. 17. The component had three sub-components: (i) Support to CDD implementation; (ii) Community Capacity Building; and (iii) Capacity Building for Backstopping Services. Actual component expenditures were about US$24.9 million, including US$19.23 million for micro-projects (of which US$5.96 million financed by communities and US$13.27 million by IDA), and capacity building and backstopping services of US$5.67 million. The component was implemented only in the 11 arid districts. Several improvements were introduced, in comparison with the CDD component of the first phase of ALRMP. The main changes were to move to a two-stage CDD process, involving ‘basic CDD’ for new communities and ‘extended CDD’ for those with CDD experience, shifting greater responsibility for procurement to communities, and increasing the community contribution for income generating projects, including a greater cash contribution. 18. Expected outputs/results. ALRMP II worked with 614 communities (331 ‘basic’ and 283 ‘extended’ communities) in the 11 arid districts. It supported 2,477 training activities for about 150,000 people (52 percent women), in Participatory Integrated Community Development (PICD), PICD updates or retraining, Finance and Community Procurement, training on Cross-cutting Issues (e.g., gender, proposal writing, resource mobilization, HIV/Aids, environmental management, conflict resolution and subjects identified during the needs assessment carried out by ALRMP II field officers). According to the Borrower’s ICR and project files, the 614 communities have implemented about 3,000 micro projects, benefiting a total of some 1.9 million people (55 percent women). The most common community micro-projects included water, 54 agriculture, education, health and sanitation sectors. Livestock, access roads and other diverse income-generating activities were also financed. 19. While these outputs relate to investments under Component 2, it is also the case that some similar small investments took place under Components 1 and 3, for which communities participated in the sense of identifying priorities but did not take the lead in managing implementation, nor were they required to make local contributions as they did for the CDD micro-projects. This occasionally created confusion at the local level, and in any future project it would be preferable to have a clearer delineation of the kinds of investments that require beneficiary contributions, whether contributions could vary depending on the nature of the investment or relative poverty of the group. Although communities did meet, and sometimes exceed, their share of counterpart funding for Component 2 micro-projects (30 percent), this did pose hardship and/or delay the start of investments in some poorer locations. 20. ALRMP II design was largely consistent with good practice lessons drawn from other CDD projects in the Africa region at the time of appraisal in the early 2000s. Project design sought to respond to risks and issues identified in the evaluation of ALRMP I including: (i) increased community participation and capacity through sustained training; (ii) increased focus on conflict reduction; and (iii) increased focus on inclusion of marginalized groups. However the design could also have benefited from a broader look at lessons from outside the region (Indonesia, India, Bangladesh, and Brazil). Nor did revisions at the time of the AF in 2006 fully take on board emerging good practices at that stage. Some intended enhancements do not appear to have been fully incorporated into project manuals, trainings, M&E framework, budgets and supervision. Most importantly, although ALRMP II and the AF emphasized increasing community participation, introducing social accountability, reducing conflict, and increasing inclusion of marginalized groups, relevant indicators were either focused on direct project outputs or were not included. The AF also eliminated some indicators related to community participation and accountability that had been part of the original ALRMP II log frame. This may have resulted in uneven attention to these issues in project implementation, supervision and monitoring. 21. There is evidence, based on stakeholder workshops and focus groups, that communities improved their abilities to articulate needs, plan and manage local projects, and to allocate limited resources and lobby for external support. Reviews carried out by the project, Kinmetrica, Mpereeza, IDA, and during ICR preparation noted several achievements: (i) increased capacity of men and women in remote areas, including indigenous and marginalized communities, to engage in development discussions and actions; (ii) CDD related structures played a central role in training credible leaders (civic and Parliamentary leaders, provincial administration chiefs and sub-chiefs); (iii) numerous CDCs succeeded in accessing and managing external resources under their Community Action Plans; (iv) many PCU and district project staff demonstrated high levels of motivation and commitment, and established strong relationships and trust with the communities; and (v) DSGs formed under the project are perceived by government, communities, and external partners to have played an important coordinating role. 55 22. These reviews also noted challenges in the implementation and supervision of the CDD model, including: (i) limited clarity of project rules at community level; (ii) limited resources for and focus on sustaining wider community involvement beyond the PICD; (iii) lack of distinction between approaches for public and private goods; (iv) issues with checks and controls and uneven implementation and supervision of social accountability measures; (v) weak M&E and learning systems; and (vi) inadequate field facilitation and implementation support to communities. Also, as noted in supervision reports, the performance of METs, comprising officers from various line departments and NGOs, was affected by high turnover, expenditures on fuel and per diem for METs, and limited accountability of MET staff because they continued to report to their respective line departments. 23. The PAD, PIP, and CDD manual include recommendations on community involvement and social accountability measures, including geo-referencing and mapping of all ALRMP II investments, more systematic public disclosure of project activities and financial information at district and community levels, and a revamped communications strategy. However, while some measures were done for the whole project (e.g., geo- referencing), it appears that others were not as systematically implemented. At an aggregate level, project reports did not enable ready reconciliation of project activities with financing. For example, until the geo-referencing was completed later in the project period, reporting formats did not enable one to see a regularly updated and comprehensive list or map of all CDD micro-projects, arranged by community and district; a view of all of the different project activities (NRM, Drought Contingency Fund, CDD, Support for Local Development) in a given community or district, and the financing for each. It should be noted that the absence of comprehensive geo-referenced data for all interventions has been a common problem not confined to Kenya, and project teams in all countries have been working actively to correct this. 24. With the benefit of hindsight and improvements in CDD methodology, the initial design included several shortcomings that affected project implementation. These included: (i) a strong emphasis on an initial participatory CDD identification and prioritization exercise (the PICD), but limited guidelines and training to sustain broader community participation beyond the PICD; (ii) lack of clarity on project rules, and limited understanding of community members of project rules; (iii) lack of clarity on how project resources were allocated for various sub-projects; and (iv) lack of robust monitoring indicators to track progress on the CDD component. In retrospect, another design shortcoming was that social accountability measures related to transparency and community monitoring were not well reflected in the financial management, procurement, and M&E guidelines of the project, and it does not appear that they were implemented uniformly by GoK (see below) or supervised very closely by IDA. 25. In response to IL-ICR team questions, a quick survey of Drought Management Officers, detailed district-by-district social accountability practices for all local ALRMP II investments (i.e., not only CDD micro-projects but also those financed under Components 1 and 3) regarding branding (use of placards to identify the investment as 56 having been financed under ALRMP) and public display of work plans and budgets. Branding was fairly widespread, though not uniform: 6 districts have 100 percent of projects branded; another 9 have 90-99 percent branded; 5 have 80-89 percent; 7 have 60- 79 percent; and only one has less than 60 percent. All but 1 district have public disclosure notice boards, though usually at the ALRMP district office rather than at community level. 26. The Ministry of Finance has recently taken the initiative to develop a new policy to ensure effective governance structures and good financial management and procurement systems and guidelines for devolved funds at national, county, facility and community levels. The policy would also define a framework for management of community funds through NGOs, community-based organizations and other mechanisms, and develop robust social accountability mechanisms including community participation in governance, financial management and procurement of devolved CDD resources, public reporting, complaints handling and recourse mechanisms. A joint government task team has been established and IDA is cooperating closely, and will reflect results in CDD manuals for ongoing projects and in arrangements for a follow up to ALRMP II. IDA has also put together a multi-unit Task Team comprising of Africa Regional specialists in social development, rural development, economic management, financial management and procurement, to coordinate its work on this topic, and is exploring possible south-south technical assistance that would draw on experiences from other regions and countries where CDD social accountability arrangements are more developed and working well. 27. Areas for improvement in the results framework. The framework would have benefited from some indicators that would measure changes in participation/empowerment (along the lines employed by the ILRI-IE; other examples could be drawn from good practice CDD operations elsewhere), with particular attention to gender and marginalized groups. 57 Component 3: Support to Local Development Link with Output/IOIs. The five IOIs most directly influenced by Component 3 were: (i) National Pastoral Policy formulated through participatory process (IOI-12: Achieved); (ii) mobile schools and nomadic education centers established in 3 project districts (IOI-13: Achieved); (iii) at least 600 communities reached with awareness raising program on human health and HIV/AIDS (IOI-14: Achieved); (iv) guidelines for emergency livestock off-take prepared and articulated in all project districts (IOI-15: Achieved); and (v) options for improved financial services delivery considered and implemented where possible within the project area (IOI- 16: Not Achieved). Link with PDO and Indicators. Component 3 was most directly relevant to two PDO Indicators: (i) increased number of people with access to basic services (PDO4: Achieved); and (ii) increased people’s participation in the project districts in local and national development as demonstrated by the reflection of arid lands concerns in the economic recovery strategy and other relevant national policies (PDO5: Partially Achieved). Together with other components, it also indirectly supported two other PDO indicators: (i) reduction in the number of people assessed as needing food aid (PDO 1: Achieved); and (ii) improved nutritional status of children under 5 years of age (PDO 3: Achieved). 28. This component had three sub-components: (i) Policy, Advocacy and Research; (ii) Local Services Development; and (ii) Piloting Financial Services. Actual component expenditures were US$11.81 million. The component was implemented only in the 11 arid districts (except for the policy, advocacy and research activities, which were not area specific). 29. Expected outputs/results of the above three sub-components were to (i) establish a policy framework for local level development; (ii) strengthen institutions and organizations for local development; and enhance existing sector delivery capacity in water, education, human & animal health, marketing, agricultural research & advisory services; and (iii) pilot a viable approach to deliver rural financial services in the arid lands and explore an expansion strategy. 30. Sub-component (i): ALRMP II sought to increase the voice of the ASAL population in decision-making on issues that affected their well-being and livelihoods. The project was involved in the formulation of nine draft policies of direct relevance to the ASALs. This was part of ALRMP II’s policy and advocacy strategy to influence Government commitment by strengthening relevant parts of the national policy framework. The Economic Recovery Strategy for Wealth and Employment Creation (Kenya’s PRSP, 2004) includes a full chapter influenced by the ALRMP II PCU and project beneficiary communities at the national and local levels. Among other policies and sessional papers to which ALRMP II provided important inputs are the National Policy for Sustainable Development of Arid and Semiarid Lands of Kenya, National Land Policy, National Disaster Policy, Policy Framework for Nomadic Education in 58 Kenya, National Policy on Land Reclamation, National Policy on Peace Building and Conflict Management, National Food Security and Nutrition Policy, National Livestock Policy, and National Irrigation and Drainage Policy. On the other hand, some of these policies were still in draft form at the time of the IL-ICR team, which poses risks for sustainability of the achievements if they are not formally adopted. On the other hand, ALRMP II advocacy on behalf of ASAL needs led to the creation of the Ministry of State for the Development of Northern Kenya and Other Arid Lands (April 2008) and to the recent establishment of the new National Drought Management Authority (November 2011) after project completion. Kenya’s Vision 2030 (successor to the Economic Recovery Strategy) also reflects ASAL concerns, and the experience and knowledge generated by ALRMP II has contributed to the formulation of Kenya’s official thinking on climate change adaptation. 31. In the area of research, the project engaged several consultancies to study issues that had been problematic under the ALRMP I and/or on which guidance needed to be given to DSGs and DCUs during the second phase: gender, mobile education, livestock survey, emergency livestock off-take, veterinary drugs delivery system, livestock marketing, livestock disease free zones, agricultural services, and river water user associations. ALRMP II also engaged a consultant to analyze the contribution of the ASALs to the national economy in the areas of livestock and livestock products, agriculture, tourism, energy, gums, resins, mineral extraction, bio-diversity and world heritage sites, among others. Much of this research was of a practical nature (obtaining information on the ground, exploring intervention models that could be adapted to the ASALs, etc.). Quality varied but for the most part the consultancies contributed positively to ALRMP’s own investment activities or to its policy advice and advocacy. 32. Sub-component (ii): ALRMP II sought to increase the availability and quality of services, by empowering ASAL communities to better articulate their needs to service providers. To assess achievement of this community empowerment objective, the ILRI- IE conducted a social network analysis and found that there had been a 54.2 percent increase in the communities’ ability to seek services from service providers rather than wait for them to be received. The change was greater in arid districts, where the increase was 95.2 percent between 2004 and 2009. 33. ILRI-IE also found that community interactions and bonds with service providers are stronger now than in 2004. Overall, 44 percent of the arid districts recorded an improvement in service provider closeness to the community. Communities also shared their perceptions of the change in quality across all services by attributing scores to each, rating the quality of health and education services to be the most improved. 34. ALRMP II investments under this sub-component included water (which supported livelihood and food security objectives, human and animal health, and reduced time spent by families – particularly women – to seek water); education (including class rooms, mobile schools, dormitories, kitchen and dining halls, and sanitary facilities in primary schools, training and education stakeholders’ forums/meetings); human health (infrastructure and training on public health, HIV/AIDS, traditional birth attendants, 59 female genital mutilation, drug abuse); livestock (breed improvement, fodder production and milk processing); fish farming, beekeeping and various disease control activities; and Disease Control activities. 35. The ILRI IE found that treatment (project) communities had better access to quality water sources; primary, secondary and adult education; and to veterinary medicines than at the start of ALRMP II. The percent of households consulting medical professionals, using bed nets, and seeking veterinary extension services also increased, and negative health-related outcomes (prevalence of child diarrhea and livestock mortality) decreased significantly. The IE also noted that this may understate project impact as improvements also occurred in control communities due to good inter-agency coordination and use of DSG plans prepared for the project), and because ALRMP II advocacy on behalf of the ASALS helped to improve service delivery in both intervention and control communities. On distance to water, the IL-ICR team estimates that distance to boreholes decreased from 6.4 to 4.3 kms, to pans and dams from 3.8 to 3.0 kms, and to shallow wells from 5.2 to 2.9 kms in the 28 arid and semi-arid districts, but was not able to determine attribution, although the project did make a significant number of water-related investments. 36. Sub-component (iii): Activities in piloting financial services were very limited during the project period. ALRMP II financed a micro-finance workshop and explored the scope for supporting financial services associations and village banks in three districts. However, village banks, although viable for limited savings, had a low capital base that made them generally unable to engage in lending activities. Preparation of the sub- component did not build on adequate knowledge of the micro-finance sector; objectives were only vaguely stated; and attention during implementation was more limited, both by GoK and IDA. While the lack of success in developing financial services during the project period did not keep ALRMP II from achieving its overall PDO, this should be an important agenda item for future projects directed towards the ASALs. 37. Areas for improvement in the results framework: The fact that ALRMP II institutional arrangements (DSGs, DPCs, Drought Contingency Fund) served a wide array of donors, GoK line agencies, NGOs and other actors in the ASALs was emphasized strongly in every single meeting the IL-ICR team held, both in Nairobi and in the field. Given the strength of this qualitative feedback, the IL-ICR concludes that this was a case of best practice inter-agency coordination, on one hand, and that because of this achievement; on the other hand, ALRMP II’s indirect impacts have probably been quite strong. It would have been useful if the results framework had included indicators to: (i) assess the extent and effectiveness of this coordination; and (ii) measure the incremental investments in the ASALs made by others because they were able to use ALRMP II institutions and planning activities (e.g., PICDs). 60 Annex 3. Economic and Financial Analysis 1. The PAD does not provide an Ex-ante Economic Analysis for the entire project. The reason given is that the micro-projects have heterogeneous impacts that depend on several factors that cannot be measured easily and therefore it is not possible to calculate an overall Net Present Value (NPV) of Economic Rate of Return (ERR). However, the PAD assesses the financial viability of several individual activities (irrigation, rain-fed agriculture, petty trade, livestock trade and bee keeping) and finds that these were all financially viable. 2. This annex quantifies the benefits that result from two sets of mechanisms that can be attributed to the project. First, the project financed micro-projects investments. These investments led to increases in income as a result of new or increased sales of pure and mixed-breed livestock and livestock products, milk, crops and fish. Second, the project invested resources in the water and livestock sectors (e.g., livestock restocking, veterinary services and improved access to animal water sources). These investments resulted in (counterfactual) gains in livestock wealth. As a consequence of the fact that livelihoods in the project areas depend on livestock - both as a source of milk that is a valuable nutrient and an important source of income, and in addition, as an instrument that aids in mitigating income shocks – the (counterfactual) gains in livestock wealth are most closely related to the project’s development objective. 3. The project results framework as revised in the AF amendment did not include any indicators on returns to micro-project investments, and therefore the ILRI IE did not cover this subject. The ICR mission therefore conducted Cost-Benefit Analyses for 4 randomly selected micro-projects during its visits to the field in connection with the stakeholder workshops. The micro-projects were selected from a roster of geo-referenced project investments provided by the implementing agency. They include dairy breeding, fishponds, water investments in sand dams and irrigation. In the aggregate, these activities account for 65.5 percent of the total number micro-projects and 43.4 percent of the value of investment in all micro-projects supported by ALRMP II. 4. The discounted net benefit streams of the four micro-projects analyzed are summarized in Table 3.1. Table 1: Benefit Cost Ratio, ERRs and Net Present Value of Micro-Projects Analyzed by IL-ICR Mission Micro-Project Benefit-Cost Economic Rate of Net Present Ratio Return Value (KES) Dairy breeding 2.1 78.8 1.040 m Fish ponds 1.1 24.0 20.296 m Sand dams 1.2 38.0 2.491 m Irrigation 1.24 55.5 14.800 m 61 5. To arrive at an ERR for all micro-projects, the above returns were weighted in proportion to overall project expenditures, yielding an estimated overall ERR on micro- project investments of 37 percent. The mission assumed a discount rate of 15 percent, and that project overheads and administrative costs were equivalent to 30 percent of the costs of micro-investments. 6. To evaluate gains in livestock wealth the ICR had access to historical weather station rainfall estimates in Garissa. Figure 1 describes historical changes in livestock as well as cumulative rainfall in Garissa. Figure 1: Historical Changes in Livestock and Rainfall in Garissa 16 400000 6000 Drought (1987) Cumulative Precipitation (ml) Total Livestock (Garissa) 300000 5000 Drought (1992-94) Drought (1999/00) Severe Drought Conditions 200000 4000 100000 3000 2000 0 1985 1990 1995 2000 2005 2010 Year Livestock (Tropical Livestock Units) Precipitation (ml)) 16 Sources: National Oceanic and Atmospheric Administration (NOAA), Kenya Ministry of Livestock. Rainfall estimates represent a weighted average of the current year and two lagged years with weights given as (0.4, 0.4, and 0.2). Missing months of rainfall are interpolated using a regression with monthly dummies and values of non-missing months. From 1985 through 2009, 19 monthly observations were missing (8 observations for 1999). 62 7. The relationship between cumulative precipitation and livestock is given in Table 3.2. Table 2: Relationship between cumulative precipitation and livestock Drought % Change in % Change in Rainfall (ml) Livestock Livestock (TLU) : Actual (TLU) : Counterfactual 1987/8 -18.3 -18.3 2,357 1993/4 -39.2 -39.2 2,828 1999/2000 -26.8 -26.8 3,028 2005/6 -7.1 -18.3 2,238 2008/9 7.0 -18.3 2,117 Median (1985- 1.7 0.2 3,576 2009) 8. The period after 2005 was characterized by the most severe drought, yet had increases in livestock. Therefore this ICR makes the conservative assumption that, absent project interventions, livestock in Garissa would have fallen by 18 percent in 2005 and 2008. This corresponds to the smallest declines in livestock across previous droughts. Figure 2 describes the counterfactual declines in livestock as a result of project investments in the livestock and water sectors. Figure 2: Observed and Counter-factual Changes in Livestock in Garissa (1985- 2009) 400000 300000 200000 100000 1985 1990 1995 2000 2005 2010 Year Actual TLU : Garissa Counterfactual TLU : Garissa 63 9. To be able to extrapolate effects in Garissa across all districts it is worth checking whether other districts also witnessed decreases in rainfall and increases in livestock during the project period. Figure 3 shows that most districts that received large investments in the water and livestock sectors also experienced declines in rainfall and increases in livestock. Figure 3: The Relationship between changes in Livestock and Precipitation across Districts17 50 samburu ijara moyale west pokot taveta taita turkana mwingi lamu garissa makueni kwale isiolo kilifi wajir meru marsabit north kajiado narok mbeere machakos malindi baringo 0 kitui laikipia river tana mandera -50 -20 0 20 40 60 % Change in Total Precipitation : (2004/8 and 2000/3) Smaller Investment in Livestock and Water Larger Investment in Livestock and Water 10. The ICR did not have access to data for prior drought years across all project districts. However, based on data available during the project years, it is likely that Garissa is broadly representative of other project districts on metrics related to changes in precipitation and livestock. Therefore in order to estimate gains in livestock wealth the ICR assumes (conservatively) that there was a 10 percent (counterfactual) increase in livestock due to project investments and includes the value of those gains in estimating the economic benefits that can be attributed to project investments. 18 The gains from 17 Data sources: National Oceanic and Atmospheric Administration (NOAA), Kenya Ministry of Livestock and ALRMP expenditure data. 18 The price of one Tropical Livestock Units (TLU) is calculated based on 2004 prices of Cattle and Goats from FEWSNET. The TLU conversion is done using weights given by the United Nations (Food and Agricultural Organization). 64 investments in the livestock and water sectors were estimated to have an ERR of 28.2 percent. 11. Based on the analysis above, the ICR estimates the project’s overall Economic Rate of Return to be 31.1 percent. This assumes that 20 percent of the funds were misused and as a consequence the randomly sampled projects overestimate the total project benefits. It is also worth examining the project returns under alternative scenarios related to the fraction of funds misused. Table 3 below summarized the results of this exercise. Table 3: The Economic Rate of Return under Alternative Assumptions of Elite Capture Funds Misused (%) ERR (%) 0 40.4 10 38.5 20 31.1 30 26.2 40 21 50 15.6 60 9.7 12. Therefore, the ICR concludes that the project was economically viable and the costs incurred are justified, even under conservative assumptions. 65 Details by Micro-Project Activity Activity I: Income Generating Activity- Dairy Breeding 1.1 Assumptions (i) Discount Rate is assumed to be 15 percent; (ii) The Opportunity cost of the time to get water is valued at the prevailing wage rate (150 KES per day); (iii) The /yearly maintenance cost are assumed to be 20 percent of the cost of the initial fixed cost of setting up the shed and other facilities. 1.2 The incremental revenues from four benefits streams are considered – sale of pure and mixed breeds, sale of milk and the value of milk for home production. Revenues with the Project 2008 2009 2010 2011 2012 2013 2014 Sale of Pure Breeds Quantity 0.0 5 7 8 9 10 12 Price (KES per goat) 0.0 10000 10000 10000 10000 10000 10000 Revenue per person 0.0 Total Revenue 0.0 50000 70000 80000 90000 100000 120000 Sale of Mixed Breeds Quantity 0.0 10 12 14 16 19 22 Price 0.0 50 50 50 50 50 50 Revenue per person 0.0 Total Revenue 0.0 500 600 700 800 950 1100 Sale of Milk Quantity 0.0 22500 22500 22500 22500 22500 22500 Price 0.0 0 0 0 0 0 0 Revenue per person 0.0 Total Revenue 0.0 1500 1500 1500 1500 1500 1500 Milk Own Consumption Quantity 0.0 22500 22500 22500 22500 22500 22500 Price 0.0 20 20 20 20 20 20 Revenue per person 0.0 Total Revenue 0.0 450000 450000 450000 450000 450000 450000 Dropping Used as Fertilizer 0.0 Total Revenue With the Project 0.0 502000 522100 532200 542300 552450 572600 Incremental Benefits From The -133.3 501867 521967 532067 542167 552317 572467 Project Present Value of Gross Project -133.3 436406 394682 349842 309986 274599 247493 Benefits 66 1.3 The micro-project costs include ALRPM investments as well as operating costs incurred by the beneficiaries. Costs With the Project 2008 2009 2010 2011 2012 2013 2014 Initial ALRMP2 Investment 349451 Inputs Time to Clean Time to get Water 52500 52500 52500 52500 52500 52500 52500 Vet Costs 3000 3000 3000 3000 3000 3000 3000 Feed Costs 3000 3000 3000 3000 3000 3000 3000 Maintenance Costs Sheds 69890 69890 69890 69890 69890 69890 69890 Other Costs of Other Projects Total Costs With the Project 478841 129390 129390 129390 129390 129390 129390 Incremental Costs Due to the 478811 129360 129360 129360 129360 129360 129360 Project Present Value of Incremental 478811 112487 97815 85056 73962 64315 55926 Costs 1.4 This project was economically viable. The Net Present Value was 1.04 million KES. The Benefit Costs ratio was 2.1, and the ERR was 78.8. Activity 2: Income Generating Activity - Fish Pond 2.1 Assumptions (i) The discount rate is assumed to be 15 percent (ii) Total project costs were estimated at 6 times the ALRMP2 investment. (iii) Water costs assumed water needs for 10 months of the year. 2.2 This is a new type of activity. Therefore, there were no revenues before the project. There were two sources of revenues – Tilapia sales and Fingerlings sales. 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Revenues without the Project Total 0 0 0 0 0 0 0 0 0 0 0 Revenues with the Project Tilapia 67 Quantity 600 600 600 600 600 600 600 600 600 600 Price 50 50 50 50 50 50 50 50 50 50 Revenue 30000 30000 30000 30000 30000 30000 30000 30000 30000 30000 Fingerlings Quantity 1000 1000 1000 1000 1000 1000 1000 1000 1000 1000 Price 10 10 10 10 10 10 10 10 10 10 Revenue 10000 10000 10000 10000 10000 10000 10000 10000 10000 10000 Gross 40000 40000 40000 40000 40000 40000 40000 40000 40000 40000 Project Benefits Present 40000 34782.61 30245.75 26300.65 22870.13 19887.07 17293.1 15037.48 13076.07 11370.5 Value of Gross Project Benefits 2.3 Total operating costs including project investments and complementary investments as well as operating costs. Costs 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 Initial ALRMP2 19,167 0 0 0 0 0 0 0 0 0 Investment Inputs Water 20000 20000 20000 20000 20000 20000 20000 20000 20000 20000 Fish Feed 3200 3200 3200 3200 3200 3200 3200 3200 3200 3200 Costs of Other 57500 0 0 0 0 0 0 0 0 0 Projects Gross Project Costs 99867 23200 23200 23200 23200 23200 23200 23200 23200 23200 Present Value of Total 99867 20174 17543 15254 13265 11535 10030 8722 7584 6595 Associated Costs 2.4. This project was economically viable. The Net Present Value was 20.296 KES. The Benefit Costs ratio was 1.1, and the ERR was 24 percent. Activity 3: Water Investments – Sand Dams 3.1 Assumptions (i) The discount rate is 15 percent; (ii) The revenues without the project are calculated based on the prevailing daily wage (150 KES) from subsistence activities; (iii) The opportunity cost of labor used was valued for the activity was valued at the same rate; (iv) Village marketers (typically women) received about 10 percent of gross revenues through sales of fruits and vegetables produced as a result of the activity. 68 3.2 Revenue gains accrue from the production and sale of 5 vegetables (Kale, Tomatoes, Onions, Spinach and Green Peppers). 2007 2008 2009 2010 2011 2012 2013 Revenues without the Project Total (Subsistence Income for 70 1512000 1512000 1512000 1512000 1512000 1512000 1512000 members) Revenues with the Project Kale Quantity (seedlings) 8000 8000 8000 8000 8000 8000 Price (KES per seedling) 80 80 80 80 80 80 Total 640000 640000 640000 640000 640000 640000 Benefits from Other Projects Tomatoes Quantity (seedlings) 80,000 80,000 80,000 80,000 80,000 80,000 Price (KES per seedling) 60 60 60 60 60 60 Total 4800000 4800000 4800000 4800000 4800000 4800000 Benefits from Other Projects Onions Quantity (kg) 500 500 500 500 500 500 Price (KES per kg) 70 70 70 70 70 70 Total 35000 35000 35000 35000 35000 35000 Benefits from Other Projects Spinach Quantity (seedlings) 20,000 20,000 20,000 20,000 20,000 20,000 Price (KES per seedling) 100 100 100 100 100 100 Total 2000000 2000000 2000000 2000000 2000000 2000000 Benefits from Other Projects Green Peppers Quantity (seedlings) 10,000 10,000 10,000 10,000 10,000 10,000 Price (KES per seedling) 150 150 150 150 150 150 Total 1500000 1500000 1500000 1500000 1500000 1500000 Benefits from Other Projects Gross Project Benefits 7463000 7463000 7463000 7463000 7463000 7463000 Incremental Project Benefits -1512000 5951000 5951000 5951000 5951000 5951000 5951000 3.3 The total costs from the activity included ALRMP II investments, land rental costs, costs of labor as well as the costs of agricultural inputs. 3.4 This project was economically viable. The Net Present Value was 2.491 million KES. The Benefit Costs ratio was 1.2. The Financial Rate of Return was 33.2 percent. 69 This was less than the Economic rate of Return of 38 percent because of positive spillovers in the village (to the marketing agents). Activity 4: Agricultural Investments – Irrigation 4.1 Assumptions (i) The discount rate was 15 percent; (ii) The subsistence income from a pastoral livelihood was assumed to be 3,000 (KES) annually. This is based on information from interviews with beneficiaries; (iii) Given the proximity to a major town (i.e., Garissa) and the fact that some households had houses there, the opportunity cost of labor was valued a higher wage rate (5,000 KES p.a.). 4.2 There were 3 main sources of revenue – sale of bananas, mangoes and vegetables. 2007 2008 2009 2010 2011 2012 Revenues without the Project Total (Subsistence Income for 32 1152000 1152000 115200 115200 1152000 115200 members) 0 0 0 Revenues with the Project Bananas Quantity (kg) 480000 480000 480000 480000 480000 Price (KES per kg) 15 15 15 15 15 Total 7200000 720000 720000 7200000 720000 0 0 0 Benefits from Other Projects Mangoes Total Revenue 7680000 768000 768000 7680000 768000 0 0 0 Benefits from Other Projects Vegetables Total Revenue 900000 900000 900000 900000 900000 Benefits from Other Projects Gross Project Benefits 1578000 157800 157800 15780000 157800 0 00 00 00 Incremental Project Benefits 1462800 146280 146280 14628000 146280 0 00 00 00 Present Value of Incremental Project Benefits 1272000 110608 961814 8363606. 727270 0 70 7 5 1 70 4.3 The main costs of this activity included the ALRMP II investments, the costs of an irrigation pump, costs of government extension services as well as operating costs (land, labor, fertilizer and pesticides). 4.4 This project was economically viable. The Net Present Value was 14.8 million KES. The Benefit Costs ratio was 1.24, and the ERR was 55.5 percent. 71 Annex 4. Bank Lending and Implementation Support/Supervision Processes (a) Task Team members Names Title Unit PREPARATION/ APPRAISAL Angostini. Astrid* Natural Resource Management Economist FAO Bagnall-Oakeley, Hugh* Consultant, Agronomist FAO Beekhuis, Geert* Economist, Financial Analyst FAO AFTR1/AFTS2 Cornelius, Christine E. *** Lead Operations Officer (TTL) De Haan, Cees* Peer Reviewer RDV Gachukia, Catherine** Operations Officer AFTS2 Githagui, Nyambura*** Senior Social Development Specialist AFMKE Fowler, Michael* Disbursement Officer LOAG2 Fox, John* Consultant, PIM FAO Hickson, Robert* Consultant, Microfinance and Enterprise Specialist FAO Kaguamba, Richard* Environmental Specialist AFMKE Karanja, Andrew Mwihia** Agricultural Economist AFTS2 Consultant, Natural Resource Management Lewis, Jeffrey G.* FAO Specialist Mansour, Tanya* Consultant, Environmental Management Framework FAO Mearns, Robin* Peer Reviewer EASDR Nawaz, Mohammad* Legal Counsel LEGAF Nyaga, John* Financial Management Specialist AFMKE Salmen, Lawrence F.* Peer Reviewer SDV Sharif, Mohamed Taki** Senior Financial Analyst AFTS1 Smith, Doug* Consultant, Environmental Management Framework FAO Tawonezv, Patrick* Consultant, Livestock Specialist FAO Wairagu, Michael* Consultant, Environmental Management Framework FAO Wales, Michael* Senior Agricultural Economist FAO Warsame, Dahir Elmi*** Senior Procurement Specialist AFMKE/AFTPC Wasike, Moses** Senior Financial Management Specialist AFTFM Watkins, Benjamin* Consultant, Drought Early Warning FAO/WFP * Original Project ** Additional Finance Only SUPERVISION/IL-ICR Amuguni, Henry Amena** Sr. Financial Management Specialist AFTFM Angostini, Astrid Natural Resource Management Economist FAO Ayres, Wendy Schreiber Sr. Economist, Monitoring and Evaluation AFTUW Belle, Arati Environmental/Natural Resource Economist AFTEN 72 Chopra, Tania Social Development, Conflict Management LEGJR Coirolo, Luis* Consultant, Lead Rural Development Specialist AFTAR AFTS2, AFTAR Cornelius, Christine E. Consultant, Lead Operations Officer 1st TTL (thru 6/30/10) Enghoff, Martin Natural Resource Management Specialist FAO Finch, Christopher* Senior Social Development Specialist AFTCS Fitwi, Efrem** Procurement Specialist AFTPC Forman, Stephane Livestock Specialist AFTAR Fye, Serigne Omar Consultant, Environment Specialist AFTS1 Gilgan-Hunt, Edeltraut Consultant, Environmental Specialist AFTAR Githagui, Nyambura Senior Social Development Specialist AFTCS Karanja, Andew Mwihia Senior Agricultural Economist AFTAR Kipuri, Naomi Ntatai Consultant, Social Scientist AFTCS Kshirsagar, Varun Consultant, Economist AFTAR Legesse, Assaye Sr. Agriculture Economist AFTAR Lewis, Jeffrey G. Consultant, Natural Resource Management AFTS2-HIS Mistiaen, Johan A. Senior Economist/Statistician DECDG Mitchell, Paul D. Development Communications Specialist EXTOC Moeller, Markus Consultant, Water Management AFTAR Moens, Marc Livestock Specialist FAO Mollard, Ingrid Marie Pierre Consultant, Research Analyst AFTAR Mozammel, Masud Senior Communications Officer EXTOC Nikiema, Emmanuel Y. Senior Natural Resources Management Specialist AFTEN Muchemi, Julius Githinji* Consultant, Environmental Specialist AFTAR Mwikya, Jame Matheka Temporary AFCE2 Munyori, Joel Buku Procurement Specialist AFTPC Ndwiga, David Ireri Temporary AFCE2 Okuny, Michael Consultant, Financial Management Specialist AFTFM Owiyo, Tom Mboya Consultant AFCE2 Pfeiffer, Hermann Agricultural Economist FAO Seevinck, Julia Economist FAO Setlur, Banu Operations Analyst, Environmental Safeguards MNSSD Sperling, Frank Environmental Specialist AFTEN Steel, William F. Consultant, Financial Services and Microfinance AFTAR Stephens, Tim Rural Infrastructure Specialist FAO Uquillas-Rodas, Jorge Consultant, Senior Sociologist AFTQK Warsame, Dahir Elmi ** Senior Procurement Specialist AFTPC Wasike, Moses Sabuni Sr Financial Management Specialist OPCFM Watkins, Benjamin Peter Consultant AFCE2 Williams, Melissa Operations Officer SASDA AFTAR Woelcke, Johannes** Senior Economist 2nd TTL (since 7/1/10) * IL-ICR only ** Supervision and IL-ICR 73 (b) Staff Time and Cost Staff Time and Cost (Bank Budget Only) Stage of Project Cycle US$ Thousands Lending 1/ 2/ No. of staff weeks (including travel and consultant costs) FY03 36.18 95,892 FY04 0 18 Total: 36.18 95,911 Supervision/IL-ICR 2/ 3/ FY04 6.06 42,973 FY05 35.68 130,869 FY06 22.41 110,355 FY07 14.47 115,797 FY08 20.41 77,303 FY09 28.10 144,106 FY10 36.50 212,233 FY11 21.18 129,321 FY12 2.30 44,840 Total: 1,007,797 187.11 Grand Total 223.29 1,103,708 1/ Some LEN costs of ALRMP II may have been included in costs of missions for SPN and ICR of the first phase project (ALRMP), which was still under implementation at the time. Excludes time and costs of FAO staff. 2/ Excludes time and costs of FAO staff and consultants. 3/ In the Africa Region, Safeguards staff charge project codes when they are members of a team, but all review work is charged to a central Internal Order mapped to the Operations Services Department, and therefore staff time and costs are not reflected in this table. Also, as of FY10, Procurement and FM also have their own Work Program Agreements and do not charge individual project codes, so the staff time and costs for FY10-12 are not included in this table. 74 Annex 5. Stakeholder Workshop Report and Results 1. The IL-ICR team conducted two separate rounds of stakeholder consultations, both to obtain feedback and to validate the team’s preliminary findings. The first round involved individual and focus group meetings with a wide range of development partners at the national level, mostly those using the EWS, drought management and district coordination institutional arrangements created/supported by ALRMP II. These included: (i) UN agencies (WFP, UNDP, UNICEF, FAO, OCHA); (ii) bi-lateral donors and NGOs (EU, ECHO, DANIDA, USAID/FEWSNET, OXFAM); and (iii) officials and authorities in the Ministry for the Development of Northern Kenya & Other Arid Lands, and the Ministry of State for Special Programs (Office of the President). 2. The second round of consultations involved three field-level stakeholder workshops held on September 23, 26 and 28, 2011 in the following project districts: Kitui (semi-arid), Garissa (arid), and Tana River (arid). Summary of Participants in Stakeholder Workshops Location/Category Garissa Tana R. Kitui Total Beneficiaries No. community groups 3 4 2 9 No. participants 29 38 9 76 Line/sector agencies No. line agency groups 2 1 2 5 No. participants 12 9 12 33 Partner organizations No. partner organization groups 1 1 1 3 No. participants 10 6 2 18 District coordination unit 1 1 1 3 No. participants 6 10 5 21 Total No. groups 7 7 6 20 No. participants 57 63 28 148 3. Participants were divided into groups according to whether they were community beneficiaries, representatives of line/sector agencies, partner organizations, or ALRMP II district level staff. In total, 148 people participated, of which slightly more than half were beneficiaries, 22 percent were line/sector agency staff based in the districts, 12 percent partner organizations, and 14 percent district level staff of ALRMP II. Beneficiaries were divided into several groups in each workshop, both to keep numbers 75 manageable and to be able to cross check beneficiary views across groups that had not had the opportunity to listen to each other’s discussions. In two districts there was a sufficient number of line/sector agency representatives to be able to divide them also into separate groups. Line/sector agencies represented included: Agriculture; County Council; Education; Health; Horticultural Crops Development Authority; Land Reclamation; Livestock; Veterinary Services; Security; Social Services; Water and Irrigation; and Planning, National Development and Vision 2030. Partner organizations represented included CARE Kenya, Constituency Development Fund, District Food for Assets, Kenya Red Cross, UNDP, UNICEF, World Food Program, Women Concern, Womankind Kenya, and EU. 4. After introductory discussions with all participants, the groups worked in separate break-out sessions, and then reported back to the plenary. They were asked to focus on a common set of questions covering (i) their perceptions on the extent to which food aid needs; (ii) response times had changed over the period between the last major drought before ALRMP II had started (2000/01 drought year) and the last major drought while the project was under implementation (2008/09 drought year – although many participants also commented on their more recent experience with the still ongoing drought of 2011); (iii) the effectiveness of institutional arrangements; (iv) ALRMP II’s impact on women; (v) ALRMP II conflict resolution activities; (vi) changes in access/quality of services and the role ALRMP II had played; and (vii) comments on any other subject they consider relevant and lessons learned. 5. The team also held discussions in situ with community members involved with several randomly selected micro-projects in water, agriculture, livestock, education, and health. 6. The key feedback obtained from the two rounds of stakeholder’s consultations is summarized below: (i) UN agencies, international donors and NGOs. There was unanimity on: (i) the relevance of ALRMP II; (ii) the high level of confidence/trust in the information provided by the project EWS, which all of them use in their decision-making process in relation to drought and food aid matters; and (iii) the ALRMP II institutional set-up, in particular the district-level DSG structure. Before the creation of DSGs, each institution had been conducting drought management and relief activities independently, in an uncoordinated manner and in many instances with duplication. The DSGs have provided for: (i) knowledge sharing and a forum for joint discussion on specific needs at the local level; (ii) identification and clarification of the comparative advantage of the respective development partners; and (iii) harmonized and coordinated action on the ground. There was also agreement on the importance of the project contingency fund, which allowed for quick response in case of emergencies. These institutional arrangements resulted in increased efficiency, reducing the response time in drought emergencies, and had a positive impact in reducing the proportion 76 of people needing food aid and improving to access to basic services. The project was considered a best practice case that should be replicated in other African countries, and some agencies have already brought other country delegations to visit ALRMP II. The main concerns expressed by these partners related to: (i) IDA’s suspension of disbursements in the midst of a serious drought (given the leading role ALRMP plays in the ASALs, this also affected the work of other agencies); (ii) the uncertainty about a possible follow-up project to consolidate gains and ensure sustainability; and (iii) the need to create a National Drought Management Authority and a National Drought Contingency Fund. (ii) ALRMP II Management. High level of frustration continues among project staff and Ministry officials about: (i) IDA’s suspension of disbursement; (ii) the halt to the preparation of a new project; and (iii) the protracted INT investigation into allegations of fraud and corruption. The IL-ICR team was informed, both in Nairobi and in the field, about the negative impact these factors had on the way GoK was able to deal with the ongoing drought emergency in 2011. While the EWS and structures created by the project have continued to function and GoK has tried to maintain the staff, activities are taking place in a more limited way and this is having an impact on the ground in terms of how the drought is affecting people and livestock. (iii) Communities and Other Civil Society Organizations. There was consistent feedback that the structures created by the project (especially the DSGs) were beneficial to communities, performed proactively in responding to drought and related emergencies, and provided access to other organizations for assistance and investments to which they had either no or only limited access in the past. Time response to drought emergencies had decreased (project field monitors reported for example on time taken to respond to water and child malnutrition issues) and there was better focus and transparency in food aid needed. Training and capacity building, and project investments in water, agriculture, livestock, education and health have improved incomes, resilience (through asset creation) and drought preparedness, thereby reducing the share of people needing food. The project has improved access to basic services, including health and nutrition and access to education, particularly for girls. The impacts on empowerment (and participation of women) and conflict resolution and prevention of violence have been very important. Concerns included: (i) the increasing frequency of droughts; (ii) the lack of adequate resources during the ongoing 2011 drought; (iii) the need for continuation of IDA support; and (iv) the need for additional focus on IGA activities, health, education and women’s groups. (iv) Line agencies and partner organizations. These agencies echoed the view that one of the most important contributions of ALRMP II was the institutional structure (particularly the DSGs) and the positive impact in 77 helping to coordinate activities on the ground (by including in its structure all key ministries, NGOs, international organizations), avoiding repetition, providing access to communities, with preparatory identification of investment needs and priorities. Capacity building, investments in water, agriculture (better farming practices), livestock, and rapid response to emergencies were the most commonly cited interventions of importance. They also emphasized the importance of the contingency fund as an agile instrument that allowed very fast action at district level (e.g., for activities like water trucking). Asset building and drought preparedness and knowledge had helped to reduce the share of people requiring food aide (comparing the last major drought during the project implementation period, i.e. 2009, with the last major drought before ALRMP II started, which was in 2001). Women’s empowerment through capacity building, IGAs, their roles in CDCs, gross reduction of Female Genital Mutilation (FGM) practices, and girl’s education were mentioned. Line ministry participants stated that vaccination campaigns, and other complementary investments in health and education helped their work very much and had important impact on local populations. Creation of DPCs and cross border peace committees, together with training and facilitated meetings among communities had important impacts in reducing conflicts, particularly those related to access to water and pastures. ALRMP II had proved that food relief can be reduced, and communities can increase their level of self-sufficiency. Concerns expressed included: (i) limited resources; (ii) the need for sustainability of present institutional arrangements; (iii) the need for additional staff in line agencies to supervise projects on the ground; and (iv) the need to expand training of communities to ensure maintenance and sustainability of micro- projects. 78 Annex 6. Borrower's Comments on Draft IL-ICR19 [The following comments were submitted to IDA by the Borrower’s Ministry of Finance on December 20, 2011.] Arid Lands Resource Management Project Phase Two 1. After perusing the Draft Implementation Completion and Results Report on the Arid Lands Resource Management Project Phase Two Credit 3795 – KE released on the 5th December 2011, the view of the Kenya Government is that the Intensive Learning Implementation Completion Report (IL-ICR) Team made independent observations that are factual situation on the ground. For a Project of this magnitude to disburse 95 percent of the US$ 120.0Million for the period between September 2003 and June 2010, with the numerous challenges notwithstanding, this would be considered commendable. The IL- ICR Team also observed on the Section under “Financial Management - Flow of Funds� page 10 that – (quote): - “…the disbursement rate of the Project was one of the highest in the Kenya portfolio.� The fact that the International Development Association (IDA) provided AF to the Borrower to deepen and widen the scope of the Project following the mid-term review, was in itself a show of trust in the capacity of the Government to implement the Project towards delivery on the intended objective. The Government counterpart funding was adequate and timely throughout the life of the Project phase. This fact is attested by the IL-ICR Team’s observation that the ALRMP II was able to achieve fully on four of the five Key Performance Indicators of the Project Development Objective (PDO) and partially on the fifth KPI where the ALRMP II Team did its best to participate in Policy formulation but had no influence on the pace of fast tracking finalization of appropriate policies that would have enabled institutionalization of the Drought Management Institutions by the Policy Makers. 2. The creation of the Ministry of State for the Development of Northern Kenya and Other Arid Lands by the coalition Government would however be considered a culmination of the advocacy activities of all stakeholders in the Arid and Semi Arid Lands (ASAL) including the Project on the need for the GoK to address the plight of the marginalized communities of the ASALs in a sustainable way. The World Bank conducted periodic review missions and provided necessary technical back stopping to the Project Team. The Missions encouraged interaction between the Project Coordinating 19 79 Unit (PCU) and Mission Consultants laden with diverse skills at the times of the Review Missions as they assessed the performance of this project during the entire life of the Project. The informal suspension (i.e. not processed through the Legal Department as per OP/BP 13.40 as shown on Page 32 of the Draft report) cut approximately $5.7 million of IDA financing and some $5.9 million of EU drought financing and had a depressing impact on the final six months of the ALRMP II especially at a time the country was going through a severe drought period. Despite this, leading Development partners continued to express confidence in the ALRMP II institutional arrangements and continued to provide funds to mitigate the full impact of the drought. 3. The first phase of ALRMP had been in the forefront in piloting Community Driven Development (CDD) in Kenya, and ALRMP II would continue this effort. The changes in the Government focus towards decentralization offered opportunities for ALRMP II to scale up the agenda that had been initiated under the first project, i.e., building capacity and empowering pastoralist and small farmer communities in the ASALs to participate more actively in identifying and advocating for their own development priorities – centered very much around reducing risks and vulnerability to droughts. With the promulgation of the new constitution for this Country, the main focus of attention is devolution of governance to the County level where the communities are expected to have a greater say in development at the grass root level. Communities in the Arid Districts that have had contact with the CDD Component of the Project have an upper hand and are better placed to internalize this new direction and hence champion faster development of their region, thanks to the support of the Bank. 4. The observations on the overall performance of the ALRMP II by the IL-ICR Team were well captured and articulated in the Draft IL-ICR Report. The fact that the original PDO and certain KPIs had to be changed in 2006 is a clear indication of a problem in the design of the Project during formulation. The IL-ICR Team observed that the strategy for obtaining data for some of the KPIs was never clearly clarified in the Project documents, a fact that was in itself a challenge to the Implementer to achieve satisfactorily on the KPIs. After the PDO, the KPIs and nearly all the Intermediate Outcome Indicators were amended, there was no deliberate effort on the part of the Bank and the Borrower to re-align the M&E framework to address data collection strategies for the new indicators and this was left to the interpretation of the implementer to design own strategies. The Project Appraisal Document for the Additional Finances was also not clear on which areas of the original PAD that needed to be retained and which needed to be discarded. But despite this, the Implementing Team still continued to adhere to the original PAD and generated data for all concerns of this PAD. The IL-ICR Team on Section 2.3: “Monitoring and Evaluation, Design, Implementation and Utilization� and under the section on “Implementation� paragraph on pg 10 emphasized this same observation- (Quote: The absence of a clear analytical structure to guide data collection, entry and analysis means, however, that much of the information on hand may be difficult to reconcile/compare in a statistically robust manner). This observation is further strengthened on Section 5.1 on “IDA performance� on the subsection highlighting the shortcomings of the Project design and Section 6 part (ix) under Lessons Leant where it is stated that “the Task Teams need stronger M&E specialist support and the 80 recommendations made where the regional Bank units should have small groups of M&E specialists with proven practical experience in design of simple measurable results and risk frameworks and indicators�. The Government also acknowledges the observation made by the IL-ICR team on page 23 on Conflict resolution. The design of the ALRMP II highlighted the linkages between conflict management and natural resource management, and built on a combination of customary and formal institutions. Peace building and conflict management continued to take centre stage during implementation of the NR & DM components due to resource competition. The Project committed substantial resources to minimize conflicts, the success of which attracted other partners like Oxfam and the UNDP. UNDP availed resources for the establishment offices in six out of 11 arid districts and one at the Project Headquarters to strengthen Conflict management activities and also partner with the National Committee on Peace Building domiciled in the Office of the President. 5. The Government acknowledges the IL-ICR Teams observations on Intermediate outcome Indicators 12 to 16 on Page ix of the draft ICR Report. However, it is worthwhile to note that the Pastoral Policy referred to in the indicator number 12 was expanded from the Pastoral Policy to the “National Policy for the Sustainable Development of Arid and Semi-Arid Lands of Kenya� which was also formulated but has yet to be officially adopted. The Intermediate Outcome Indicator number 15 on the Guidelines for emergency livestock off-take prepared and articulated in all project Districts was not only successfully prepared but the result also uploaded into the website: www.aridland.go.ke/sld/emergency-livestock-offtake.pdf for use by other stakeholders. 6. Under the Section on the INT Review on page 16, the Government concurs with the views of the IL-ICR Team where they observed that despite the INT investigations having commenced on 2009, a preliminary written report was given by the INT in April 2011 with the same being published on the INT website on July 15th despite a GoK request to allow the GoK until 15th August 2011 to comment. The GoK, IDA and other partners expected there would be a follow-up project phase and substantial preparatory work had been completed. But with the release of the INT preliminary report, the IDA Management stopped further preparation of the follow-up Project leaving the GoK and other Development Partners in limbo. The IL-ICR Team requested the PCU to organize for a national level stakeholders and partners meeting with the IL-ICR Team as recorded on Section 3.6 of the draft ICR Report on Page 28. The same IL-ICR Team requested for a visit to Project Districts and a chance to hold workshops with various stakeholders at the District Level. Having accompanied the IL-ICR Team to the field and witnessing a random selection of project sites to be visited, the Government confirms the observations made by the IL-ICR Team (See Annex 6 on Page 65). 7. The main concerns raised by the UN Agencies, international and local NGOs during their meeting with the IL-ICR Team related to (i) IDA’s suspension of disbursement in the midst of a serious drought which affected the work of other stakeholders; (ii) Uncertainty about a possible follow-up project to consolidate and ensure sustainability of ALRMP II gains, and (iii) the need to create a National Drought Management Authority to institutionalize National Drought Contingency Fund. While all 81 these concerns raised were legitimate, the GoK wishes to report that despite the suspension of funding to the Drought Management Structure in the Country by the IDA, the GoK continued to sustain the structure and also pursue the Establishment of the National Drought Management Authority. The President did authorize through a decree the establishment of this Authority as per Section 3 of the Legal Notice Number 171 of 24th November 2011 under the State Corporations Act Cap 446, Laws of Kenya. The NDMA is hence established and in place. On the other hand, the Sessional Paper for the establishment of the NDCF has received necessary approval from the relevant stakeholder Ministers and is awaiting approval from the Cabinet and subsequent operationalization by the Ministry of Finance. It is hoped that with the establishment of the NDMA, the approval for the establishment of the NDCF will be fast tracked, as the use of NDCF happens to be a major component of the NDMA’s operations. 8. The Government faults the observation made by the IL-ICR Team on the status of the Intermediate Outcome Indicator number 11 on Page viii of the Draft ICR Report and Page 47 Component 2 “Link to output/Intermediate Outcome Indicators�. While the Indicator required that the Project Team would assist at least 1000 groups in developing savings capacity through top-up savings grant, by the closure of the ALRMP II, the Project Team in the CDD component had through 1588 events/transactions managed to support 57,095 males and 83,790 females to undertake Income Generating Activities through the top-up savings grants (Matching Grants). The CDD Component team was able to disburse a total of Kshs 186,203,452 as matching grants from the Project while the community groups contributed Kshs 99,086,648. Considering that some of these groups qualified for two tranches of support having graduated from the first level of funding, on the average the Project supported more than the proposed 1000 groups with matching grants for income generating activities. The Indicator number 11 in our view was hence over achieved. 9. The IL-ICR Team’s observation on the Section under “Financial Management - Flow of Funds� page 13 that disbursement to some community groups was made without requisite community contribution being made as specified in the PAD, and that the PCU advanced reasons of some communities being too poor to make the contribution; was inaccurate. The only explanation offered by either the DCU or the PCU at any one point may have been that contributions from communities were at times delayed due to prevailing drought conditions or conflicts but never were the communities exempt from making the contribution. This condition is clearly stipulated on Section 4.5.1 of the Project Implementation Plan (PIP) under the section on “Implementation Modalities for basic CDD�, page 69 and also Section 4.8.3 under the section on Community Contribution and Sustainability of the PIP. 10. While on the same subject above, it should be noted that, under “Accounting and Financial Reporting� on page 15, the fact that IFRs did not indicate status of accountability of funds once disbursed was also not accurate as IFRs are meant to detail disbursement of funds made within a particular quarter and are hence exhaustive as per the IFR design as they contain both technical and financial details for each station of disbursement. The detailed support documents for these IFR are filed at the national 82 office, as they would make the same too bulky for dissemination to the World Bank. It should be noted that the arrangement for CDD in the second phase compared to the first phase was to empower community members to take full control of managing implementation of their micro projects. The responsibility of the Project was to transfer funds to the CDC Account in tranches and demand for accountability of the same before the community could qualify for a second tranche. The METs were expected to verify the records and documents filed and maintained at the community level for Audit purposes. The Government Accounting System does not provide for community accounting documents to be part of GoK finance documentation. Parallel files on community bank transactions were also maintained at the Drought Management Offices. Part of the Kshs 340 million quoted by the joint IAD and INT report on Page 16 actually contains bulk of the funds transferred to the communities as provided for in the PAD. 11. The observation by the IL-ICR Team on Section 5.2 (b) on the Implementing Agency or Agencies Performance that “the PCU could have detected earlier the need to clarify decision-making responsibilities and accountabilities at community and district level and ensuring adequate O&M arrangements� on page 35 of the Draft IL-ICR needs to be clarified. The CDC as a management structure at the community level is highly respected and recognized. This committee makes key decisions for the community and part of their key responsibilities is undertaking expenditure and accountability both of which are closely monitored by the Mobile Extension Teams (METs). The O&M arrangements were the responsibilities of the community as part of their contribution to facilitate the CDC subcommittees undertake assigned functions like floating of quotations and/or transacting banking (Refer to Section 2.3.6 of the Project Implementation Plan on the functions of the CDDC on Page 31). 12. The observation made by the IL-ICR Team on Page 49 of the Draft report that the Project Reporting formats did not enable one to see a regularly updated and comprehensive list of maps of all CDD micro-projects arranged by community and districts and by components is incorrect. The ALRMP II MIS database is able to sort out the data by sector, component, year and even by activity all of which are geo-referenced. The development of maps is generated from a different platform not directly linked to the MIS though each activity is geo-referenced. Government relationship with the Bank 13. The Arid Lands Resource Management Project Phase Two (ALRMP II), a community-based drought management Project of Kenya Government was the latest stage in a nearly two-decades‟ long partnership between IDA and the Government (GoK) to strengthen drought management and reduce vulnerability of pastoralists and small-scale farmers in Kenya’s very poor and fragile arid and semi-arid lands (ASALs). The partnership evolved from: (i) emergency support after a severe drought in the early 1990s (Emergency Drought Recovery Project, EDRP); to (ii) the development under the first phase of ALRMP of a community-based drought early warning system (EWS) for the arid districts; contingency planning at the district level; institutional and financial arrangements to reduce response times and improve drought management; small 83 investments to reduce livelihoods vulnerability; and finally (iii) to ALRMP II, which aimed to continue working in the arid districts while extending parts of the program to the semi-arid districts, enhance and substantially scale up the EWS and other institutional, financial and inter-agency coordination arrangements, expand investments to improve livelihoods, and intensify advocacy for a much stronger policy focus on the ASALs. One of its most important functions has been to operate the drought management and early warning system, a system, which, thanks to the Bank’s investment has gained international support. The Bank has over time stated its satisfaction with the operations of phases of the ALRMP. The Bank has also accepted the financial management reports of the Project and rated its financial accounting and reporting as either “Highly Satisfactory� or “Satisfactory� and confirmed its high regard for the competence of the PCU Team, repeatedly referring to its exemplary performance. 14. The World Bank’s approach through ALRMP was widely recognised as the embodiment of good development practice in pastoral and marginalised areas – for example in its commitment to long-term engagement, its integration of drought management within project design, its multi-sectoral scope, and its championing of community empowerment. It was the Bank which promoted the use of the Community Driven Development (CDD) approach and it was the same Bank which encouraged the Government to adapt its financial management systems so that direct disbursements could be made to communities. The World Bank’s approach through ALRMP was widely recognised as the embodiment of good development practice in pastoral and marginalised areas – for example in its commitment to long-term engagement, its integration of drought management within project design, its multi-sectoral scope, and its championing of community empowerment. It was the Bank that promoted the use of the Community Driven Development (CDD) approach and it was the same Bank which encouraged the Government to adapt its financial management systems so that direct disbursements could be made to communities. In deciding to finance ALRMP, the Bank was not blind to the challenges it would face. The Project Appraisal Document for ALRMP II, prepared in 2003, acknowledged the risks of operating in the kind of environment that characterises Kenya’s arid lands, but concluded (rightly) that the risks of inaction were even greater. 15. In assessing Quality of Supervision by IDA, the IL-ICR team took into account not only performance by the task team, but also IDA management, and other parts of the Bank Group that had an impact on IDA’s overall oversight of ALRMP II. The IL-ICR team found that supervision by the task team was satisfactory in many respects. (i) The skill range of the team was quite strong, considering not only the core members, but also IDA consultants and FAO staff. (ii) There was a good mix of Headquarters and Country Office-based staff, and considerable continuity of team members. (iii) Supervision missions took place regularly at roughly six-month intervals. ISRs and aide memoires were timely and informative. It therefore is extremely suspect to witness one out of 16 bank rating of the performance of the Project as “Moderately Unsatisfactory� while all others were rated “Satisfactory�. 84 16. The Government convened a series of workshops for Permanent Secretaries that culminated in the identification of various categories of risks in their day-to-day activities and development of Risk Management Frameworks at the Ministry level. This was in line with the Bank requirement on governance and accountability. The Project on the other hand engaged a Consulting Firm to assist in the development of a risk framework as well as a risk register. The AF PAD also stressed the need for transparency and accountability that triggered the need for all micro projects being branded with the establishment of notice boards at the Drought Management Offices on which the annual work plans and budgets as well as progress reports were posted. The Government also introduced the use of “Anti-corruption� boxes, “Suggestion� boxes as well as “Complaint� boxes, which were checked periodically. Towards the end of the ALRMP II, the Project team using key lessons from implementation of the Project phase developed a comprehensive Governance and Accountability Action Plan (GAAP), which they shared with the Bank. These initiatives were all measures aimed at reducing the possibility of fraud and corruption. There is no doubt that corruption remains a major development challenge world over. However when fraud and corruption is suspected or alleged within a Project/Programme, there is no question that the Bank and Borrower need to act quickly and decisively to investigate and take appropriate measures in a professional manner. 17. The forensic audit of ALRMP II started in April 2009. One of the guiding principles of the World Bank is that of ‘working in partnership’ with the Borrower. In its conduct and management of this audit, the Bank failed to live up to this principle of partnership, on several counts. First, it failed to keep the Government of Kenya adequately informed. Both the informal suspension of ALRMP II and the decision to suspend planning for a new programme to replace it were taken unilaterally and communicated verbally. Nothing was ever given in writing to substantiate the original (and subsequently disproved) percentage of suspected fraudulent expenditure, which was the trigger for the project’s suspension. Second, the Bank has shown a lack of transparency. Between October 2009 and March 2011 (i.e. throughout the audit period) officers from the Government and from three other development partners, as well as representatives of stakeholders, participated in good faith in a series of six design missions for a new joint arid lands programme. Even in February 2011, the Bank was still expressing its support for the programme. This support was abruptly withdrawn in May 2011. Third, the Bank reneged on an agreement with the Government of Kenya. On 13 June 2011 the Vice Presidents for Africa and for INT and the Prime Minister of Kenya agreed on a process for establishing ineligible expenditure and for managing communication of the same. The Bank failed to honour this by prematurely publishing the INT report on its website without the benefit of a detailed response to the allegations – a response which the Government was unable to provide because INT did not make key documents available. Once allegations are in the public domain, it is hard to recover from the reputational damage caused. Fourth, the Bank has abused the trust and goodwill shown by the Government of Kenya. Its stated approach is ‘to constructively support those in government who are making a good-faith effort to bring change to Kenya via the fight against corruption’. The Ministry on the other hand repeatedly expressed its willingness to cooperate with INT and to act on any evidence of corrupt practice. The management of the original allegations and of the audit itself by the Bank has not only 85 been damaging to the reputation of the World Bank but also to the reputation of the Government of Kenya, and, above all, to poor people in a vulnerable part of Kenya. It has contravened the Bank’s own values and principles. The most important of these principles is essentially ‘don’t make the poor pay twice’. In this case the informal suspension of the project and the immediate resultant effect on these poor communities. In our opinion, the performance of INT throughout the audit has been very poor. At the most basic level, the audit has taken far longer than the Bank’s target of one year for most investigations and 18 months for complex cases. INT officers have demonstrated significant shortcomings – for example, in failing to brief themselves about Government procedures and about the Bank’s own disbursement procedures; in failing to verify allegations which have been easily disproved by IAD; and in failing to ask for documents which might have assisted them in their investigation. Had INT’s officers engaged appropriately with project staff (or with IAD) rather than presuming their guilt and keeping them at arms-length, many problems could have been avoided. There is obviously huge pressure on INT (and on our own IAD) to demonstrate results in the fight against corruption. Even so, it is still important that conduct is above-board and professional at all times. 18. In conclusion, all effort should be made by the Bank through INT and the Regional Bank Offices to assure that Projects do not go into an indeterminate status of informal suspension. Informal suspensions do not allow for careful review by the Legal Department of evidence justifying the remedy; there is no formal communication to a Borrower detailing the reasons for suspension and what, if any, actions can be taken to correct the situation; the Board and co-financiers are not formally advised; the Bank does not hold itself to any clear timetable for follow-up; and the process generally lacks the transparency that should accompany such serious actions. Future designs of Projects should factor among the possible risks that may stall project implementation, the “INT� entity as has been witnessed by the ALRMP II. 86 Annex 7. Comments of Co-financier [The following comments were submitted to IDA by the Delegation of the European Union to Kenya on December 15, 2011.] 1. The EU Delegation to Kenya agrees on the important leadership role that has been provided by ALRMP in the development of the Kenya ASAL areas during the past decade. We also note the impressive progress registered during its activity period, in particular in the fields of drought preparedness and systems for early warning systems. 2. The EU Delegation also share the findings on the strong negative impact that the suspension of the IDA funding and the consequent interruption of support from IDA to sustain and scale up the previous results in a new joint programme with several development partners. That this happened at a time when the Kenya ASAL faced a very severe drought compounded this negative outcome. 3. The EU Delegation agrees that following this withdrawal, the implementation of the EU assistance in ASAL was negatively affected, in particular the disbursement of the EU contingency fund (under 9th EDF) delivered through the World Bank but also the setting up of the contingency fund planned on the Kenya Rural Development Programme (under 10th EDF). In addition, the foreseen jointly financed basket fund that was supposed to improve the coordination between the main stakeholders in contingency funding remains a key issue in the efficiency of the emergency response to droughts. 4. The EU Delegation to Kenya would however like to highlight that the EU continues its institutional support to the Ministry of Development of Northern Kenya and other Arid Lands, although with a strong reduction in the scope of the activities funded and focusing on absolute priorities. This reduction was decided in order to limit the possible fiduciary risks related to the management at community level, pending the final outcomes of the ongoing INT investigations. This continuity makes EU able to accompany the current setting up of the National Drought Management Authority (NDMA), following the signature on 17/11/2011 by the President of the Republic of Kenya of the executive order to establish the NDMA. This is a major achievement in the process of Government of Kenya co-ordination efforts and its ownership of the drought response initiatives. 5. The EU Delegation believes that this momentum is a very good opportunity to strengthen the partnership between the Government and the development partners and to improve the coordination between the development partners. In the light of this, the EU Delegation will be open to a renewed partnership with the WB on the drought management activities in Kenya. This renewed cooperation will of course have to take into account the final conclusions taken by the World Bank on the basis of the INT report as far as their part is concerned as well as taking into account the recommendations and conclusions given. 87 Annex 8. List of Supporting Documents ALRMP II (World Bank) 1. Integrated Safeguards Data Sheet for ALRMP II, Appraisal Stage (Report 53596) 2. Kenya Community Driven Development: Challenges and Opportunities. World Bank, Africa Region. June 27, 2002 3. Africa Online, Social Assessment Report for ALRMP II 4. Project Appraisal Document, Report No. 25641, May 23, 2003 5. Development Credit Agreement No. 3795-KE, June 26, 2003 6. Project Paper for Additional Financing, Report No. 36655-KE, July 10, 2006 7. IDA Agreement Amending DCA (Cr. No. 3795-1-KE, October 25, 2006) 8. Implementation Status Reports (ISRs) Nos. 1-16 and related Aide-Memoires (including Mid-term Review Aide-Memoire, November 23, 2006) 9. Country Assistance Strategy for Kenya, 2004-07, Report No 29038, May 19, 2004 10. Country Partnership Strategy for Kenya, 2010-2013, Report No. 52521, March 23, 2010 11. Integrity Vice Presidency. Forensic Audit Report on Kenya Arid Lands Resource Management Project Phase 2. July 2011 12. Drought and Food Crisis in the Horn of Africa, September 2011 13. Back to Office Report of Kenya Drought Response Mission. August 31, 2011 14. IDA and AusAid. Kenya Economic Update, Edition No. 5. IDA and December 2011 15. Project Information Document (PID) No. AB5561 (March 10, 2010) for Arid and Semi-Arid Lands Sector Wide Program and other documentation on proposed ALRMP II follow-up operation 16. Social Accountability Concept Note, February 2011 (and peer review comments and Decision Note) 17. Munshi, Meena. CDD Diagnostic Note on ALRMP II, October 7, 2010 (Government of Kenya) 18. Project Manuals for ALRMP II (various) 19. ALRMP Website (http://www.aridland.go.ke/index.php) 20. Borrower’s Completion Report for ALRMP II, September 2011 21. Annual Environmental Audits for ALRMP II 22. Environmental Management Framework for ALRMP II 23. Vision 2030 24. Economic Recovery Strategy (PRSP) (Other) 25. European Commission TF 070845 for Drought Management Initiative (DMI); various grant monitoring documents, including final report by IDA to the EU,June 28, 2011 26. ILRI, The Impacts of the Arid Lands Resource Management Project (ALRMPII) 88 on Livelihoods and Vulnerability in the Arid and Semi-Arid Lands of Kenya (2011) 27. ILRI, An Assessment of the Response to the 2008-09 Drought in Kenya: Report t the European Commission. May 12, 2010 28. Watkins, Ben and Mwangi Margaret. Evaluation of the World Bank’s Arid Lands Resource Management Project. Report prepared for Danida. Kimetrica, March 17, 2009 29. Danida Kenya Natural Resource Management website, various documents: http://www.ambnairobi.um.dk/en/menu/Development/ NaturalResourceManagement/ 30. Mude et al. Empirical Forecasting of Slow-Onset Disasters for Improved Emergency Response: An Application to Kenya’s Arid North.. Food Policy 34 (2009) 31. Naschold, Felix and Barrett, Christopher. A Stochastic Dominance Approach to Program Evaluation with an Application to Child Nutritional Status in Arid and Semi-Arid Kenya, April 30, 2010 (Paper for American Agricultural Economics Association meeting, 2010) 32. Batch Associates Ltd, Draft Technical Audit of ALRMP II Community Level Infrastructure Investments (Community-Driven Development Component), July 2010 ALRMP I (IDA 2797-KE) 1. Implementation Completion Report (No. 27560, December 30, 2003) 2. Project Performance Audit Report (No. 34052, October 31, 2005) 3. Appropriate Development Consultants Limited. Beneficiary Assessment Study of the Arid Lands Resource Management Project. August 2002* Emergency Drought Recovery Project/EDRP (IDA 260-KE) 1. Implementation Completion Report (No. 18090, June 29, 1998) 89 IBRD 38900 KENYA ALRMP II PROJECT DISTRICTS PROJECT DISTRICTS: MAIN CITIES AND TOWNS DISTRICT BOUNDARIES ARID PROVINCE CAPITALS PROVINCE BOUNDARIES SEMI-ARID NATIONAL CAPITAL INTERNATIONAL BOUNDARIES 34°E 36°E 38°E 40°E 42°E SU D A N Karungu ETHIOPIA Lokichokio 4°N 4°N Lake Ramu Mandera Kakuma Turkana Sololo North Horr Moyale Lodwar Buna El Wak ND UGANDA Lokichar South Horr Marsabit Tarbaj 2°N Kangatet 2°N Wajir SO M A L I A LI Maralal Kitale Kapedo Archer’s WESTERN R I F T VA L L E Y Post Garba Mando Tula Gashi Eldoret Marigat Isiolo Mbalambala NORTH Butere Kakamega Nyahururu EASTERN EASTERN Falls Nanyuki 0° 0° Kisumu Nakuru N YA N Z A Kericho Nyeri Garissa Gilgil Embu CENTRAL Nguni Lake Karungu Narok Thika Bura Victoria Lolgorien NAIROBI Kolbio NAIROBI Kitui Machakos AREA Konza Magadi Bodhei 2°S Ikutha 2°S Garsen Lamu Kibwezi Namanga C O A S T Tsavo TANZANIA Malindi INDIAN Voi Mackinnon OCEAN Park KENYA 4°S Mombasa This map was produced by the Map Design Unit of The World 0 40 80 120 160 200 Kilometers Bank. The boundaries, colors, denominations and any other Shimoni information shown on this map do 0 40 80 120 Miles not imply, on the part of The World Bank Group, any judgment on the legal status of any territory, or any endorsement or 34°E 36°E 38°E 40°E acceptance of such boundaries. OCTOBER 2011