68284 The World Bank Group WORKING FOR A WORLD FREE OF POVERTY The World Bank Group consists of �ve institutions – the International Bank for Reconstruction and Development (IBRD), the International Finance Corporation (IFC), the International Development Association (IDA), the Multilateral Investment Guarantee Agency (MIGA), and the International Centre for the Settlement of Investment Disputes (ICSID). Its mission is to �ght poverty for lasting results and to help people help themselves and their environment by providing resources, sharing knowledge, building capacity, and forging partnerships in the public and private sectors. The Independent Evaluation Group IMPROVING THE WORLD BANK GROUP’S DEVELOPMENT RESULTS THROUGH EXCELLENCE IN EVALUATION The Independent Evaluation Group (IEG) is an independent unit within the World Bank Group. It reports directly to the Board of Executive Directors, which oversees IEG’s work through its Committee on Development Effectiveness. IEG is charged with evaluating the activities of the World Bank (the International Bank for Reconstruction and Development and the International Development Association), the work of the International Finance Corporation in private sector development, and the guarantee projects and services of the Multilateral Investment Guarantee Agency. The goals of evaluation are to learn from experience, to provide an objective basis for assessing the results of the Bank Group’s work, and to provide accountability in the achievement of its objectives. It also improves Bank Group work by identifying and disseminating the lessons learned from experience and by framing recommendations drawn from evaluation �ndings. The Matrix System at Work The Matrix System at Work An Evaluation of the World Bank’s Organizational Effectiveness ©  2013 Independent Evaluation Group The World Bank Group 1818 H Street NW, Washington DC 20433 Telephone: 202-458-4497 Internet: http://ieg.worldbankgroup.org E-mail: ieg@worldbank.org Some rights reserved 1 2 3 4 15 14 13 12 This work is a product of the staff of the Independent Evaluation Group (IEG) with the exception of the “Management Response� and the “Chairperson’s Summary.� Note that IEG and the World Bank Group do not necessarily own each component of the content included in the work. IEG and the World Bank Group therefore do not warrant that the use of the content contained in the work will not infringe on the rights of third parties. The risk of claims resulting from such infringement rests solely with you. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of IEG, the World Bank Group, its Board of Executive Directors, or the governments they represent. IEG and the World Bank Group do not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of IEG and the World Bank Group concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or waiver of the privileges and immunities of IEG or the World Bank Group, all of which are specifically reserved. Rights and Permissions This work is available under the Creative Commons Attribution 3.0 Unported license (CC BY 3.0) http://creativecommons.org/licenses/by/3.0. Under the Creative Commons Attribution license, you are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the following conditions: Attribution—Please cite the work as follows: IEG. 2013. The Matrix System at Work: An Eval- uation of the World Bank’s Organizational Effectiveness. Washington, DC: World Bank. doi: 10.1596/978-0-8213-9715-2 License: Creative Commons Attribution CC BY 3.0 Translations—If you create a translation of this work, please add the following disclaimer along with the attribution: This translation was not created by the Independent Evaluation Group or the World Bank Group and should not be considered an official World Bank translation. IEG and the World Bank Group shall not be liable for any content or error in this translation. All queries on rights and licenses should be addressed to the IEG, 1818 H Street NW, Washing- ton, DC 20433, USA; fax: 202-522-3125; e-mail: ieg@worldbank.org. ISBN: 978-0-8213-9715-2 e-ISBN: 978-0-8213-9716-9 DOI: 10.1596/978-0-8213-9715-2 Cover photo: Planetary matrix. ©  Ocean/Corbis Library of Congress Cataloging-in-Publication data has been requested. Contents ACKNOWLEDGMENTS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . x OVERVIEW . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xi MANAGEMENT RESPONSE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxi CHAIRPERSON’S SUMMARY: COMMITTEE ON DEVELOPMENT EFFECTIVENESS . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxvi ABBREVIATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxxviii GLOSSARY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xl 1. INTRODUCTION AND CONTEXT . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Rationale and Purpose . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 Scope and Evaluative Questions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8 Literature Review on Matrix Organization . . . . . . . . . . . . . . . . . . . . . . 12 Evaluation Methodology . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 2. THE STRATEGIC ALIGNMENT OF SECTOR AND COUNTRY PRIORITIES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 The Relevance of Strategic Alignment . . . . . . . . . . . . . . . . . . . . . . . . . 22 Responsiveness and Effectiveness of Country Programs . . . . . . . . . . . . . 26 Alignment of Country and Sector Strategies . . . . . . . . . . . . . . . . . . . . . 31 Efficiency and Effectiveness of Strategic Alignment . . . . . . . . . . . . . . . 37 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39 3. THE PROMISE OF A KNOWLEDGE BANK . . . . . . . . . . . . . . . . . . . . . 41 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 42 Investing in Knowledge Production . . . . . . . . . . . . . . . . . . . . . . . . . . . 45 Dissemination and Use of Knowledge . . . . . . . . . . . . . . . . . . . . . . . . . 48 Tacit Knowledge Does Not Flow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Efficiency and Effectiveness of the Knowledge Bank . . . . . . . . . . . . . . 60 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 65 4. DELIVERING QUALITY SERVICES . . . . . . . . . . . . . . . . . . . . . . . . . . 67 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 68 v Contents Institutional Incentives for Quality . . . . . . . . . . . . . . . . . . . . . . . . . . . 72 Managing for Quality Under the Matrix System . . . . . . . . . . . . . . . . . . . 74 Staff Incentives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 80 Conclusions on Efficiency and Effectiveness of Quality Assurance Systems . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 84 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 85 5. INSTITUTIONAL ISSUES AND ORGANIZATIONAL STRUCTURE . . . . . . 87 The Sector Board Model . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 89 Span of Control in the Regions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 94 The Effects of the SDN Merger on the Matrix System . . . . . . . . . . . . . . . 96 Budgetary Incentives as a Driver of Matrix Behaviors . . . . . . . . . . . . . 101 Quality Assurance and Conflicts of Interest . . . . . . . . . . . . . . . . . . . . 107 Key Findings and Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 108 Notes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 109 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 111 6. CONCLUSIONS AND RECOMMENDATIONS . . . . . . . . . . . . . . . . . . . 113 Recommendations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 117 References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118 APPENDIXES APPENDIX A: MATRIX MANAGEMENT AND ORGANIZATION OF THE WORLD BANK IN HISTORICAL PERSPECTIVE . . . . . . . . . . . . . . . 119 APPENDIX B: EVALUATION METHODOLOGY . . . . . . . . . . . . . . . . . . . 129 APPENDIX C: RESPONSIVENESS AND EFFECTIVENESS OF THE BANK’S COUNTRY PROGRAMS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135 APPENDIX D: EXPERIENCE IN THE TRENCHES: RESULTS FROM THE STAFF SURVEY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147 APPENDIX E: KEY FINDINGS FROM SECTOR MANAGER INTERVIEWS AND COUNTRY DIRECTOR/MANAGER INTERVIEWS . . . . . . . . . . . . . . . 167 APPENDIX F: FOCUS GROUPS DISCUSSIONS: THE KEY FINDINGS . . . . . 183 APPENDIX G: COMPARISON OF THE SIX REGIONAL MATRICES . . . . . . . 189 APPENDIX H: WORKING ACROSS ORGANIZATIONAL BOUNDARIES: vi The Matrix System at Work THE USE OF CROSS SUPPORT . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 197 APPENDIX I: BUDGETARY INCENTIVES . . . . . . . . . . . . . . . . . . . . . . . 211 APPENDIX J: WORKING ACROSS SECTORS AND THEMES . . . . . . . . . . . 225 APPENDIX K: WORLD BANK SECTOR STRATEGY PAPERS . . . . . . . . . . . 233 APPENDIX L: INTERNATIONAL FINANCE CORPORATION 2013 CHANGE INITIATIVE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 239 APPENDIX M: STAFFING DATA . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 247 APPENDIX N: MANAGERS INTERVIEWED . . . . . . . . . . . . . . . . . . . . . 253 BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 257 BOXES Box 2.1. Sector Strategy Papers at What Cost? . . . . . . . . . . . . . . . . . . . . 38 Box 3.1. Sector Manager Perspectives on Primacy of Lending . . . . . . . . 46 Box 3.2. Social Protection: A Model for Global Knowledge Management . . 51 Box 5.1. The New FPD Model: Creating “Global Technical Practices� . . . . . 93 Box 5.2. Views on Matrix Issues of East Asia and the Pacific Staff Based in Sustainable Development Network Units . . . . . . . . . . 101 Box 5.3. The International Finance Corporation’s 2013 Change Initiative . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103 FIGURES Figure 1.1. Bank-Wide Matrix Structure (2012) . . . . . . . . . . . . . . . . . . . 7 Figure 1.2. Matrix-Related Bank Budget Allocation and Bank-Executed Trust Fund Mobilization (FY10 Data) . . . . . . . . . . . . . . . . . . 8 Figure 1.3. Matrix Management Evaluative Framework . . . . . . . . . . . . . 11 Figure 3.1. The Production and Application of Knowledge at the World Bank . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 47 Figure 3.2. The Composition of Expenditures on Knowledge for External Clients Has Changed Substantially With Non-Lending Technical Assistance Overtaking Economic and Sector Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 48 Figure 3.3. Cross-Support from Network VPUs to the Regions, FY02-FY10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 59 Figure 3.4. Cross Support Under the Matrix System in FY10 . . . . . . . . . . 59 Figure 3.5. BETFs Now Finance More Than 40 Percent of the Bank’s Core Knowledge Products . . . . . . . . . . . . . . . . . . . . . . . . . . 61 vii Contents Figure 4.1. Outcomes of Bank-Financed Operations by Fiscal Year . . . . . . 70 Figure 4.2. Investment Lending Quality and Outcomes by Fiscal Year of Approval . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71 Figure 4.3 The Quality Assurance System Has Little Time to Influence many Studies . . . . . . . . . . . . . . . . . . . . . . . . . . . 77 Figure 4.4. Average Share of Investment Loans Allocated to Non-affiliated Sectors and Themes . . . . . . . . . . . . . . . . . . . . 83 Figure 5.1. Regional Sector Units Are Delivering More With a Smaller Share of Regional Bank Budgets as Overheads Grow . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 105 Figure 5.2. Network Anchors and WBI Receive a Much Higher Proportion of Trust Funds (BETF) Than Bank Budget . . . . . 106 TABLES Table 2.1. Assessment of the Degree of Implementation of Sector Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25 Table 2.2. Key Features of Country Strategies and Sector Strategies . . . . 26 Table 2.3. Basis for Rating Each Component . . . . . . . . . . . . . . . . . . . . . 28 Table 2.4. Percentage of Country Programs Rated Moderately Satisfactory or Higher Over Time . . . . . . . . . . . . . . . . . . . . . 28 Table 2.5. Country Program Variables That Require Attention . . . . . . . . 29 Table 2.6. Integration of Country and Regional Experience in Sector Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Table 2.7. Alignment Among Country Strategies and Operations and Sector Strategies . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 33 Table 2.8. Greatest Weakness of the World Bank . . . . . . . . . . . . . . . . . . 39 Table 3.1. Dissemination of Knowledge Within the Bank . . . . . . . . . . . . 49 Table 3.2. Most Knowledge Generated in a Unit is not Shared With Other Units . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 50 Table 3.3. Few Staff Use Knowledge Produced Elsewhere . . . . . . . . . . . . 53 Table 3.4. Most Units Do Not Use a Substantial Amount of Knowledge Produced Elsewhere . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 Table 3.5. Proportion of Headquarters-Appointed GF-GI Staff in Unit for More Than Five Years . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 Table 3.6. Total Cross-Support Sold by Regions and Network Anchors, FY10 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 Table 4.1. Percentage of Anchor and Sector Staff Indicating They Received Substantial Encouragement to Carry Out the Following Behaviors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81 viii The Matrix System at Work Table 4.2. Percentage of Respondents Indicating That Incentives Influence Their Behavior to a Substantial Extent . . . . . . . . . . 82 Table 5.1. Selected Indicators for Networks . . . . . . . . . . . . . . . . . . . . . 97 Table 5.2. Staff Head Counts of Network VPUs and Selected Central VPUs, FY11 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 98 ix Contents Acknowledgments This evaluation of the World Bank’s organizational effectiveness was prepared by an IEG team led by Anis Dani. The evaluation was carried out under the guidance of Cheryl Gray (former Director) and Mark Sundberg (Manager), and the direction of Caroline Heider (Director General) and Vinod Thomas (former Director General). Team members included Karen Coulibaly, Brett Libresco, Manuel Penalver- Quesada, Bahar Salimova, Salvatore Schiavo-Campo, Jesse Torrence, and Bowen Uhlenkamp. Daniela Gressani, Hans-Martin Boehmer, and Christine Wallich provided valuable feedback at several stages of the evaluation. The team also acknowledges the advice received from Amit Banerji on IFC’s matrix system and Geeta Batra on the recommendations and Management Action Record. Peer reviewers were Nicholas Argyres (Senior Associate Dean for Faculty and Vernon W. and Marion K. Piper Professor of Strategy, Olin Business School, Washington University in St. Louis), Nancy MacPherson (Rockefeller Foundation), Praful Patel (former vice president, South Asia Region) and Joanne Salop (former vice president, Operation Policy and Country Services). IEG is grateful to World Bank management and staff for providing valuable time, information, and feedback required to conduct this evaluation. In particular, IEG thanks the Managing Directors, the Matrix Leadership Team, and the Internal Reform Secretariat for feedback during the consultations for finalizing the evaluation report. IEG also acknowledges the assistance received from HR Analytics and Corporate Finance for their help in procuring essential staffing and budget data, and from the FPD Network and IFC for information regarding their organizational systems. IEG would also like to acknowledge the insights from country and sector managers, and the several hundred staff who took the time to respond to the survey to help understand the complexities and challenges of the World Bank’s matrix system. William Hurlbut and Cheryl Toksoz edited the report and Maria Gabriela Padrino helped with the formatting of the report, Melvin Vaz provided analytical assistance during the initial phase of the evaluation, and Aline Dukuze provided administrative support to the team and helped with production of the report. x The Matrix System at Work Overview The Matrix System at Work Context The 1997 Bank reforms that introduced the matrix management concept aimed to adapt the organization to changing circumstances and address concerns among external stakeholders about the role of aid in development. The reforms were motivated largely by widespread recognition that the Bank’s development programs were excessively driven by a culture of lending, with insufficient attention to client needs and the quality of results, which are crucial to development effectiveness. A previous round of reforms in 1987 had strengthened the country focus, but quality remained a concern. Furthermore, access of developing countries to development finance from the private sector had increased significantly, leading to a decreasing share of official development aid, including Bank financing, in total flows to developing countries. This trend has continued after slight interruption by the Asian financial crisis. In 1987, World Bank lending represented 15 percent of all external financing for developing countries. By 2002 Bank lending had declined to 4 percent of external financing (Organizational Effectiveness Task Force: Final Report, 2005). By the mid-1990s, pressure for change was acute. The 1997 reforms tried to address these challenges through a new set of organizational arrangements, increased decentralization, and matrix management. The most frequent rationale for a matrix structure is to balance competing priorities, combine capabilities for market advantage, share resources for efficiency, and retain flexibility to redeploy resources in the face of changing priorities and a diversified client base. Typically matrix systems involve a creative tension between technical product delivery units and the units that service clients. The matrix system—a dual matrix, Bank-wide between the six Regions and four networks, and in each Region between Country Management Units (CMUs) and Sector Management Units (SMUs)—was to be facilitated by dual accountability for technical quality and an internal labor market for staff renewal and mobility. Figure 1 depicts the current matrix structure. It has been more than a decade since the 1997 reorganization, and concerns that the matrix system is not delivering on its promise persist. The objectives of the 1997 reforms were to create a new Bank culture through increased “client responsiveness� and delivery of “quality services.� Board documents indicate that client responsiveness was to be improved through a new country compact and strong country managers who would be closer to the client, hold budget authority, and be accountable for delivering the country program and ensuring results. High-quality services were to be achieved through strong sector groups working together as a network, sharing knowledge and expertise across sectors and the World Bank. These groups function through product teams providing client solutions and results, xi Overview Figure 1 Bank-Wide Matrix Structure (2012) President MD MD MD Networks Regions Central Vice Strategy Development Presidencies Delivery of Services to Clients Quality TENSIONS Lending AAA • Country vs. Corporate agenda Partnerships • Country directors vs. sector directors • Intersectoral tensions ar ising from sector silos Regional Matrix Knowledge • Knowledge dispersion vs. client demands for top quality knowledge HR Development CDs SDs Source: IEG. building the Bank’s global knowledge base, linking country and regional experience to global experience, and ensuring quality and accountability. The 1997 reforms were accompanied by a Strategic Compact with the Bank’s shareholders who approved an incremental $250 million for organizational renewal over 30 months. An external evaluation of the Strategic Compact found the design of the matrix to be a good one but identified several implementation weaknesses. These included variations in institutional design and procedures in the regional matrix systems leading to institutional complexity, and insufficient follow through on organizational design elements after the launch of the networks. Teamwork behavior, integral to the matrix, was being adversely affected by the internal market, generating competition for tight budgets. Despite actions taken in 2001 to address these constraints, many of the institutional and implementation gaps identified persist. Changes in the external environment indicate that the matrix system is even more relevant today than when it was introduced. Client needs have diversified, with greater differentiation among countries, even within the Regions; the growth of global public goods and corporate priorities is creating tensions and has given rise to new challenges which need to be reconciled with the country model; demand for cutting-edge knowledge is growing, both to enhance quality of lending and as a business line for policy and program advice to clients; and new global practices have emerged to meet needs such as information, communication and technology, and disaster management. The Bank’s ability to renew itself and function as a truly global Bank is critical to its success. xii Overview Evaluation Objectives and Methodology The objective of the evaluation is to assess the extent to which the dual objectives of the matrix system—enhancing client responsiveness and establishing strong technical networks to deliver quality services—have been attained, and have enhanced the Bank’s development effectiveness. The evaluation focuses on implementation of the current matrix system rather than on the 1997 matrix design and follows an objectives-based approach to assess the relevance and effectiveness of the matrix reform until 2010. To the extent feasible, the evaluation also examines the efficiency of matrix arrangements. The framework for the evaluation (Figure 2) depicts the theory of change and the three interrelated dimensions of the matrix reform. The system of incentives, structures, and processes were to support the country units and technical networks in achieving three main outputs: strategic alignment, a global knowledge bank, and accountability for quality, each with a set of subsidiary outputs. These outputs should lead to three intermediate outcomes on strategy, knowledge, and quality and, if these three are achieved, to two longer-term outcomes: (1) country programs and services that are timely and responsive to client needs; and (2) high-quality operations and advisory services and results based on global best practices. The evaluation follows the logical framework in Figure 2, focusing first on the three dimensions the matrix reform was intended to address and, to the extent that these are achieved, on the associated intermediate outcomes. It then goes on to examine the organizational and institutional impediments to the implementation of the reform. Evidentiary Basis In addition to a literature review and commissioned background papers, evidence for this evaluation comes from a wide range of available documents and data sources, including relevant Bank management assessments of the matrix, portfolio data, human resources data, and budget data. The evaluation draws on findings of previous IEG assessments, including Country Assistance Evaluations (CAEs), Country Assistance Strategy Completion Report (CASCR) reviews, and sector and corporate evaluations to identify matrix issues that had a bearing on the evaluation questions. The portfolio review covers the period 1991–2010. The analysis of trends in the sector and thematic composition of the portfolio are based on data from 5,280 lending operations and supplements approved in fiscal years (FY) 1995–2010, with a total commitment amount of $422 billion. The analysis of cross support and budget support covers FY02-10, since comparable data prior to FY02 does not exist. The commitment amount for projects during this later period, when the fully developed matrix system was in place, is $213 billion. Data on country programs is based on the analysis of all 96 CASCRs available as of December xiii Overview Figure 2 Matrix Management Evaluative Framework Incentives, Structures, Processes Decentralized CMUs/ HR processes for Quality Budget Reform and Networks/ Sector SMUs arrangements recruitment, deployment, assurance processes �nancially with Inputs empowered Boards regional and processes performance by QAG, regions CDs variations evaluation and networks Global Delivering Strategic knowledge quality alignment Bank services Sector strategies Knowledge produced Clear mandates and reflect ef�ciently by accountability ensure country and regional regions and effective quality priorities network anchors assurance Outputs Country strategies and Networks facilitate Budget incentives programs integrate access and use of promote effective sector and Bank and knowledge and corporate strategies global knowledge technical services Global and regional High quality expertise Matrix fosters priorities integrated mobilized across cross-sector in country programs regions and sectors collaboration and teamwork • Strategy: Country strategies and programs aligned with Bank priorities and client needs Inter- • Knowledge: Innovation in the �eld is well captured across the institution and country operations draw ef�ciently on mediate global knowledge to meet client needs Outcomes • Quality: Incentives and accountability ensure teamwork and high quality country dialogue and client services • Country programs and services are timely and responsive Long-term to client needs Outcomes • High quality operations, advisory services, and results based on global excellence Impact Greater development effectiveness in developing countries Source: IEG. xiv Overview 2010, spanning all Country Assistance Strategies (CASs) for 80 countries since introduction of the matrix system. Given the absence of baseline data, the evaluation undertook comparative analyses of findings from the universe of CASCR reviews to assess the responsiveness and effectiveness of country programs; time series data on budgets, staffing, and cross support, and thematic composition of the portfolio; incentives and constraints of regional matrix systems; data from manager interviews, a staff survey, and focus groups; and perceived strengths and weaknesses of the matrix system compared with those of the pre-1997 system. Limitations of the Evaluation Attribution of impacts to the matrix is problematic because of changes in the external environment and the overlapping effects of other elements of the renewal strategy—decentralization, business process renewal for budget and human resources management, restructuring of network Vice Presidential Units (VPUs) in 2006, and the quality assurance mechanisms outside the sectors and networks— that have evolved in parallel but have a bearing on the matrix system. Isolating the effects of the matrix system from these overlapping reforms is not feasible. While focusing primarily on matrix issues, the evaluation has acknowledged and dealt with these collinear effects to the extent that they have had an impact on the effectiveness of the matrix system. Strategic Alignment Alignment of client priorities with sector and corporate priorities is a relevant objective for maximizing the Bank’s development effectiveness. Central to the question of alignment of sector and corporate priorities with country programs is the relative power of country directors and sector managers and the differences in scope, time frame, and audience between sector and country strategies. By design, the country-based model is more responsive to short-term client demands within the three- to four-year time frame of a CAS than to longer-term sector and corporate priorities that take into account and should benefit clients with the global perspective of the Bank. In contrast, sector and corporate strategies are generic because of their global coverage and longer time frame, inhibiting alignment with country programs. Responsiveness and Effectiveness of Country Programs The Bank’s record of delivering responsive and effective country programs is moderately satisfactory or better two-thirds of the time. On aggregate, CAS outcomes since the introduction of the matrix system (covering country programs from FY1998-FY2008) have been moderately satisfactory or better in nearly two-thirds of country programs. The Bank has been most effective in responding to client government demands but less so in establishing strategic priorities that are realistic and have broader ownership or in integrating corporate priorities, such as gender and environment within country programs, xv Overview which from a global perspective have proven to be important to achieve development results. There is considerable variation in the performance of different components of country programs. During the entire period reviewed, country programs were better in terms of alignment with national priorities (96 percent moderately satisfactory or better) and flexibility (67 percent) when country circumstances changed. The high rating of country programs for alignment reflects responsiveness to the government’s demands or strategic priorities. However, realism (32 percent), the quality of the results framework (28 percent), and ownership (36 percent) lag behind, revealing systemic weaknesses in country programs which the Bank should address. About half of the country programs also suffer from weaknesses in implementation (54 percent). The low ratings on realism and broad country ownership in the CASCR reviews due to underestimation of country risks indicate significant gaps in identifying the countries’ development needs and constraints, and priorities of other key stakeholders. Overall performance of country programs improved over time. There were improvements in some components, while others remained flat or worsened slightly. CASCR reviews covering the period FY04-08 had an overall outcome rating of 72 percent, noticeably better than the outcome rating of 59 percent for FY1998-2003. Over this period, ownership improved from 18 to 55 percent. In the later period, ownership in International Development Association (IDA) countries was 71 percent, compared to 54 percent for International Bank for Reconstruction and Development (IBRD) countries, when fragile and conflict- affected states were excluded. Quality of the results framework also improved from 20 to 36 percent, but there was no improvement in realism and quality of portfolio implementation. Bringing a global perspective to delivery of country services The Bank invests considerable resources in developing sector and corporate strategy papers (SSPs). Actual expenditure on sector strategy development during FY02-10 was $118 million. During this period the Bank produced 11 sector strategies and two corporate strategies, yielding an average expenditure per sector and corporate strategy of almost $9 million. Of this, $3.8 million was expended by network anchors and the rest primarily by the Regions. With the CMUs having most of the operational decision-making power in the Regions, the sectors have tended to develop more elaborate SSPs with broader stakeholder involvement in an attempt to influence country programs. Certain characteristics of Bank-wide sector strategies inhibit their direct application in country programs. By design, SSPs are generic documents rather than specific to the diverse needs of partner countries because of their global coverage. They need to be complemented by regional sector strategies to be operationally relevant. Most SSPs do not include a typology of countries xvi Overview and guidance on how to apply strategic priorities to different country contexts. Consequently, they have little direct influence on the sector portfolio in country programs and adapting global sector knowledge into country perspectives remains challenging. Country and sector strategies have differences in their time horizon and audience. SSPs have a longer time frame (8–10 years) than the CAS cycle (3–4 years). SSPs typically discuss both long-term strategic issues and medium- term priorities, while country clients vary widely and expect customization of sector priorities to the country’s context and government priorities. More recent SSPs are better at internalizing country and regional experience than the first generation of SSPs. The review of SSPs reveals that recent SSPs have often been supported by regional sector strategies developed as inputs to the SSPs, which are better at reflecting country and regional experiences than were the first generation of SSPs. SSPs typically describe the state-of-the-art in that sector and serve four functions: providing a conceptual framework for the Bank’s work, stocktaking of experiences, generating shared understanding on sector priorities among anchor and regional staff, and communicating with external partners. Implementation at the country level is contingent upon the extent to which sector priorities are integrated in CASs since SSPs are not accompanied by resources for implementation. The country-based model makes it difficult to integrate sector and corporate strategies with country strategies. The Regions contribute to preparation of SSPs, but for the most part, CASs and country programs do not draw on them. SSPs lack ownership among country directors who exercise decisions over operational priorities and budget allocations. Country directors ration country budgets among different sectors based on their understanding of client needs during the CAS period, which may or may not coincide with the longer-term strategic priorities articulated in particular SSPs. And country directors have little incentive to finance cross-country work or strategic economic and sector work (ESW) with a longer time horizon. The principal means through which sector priorities are reflected in the CAS is through the direct involvement of sector staff members who contribute to CAS development. Less than a third of staff and managers report that SSPs have had significant impacts on CASs and country programs. Global and regional programs are not well integrated into CASs and country programs. In the 14 focus countries reviewed, global public goods were integrated in CASs to a very limited extent with a substantial presence in only four. IEG’s 2008 Annual Review of Development Effectiveness refers to the country-based model as a double-edged sword because historically global public goods are incorporated in country programs only when domestic and global benefits are aligned. The country- based model does not incentivize staff to internalize corporate goals. What is needed is a balance between what the client wants and the global perspective the Bank brings. When global and country interests diverge, additional incentives may be needed to increase client interest in global public goods. xvii Overview About one-third of staff and managers reported integration of global programs and an even smaller proportion of staff reported integration of regional issues in CASs and country programs to a substantial extent. The uptake of global and regional programs has been catalyzed by availability of trust funds or, in Africa, by earmarked IDA funds for regional programs. In some areas, such as climate change, there is a growing demand for Bank support. In others, such as governance and gender, there is a need for appropriate tools and support to adapt global agendas to country circumstances. The Promise of a Global Knowledge Bank In 1996, the concept of a Knowledge Bank was put forward both as the means to enhance the quality of World Bank lending and to provide a new stream of knowledge services to developing countries. The concept implied a shift toward capacity building and knowledge creation drawing on global and local knowledge to improve the quality of client services and achieve better results. The evidence from the flow of virtual knowledge and tacit expertise indicates that, rather than functioning as a global organization, the Bank is at risk of evolving into a group of regional banks interconnected by fraying ties among the Regions and between the Regions and networks. Decentralization has created further impediments to the flows of knowledge and expertise, and until the underlying incentives and constraints inhibiting effective knowledge flows are addressed, further decentralization will likely aggravate these problems. Management’s findings from their FY10 review of the matrix are that “innovation in the field is not well captured across the institution;� “not all country programs are able to draw on available knowledge;� “anchors and Regions are increasingly disconnected;� and “the network is not working.� Management found cross support between VPUs to be 3 percent of staff time, and labeled it “inadequate.� IEG’s findings on the Bank’s past performance as a knowledge bank, described further below, are largely consistent with these messages. The impact of analytical and advisory activities (AAA) is undermined by less effective dialogue and dissemination. Overall results from country AAA assessments undertaken from FY2000-09 by the former Quality Assurance Group (QAG) show that dialogue and dissemination, and consequent likely impact of AAA on government policy is significantly lower than strategic relevance and internal quality of AAA. Similarly, IEG found that only 55 percent of impact evaluations conducted by the Bank Group are disseminated to governments and clients (World Bank Group Impact Evaluations: Relevance and Effectiveness, 2012 forthcoming). Dissemination to clients was found to be a critical determinant of their influence on policy dialogue and on the design of subsequent operations. Further, of the 75 technical assistance activities in the QAG sample, 36 could not be evaluated, indicating substantially weaker management oversight of technical assistance products and services. xviii Overview Most staff, particularly those in the Regions and country offices, are unable to draw efficiently on knowledge generated inside and outside the Bank. Knowledge products are not stored in an easily searchable and retrievable form and are rarely used by staff outside the units where they are produced. As a result, and notwithstanding the analytical quality of the Bank’s AAA, much of the Bank’s knowledge has limited shelf life and use value. The knowledge produced by the Development Economics VPU (DEC) is widely disseminated to a global audience. But only 7 percent of operational staff report making direct use of DEC’s knowledge in operational work. The Bank generates a large volume of knowledge embedded in its operations but lacks the ability to efficiently capture and share that knowledge. The knowledge report (Knowledge for Development 2011: The State of World Bank Knowledge Services) estimates that the Bank spends $2.5 billion a year on embedded knowledge, four times the amount spent on core knowledge products. Embedded knowledge suffers from the same production inefficiencies as explicit knowledge, and is mostly buried within project documentation. Management does not have a system in place to capture, share or even track embedded knowledge across the Bank. Management of embedded knowledge is thus neither efficient nor effective. The Bank’s expenditures on core knowledge products increased by 50 percent from FY02 to reach $606 million in FY10, driven largely by reliance on trust funds, which has increased inefficiencies and risks. The Bank’s budget for knowledge remained steady as a share of the Bank’s overall budget over the FY02-10 period. In contrast, trust funds for core knowledge grew from $107 million to $245 million over the same period adding to the imbalance between country and sector units in support from the Bank’s own budget, owing to the greater reliance of sector units on trust funds. IEG’s evaluation of trust funds (IEG 2011) also raised concern over the growing use of Bank- executed trust funds as they are not allocated through the Bank’s strategic budget allocation process, which relies on the tensions between country and sector units to balance the needs of both sides of the matrix structure. Staff also report high transaction costs both in soliciting funds from multiple trust funds and in tracking and reporting separately on a myriad of different trust funds that have their own grant-making procedures and reporting requirements. Trust funds also play a significant role in the implementation of corporate priorities such as governance and gender, and global public goods such as climate change whose externalities extend beyond individual client countries and which may not otherwise be integrated into country programs. Sector silos persist, but regional and network silos are even stronger. The flow of tacit knowledge among staff through cross support has been limited, is declining in volume, and is constrained by structural boundaries. First, there has been a decrease in inter-VPU cross support from 7.6 percent of staff time in FY02 to 5.2 percent in FY10. Second, the cross support from network anchors to the Regions has fallen from 23 percent of network anchors’ staff time in FY02 to 12.5 percent in FY10. Third, less than 1 percent of regional staff time goes xix Overview to support other Regions. Fourth, even within the Regions, cross-sector work is not always carried out by cross-sector teams; cross support is less than half the volume of cross-sector work reported in project design suggesting that sector silos persist. Finally, cross support from the network anchors to the Regions and from the Regions to the network anchors, occurs within network silos (Figure 3). Innovation in the field is not well captured across the institution, and country operations do not draw efficiently on global knowledge to meet client needs. Operational staff in the Regions, especially those in country offices, complain about the ineffectiveness of the Bank’s role as a connector of knowledge. Innovations and knowledge generated at the country level, is rarely disseminated further. Thematic groups and communities of practice are the primary vehicles through which regional staff have been able to share tacit and embedded knowledge. Rather than improving their governance, most of them have been starved of resources and been allowed to attenuate. In the absence of efficient search engines for knowledge retrieval and mechanisms for knowledge sharing, decentralization creates further challenges. Decentralization has helped knowledge become more closely tailored to client needs and increase its impact. However, problems arising from weak knowledge flows have become more acute, with 54 percent of technical staff mapped to the Regions now being country-based. The dissemination of operationally relevant external knowledge was considered to be substantial by only 25 percent of staff surveyed, and internally generated knowledge by only 40 percent of staff. Managers interviewed were not more encouraging. Recognizing that personal networks are currently the most effective mode of Figure 3 Cross Support Under the Matrix System in FY10 0.9% Other Regions Other units 15.6% Regions in Region 18.3% total 0.6% Other VPUs 1.2% 11.4% 0.9% Other units 8.6% in Network Networks 21.1% total Other Networks 0.2% Source: World Bank. Note: The system records an additional 0.6% inter-regional cross support that in practice is cross support to a staff member’s previous unit within a year of departure. Denominator for Regions is total staff weeks in Regions; for networks it is total staff weeks in networks. xx Overview knowledge sharing, staff and managers alike lamented the loss of networking opportunities due to the scaling back of the annual or biannual gathering of sector staff known as “sector weeks.� In the absence of effective mechanisms for sharing of virtual knowledge, sector weeks have served as the primary forum for sharing new initiatives and cross-fertilization of operational knowledge, especially for country-based staff. They are also the most valued opportunity for country-based staff to network and demonstrate their ability to their sector colleagues, a prerequisite for mobility across the Bank. Both in efficiency and effectiveness the intermediate outcome of the Bank’s knowledge work—capturing innovation across the institution, and drawing on global knowledge to meet client needs—has not yet been achieved. Notwithstanding the quality of individual knowledge products, until the end of 2010, the matrix system had been unable to achieve the objectives of the global knowledge bank. Feedback from country surveys indicates that some clients value the Bank’s knowledge and policy advice even more than its financial assistance. Improving its knowledge portfolio is critical to the Bank’s future. The Bank has recently launched several initiatives to rectify these shortcomings. The creation of the Knowledge and Learning Council is an attempt to redress the weaknesses in the Bank’s knowledge agenda. The Council has increased corporate oversight of knowledge activities and launched several initiatives including Global Expert Teams and Knowledge Platforms. The restructuring of the Financial and Private Sector Development Network (FPD) into Global Practice Teams is also aimed at improving knowledge flows. The Council also oversaw production of the first-ever Knowledge Report, following which management has committed to actions to fill gaps in governance arrangements of individual knowledge products, establish a results framework for all core knowledge products, and establish mechanisms to promote greater connectivity and support management of knowledge as a portfolio. A more detailed implementation plan is under preparation. While these are steps in the right direction, the underlying incentives that effect demand for and use of knowledge will also need to change if these initiatives are to address the constraints that have hampered the knowledge bank. The Bank’s role in development knowledge is equally vital to the quality of its lending portfolio. In addition to knowledge services that have a direct bearing on policy advice and institutional development, there has always been a complementarity between knowledge and lending. IEG’s evaluation of the Bank’s use of knowledge to improve development effectiveness found that economic and sector work (ESW) and technical assistance was effective in shaping Bank lending and was statistically associated with better loan design (IEG 2008). The Bank’s ability to produce high-quality knowledge needs to be preserved to improve the quality of Bank lending, and its ability to access and customize cutting-edge global knowledge is essential to maintain its relevance to the increasingly complex demands of its client countries. Overview xxi Delivering Quality Services The matrix system was designed to strengthen the Bank’s ability to deliver quality services that combine relevant local solutions with global knowledge. This was to be achieved by the combination of country priorities articulated by country management and global technical expertise brought to bear by sector management and the Bank-wide networks. Outcomes of Bank-financed operations as measured by year of exit from the portfolio improved although the system is under increasing stress in recent years. The aggregate share of operations rated moderately satisfactory or better at exit after FY97 is significantly higher than before the matrix was introduced. But outcome ratings by exit year have declined from a peak of 79 percent moderately satisfactory or better in FY05-07 to 74 percent in FY08- 10 (Figure 4), which is statistically significant. Reporting outcomes by year of exit is a lagging indicator of quality at entry, reflecting performance of projects prepared an average of seven years earlier and affected in part by Bank supervision. When project outcomes are compared by year of approval, the most significant improvement in the share of Bank-financed operations with moderately satisfactory or better outcomes occurred before the introduction of the matrix, and a downward trend is evident after a peak in FY01 (Figure 5). Outcome ratings by year of entry show that the most significant increase occurred in projects approved before the introduction of the matrix. The slight gain for projects approved during the early years of the matrix (FY1998-01) was lost by projects approved after FY01. Using a three-year moving average, the share of operations rated moderately satisfactory or better among those approved during FY01-03 was 79 percent, compared to 70 percent among operations approved during FY04-05, a drop that is statistically significant with 95 percent confidence. The quality of supervision and quality at entry have shown a declining trend since the introduction of the matrix. In addition to assessing the outcomes of operations after they have closed, IEG assesses Bank performance in ensuring quality at entry and the quality of supervision. Data from the portfolio of investment projects from FY1990-2010 reveal that quality at entry peaked in FY96-98, while quality of supervision peaked in FY97-99. Portfolio indicators of supervision quality tracked by management also show a decline in Bank performance, despite recent increases in supervision budgets. These may be interpreted as leading indicators of portfolio quality. Multiplication of quality control layers for lending has increased neither efficiency nor effectiveness in ensuring quality. Accountability for quality has become more diffused with more cumbersome review procedures. Sector units, country units, and regional functions units, which undertake quality reviews on behalf of regional vice presidents, each have review processes, xxii Overview Figure 4 Outcomes of Bank-Financed Operations at Exit by Fiscal Year Percent moderately satisfactory 90 85 3 year average 80 75 or better 70 65 60 Launch of matrix 55 system 50 90 91 92 93 94 95 96 97 98 99 08 10 00 01 02 03 04 05 06 07 09 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 Figure 5 Outcomes of Bank-Financed Operations by Fiscal Year of Approval 90 Percent moderately satisfactory 85 80 75 or better 70 65 3 year average 60 Launch of matrix 55 system 50 90 91 92 93 94 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 19 19 19 19 19 19 19 19 19 19 20 20 20 20 20 20 20 20 20 20 20 Source: IEG ICR review database. sometimes with multiple levels. However, most of these reviews are conducted by generalists who do not have expertise in the relevant sector. Their role needs to be balanced by the role of sector units and network anchors. Systems for accountability and quality control are skewed toward avoiding risks. Regional functions units have grown substantially as they have strengthened their fiduciary and safeguards teams. For example, fiduciary staff in the Regions grew from 53 in FY02 to 407 in FY10. The growth during the first two years can be attributed to remapping of country unit staff to Operational Overview xxiii Service units reaching 158 staff in FY05 but the subsequent increase reflects heightened attention to fiduciary areas. A few high profile investigations of corruption and safeguards have resulted in widespread defensive behavior, leading many country directors and some sector units to add staff to undertake additional quality reviews focused on operational risks that could damage the Bank’s reputation. Feedback from staff and managers indicate that the Bank’s quality control systems are more concerned with fiduciary and safeguard risks than with other aspects of quality. These aspects of quality do not have to be mutually exclusive but the role and accountability of sector and network anchors would need to be enhanced to strengthen quality. There are insufficient incentives to focus on quality and results. Despite the intent to move away from a lending bank, the incentive to prioritize lending delivery for managers and staff alike is incontestable. Staff are clear that they receive more encouragement to meet lending targets than to focus on results. The pressure for end-of-fiscal-year delivery leaves little time to address quality issues identified through the review process. The peer review process to enhance the quality of projects and ESW is also affected by similar incentives. The selection of peer reviewers is not well managed and the peer reviewers are aware of the need to deliver products under tight deadlines. Final review meetings for ESW are often scheduled toward the end of the fiscal year, which undermines the potential value of peer review feedback, as there is hardly enough time to address substantial comments. The matrix system was put in place in part to ensure delivery of quality services by Bank-wide networks of sector specialists. Quality enhancement reviews were introduced by QAG in FY00 to focus on quality by enhancing the role of sector specialists. They were subsequently transferred to the networks and Regions but have become optional and are often not timely. The checks and balances created by a mechanism, independent of line management, to take periodic stock of the robustness of the Bank’s quality assurance system, disappeared with the dissolution of QAG in 2010. An alternate mechanism has not yet been put in place. Staff lack incentives to focus on quality since outcomes of operations and knowledge work are not linked to the Bank’s performance management system. Quality assurance systems under the matrix system continue to be more robust for lending operations than for ESW and weakest for non-lending technical assistance. Non-lending technical assistance has widely varying quality. QAG reviews of AAA are the main source of evidence on AAA quality used in the Knowledge Report. With QAG disbanded, the Bank has lost its sole mechanism for autonomous review of AAA quality. Sector managers have to cope with excessive span of control. The span of control of regional sector managers—managing 37 staff on average, compared to 7–15 in comparator external organizations and 19 in IFC—is unrealistic, making it impossible for them to provide quality control for all lending and non-lending operations produced in their unit. They are forced to rely on sub- xxiv Overview regional or country-based shadow managers, introducing an extra layer of non- technical review, but they are not always experts within the sector. Span of control issues are most acute in the Africa Region (54 staff per sector manager) and in the Sustainable Development Network (47 staff per sector manager). In FY12, management has announced its intention to reduce the span of control of sector managers to a maximum of 35 staff per manager. While this will reduce the workload of the most overburdened managers, it still remains two or more times higher than that of comparator organizations. Budget pressures are also adding risks to delivery of quality services. The first indicator on organizational effectiveness and modernization in the corporate scorecard released by management at the end of FY11 (World Bank for Results 2011) reports that the Bank spent 61 percent of its budget on “client services,� appearing to meet the Strategic Compact’s target of 60 percent on frontline services. This definition of client services includes trust funds and the cost of resource mobilization and management; without those the share consists of 47 percent of the Bank’s budget. Expenditures on more narrowly-defined country services—direct costs of lending and non-lending products delivered to country clients—are 32 percent of the Bank’s budget, or one in three dollars. The regional sector units, responsible for operational delivery and quality, are receiving a decreasing share of regional resources. While the Regions continue to utilize two-thirds of overall Bank resources, within the Regions, the units responsible for country services are under increasing budget pressure. Based on end-of-year expenditures, the share of regional sector units declined from 66 percent of regional Bank budgets in FY02 to 61 percent in FY10. The squeeze on sector unit budgets increases the risks to quality, especially in the context of escalating lending targets and the growth of more demanding, middle-income countries in recent years. To fill this budget gap, the Bank has become more reliant on external financing. A larger share of trust funds finances Bank instruments with weaker quality assurance systems—particularly non-lending technical assistance— compounding the quality challenge for knowledge and technical services. The Bank’s budget processes have not been effective in promoting cross- sector collaboration and teamwork under the matrix system. Financial and budgetary incentives play a significant role in discouraging cross support and cross-sector collaboration. The inability to share budgets across organizational units discourages collaboration across sector units. Budget rules create a cross support penalty and budget settlement rules vary across the Regions, creating further uncertainty. Delayed budget settlement and different norms for staff costs compared to consultant costs also discourage cross support. Institutional Issues and Organizational Structure The matrix is imbalanced. At first glance it appears that one of the two objectives— enhancing client responsiveness—has been achieved but that the sector side of the matrix did not succeed in creating global technical practices with world-class Overview xxv knowledge systems. This perception needs to be qualified by the fact that the very strength of the country units and, by extension, the Regions, created an imbalance in the matrix. Strong regional structures and incentives weakened the Bank-wide matrix, undermining the ability of the Bank-wide networks to mobilize the best global knowledge and expertise to serve regional operations. At the regional level, the imbalanced relationship between country and sector units gave primacy to short- term client demands over strategic country needs. This is reflected in the weak results on realism and ownership in country programs. Imbalances between the Regions and networks are also found in the Bank- wide matrix. The design of Sector Boards appears to have been flawed due to the inadequate division of labor between the sector managers and sector directors in board governance and quality. Accountability for quality is high for sector managers, but far less for sector directors. Regional sector managers are overburdened with a wide range of regional responsibilities and have little incentive, time, or resources to perform the multifarious corporate roles expected of them. Operational demands in the Regions make the functioning of the Bank- wide matrix a challenge. Incentives of managers are skewed toward the operational Regions where they are accountable for delivery of products and services. Sector managers and sector directors are paid for solely by, and accountable largely to, the Regions, with the exception of the newly created dual directors in the FPD network, who are co-financed by the network anchors. Although non-FPD sector directors have Bank-wide responsibilities toward the Sector Boards and Network Councils, those responsibilities take a backseat to their regional deliverables. Reporting lines for sector managers are primarily to their respective Region. Sector directors have dual reporting lines, but their primary responsibilities and accountabilities are also to their Region. Furthermore, the sector anchors have taken on additional corporate and global tasks that diverted attention from their ability to support frontline operations in the Regions. A more realistic sharing of responsibilities and accountabilities between the sector directors and sector managers might have enabled a more effective design of the sector units. Rather than functioning as a global institution, the Bank is at risk of evolving into six regional banks. Regional variations in structure and processes compound the problems arising from insufficient inter-regional flows of knowledge and expertise across the Regions. Sector configuration has varied across Regions and over time (for example environment is combined with water in one, urban with water in others) and the cost centers, some recording budgets at SMU level, others at sector department level. Norms and procedures for quality assurance also differ significantly. In East Asia, until July 2011, country directors chaired concept review meetings and sector directors chaired decision meetings; in other Regions sector directors had no formal role in quality assurance. Some Regions (such as South Asia and Africa) rely on quality enhancement reviews more systematically, others infrequently. xxvi Overview Subsequent changes added stress to the matrix design. The rigidity in Sector Board structures proved to be an impediment to the flexibility needed to address issues arising out of emerging operational priorities and subsequent organizational restructuring adding to the stress arising from inadequate knowledge flows and accountability for quality. The creation of the Sustainable Development Network (SDN) by merging the two largest networks created additional organizational challenges. To manage a network of this size, SDN added more layers to an already complex management structure but these differed across the Regions. SMUs were merged in some Regions aggravating the span of control issue for sector managers (Table 1). SDN also introduced an unrealistic range of sector responsibilities for sector directors and country sector leaders/coordinators. Since it is impossible for any individual to provide technical support over such a large range of sectors, a drawback that applies equally to sector directors, the quality oversight over SDN sectors in each CMU and the policy dialogue with clients tends to be undertaken by a generalist rather than a technical specialist. This is less true of other networks where the leadership role is exercised by technical staff over a smaller range of sectors that they are able to represent and oversee more effectively. Finally, as evidenced by IEG’s safeguards evaluation, in SDN incentives are geared toward heightened attention to safeguard issues compared to other aspects of environmental and social sustainability. In contrast, FPD—created at the same time as SDN through the merger of the two smallest networks—recognized the constraints on knowledge flows and quality. FPD was restructured at the end of FY11 into Global Practices designed to address these issues. While it is too soon to evaluate the effects of the Table 1 Selected Indicators for Networks VPU FPD HDN PREM SDN Total regional and anchor staff 565 904 827 2,308 (as of June 2011) Share of anchor staff/total staff 48% 23% 21% 30% Span of control 28 40 34 47 (staff/sector manager) Scope of control 6* 3 4 8 (Sector Boards/Global Practices) Network Anchor Expenditures 68 64 47 208 (FY10 BB + BETF actuals) (in US$ million) Share of BETFs in Anchor Expenditures 39% 43% 32% 52% Source: World Bank databases. Note: Data exclude the financial management and procurement sector boards of the OPCS Network. *FPD reorganized itself into six Global Practices at the end of FY11. FPD staff data include only IBRD head counts. Overview xxvii reorganization, or to assess its replicability across networks, FPD’s efforts to recognize and confront these issues are to be commended. Recommendations The many positive aspects of the 1997 reorganization in improving client responsiveness and country-level outcomes were not matched during more recent years by similar achievements in balancing knowledge and lending; country and sectors; and the Regions and networks. The matrix system, which was predicated on collaboration and teamwork across Regions and sectors, left intact or strengthened the incentives and underlying organizational culture, maintaining silos and competition among them. Changes in the global environment—diversification of client needs, a growing acknowledgement of the importance of global public goods and corporate priorities such as governance and gender, demand for cutting-edge knowledge for policy and program advice to clients, among others—suggest that a matrix system that marries a client focus to technical excellence is even more relevant today than when it was introduced. Achieving the objectives of the matrix system requires the removal of constraints to its implementation and the establishment of clear rules and incentives to improve development effectiveness. The Bank-wide matrix, between the network anchors and the Regions, has become increasingly weak and needs substantial strengthening, with some rebalancing needed within the Regions also. And the Regions need to be linked together more systematically by strengthening the flows of knowledge and expertise—such as among middle-income countries and fragile and conflict situations—to maintain the Bank’s comparative advantage as a global institution. The overarching IEG recommendation is to make substantial changes to the matrix system to realize its potential and reduce its downside. This should be achieved by restoring the creative tensions among the Regions and networks while allowing operational flexibility to support the development challenges of its clients and remaining responsive to external demands for global leadership. To this end, the sector and network architecture needs to be revisited and revitalized, and the distribution of responsibilities and accountabilities of sector directors and managers need to be reconsidered to ensure a balance between corporate sector responsibilities and regional needs. The relationship between country and sector units also needs to be rebalanced to ensure that the country-focused model is complemented by appropriate attention to regional and global priorities. And incentives for delivering quality services and results need to be strengthened. Management has recently introduced a series of steps to tackle many of these organizational constraints under the Matrix Leadership Team and the xxviii Overview Knowledge and Learning Council. A Managing Director has been tasked with providing leadership for the modernization agenda and several initiatives are underway. The empirical evidence in this evaluation is relevant for that decision-making process to identify additional actions needed to achieve the Bank’s strategic objectives. The unfinished task of designing incentives and operational processes that support operational needs must be addressed urgently. The main adjustments proposed by IEG are described below. Incentives Institutional incentives: Under guidance of the Knowledge and Learning Council, the Bank should enhance the incentives and resources for effective sharing and use of internal and external knowledge across the Bank. Develop and implement an integrated approach to enhance resources and incentives for documentation and sharing of key lessons from country experience, including through thematic groups, and for field-based staff to access and share knowledge. Develop and implement monitoring systems to track use of knowledge from relevant operations and AAA from other Regions, and from DEC, to assess AAA effectiveness and monitor its use by clients. Staff incentives: Management should revisit the performance management system to transparently reflect performance that goes beyond traditional deliverables by ensuring that staff are recognized for (a) sharing of knowledge and collaboration across sectors, and (b) the quality and results of operations for which they are responsible. Processes Redefine the purpose, process, and frequency of sector strategy papers as a more modest five-yearly report on the “state-of-the-sector� and provide adequate resources for strategic ESW and business development. Reflect in sector reports emerging good practice and operational experience in different country contexts to staff and external partners. Building on the experience of the Latin America and Caribbean Region, provide centralized funding within regional vice presidencies for strategic ESW, business development, and global and regional public goods independent of CMU budgets and work program agreements (either from a more substantial pool with the chief economist, or at sector department level). Quality: Strengthen mechanisms to enhance the focus on quality by (a) strengthening the peer review process, quality enhancement reviews, and mid- term reviews for quality through enhanced collaboration between Regions and sector anchors; (b) ensuring dual accountability of country and sector directors for quality; (c) strengthening quality assurance mechanisms for non-lending technical assistance; and (d) establishing a quality assurance mechanism for feedback on quality, independent of line management. Budget: To encourage cross sector and cross-regional collaboration, management should streamline and simplify the cross support budget Overview xxix settlement process and introduce flexibility in budgetary rules to allow sharing of tasks and budgets across organizational units. Structure Bank-wide matrix: Management should, by the end of 2012, develop and initiate implementation of a plan to enhance matrix effectiveness, address the existing constraints in the sector and network architecture, and strengthen the links among the Regions and between the Regions and anchors. This would need to address: fragmentation of sectors by regrouping Sector Boards with overlapping mandates under the same networks; excessive burden on sector managers by shifting greater responsibility and accountability for sector governance toward sector directors; inadequate accountability of sector directors/managers to the networks by strengthening incentives and the dual reporting system of sector directors/managers to both networks and Regions; and the inadequate role of sector/network anchors in supporting and ensuring quality of regional operations. Regional matrix: Rebalance the relationship between country and sector units and regional functions to ensure appropriate attention to lending, knowledge, and quality. This would need to address: the span of control of regional sector managers by reducing the number of staff per sector manager, learning from the experience of comparator organizations, and recognizing actual costs of shadow managers; and the decline in regional sector unit budget shares by enhancing resources for frontline services. Network rebalancing: Address key organizational constraints in the Sustainable Development Network. This includes: (a) excessive span of control of SDN sector managers; (b) dissimilarity in structures across the Regions; (c) unrealistic range of sector responsibilities of sector directors and country sector leaders/coordinators; and (d) removing the separation between safeguards and non-safeguards work by paying equal attention to both aspects of environmental and social sustainability. References IEG (Independent Evaluation Group). 2008. Using Knowledge to Improve Development Effectiveness: An Evaluation of World Bank Economic and Sector Work and Technical Assistance, 2000–2006. Washington, DC: World Bank. ———. 2011. Trust Fund Support for Development. Washington, DC: World Bank. xxx Overview Management Response Management welcomes the opportunity to engage with the Independent Evaluation Group (IEG) around management’s on-going modernization efforts, especially those concerning the reform and strengthening of the Bank’s matrix structure. The IEG review of the matrix concludes that the matrix system may be even more relevant today than when it was created in 1997. This conclusion reflects changes over the last 15 years, including the diversification of client needs, increased differentiation among countries and within Regions, the increasing role of global public goods (GPGs) and corporate priorities, and increasing demand for knowledge and cutting-edge mechanisms for transmitting knowledge. Management agrees with this conclusion and supports the general assertion that despite the effectiveness of the matrix system in improving country-level results, the matrix needs to be further strengthened to be even more effective in supporting the Bank’s clients. The IEG review of the matrix addresses challenges that were surfaced and analyzed extensively by management in recent years. In large part, it gives rise to the ongoing work on the modernization and knowledge agendas and provides welcome additional analysis and suggestions for further reform. It is most closely related to reforms under way to improve talent management, global knowledge and staff connectivity, and manager span of control. We appreciate that IEG has recognized the broader modernization efforts in this evaluation. As management has noted on previous occasions, this evaluation appears to significantly stretch IEG’s mandate of directly assessing development effectiveness. It focuses on the organizational structure as opposed to activities with explicit direct developmental goals and outcomes. As we noted then, this kind of evaluative framework would likely lead to recommendations that cross the line between the independent evaluation function and management’s function. Many of these fall under management’s purview and are closely related to an ongoing reform program. As agreed, management will be regularly discussing progress in matrix reforms with Executive Directors through its regular business modernization updates. In this broader context, we will continue to report periodically on implementation of the evolving modernization and knowledge agendas. This Management Response covers key recommendations of the evaluation but does not repeat the detailed analysis of key issues that gave rise to the modernization agenda.1 Strategic Alignment of Sector and Country Priorities The challenge of an effective matrix organization in an international institution is twofold. First, how can we most effectively bring corporate issues, notably GPGs, to the country level, especially when appropriate GPG economic incentives for countries are in short supply? The second is how to ensure 1  See Modernizing the World Bank Group—An Update (DC2011-0005), April 4, 2011. A further update is forthcoming in April 2012. Management Response xxxi the kinds of strategy discussions that lead to a cost-effective and selective support program—including a mix of support for a country-owned agenda and analysis and dialogue to bring objective evidence on issues identified as key for building interest and ownership over time (recognizing that the differences in fact are not actually that black and white). Those challenges predate the matrix and, in fact, the inherent tensions of the matrix have led to a better approach to these challenges. The re-balancing of the Bank’s development work since 1997 has given greater prominence to agendas defined by the Bank’s borrowers (the primary clients of our development cooperative). The Bank has aligned much of its work with this changed world and has helped bring developing country concerns more directly into the mainstream of global discussions. However, at the operational level, management agrees that there is room for improvement in the interaction of sector, country, and project teams in the performance of quality assurance and support functions (discussed in more detail below). These improvements will be sought in a variety of ways, which will be reflected in the definition of network accountabilities and their monitoring, as well as in the strengthening of quality assurance processes at the project level. The key accountabilities of sector management and country management for critical control points in the processing of investment lending and other products have been clear, although hampered by issues of span of control of the sector managers. Management has now addressed this issue, with the implementation across all Regions of the decision to limit the span of control of sector managers to no more that 35 staff. Management will assess as part of its work on corporate quality assurance whether further reconsideration of accountabilities—for example assigning a more direct accountability to sector directors at specific points of operations—might be needed. Management agrees that it is a good time to take stock of the Bank’s experience with sector strategy papers. That said, there have been a number of changes made in recent years to the process and content of sector strategies and it would have been useful for IEG to have reported on and analyzed this modernization work. Management will undertake a review of sector strategies and develop a proposal for their future, including further content and process changes. The IEG suggestion to examine the “state of the sector� and capture emerging good practices and operational experiences will be considered in that context. Knowledge As acknowledged in the IEG review, management agrees that appropriate incentives to improve internal knowledge flows require changes to management practices, business priorities, and resource allocation. The Bank is working on such changes in incentives as part of the implementation of the first Knowledge Report and the Knowledge Strategy. The main reform is system changes to economic and sector work (ESW) and technical assistance (TA) xxxii Management Response to enable the capture of lessons learned across the Bank. In addition, building on existing databases (for example Knowledge Notes and the Development Impact Evaluation initiative database), the Bank is currently developing a database of lessons learned in operations. Management has formed a working group to map existing knowledge organizational structures that will inform the implementation of this work. Management has also decided to implement a 10-point program for strengthening Sector Boards and global practices that focuses on ways to improve the sharing of and access to knowledge. A sector practice survey of all sector-mapped staff is being used to prioritize the key areas of change. Management views decentralization as critical for the Knowledge Strategy. The IEG report does not provide convincing evidence that decentralization has impeded flows of knowledge. In fact, decentralization ensures that knowledge work is closely aligned with client needs and improves client ownership and thus impact. Management stresses that knowledge flows are not only North to South and the Bank has done a great deal to strengthen the focus and capacity to facilitate those important and growing South-South flows. Incentives and Accountability for Quality Management continues to give the highest priority to technical quality. There is room for improvement of the peer review and quality enhancement review processes, for example, by making these reviews timelier, involving external peer reviewers, clarifying sector accountability for technical quality, and supporting the organization of these reviews in accessing appropriate technical skills. Management also agrees with the recommendation to improve quality assurance mechanisms for ESW and non-lending TA and is actively pursuing actions in this direction. The corporate quality assurance approach that management is now developing will soon (within the next year) include all knowledge work. Management has clarified responsibility for overseeing, coordinating, and facilitating the implementation of the knowledge management agenda, including strengthening the governance and coding for these products. This builds on the decisions coming out of the ESW/TA review, already under implementation, which has defined control points for the processing of non- lending outputs, with clearance now required from sector and country management, both sides of the matrix. With regard to a quality assurance mechanism for feedback on quality independent of line management, it is not foreseen at this point to introduce a new “independent� office or function for the purpose of quality assurance. Experience has shown that such a unit was not effective in providing the needed just-in-time support that truly enhances quality. Instead, attention will focus on continuing to strengthen and support the front line in delivering quality across the board in Bank support. Nevertheless, as part of the ongoing review of Corporate Quality Assurance, management is considering an appropriate Management Response xxxiii framework for an accurate and timely gauge of operational quality for use at the corporate level. Management agrees that staff rewards should be based on results achieved, but has reservations about tightly linking staff performance with the achievement of operational development outcomes. Because the Bank supports clients’ policies and programs, the recommendation raises an attribution problem and potentially discourages staff from taking risks or working in fragile and conflict-affected situations. Management notes that it is already planning an extensive reform of the performance management system—including replacement of the current overall performance evaluation (OPE) form—that will reflect a realignment of incentives with due attention paid to knowledge. Also, the recent changes to the governance of analytical and advisory activities (AAA) will focus on well-defined objectives and outcomes of AAA. This includes how these will be assessed, for example, through consultation with client countries. Systematic capture of these aspects in the system will form the basis for developing processes to recognize staff contributions and provide incentives. Organizational and Institutional Issues Management agrees that more clarity is needed on the roles of sector directors and sector managers. Management is examining their roles and responsibilities to ensure a more balanced division of labor, particularly with regard to sector governance. Building on the work of the Finance and Private Sector Development (FPD) pilot, management has agreed that sector directors and managers will report to both regional and network anchor sides of the matrix. Management considers that all managers (indeed all staff) have a corporate role and that sector managers have a uniquely important role in this regard. Sector manager Bank-wide governance functions are critical to improving horizontal management, effective talent management, strategy development and implementation, as well as effective knowledge sharing and learning. Management strongly supports the idea that we need to develop a more corporate management of talent mobility through Sector Boards or global technical practices, an approach that encourages knowledge transfer. Management has been working on ways to make the process of budgeting work across Regions and networks simpler and more flexible. Any simplification of the budgeting of this kind of work also hinges on critical inputs such as: (i) clear accountability frameworks for networks and strengthening the role of Sector Boards: (ii) evaluation of the FPD network pilot, which includes cross support changes; and (iii) clear incentives for managers and staff. At this time, these critical inputs are still being developed. However, management intends to make changes designed to facilitate work across Regions and networks as soon as feasibly possible. xxxiv Management Response Management believes that the current Sector Board configuration serves current institutional purposes. However, management also recognizes that some changes may need to be made as business lines change. As mentioned above, management is examining the current structure of the various knowledge entities in the Bank (for example, Global Expert Teams, Communities of Practice, and Thematic Groups) to see how they fit together and how they should evolve over time. A working group will develop an integrated approach with respect to four key issues: (1) governance rules; (2) financing; (3) incentives and resources for staff; and (4) synergies and alignment across these knowledge structures. Management Response xxxv Chairperson’s Summary: Committee on Development Effectiveness The Committee on Development Effectiveness (CODE) considered the Independent Evaluation Group (IEG) report entitled The Matrix System at Work: An Evaluation of the World Bank’s Organizational Effectiveness and the Draft Management Response. An Approach Paper entitled The Matrix System at the World Bank: An IEG Evaluation was endorsed by CODE on January 11, 2011. Summary CODE members welcomed IEG’s timely and informative evaluation, and agreed with many of IEG’s conclusions and recommendations. They noted the broad alignment between management and IEG positions on the general assertion that, the matrix system has been effective in improving client responsiveness, but needs to be further strengthened through better integration of corporate, regional, and sector priorities with country programs to serve client needs even better. They acknowledged that the findings are consistent with the conclusions of the matrix review previously undertaken by management, which was the basis of the Bank’s Modernization Program and Knowledge Agenda currently under implementation and previously endorsed by the Board. Members encouraged management to proceed vigorously with these reforms and adjust them to take into account IEG’s findings and recommendations. In particular, they requested that the IEG evaluation’s recommendations be reflected in the next update on the Modernization Program. The importance of innovation through new instruments, such as Program-for-Results (PforR), was noted. Members noted that the matrix is influenced by the underlying incentive structure, decision-making processes, systems, and the culture of the Bank and reinforced the importance of management’s actions in these areas to make the complex matrix system more effective in the broader context of modernization. IEG emphasized that the matrix system is even more relevant to the Bank’s mission today than when it was introduced in 1997, but requires fine-tuning. The IEG evaluation found improvement in the responsiveness of country programs, but found that the goal of creating a global knowledge bank has fallen short, and country operations do not draw sufficiently on global knowledge to meet client needs. Regional and network silos have replaced sector silos, with incentives skewed toward the Regions, and portfolio quality has been declining. Some incentives and processes continue to be misaligned with the matrix objectives, which inhibits development effectiveness. IEG acknowledged management’s initiatives to address the imbalances in the matrix, which risk undermining the Bank’s comparative advantage, and indicated that the recommendations of the evaluation are meant to support and sustain momentum of management’s reform agenda. xxxvi Chairperson’s Summary Management broadly agreed with many of IEG’s conclusions, which provide valuable input to the Bank’s ongoing Modernization and Knowledge Agendas. Management cautioned that the matrix should not be considered in isolation, and that management is implementing measures that reach far beyond the matrix, which are designed to create a more agile, responsive, and innovative Bank culture. The essential building blocks of such cultural transformation are talent and performance management; global mobility and connectivity; uniformity across the Regions; lending reforms; review of delegation of authority to establish clear accountabilities and uniformity across the institution; consolidating processing issues to more clearly define the roles of the anchors; and aligning sectors across the Regions. The Knowledge report, considered by CODE in September, sets out the agenda for changing the management of knowledge products.  Anna Brandt Chairperson’s Summary xxxvii Abbreviations AAA Analytical and advisory activities AFR Africa Region ARDE Annual Review of Development Effectiveness ARPP Annual Report on Portfolio Performance BETFs Bank-Executed Trust Funds CAE Country Assistance Evaluation CAS Country Assistance Strategy CASCR Country Assistance Strategy Completion Report CASCRR Country Assistance Strategy Completion Report Review CD Country Director CFR Corporate Finance and Risk Management CM Country Manager CMU Country Management Unit CODE Committee on Development Effectiveness DEC Development Economics DPL Development Policy Loan EAP East Asia and the Pacific Region ECA Europe and Central Asia Region ESSD Environmental and Socially Sustainable Development Network ESW Economic and Sector Work FCS Fragile and Conflict-Affected States FPD Financial and Private Sector Development Network FY Fiscal Year GEF Global Environment Facility GET Global Expert Team GPG Global Public Good HDN Human Development Network HR Human Resources IBRD International Bank for Reconstruction and Development ICR Implementation Completion and Results Report ICRR Implementation Completion and Results Report Review IDA International Development Association IEG Independent Evaluation Group IFC International Finance Corporation IL Investment Loan INF Infrastructure Network ISR Implementation Status and Results Report KLC Knowledge and Learning Council LCR Latin America and the Caribbean Region xxxviii Abbreviations M&E Monitoring and Evaluation MD Managing Director MNA Middle East and North Africa Region MLT Matrix Leadership Team NGO Nongovernmental Organization OETF Organizational Effectiveness Task Force OPCS Operations Policy and Country Services OS Operations Services PREM Poverty Reduction and Economic Management Network PPAR Project Performance Assessment Report PSG Public Sector Governance RETFs Recipient-Executed Trust Funds SAR South Asia Region SB Sector Board SD Sector Director SDN Sustainable Development Network SM Sector Manager SMU Sector Management Unit SSP Sector Strategy Paper QAG Quality Assurance Group QERs Quality Enhancement Reviews VP Vice President VPU Vice Presidential Unit WPA Work Program Agreement WBI World Bank Institute Abbreviations xxxix Glossary Term Definition Anchor sector The head of a sector anchor who also chairs the respec- director tive sector board; in most instances the director of a sector anchor heads a sector department, however, in the Sustainable Development Network the mandate of anchor sector directors extends beyond a single sector Bank-wide matrix The relationship between regional vice presidencies and network vice presidencies, sometimes referred to as the “big matrix� Country director The head of a Country Management Unit (CMU) re- sponsible for the Bank’s strategy and program in one or more countries and having budgetary authority over all operations for those countries; the country director reports to the regional vice president and is a member of the regional management team Country manager The administrative head of a country office reporting to the country director of the relevant CMU Country management The administrative unit for one or more countries unit tasked with preparation and implementation of the country assistance strategy, policy dialogue, budget- ary oversight, approving work program agreements with Sector Management Units (SMUs), and operation- al delivery to the countries under that CMU Country program The principal liaison for the CMU who performs vari- coordinator ous country director functions in Washington, and may have shared responsibility for strategy formula- tion, implementation, country team coordination, and relations with clients and partners Country sector Coordinator of all the sectors under the jurisdiction leader/country sector of a regional sector department for a CMU. Co-located coordinator (CSL/CSC) with the country director mostly in the field, the CSL/ CSC has oversight over more than one sector and re- ports to the regional sector director in most instances; however, there are notable regional differences1 1. The Latin American and Caribbean region, where this model is the most developed as it was introduced at the inception of the matrix, has “sector leaders� who are responsible for work program agreement (WPA) coordination, quality assurance and policy dialogue, and mentor- ing and staff oversight; in the East Asia and Pacific region for Sustainable Development Net- work sectors the same three functions are performed by sector managers in the three larger CMUs and sustainable development leaders in the three smaller CMUs, while the Poverty xl Glossary Cross support The staff time of an expert or specialist purchased from outside the responsible unit for a specific task; cross support is a measure of matrix functionality and organizational flexibility to achieve flows of knowl- edge across different units Cross support sold The amount of cross support provided by the staff of one administrative unit for tasks managed by other responsible units Cross support The amount of cross support purchased by a respon- purchased sible unit from staffs of other units Dual directors Directors of the Financial and Private Sector Development (FPD) Network, each of whom has respon- sibility for one region and one Practice Group. Each di- rector reports to and is funded in equal parts by the FPD Network Vice Presidency and the relevant regional Vice Presidential Unit (VPU) to ensure balanced atten- tion to global knowledge and operational agendas Extended Leadership The management team of each region, chaired by Team the regional vice president (RVP) constituted by the Regional Management Team and the sector managers and country managers of that region Global Practices (GP) New organizational structure of FPD Network organized around 6 technical areas. The pillars of the GP design are new knowledge management, information technol- ogy, and talent management systems, coupled with in- creased staff mobility, the elimination of current silos, and a stronger focus on results monitoring. Inter-VPU cross Cross support provided by staff of one VPU to respon- support sible units in any other VPU Intra-VPU cross Cross support provided by staff to other responsible support units within a single VPU Network An aggregation of all the sector families combined under a network vice presidency, including the entire staff community housed in the network anchors and the staffs under the regional sector departments Reduction and Economic Management (PREM) Network has a lead economist and the Human Development Network (HDN) has CSLs in each CMU; the Europe and Central Asia, the Middle East and North Africa, and South Asia regions fall in between with the CSL/CSC’s role largely confined to WPA coordination. Africa is the most varied with different sector departments organized differently. When the CSL/CSC functions only as a part-time coordinator they may report to their own regional sector manager rather than the regional sector director. See Ap- pendix G for a detailed comparison of the regions. Glossary xli Network anchor All staffs housed in a network vice presidency includ- ing the network vice president’s (NVP)’s front office and the sector anchors for sectors falling under that NVP Network Council The governing body of a network, constituted by the sector directors of regional sector departments and di- rectors of sector anchors; the Network Council is chaired by the relevant NVP Network vice The administrative head of a network who also chairs president the Network Council; currently the Bank has four NVPs Network Vice Same as network anchor, which is sometimes called the Presidential Unit network VPU The PM serves as the chief operating officer of the Practice managers (PM) Global Practice. She or he serves as the “right hand� of the FPD director for all global practice activities. The PM helps formulate practice strategy, conducts portfo- lio reviews, handles corporate mandates, serves as a broker for project staffing, and serves as a “secretariat� for the Practice Management Team (PMT) Region One of six geographic clusters of countries which serves as the primary organizational unit responsible for op- erational delivery Regional Management The management team of each region, chaired by the Team RVP constituting all the directors of CMUs, sector de- partments, front office units, and the chief economist of that region Regional matrix The relationship between country management units and sector management units within a region, some- times referred to as the “small matrix� Regional sector The administrative unit containing a front office and department all the relevant SMUs that fall under the jurisdiction of a network Regional sector The head of a regional sector department; the RSD re- director (RSD) ports both to the NVP and the RVP and is a member of the Network Council and the Regional Management Team Regional sector The head of a SMU responsible for the delivery of the manager SMU’s work program, the quality of all SMU operational products and services, and management of SMU staff; the regional sector manager is a member of the rel- evant Sector Board, representing his or her region on the Sector Board, and is also a member of the Extended xlii Glossary Leadership Team in the region; in some cases sector managers have responsibility for more than one sector Regional vice The administrative and operational head of a region president (RVP) who also chairs the Regional Management Team; cur- rently the Bank has six RVPs Regional vice Same as the region, sometimes called the regional VPU president unit Regional functions Responsibilities and budgets under the RVP, not man- aged by the CMUs, SMUs or regional sector depart- ments; this includes at a minimum the front office, chief economist’s office, strategy and operations, and the operational services and quality unit Sector anchor Secretariat of the Sector Board located within the net- work anchor Sector Board The governing body of a sector, constituted by the managers of the regional sector management units and the sector anchor, as well as representatives of other relevant vice presidencies; the sector board is chaired by the director of the relevant sector anchor and is responsible for the sector strategy, quality includ- ing knowledge, learning and operational quality, and human resources management of sector staff includ- ing staff recruitment, mentoring, deployment, and promotion Sector family All the technical staff working in a sector housed in the sector anchor or the regional sector management unit responsible for that sector, or working on that sec- tor within one of the central vice presidencies Sector management The administrative unit in a region responsible for the unit work program, staff, and operational delivery of one, or sometimes, more than one sector; most SMUs cover the entire region, however, in East Asia and the Pacific, three SMUs cover all eight sectors of the Sustainable Development Network in individual CMUs Service line SLMs manage teams of FPD experts (thought leaders managers (SLM) and “gurus�) who are expected to (a) provide global knowledge and best practice in the design and struc- turing of operations; (b) create and manage a technical community of experts, within FPD and beyond; and (c) are accountable for FPD quality globally Glossary xliii Shadow managers Senior staff acting as deputies who have been delegat- ed responsibilities by regional sector managers; they may have geographic responsibility over a sub-region (as in Africa HDN) or across the region for regional sec- tor managers with responsibility for multiple sectors (as in the Sustainable Development SMUs in Africa and East Asia and Pacific) Task team leader The person responsible for management and delivery (TTL) of individual lending or non-lending tasks and who is accountable for the quality and timeliness of product delivery Vice president unit The administrative unit overseen by a single vice presi- dent including all the staff, budgets, and operations under their purview Work Program Agreement between SMUs and CMUs that specify the Agreement lending and non-lending deliverables and supervi- sion tasks for which the SMUs will be accountable and the budget they are allocated by the CMUs for those deliverables xliv Glossary Chapter 1 Introduction and Context Chapter • Highlights • Many independent evaluations have found that internal World Bank constraints are barriers to achieving development outcomes. • Organizations adopt matrix systems to help balance competing objectives and to better manage interdependent activities; successful matrix organizations continuously review their organizational effectiveness and readjust internal power imbalances as necessary to enhance their ability to achieve their strategic goals. • For the Bank, these competing objectives are responsiveness to client needs on the one hand, and technical excellence in operations, analytical work, and advice on the other. • The institutional aspects of reform—incentives, processes, and organizational culture—are far more challenging than changes in organizational structure, even in the private sector. • The objective of the evaluation is to assess the extent to which the twin objectives of the matrix system—enhancing client responsiveness and establishing strong technical networks to deliver quality services—have been attained and have enhanced the Bank’s development effectiveness. • The evaluative framework for the matrix system focuses on outputs and intermediate outcomes along three dimensions: strategic alignment, the global knowledge base, and delivering quality services. Rationale and Purpose The World Bank’s ability to deliver effective country operations and services is contingent upon responsiveness to the client and technical excellence. The Bank has made several attempts, most notably in 1987 and 1997, to strengthen its ability to achieve these twin objectives. The 1997 renewal was motivated largely by recognition that the Bank’s development programs were excessively driven by a culture of lending, with insufficient attention to client needs and the quality of results, which are crucial to development effectiveness.1 The reforms included an organizational restructuring and change management process and brought the word “matrix� into the Bank’s organizational vocabulary. The resulting structure consists of a dual matrix— Bank-wide between the six regional vice presidential units (VPUs) and network VPUs, and in each Region between Country Management Units (CMUs) and Sector Management Units (SMUs). Several recent Independent Evaluation Group (IEG) evaluations found that organizational constraints stood in the way of better development outcomes. For instance, evaluations of the Bank’s work on social safety nets (IEG 2011), agriculture and agribusiness (IEG 2010a), climate change (IEG 2009a), and health, nutrition, and population (IEG 2010b) indicated that there are few incentives for teams to work across sectors, and the “narrow subsectoral view� adopted by staff leads to weaker development outcomes. The problem is not new—the same issue was cited in the 2006 Annual Review of Development Effectiveness that looked back on previous IEG reports. Other 2 The Matrix System at Work evaluations—including the 2008 Annual Review of Development Effectiveness and evaluations on the environment (IEG 2009b) and the global economic crisis (IEG 2010)—highlighted the tension that arises when the country-based model is called upon to contribute to the supply of global public goods (GPGs) and address other cross-country issues. Evaluations also raised concern about the internal pressure to lend and resulting preference for operations with large commitments, noted in evaluations of governance and anticorruption (IEG 2011) and climate change (IEG 2009a). This lending pressure has also resulted in an insufficient focus on the sharing of knowledge, as noted in evaluations of GPGs, agriculture and agribusiness (IEG 2010a), transport (IEG 2007), and knowledge (IEG 2008). IEG is evaluating the Bank’s matrix system, at the request of the Committee on Development Effectiveness (CODE),2 as a contribution toward strengthening the Bank’s effectiveness in assisting its client countries. The primary motivation is to build on the findings from IEG’s previous evaluations to assess the relevance and effectiveness of the Bank’s matrix system and its ability to provide high- quality services to its clients in order to inform the Bank’s ongoing reforms. The objectives of the 1997 reforms were to create a new Bank culture through increased client responsiveness and delivery of quality services. Client responsiveness was to be improved through a new country compact and strong country managers who are closer to the client, hold budget authority, and are accountable for delivering the country program and ensuring results. Quality services were to be achieved through strong sector groups working together across the World Bank. These sector groups would deliver quality through product teams tasked with providing client solutions and results, building the Bank’s global knowledge base, and ensuring quality and accountability.3 The evaluation examines the extent to which the twin objectives of enhancing client responsiveness and establishing strong technical networks to deliver quality services have been attained and have enhanced the Bank’s development effectiveness. The evaluation focuses on the implementation of the current matrix system rather than on the 1997 matrix design, and it follows an objectives-based approach to assess the relevance and effectiveness of the matrix reform. To the extent feasible, the evaluation also examines the efficiency of matrix arrangements. In this evaluation, the terms “matrix system� and “matrix reform� refer to both the organizational structure and the accompanying institutional rules and behaviors. Adapting the Bank to Changing External Realities The 1997 reforms were aimed at adapting the organization to changing circumstances and addressing concerns among external stakeholders about the role of aid in development. External circumstances changed dramatically in the 1990s with the increasing importance of private financing of development leading to a decreasing share of official development aid, including Bank financing, in total flows to developing countries. This trend has continued after Introduction and Context 3 a slight interruption by the Asian financial crisis. In 1987, World Bank lending represented 15 percent of all external financing for developing countries. By 2002, Bank lending had declined to 4 percent of external financing. Other changes included the opening of previously closed economies as well as a shift from a demand for development financing toward knowledge services to help client countries deal with increasingly diverse challenges across Regions, and even within the Regions. In hindsight, the introduction in 1997 of a “matrix� organization to manage the different priorities of the country and sector units was only to be expected. The previous 1987 reform had strengthened the country focus, but quality remained a concern.4 Country Departments housed technical staff under division chiefs. Hierarchical relationships between country and sector units and vertical Country Department structures (“silos�) hindered knowledge flows and affected operational quality. By the mid-1990s, pressure for change was acute. The 1997 reforms tried to address these challenges through a new set of organizational arrangements, increased decentralization, and matrix management arrangements that featured dual reporting in the sector units and budget authority with the country units. Thus, the introduction of matrix management was a major element of the 1997 changes, but not the only change (see Appendix I). It has been more than a decade since the 1997 reorganization, and concerns that the matrix system is not delivering on its promise persist. The fundamental questions that this evaluation hopes to elucidate are whether the matrix system remains suited to today’s circumstances, and whether current incentives and accountability systems support the aims of the Bank’s organizational architecture. The 1997 Matrix System Matrix organizations are characterized by their managerial arrangements, which are divided into product delivery units and client units. Individuals in matrix organizations are grouped by functions—for instance, sector experts at a development bank—and report to a product manager responsible for that function or sector. Staff members are identified for particular products and services to be delivered to a client—a government implementing a project or reform program—which is overseen by a client manager. Individuals thus have dual reporting lines, accountable to both the product managers and client managers. The purpose of this arrangement typically is to allow for better information sharing across task boundaries, better customization of services to the needs of clients, and better management of career progression in particular product areas. In the Bank’s matrix system, the client managers are called country directors and country managers; the product managers are called sector managers.5 The 1997 matrix reform presumes that to be effective, the Bank’s operational work must be tailored to client needs and based on global technical 4 The Matrix System at Work excellence to provide quality services. Country units for understanding client needs and sector units to deliver on global technical expertise represent the two sides of the Bank’s structure. Encouraging conducive “matrix behaviors�— flexibility to work across Regions and sectors, dual reporting lines for sector staff and managers, responsiveness to diverse client needs, and the need to “get the plumbing right� through incentives and supporting budget and human resources management systems—are prerequisites for the matrix to deliver on its promise. The matrix system was an essential element of the 1997 renewal strategy, intended to ensure consistency between reform objectives and the organization’s delivery system for operations (World Bank 1997, 2001). Four technical networks were created, comprising 19 Sector Boards to manage sector staff and programs. Each Sector Board was supported by a secretariat, known as the “sector anchor,� in the network VPU, which was usually, but not always, a separate sector department.6 The matrix system was to be facilitated by dual accountability for technical quality and an internal labor market for staff mobility and renewal. An autonomous Quality Assurance Group (QAG) was established to provide independent feedback on portfolio quality. Reform implementation was intentionally kept flexible. There was little a priori agreement on how the organizational culture would be adapted. Issues would be addressed during implementation through a participatory learning process, with indicators to monitor progress and share lessons. The design permitted variations in structure and process among regional VPUs, some of which have since been standardized while others continue to vary. The Bank’s ability to ensure a critical mass of high-quality technical expertise and essential flows of cutting-edge knowledge is vital to ensuring development effectiveness in a setting that is more decentralized. The number of Country Departments increased from 20 to 45, with almost 90 percent of country directors based in-country by fiscal year (FY) 2010, compared to 5 percent in 1997. Increasing numbers of sector staff are also co-located with the country director. The 1997 reforms were accompanied by a Strategic Compact with the Bank’s shareholders, who approved an incremental $250 million for organizational renewal over 30 months. The main purpose was that in return for a short-term restructuring investment, the Bank would be reorganized in a way that would provide cost savings and improved output to achieve greater effectiveness in its development goals (World Bank 1997; Phillips 2009: 38). The four priority areas were: • Refueling current business activity to ensure resource allocation to front- line services (which had been cut 13 percent over the previous three years) to protect the level and quality of client services. • Refocusing the development agenda on social and environmental sustainability, governance, and partnership with the private sector. Introduction and Context 5 • Retooling the Bank’s knowledge base through establishment of the sector networks and the capacity-building work of the Economic Development Institute (now renamed the World Bank Institute or WBI). • Revamping institutional capabilities through a more integrated information system, human resource strategy, financial management system, and increasing decentralization. An external assessment of the Strategic Compact found the design of the matrix to be a good one, but identified several implementation weaknesses.7 The regional matrix organizations lacked many of the common elements of institutional design and procedures that led to institutional complexity and implementation challenges. Each Region created different interfaces with the networks. For staff crossing the Regions, there were different processes and practices to which they needed to adjust. There was some concern that too many country directors were created, leading to fragmentation of the budget and over-programming as the unit of accountability became smaller. There was insufficient follow-through on organizational design elements after the launch of the networks. Accountability for contracting and quality assurance was not addressed and teamwork and the internal market appeared to be working at cross-purposes. Teamwork behavior, integral to the matrix, was being adversely affected by competition for tight budgets. Corrective actions were taken in 2001 to address these constraints, but many of the institutional and implementation gaps identified persist. The fundamental structure, including the regional structure put in place in FY97 (Figure 1.1), is still intact. CMUs are responsible for client relations, preparation of Country Assistance and Country Partnership Strategies (CASs), and allocation of the Bank administrative budget to sector units based on work program agreements (WPAs) negotiated with them. SMUs are responsible for managing lending and nonlending tasks agreed with the respective country unit and for ensuring quality. The regional and network VPUs (and the Sector Boards) are within the matrix. Operations Policy and Country Services (OPCS) is both within the matrix—it houses the financial management and procurement Sector Boards, which are in the matrix, and also oversees Country Services, the other axis of the regional matrix—and outside, due to its multiple corporate roles. OPCS also has a counterpart Operations Services (OS) unit in each regional VPU that is responsible for ensuring compliance with Bank policies and procedures and overall portfolio monitoring. The Development Economics (DEC) VPU and WBI play a significant role in knowledge products and services but are outside the matrix. With the exception of the consolidation of some networks in 2006 to form the current structure (see Appendix I) the matrix system is intact. Significance of the Matrix System Relevant, effective, and efficient systems to deliver lending operations and knowledge services are of enormous significance to the Bank’s clients. During FY 02-10, when the matrix system was fully in place and for which 6 The Matrix System at Work Figure 1.1 Bank-Wide Matrix Structure (2012) President MD MD MD Networks Regions Central Vice Strategy Development Presidencies Delivery of Services to Clients Quality TENSIONS Lending AAA • Country vs. Corporate agenda Partnerships • Country directors vs. sector directors • Intersectoral tensions ar ising from sector silos Regional Matrix Knowledge • Knowledge dispersion vs. client demands for top quality knowledge HR Development CDs SDs Source: IEG. Note: The original structure had a joint office of managing directors overseeing the Regions and the networks. Over the past five years, one managing director was designated to oversee the networks while two were responsible for the Regions, the latter two being consolidated into one in FY12. comparable budget data are available, the Bank approved $222 billion in International Bank for Reconstruction and Development (IBRD)/International Development Association (IDA) lending and grants, $17.4 billion in recipient- executed trust funds (RETFs), and used $2.7 billion on core knowledge products. With such amounts at stake, small efficiency gains can substantially enhance the Bank’s impact. Almost two-thirds of the Bank’s own budget is administered through matrix arrangements. Of the Bank’s average annual budget for FY08-10 of $2.4 billion, the regional VPUs expended 53.4 percent; the network VPUs (network anchors) accounted for 9.4 percent; and OPCS, DEC, and WBI vice presidencies expended 5.2 percent of the Bank’s budget. The remaining 32 percent is for other corporate vice presidencies and corporate functions. The flow of Bank- executed trust funds (BETFs) is quite different from Bank budget allocation (Figure 1.2). With trust funds, the share of funds administered through matrix arrangements increased to 59 percent since the regional and network VPUs together absorb 85 percent of BETFs. The matrix system directly affects 72 percent of the Bank’s staff. In FY11, of 11,402 total Bank staff, 6,235 were located in the six regional VPUs and 1,470 in the anchors of four networks (Finance and Private Sector Development, or FPD; Human Development, or HDN; Poverty Reduction and Economic Management, or PREM; and Sustainable Development, or SDN). An additional 641 staff were in DEC, WBI, and OPCS (see Appendix M).8 Introduction and Context 7 Figure 1.2 Matrix-Related Bank Budget Allocation and Bank-Executed Trust Fund Mobilization (FY10 Data) RVP Totals (BB $1,183m) (BETF $257m) Reg Functions CD/CMUs SD/SMUs (BB $213m) (BB $246m) (BB $724m) (BETF $9m) (BETF $20m) (BETF $227m) BB Total BETF Total (BB $2,462m) (BETF $508m) Regional OPCS (BB $95m) (BETF $5m) To Other Units Chief To Other Units (BB $1,069m) Economist (BETF $74m) Sector Depts Other NVP Totals (BB $210m) (BETF $177m) Source: World Bank databases. Notes: Solid lines highlight the Bank’s budget allocation system; dotted lines indicate unofficial and trust fund sources of financing that must be mobilized outside the budget allocation system. Figures in italics are amounts held within that unit as opposed to total amounts flowing through a given unit. Scope and Evaluative Questions Scope of the Evaluation The evaluation has reviewed the Bank’s recent experience with matrix management to assess its relevance, effectiveness, and efficiency,9 as well as its effects on development effectiveness. • Relevance (of objectives and design) is defined as the extent to which the objectives are consistent with current development priorities, strategies, and corporate goals, and the extent to which the design is consistent with the stated objectives. • Effectiveness is defined as the extent to which the objectives were achieved, or are expected to be achieved, taking into account their relative importance. 8 The Matrix System at Work • Efficiency is defined as the extent to which the costs involved in achieving organizational objectives were reasonable in comparison with their benefits. The evaluation has focused on current matrix arrangements Bank-wide and at the regional level, including organizational structures, procedures, and institutional behaviors. Given the primacy of the country unit as the Bank’s delivery model, the evaluation reviews both the Bank-wide matrix between the Regions and network anchors, and the regional matrix system between the country units and sector units. The 1997 matrix design was deliberately kept flexible and it has continued to evolve, adapting to internal and external changes. Given the regional variations in implementation of the matrix, the evaluation has adopted a comparative approach to the extent feasible, relying heavily on analysis of data from existing Bank databases and prior IEG evaluations, supplemented by evidence from staff and manager surveys, informed opinions of knowledgeable interlocutors, and a review of eight focus sectors and 14 focus countries that were selected because of the availability of recent IEG evaluations. The matrix and knowledge reforms introduced in 2010 are too recent for their implementation and results to be meaningfully evaluated. However, IEG has examined the issues raised by the 2010 management assessment. Although too early to evaluate its impacts, the changes and proposed reform actions have been considered and related to the findings of this evaluation. Decentralization and labor market issues are distinct from matrix issues but affect its functioning and have therefore been addressed in the context of flows of expertise and knowledge. The evaluation undertook in-depth analysis of cross support data (available for FY02-10) to assess its effects on the flow of global knowledge and cross-sector collaboration. The review of different forms of decentralization adopted by the regional VPUs was also instructive. However, the broader range of issues around decentralization and labor markets were not explicitly evaluated and would need to be addressed separately. The 1997 internal reform agenda consisted of several interrelated dimensions and was intended to include changes in incentives, organizational structures, and budgetary, human resource, and quality processes. Five areas of responsibility were specifically defined for the networks: strategy, staffing, knowledge, quality, and partnerships. Country units were responsible primarily for the country strategy and country program. Achieving better results in these areas is assumed by IEG to have been a principal objective of the reorganization. Specifically, the matrix was expected to facilitate: • Preparation of well-defined and realistic sector and thematic corporate strategies and alignment of country programs to those strategies. • Hiring, placement, management, rotation, and exit of technical staff as needed to meet the Bank’s operational needs. • Design and implementation of mechanisms to identify and manage knowledge to ensure timely and efficient access to relevant information by operational staff. Introduction and Context 9 • Enhancement of the quality of operational products. • Stronger relationships with partners to help achieve operational results. The evaluative framework for the matrix system focuses on outputs and intermediate outcomes along three dimensions: strategic alignment, the global knowledge base, and delivering quality services. The framework for the evaluation (Figure 1.3) depicts the theory of change and the three interrelated dimensions of the matrix reform. The system of incentives, structures, and processes was to support the country units and strong technical networks in achieving three main outputs: strategic alignment, a global knowledge bank, and delivering quality services. Through a set of subsidiary outputs and intermediate outcomes on strategy, knowledge, and quality, these outputs were designed to achieve two longer-term outcomes: (a) country programs and services that are timely and responsive to client needs; and (b) high-quality operations, advisory services, and results based on global best practice. These are interlinked both in the original objectives of the matrix reform10 and in their contribution to development effectiveness (Figure 1.3). The first dimension links the Bank-wide matrix to the regional matrix and the country level. The second dimension is largely the responsibility of Bank-wide networks in response to country, regional, and global demands for knowledge. The third dimension manifests itself most at the regional level, although Bank-wide networks also have a role in quality assurance. Evaluation Questions The overarching question for this evaluation is: “To what extent has the matrix system, including the organizational structure and the incentives and accountability mechanisms it created, contributed to development effectiveness by fostering strategic alignment, harnessing of knowledge, and quality assurance?� The questions to be examined under the three dimensions of this evaluation are derived from the evaluative framework and provided below: • Strategic Alignment: The matrix structure was intended to ensure both responsiveness to the client and alignment with corporate and sector objectives. The matrix was also conceived as a way of eliminating “silo� behaviors and encouraging cross-sector collaboration. The key question on strategy is “How effective is the matrix system in aligning country programs with Bank-wide sector and thematic strategies and client needs?� The sub-questions are: 1. To what extent do sector strategies reflect country and regional experiences? 2. How effectively do country strategies and programs draw on sector strategies, sector experience, and global best practice? 10 The Matrix System at Work Figure 1.3 Matrix Management Evaluative Framework Incentives, Structures, Processes Decentralized CMUs/ HR processes for Quality Budget Reform and Networks/ Sector SMUs arrangements recruitment, deployment, assurance processes �nancially with Inputs empowered Boards regional and processes performance by QAG, regions CDs variations evaluation and networks Global Delivering Strategic knowledge quality alignment Bank services Sector strategies Knowledge produced Clear mandates and reflect ef�ciently by accountability ensure country and regional regions and effective quality priorities network anchors assurance Outputs Country strategies and Networks facilitate Budget incentives programs integrate access and use of promote effective sector and Bank and knowledge and corporate strategies global knowledge technical services Global and regional High quality expertise Matrix fosters priorities integrated mobilized across cross-sector in country programs regions and sectors collaboration and teamwork • Strategy: Country strategies and programs aligned with Bank priorities and client needs Inter- • Knowledge: Innovation in the �eld is well captured across the institution and country operations draw ef�ciently on mediate global knowledge to meet client needs Outcomes • Quality: Incentives and accountability ensure teamwork and high quality country dialogue and client services • Country programs and services are timely and responsive Long-term to client needs Outcomes • High quality operations, advisory services, and results based on global excellence Impact Greater development effectiveness in developing countries Source: Adapted from IEG Approach Paper. Introduction and Context 11 3. How effectively are corporate strategic priorities, including GPG agendas and regional and sub-regional issues, reflected in country strategies and country programs? • Knowledge Production and Dissemination: The key question on knowledge is: “How effective is the matrix system in enabling the Bank to distill and use country experience and global knowledge to meet client needs?� The sub-questions are: 1. How effective have the networks been in harnessing and making available country experience and knowledge across Regions? 2. How effective have the networks been in facilitating the flow and adaptation of global knowledge to Regions and countries? 3. How effective has the matrix system been in creating the flexibility for technical staff to work across Regions and sectors, and in enabling managers and staff to work effectively across dual reporting lines? • Delivering Quality Services: One of the objectives of the matrix system was to shift the culture from inputs and lending to quality and results. The ultimate objective of the Bank is to promote development through high-quality advice and financing. The key question in this respect is: “Are incentives and accountability mechanisms aligned in a manner that rewards high quality in client services and results?� The sub- questions are: 1. Are incentives and accountabilities of country and sector managers and mandates of Sector Boards well defined and designed to ensure quality assurance and results? 2. Do financial and budgetary incentives related to the matrix system promote effective knowledge and technical services? 3. Have incentives fostered the cross-sector collaboration and teamwork needed to achieve results in country operations? Literature Review on Matrix Organization Relevant Literature on the World Bank Much of the literature on the Bank’s internal reform focuses on organizational culture as the impediment, or facilitator, of past reform efforts (Phillips 2009, Wade 2001, Weaver and Leiteritz 2005). The literature views the post-1997 Bank as more responsive to its clients (for example, through fast-disbursing loans with greater ownership and less conditionality) but falling short in other intentions, such as demonstrating development impact and sustainability (Mallaby 2004, 2005; Easterly 2001, 2002; Stiglitz 1999, 2002; Pincus and Winters 2002; Gutner 2005b) and adhering to its own mandates, such as safeguards. Organizations adopt matrix systems to help balance competing objectives and to better manage interdependent activities. Successful matrix organizations continuously review their organizational effectiveness and readjust internal power imbalances, as necessary, to enhance their ability to achieve their goals (Gottlieb 2007, Galbraith 2009). The Bank’s matrix system, 12 The Matrix System at Work launched in 1997, continued to evolve through 2001. The current budget and staff data recording systems were introduced in 2002. Significant structural changes were made in 2006, merging some networks; however, this did not involve additional change management processes. The Bank’s unique status as a global, public sector entity constrains organizational innovation. First, as an international organization, the Bank must respond to the demands of multiple stakeholders (donors, borrowers, nongovernmental organizations, and academics) that provide it with legitimacy and resources; demands that are often in conflict with each other (Nielson, Tierney, and Weaver 2006; Gutner 2005a; Wade 2005). Second, the strong internal culture of the Bank—as in most large organizations—is slow to change and requires ample time, money, and skillful leadership at all levels (Finnemore and Sikkink 1998, Burke 2002, Bebbington and others 2006). Consequently, the internal Bank reforms that mesh well with its “preexisting intellectual and operational culture� have more traction than those that do not (Weaver 2008). Insights from the Private Sector The most frequent rationale for a matrix structure in the private sector is to balance competing priorities, combine capabilities for market advantage, share resources for efficiency, and retain flexibility to redeploy resources in the face of changing priorities. Competing goals—in particular the need for globally competitive products and services and customization to meet client needs—are inherent in an organization with a global presence that simultaneously wants to service different market niches. By the 1990s, this was certainly true for the World Bank if it was to stay relevant across country clients with increasingly divergent needs. Some of its clients faced the challenges of post-conflict reconstruction; others sought help in transition from central planning to open market economies; low-income countries grappled with the vicious cycle of poverty and low growth; and middle-income countries faced second-generation challenges of emerging markets. The matrix system offered a way to reconcile global and local priorities. There are several known tradeoffs in a matrix structure. Matrix solutions involve compromised accountability due to dual reporting relations. The need for collaboration often lengthens decision-making. There is greater potential for risk avoidance. Matrix systems are complex and often require greater management oversight. Organizational effectiveness is not just about structure. When organizations attempt restructuring, all the elements that make up the organization (people, process, structure, systems) and how they interact must be considered by the design process. Organizational design efforts fail to deliver expected benefits if they are not treated as a comprehensive change process or if there is hidden misalignment. The design process must enable all of these organizational elements to be aligned to the purpose and strategy of the organization. Introduction and Context 13 The institutional aspects of reform—incentives, processes, and organizational culture—appear to be far more challenging than changes in organizational structure, even in the private sector. Incomplete implementation, when formal changes in structure are not accompanied by changes in the softer, more stubborn issues of organizational culture and incentives is a common weakness. Galbraith (2009) argues that changes in strategy, people, rewards, processes, and structure are essential for effective matrix reform. Others have pointed to unclear lines of accountability, competitiveness and internal power struggles, and proliferation of bureaucratic layers (Gottlieb 2007), as well as the inability to rebalance power across the matrix, reluctance to adopt collaborative behaviors, and lack of systems and processes to support the reforms as recurring challenges in matrix organizations. Sy and D’Annunzio (2005) identify five top challenges for matrix organizations: misaligned goals, unclear roles, ambiguous authority, lack of a matrix guardian or champion, and perpetuation of silos.11 A review of private sector experience reveals five critical success factors in implementing matrix structures. While a wide range of institutional factors and processes influence the effectiveness of any organization, the following five factors appear to underpin the success of matrix organizations: • Goals are aligned and tracked across the organization, • Responsibilities and expectations are clear to all, • Collaborative culture and leadership, • Mature resource management processes, and • A robust performance management framework. Matrix organizations thrive on transparency, accountability, and synergies across these success factors. However, these go far beyond changes in organizational structure. They require sustained attention to ensure that the changes are internally consistent and periodic rebalancing to maintain responsiveness to evolving business needs. Bank Management ’s Self-Assessment of the Matrix System Previous assessments12 by Bank management resonate with those from recent IEG evaluations.13 Since 1997, the Bank’s matrix system has been assessed by management several times. The most intractable problems were identified as: (a) the persistence of sector “silos� leading to weak cross-sector collaboration and untapped synergies; (b) disconnects between Board-approved sector strategies and country operational programs; (c) limited staff mobility and sharing of knowledge and expertise across Regions; and (d) dispersed and poorly defined managerial accountabilities, resulting in inadequate quality control. A 2001 review of the Strategic Compact led to some adjustments in budget and human resources (HR) systems to improve incentives, but integration of country and sector strategies and cross flow of people and knowledge remained problematic. In FY03-04, QAG carried out an assessment of Sector Boards.14 The QAG assessment identified several matrix issues, including: the governance and effectiveness of Network Councils and Sector Boards, the relationship between 14 The Matrix System at Work global agendas and regional operations, the proliferation of Sector Boards, and weaknesses in HR management.15 A review by an Organizational Effectiveness Task Force (OETF) in FY04 and FY05 identified similar constraints: (a) excessive fragmentation of tasks across organizational units with overlapping mandates; (b) inadequate leveraging of global knowledge; and (c) suboptimal deployment of staff and budget resources. These recommendations were not acted upon at the time.16 In FY10, Bank management undertook a review of the matrix system, which reiterated many of the same issues and added one more concern arising out of the growth of trust funded activities and global programs not well connected to country programs. Management identified five priorities for matrix reform:17 • Mobilize top talent and global knowledge for country clients. • Create strong and rejuvenated technical practices. • Increase staff mobility. • Ensure clear accountability, especially of networks and Sector Boards. • Strengthen corporate management through the establishment of a Matrix Leadership Team of regional and network vice presidents, chaired by one of them on a rotational basis, to manage matrix issues. In parallel, Bank management undertook a review of the Bank’s knowledge agenda. The review identified a series of systemic weaknesses and proposed a new knowledge strategy18 with three pillars: developing global practices to promote mobility of staff and knowledge, managing the Bank’s knowledge products as a portfolio to ensure greater impact, and strengthening the Bank’s role as a global connector of knowledge. A Knowledge and Learning Council (KLC) composed of vice presidents and headed by a managing director has been established to oversee the Bank’s knowledge portfolio and has issued a report on the state of the Bank’s knowledge.19 The KLC will also manage the annual strategic knowledge budget and ensure accountability for results. The three pillars of the knowledge agenda are integral to the original conceptual framework of the matrix. Recognizing the linkages between matrix and knowledge issues, management has also established a Reform Secretariat, housed within OPCS, to act as a joint secretariat for the Matrix Leadership Team and the KLC. This evaluation also considers knowledge issues inseparable from the matrix system, but it focuses on the effectiveness of knowledge processes rather than on the quality of the Bank’s knowledge products. Evaluation Methodology Evidentiary Basis Data sources. In addition to a literature review and commissioned background papers, evidence for this evaluation comes from a wide range of available Introduction and Context 15 documents and data sources, including relevant Bank management assessments of the matrix, Business Warehouse portfolio data, HR data, and budget data. The evaluation draws on findings of previous IEG assessments, including Country Assistance Evaluations (CAEs), CAS Completion Report (CASCR) reviews, and sector and thematic evaluations to identify matrix issues that had a bearing on the evaluation questions. The methodology is described in Appendix B. The portfolio review of project outcomes and trends in the sector and thematic composition of the portfolio are based on data from all 4,812 Bank lending operations approved in FY1995-2010, which had a total commitment amount of $441 billion. The analysis of cross support and budget support covers about half of that period, comparable data before FY02 being unavailable. The commitment amount for projects during this later period, when the fully developed matrix system was in place, is $213 billion. Comparative analysis. Given the absence of baseline data, the evaluation undertook comparative analyses of findings from the universe of CASCR reviews to assess the responsiveness and effectiveness of country programs; time series data on budgets, cross support, portfolio outcomes, and thematic composition of the portfolio; incentives and constraints of regional matrix systems; data on sector staffing; data from manager interviews, the staff survey, and focus groups; and perceived strengths and weaknesses of the matrix system compared with those of the pre-1997 system. Limitations of the Evaluation Attribution of impacts to the matrix is especially problematic in this case because of the overlapping effects of other elements of the renewal strategy that have continued to evolve in parallel. • First, most country directors were decentralized to the field (40 of 45 were decentralized by 2010), which had a significant bearing on the effectiveness of the matrix. • Second, business processes for managing budget and HR were integrated through SAP software in 2001, enabling systematic storage and retrieval of data from FY02 but generating unanticipated constraints on operational units. • Third, in 2006 two new networks—SDN and FPD—were created by combining previous networks without anticipating or assessing the impact of this organizational restructuring on the matrix. • Fourth, QAG was dissolved in 2010, placing greater burden for quality assurance on the matrix system. Isolating the effects of the matrix system from these overlapping process reforms is not feasible. While focusing primarily on matrix issues, the evaluation has acknowledged and dealt with these collinear effects to the extent that they have had an impact on the effectiveness of the matrix system. 16 The Matrix System at Work Organization of the Report The report is organized in six chapters. The next three chapters analyze the three vertical dimensions—strategy, knowledge, and quality—from the evaluative framework (Figure 1.3). Chapter two discusses the alignment among sector and global strategies and country strategies and programs, followed by the findings on responsiveness and effectiveness of country programs from the analysis of IEG’s CASCR reviews. Chapter three is an in-depth assessment of knowledge issues, including the effectiveness of knowledge services based on the staff survey and manager interviews, a detailed analysis of cross support, and an analysis of budget data relevant to the matrix system. Chapter four reviews the incentives and accountability for delivering quality lending products and nonlending services based on reviews of the Bank’s portfolio, the focus countries, the focus sectors and the regional matrix profiles; and the findings from the staff survey, manager interviews, and focus group discussions. Chapter five discusses cross-cutting organizational structure and institutional issues that are impeding the Bank’s effectiveness. Chapter six provides recommendations for strengthening the Bank’s organizational effectiveness. Additional findings from the background papers and analyses are summarized in the form of appendixes to this report. Notes 1 See Wapenhans 1992, 1994; and Pincus and Winters 2002. 2 This request was reiterated at an informal meeting of the Board on March 16, 2010, where Bank management presented its proposed matrix reform. 3 Board briefing on Change Management, January 1996. 4 The Wapenhans Report (1992) and first QAG reviews, which cover projects approved before 1997, indicated that about one-third of Bank projects were not likely to have satisfactory outcomes. 5 On the sector side, primary responsibility for management of the Sector Boards, which govern the sector groups and sector staff, rests with sector managers and not with sector directors. 6 The term “anchor� is also used loosely sometimes to refer to the network VPU, which consists of the vice president’s front office and the sector anchors within that VPU. 7 “Assessment of the Strategic Compact.� Internal World Bank document. March 13, 2001. 8 The remaining staff are in other central units including Administrative and Client Support, Communications, Concessional Finance and Global Partnerships, Controller’s unit, Corporate Finance and Risk Management, Corporate Secretariat, External Affairs, General Services, Health Services, Human Resources, Independent Evaluation Group, Information Solutions Group, Information Management and Introduction and Context 17 Technology, Integrity Vice Presidency, Internal Audit Vice Presidency, Legal and Treasury. T he International Development Research Centre recommends using 9 four measures to assess performance of development organizations: relevance, efficacy (or effectiveness), efficiency, and sustainability/ v iability (Lusthaus and others 1999). The first three dimensions are financial ­ more pertinent to the assessment of the World Bank’s matrix system. To the extent that it is relevant, the fourth dimension is endogenous to the first three. This evaluation will therefore focus on the first three dimensions. 10 While related staffing implications of those three dimensions may be touched upon, the evaluation will not attempt to assess the full spectrum of human resources management issues. 11 See Thomas Sy and Laura Sue D’Annunzio. “Challenges and Strategies of Matrix Organizations: Top-level and Mid-Level Managers’ Perspectives.� Human Resources Planning, vol. 28, No. 1, March 2005: 39–48. Compact� (2001), 12 Previous assessments include the “Assessment of the Strategic ­ ­ rganizational E the QAG Assessment of Sector Boards (2004), O ­ ffectiveness Task Force findings (November 2004), and Bank ­ management’s recent diagnostic (December 2010). 13 Past IEG evaluations have identified significant organizational and institutional constraints emanating from the matrix system that are likely to affect development effectiveness. These include: (a) difficulties in linking corporate and sector strategies with country-driven operations (Gender and Development 2009) and global programs with country programs (ARDE 2009); (b) the key role of knowledge and data, and weaknesses in sharing knowledge about what works (Climate Change and the World Bank Group: Phase I 2009); (c) the erosion of technical expertise and need for stronger staff skills, especially those relevant to new business needs (Growth and Productivity in Agriculture and Agribusiness 2010; The World Bank Group’s Response to the Global Economic Crisis: Phase I 2010); (d) inability to fully exploit cross-sector synergies for development effectiveness (Improving Outcomes for the Poor in Health, Nutrition, and Population 2009; Environmental Sustainability 2008); (e) weak accountability for implementation (Safeguards and Sustainability Policies in a Changing World 2010). 14 Quality Assurance Group, FY02 Annual Report on Portfolio Performance, p. 21. 15 See (World Bank 2004b) from Approach Paper. 16 In 2006, the ESSD and INF networks (the two largest) were merged into the SDN network, which houses almost two-thirds of regional staff and manages two- thirds of the Bank’s investment lending portfolio. 17 These priorities are summarized in the presentation, “Towards a More E ­ ffective Matrix� (World Bank 2010d) that was discussed with the E ­ xecutive Directors at the Informal Board Meeting on March 16, 2010. 18 The Matrix System at Work 18 (World Bank 2010b, 2010c) from Approach Paper. 19 See The State of World Bank Knowledge Services—Toward Open Knowledge. August 17, 2011. References Easterly, William. 2001. The Elusive Quest for Growth: Economists’ Adventures and Misadventures in the Tropics. Cambridge, Massachusetts: MIT Press. Galbraith, Jay R. 2009. Designing Matrix Organizations That Actually Work: How IBM, Procter & Gamble, and Others Design for Success. San Francisco: Jossey-Bass. Gottlieb, Marvin R. 2007. The Matrix Organization Reloaded: Adventures in Team and Project Management. Westport, Conn.: Praeger Publishers. Gutner, Tamar. 2005a. “Explaining the Gaps between Mandate and Performance: Agency Theory and World Bank Environmental Reform.� Global Environmental Politics 5 (2): 10–37. ———. 2005b. “World Bank Environmental Reform: Revisiting Lessons from Agency Theory.� International Organization 59 (3): 773-83 IEG (Independent Evaluation Group). 2008. Using Knowledge to Improve Development Effectiveness: An Evaluation of World Bank Economic and Sector Work and Technical Assistance, 2000–2006. Washington, DC: World Bank. ———. 2009a. Annual Review of Development Effectiveness: Achieving Sustainable Development. Washington, DC: World Bank. ———. 2009b. Climate Change and the World Bank Group Phase I: An Evaluation of World Bank Win-Win Energy Policy Reform. Washington, DC: World Bank. ———. 2010a. Growth and Productivity in Agriculture and Agribusiness: Evaluative Lessons from World Bank Group Experience. Washington, DC: World Bank. ———. 2010b. Gender and Development: An Evaluation of World Bank Support, 2002–08. Washington, DC: World Bank. ———. 2010c. The World Bank Group’s Response to the Global Economic Crisis: Phase I. Washington, DC: World Bank. ———. 2011a. Social Safety Nets: An Evaluation of World Bank Support, 2000–10. Washington, DC: World Bank. ———. 2011b. Governance and Anticorruption: Evaluation of World Bank Support, 2002– 08. Washington, DC: Mallaby, Sebastian. 2004. The World’s Banker: A Story of Failed States, Financial Crises, and the Wealth and Poverty of Nations. New York: Penguin Press. Introduction and Context 19 Nielson, Daniel, Michael J. Tierney, and Catherine Weaver. 2006. “Bridging the Rationalist-Constructivist Divide: Re-Engineering the Culture of the World Bank.� Journal of International Relations and Development 9 (2):107–39. Phillips, David A. 2009. Reforming the World Bank: Twenty Years of Trial and Error. New York: New York University Press. Pincus, Jonathan and Jeffrey Winters, eds., 2002. Reinventing the World Bank. Ithaca, NY: Cornell University Press. Stiglitz, Joseph E. 2002. Globalization and Its Discontents. New York: W.W. Norton. Sy, Thomas and Laura Sue D’Annunzio. “Challenges and Strategies of Matrix Organizations: Top-level and Mid-Level Managers’ Perspectives.� Human Resources Planning, vol. 28, No. 1, March 2005: 39–48 Wade, Robert H. 2001. “The World Bank as a Necessarily Un-forthright Organization.� Paper presented at G24 Meeting, 17–18 April 2001, Washington, DC. ———. 2005. “International Organizations and the Theory of Organized Hypocrisy: The World Bank and Its Critics.� DESTIN Working Paper. London School of Economics. http://www.lse.ac.uk/collections/DESTIN/workingPapers.htm. Weaver, Catherine and Ralf J. Leiteritz. 2005. “Our Poverty Is a World Full of Dreams: Reforming the World Bank.� Global Governance 11:369–388. Weaver, Catherine. 2008. Hypocrisy Trap: The World Bank and the Poverty of Reform. Princeton, NJ: Princeton University Press. World Bank. 1997. The Strategic Compact: Renewing the Bank’s Effectiveness to Fight Poverty. 13 February 1997. Washington, DC: World Bank. ———. 2001. “Assessment of the Strategic Compact.� Matrix Management. Strategy and Resource Management Vice-Presidency. March 13, 2001. Washington, DC: World Bank. 20 The Matrix System at Work Chapter 2 The Strategic Alignment of Sector and Country Priorities Chapter Highlights  lignment of client ­ • A priorities with sector and ­ priorities is a relevant ­ corporate ­ objective for maximizing the Bank’s development e ­ ffectiveness.  esponsiveness and ­ • R country needs has been ­ effectiveness in meeting ­ moderately sat- isfactory or better in nearly two-thirds of country programs since the introduction of the matrix system.  hortcomings persist in several areas, realism of the Bank’s country programs—the ex- • S tent to which programs match country capacity, estimate ownership, and are sufficiently selective—being the weakest dimension. The Bank invests considerable resources in developing sector and corporate strategies. •   ore recent sector strategy papers (SSPs) are better at internalizing country and • M regional experience than the first generation of SSPs, but they do not effectively influence country programs. • A hampered by differences in scope,  lignment among sector and country strategies is ­ ­timeframe, and ­audience.  he country-based model makes it difficult to ­ • T strategies, integrate sector and corporate ­ ­ rograms into country p and global and regional p ­ rograms. The Relevance of Strategic Alignment Under the Bank’s country model, sector and country priorities are expected to be aligned as part of a single country program. Country directors, typically based in the client country or a nearby neighbor, are expected to ensure country programs and their components are responsive to country needs and capacities. Sector staff, based at headquarters or in-country, and their sector managers are expected to ensure that sector operations and programs individually and jointly are both appropriate to country needs and technically sound, based on global excellence, and reflected in country programs. Country and sector units prepare strategy documents to guide their work. A country program is governed by a CAS. The CAS specifies the Bank’s priorities in a country program, typically over a three- to four-year period, identifies objectives to be achieved, and notionally allocates resources among lending and nonlending operations to achieve these objectives. Country programs can deviate from plans outlined in the strategy, if conditions change. SSPs are prepared by Sector Boards to specify the Bank’s global priorities in the sector at roughly eight- to 10-year intervals, strategy formulation being one of five core responsibilities of Sector Boards.1 The Bank’s matrix system is predicated on the assumption that its dual goals—increasing client responsiveness and providing quality services through technical excellence as articulated in its country and sector strategies—are interrelated and necessary to achieve good country results. 22 The Matrix System at Work The Bank’s experience shows that lending and technical assistance achieves better outcomes when global knowledge is customized to country conditions. The creative tension between those who manage country units and those who manage sector units has been purposively designed to ensure that a country program achieves both objectives. Assessing Strategic Alignment This chapter assesses the extent to which the outputs of the first dimension of the conceptual framework—strategic alignment—have been achieved by examining the following questions: • To what extent do sector strategies reflect country and regional experiences? • How effectively do country strategies integrate sector and corporate strategies? • How effectively do country programs integrate global and regional priorities? The evaluation relies on five sources of evidence in its assessment of the alignment between sector and country strategies. First, the 96 reviews of CASCRs undertaken by IEG were compared to obtain substantive evidence of responsiveness and effectiveness of country programs. Second, it draws on a set of desk reviews of country programs in a diverse set of 14 countries selected from among those with recent IEG CAEs. Reviews were based on the CAEs and other key documents typically including two or more recent CASs, CAS Progress Reports and other available documentation. Third, it uses a set of desk reviews of six sectors and two cross-cutting themes selected from among those with recent IEG evaluations. Reviews were similarly based on IEG evaluations and on key strategy documents on the sectors and themes. Fourth, the desk reviews were complemented by extensive structured interviews with more than 30 sector managers, 16 sector directors, and 14 country managers and directors responsible for the relevant sectors and countries. Finally, results from a staff survey of more than 400 technical staff and a response rate of over 40 percent, and perspectives offered at focus groups with staff are included where relevant. It should be noted that interviewees, survey respondents, and focus group participants were primarily asked to describe their own experiences in recent years on specific questions rather than to offer their opinions of the matrix system at the Bank. Based on this evidence the evaluation assesses the extent to which country programs and services are responsive to client needs. Strategic alignment is one of the outputs contributing to this outcome. Consequently, conclusions are drawn regarding the matrix system’s efficiency and effectiveness in achieving the intermediate outcome—“country strategies and programs are aligned with Bank priorities and client needs.� The Role of Country and Sector Strategies Alignment of client priorities with sector and corporate priorities is a relevant objective for maximizing the Bank’s development effectiveness, but the country-based model is not well designed to achieve this objective. The Strategic Alignment of Sector and Country Priorities 23 Central to the question of alignment of sector and corporate priorities with country programs is the relative power of country directors and sector managers. The country-based model is more responsive to short-term client demands within the three- to four-year time frame of a CAS than to longer-term sector and corporate priorities. In contrast, sector and corporate strategies are generic because of their global coverage and longer time frame, inhibiting alignment with country programs. The CAS provides the basic structure for a country program over a fixed time horizon and is accompanied by a budget to finance implementation. CASs typically specify strategic objectives that are relevant and achievable in a specific period; policy priorities and operations to achieve those objectives; and an implementation plan to apply available resources toward those objectives.2 CAS objectives are usually organized into three or four pillars, with indicative resource allocation among various sectors; among various lending instruments; and among analytical studies, just-in-time policy notes, or technical assistance. CASs are not binding documents and are often drafted to ensure flexibility in the choice and volume of lending and nonlending activities financed during the CAS period. Nevertheless, the strategic pillars and projected size of the portfolio provide the boundaries that determine the country program, and “consistency with the CAS� is a key justification for operations. By contrast, SSPs, although labeled strategies, have little visible influence on the sector portfolio in country programs. SSPs have little direct influence on country programs as they are global in scope and have to be customized to the diverse needs of partner countries to make them operationally relevant. A close review of the available SSPs in the six focus sectors and two cross-cutting corporate priorities shows that SSPs have multiple objectives and audiences, some internal and some external. While initially drafted as a single document, an SSP is usually refined through a variety of supplementary documents, including progress reports and action plans that can change directions or add new objectives. Most of the SSPs reviewed did not include an implementation plan, a results framework with indicators and targets, or a monitoring and evaluation plan. More recent SSPs (particularly those prepared after FY08) have attempted to include these elements, but even for those, the time horizon and geographic reach are too large for realistic planning and implementation. In practice, sector knowledge has to be customized by country teams to meet country needs and the influence is indirect and implicit rather than explicit. The primary influence of SSPs on future Bank activities comes from the participatory nature in which they are prepared and the direct involvement of sector staff members who contribute to CAS development. They need to be complemented by regional sector strategies to be operationally relevant and implemented through the CAS. There is typically little or no separate budget for implementation of sector strategies, unless they are able to mobilize trust funds for implementation. One constraint is that until recently, most SSPs tended to be prepared by sector staff in isolation from country units and other sectors,3 hence ownership has been largely limited to staff members of that 24 The Matrix System at Work sector. Furthermore, not all sectors have an SSP and not all sectors have trust funds for implementation. In most cases, the Bank was not able to fully implement its sector and corporate strategies. A review of 10 recent IEG evaluations of the Bank’s performance in individual sectors or subsectors found a substantial or high degree of implementation in only three, as shown in Table 2.1. This reflects less on the performance of sector staff and managers in the Regions, and more on the efficacy of the strategy itself. Given the primacy of client demand in the design of country programs, it is unreasonable to expect an unfunded global strategy to accurately predict the needs of countries as far as a decade in advance. Table K.2 in Appendix K provides greater detail on the implementation of each sector strategy. Although not strategies in the usual sense, SSPs do serve other purposes. In practice, SSPs have four functions that serve different audiences:4 • SSPs provide a conceptual framework leading to a shared sector approach and a common language among sector specialists. • SSPs take stock of experience in the Bank and sometimes outside the Bank, offering a view on “the state of the sector,� including the lessons from regional operations and insights on new approaches and instruments. • SSPs represent a shared understanding among regional sector staff and anchor staff of the strategic priorities of the Bank in that sector. • SSPs provide a forum for communication and an opportunity to improve alignment with external partners. Table 2.1 Assessment of the Degree of Implementation of Sector Strategies Negligible (2) Modest (5) Substantial (3) High (0) Gender (2010b) Governance and Agricultural productivity anticorruption (2011c) (2011, rural development) Environment Information and Social safety nets (2008b) communication (2011b, social protection) technology (2011d) Water (2010) Transport (2007) Municipal management (2009, urban) Health, nutrition, and population (2009c) Source: IEG sector and thematic evaluations. Note: Year represents year IEG evaluation was published; sector indicates the sector strategy if different than the topic of the evaluation. The Strategic Alignment of Sector and Country Priorities 25 Sector issues vary across countries, making alignment of the SSPs with individual country strategies and country programs more challenging. Sector staff and managers in the Regions do not always appreciate the role SSPs play in relation to donors and other global stakeholders and sometimes see them as supply-driven by external stakeholders. Sector staff identify more with some strategies than with others, as some are seen as reflecting the Bank’s own experience and others as impositions that add complexity and difficulty to country operations. Table 2.2 summarizes the distinctions between country strategies and SSPs. The sector strategy process has evolved with additional guidelines introduced in FY08 to increase the focus on operational relevance and results. Managers reported much greater involvement of the Regions and more elaborate consultation processes with stakeholders in the new generation of SSPs in the environment and energy sectors, for example. However, the structural constraints that limit the effectiveness of SSPs remain, warranting some rethinking of how best to provide sector services to country clients at reasonable cost. Some SSPs also serve as vehicles for addressing GPGs, such as communicable diseases in the health sector strategy. Other corporate priorities, such as governance and anticorruption, are issued as stand-alone strategy documents. Responsiveness and Effectiveness of Country Programs The Bank’s record of delivering responsive and effective country programs that are responsive to client needs is moderately satisfactory. In aggregate, CAS outcomes have been moderately satisfactory or better in nearly two-thirds of country programs. The Bank has been most effective in responding to client Table 2.2 Key Features of Country Strategies and Sector Strategies Characteristic Country strategy Sector strategy Typical timeframe 3-4 years 8-10 years Country scope One partner country All partner countries Sector scope All sectors relevant to One sector, sub-sector, or corporate partner country priority Results framework Yes No Implementation budget Annual budget allocation, No associated budget unless allocated by tri-annual IDA allocation country directors on country-by-country basis or by dedicated trust funds Self-evaluation CAS Completion Report None Source: IEG compilation from OPCS website. 26 The Matrix System at Work government demands but much less so in selecting strategic priorities that are realistic and have broader ownership. The responsiveness and effectiveness of the Bank’s country programs have improved modestly during the period of matrix implementation, though several areas require further attention. Systemic weaknesses persist in the strategic content of CASs. For instance, the Bank’s 2009 CAS Retrospective indicates that despite some effort to improve, CASs are not sufficiently selective in the choice of key objectives and sectors through which those objectives will be achieved. Since 2003, the Bank has been issuing CASCRs—self-assessments of country program outcomes at the completion of the CAS period—and IEG has reviewed the CASCRs to validate their accuracy and draw lessons from them. These reviews show that outcomes, measured as achievement of objectives by CAS pillar, were satisfactory less than two-thirds of the time over the period reviewed, compared to a Bank target of 70 percent satisfactory. As the number of completed CASCR reviews now approaches 100, it is possible to take a closer look at specific factors that underlie these aggregate results. IEG reviewed all 96 CASCRs available as of December 2010 which reviewed CASs prepared since the introduction of the matrix system—covering country programs in 80 countries. The quality of country programs was assessed against the two goals of the 1997 reform using eight variables, four reflecting responsiveness and four representing effectiveness. Each variable was assessed separately on a four-point scale using evidence from IEG’s CASCR reviews and combined into an aggregate rating for each country program, applying equal weights to each variable. The variables are defined in Table 2.3. Overall, 66 percent of country programs received an aggregate rating of moderately satisfactory or better based on these eight criteria, as shown in Table 2.4. This is very similar in magnitude to the share of programs rated moderately satisfactory or better in meeting their CAS objectives, 63 percent, as explicitly rated in CASCR reviews. Moreover, these two ratings are strongly correlated.5 Alignment of Bank country programs with national development strategies and programs received relatively high marks, as did flexibility, project outcomes, and analytical and advisory activities (AAA) outcomes. On the other end of the scale, only about one-third of programs were rated moderately satisfactory or better on country ownership and realism. The quality of the results framework was deemed satisfactory for less than 30 percent of the country programs, consistent with the Bank’s own finding that the quality of results frameworks in CASs varies wildly (CAS Retrospective 2009). And roughly half of the programs were rated satisfactory for quality of implementation. Appendix D provides detailed analysis for each of the eight components. The overall performance of country programs has improved over time due to improvements in some of the components, while others have remained flat The Strategic Alignment of Sector and Country Priorities 27 Table 2.3 Basis for Rating Each Componenta Alignment with Extent to which objectives are appropriate for the country’s development and consistent national priorities with the country’s own development strategy Ownership Extent to which there were consultations and broad commitment by stakeholders in Responsiveness design of strategy, and extent to which commitment was sustained during program implementation Realism Extent to which program was matched to country capacity, appropriately estimated government commitment, identified appropriate lending targets and outcome objectives, and was selective and prioritized Flexibility Extent to which program made appropriate midcourse adjustments due to changed country conditions (such as change of government, crisis, or natural disasters); rated when relevant Quality of results Extent to which framework includes quantified and time-bound targets; an outcome as op - framework posed to an output orientation; and specific, monitorable, adequate, relevant, and timely indicators. Effectiveness Quality of portfolio Extent to which implementation proceeded as planned, with due attention to disbursement implementation rates, delays in implementation, and reductions or increases in the share of projects and commitments at risk in the portfolio Project outcomes Extent to which operations that closed during CAS implementation have met their stated objectives AAA outcomes Extent to which the CASCR review indicates that analytical and advisory activities have had an effect on operations or policy. a. As a number of these components could overlap in scope, reviewers set clear boundaries to avoid double counting. For instance, a program might lack realism because of an overestimation of government ownership at the outset. In this case, the assessment of the program’s realism will be moderately unsatisfactory or worse based on that overestimation, but the rating of ownership will depend on the extent of the government’s (and other stakeholders’) commitment to the program, not the degree to which that commitment was appropriately assessed by the Bank. Ownership may be strong enough to be rated “above-the-line,� but the program may still be unrealistic if it relies too heavily on government commitment to implement reforms without adequate safeguards built into the program’s design. In this way, each rating was isolated from the others to the extent possible. Table 2.4 Percentage of Country Programs Rated Moderately Satisfactory or Higher Over Time AAA Aggre- CAS Start of CAS Owner- Real- Flex- Results Implemen- Project out- gate outcome period Alignment ship ism ibility framework tation outcome come rating rating FY98-FY03 94 18 33 60 20 57 71 67 59 55 (n = 49) FY04-FY08 100 55 32 73 36 51 70 81 72 70 (n = 47) Overall (n = 96) 97 36 32 67 28 54 71 74 66 63 Source: IEG reviews of CAS Completion Reports as of December 1, 2010. Note: CAS outcome rating is the rating of the outcomes of country programs as assessed in CAS Completion Report Reviews. 28 The Matrix System at Work or worsened slightly. More recent country programs had an overall rating of 72 percent moderately satisfactory or higher. When dividing the universe into two cohorts (49 CASs prepared between FY1998 and FY2003, and 47 CASs prepared between FY04 and FY08) flexibility and, most dramatically, ownership show improvement. Realism and implementation have not improved. Ratings for realism and results frameworks remain particularly low despite some improvement in the latter.6 Table 2.5 highlights the components that require the most attention. Disaggregated data from the CASCR reviews show tension between responsiveness to the government and responsiveness to country needs. Starting from the original statement of objectives in 1997, the terms “client� and “country� have often been used interchangeably in discussions of the reorganization and its impact. However, the findings from the CASCR analysis (and the more favorable feedback on responsiveness from the focus countries’ desk reviews and interviews) show that the distinction is critical: responsiveness to government demands is not necessarily coterminous with responsiveness to country development needs. The high rating for alignment in the country program in the CASCR analysis reflects responsiveness with the national government’s demands or strategic priorities. At the same time, the much lower ratings on realism and ownership in the CASCR reviews indicate significant gaps with the countries’ development needs and priorities of other key stakeholders. This gap is reinforced by a 2010 survey of Bank clients7 where alignment with development priorities of their country received an average rating of 6.3 on a 10-point scale. The Ministry of Finance and office of the head of state rated alignment higher than other stakeholders, but even they scored it 6.6, showing a considerable gap in Bank responsiveness even to their expectations. These findings underline anew the importance of realistic assessments of the quality of governance when responding to government requests. They also help explain the differences between IBRD and IDA country programs. In many IDA countries, the process of formulation of national poverty reduction strategies has been more inclusive and over time resulted in broadening ownership Table 2.5 Country Program Variables That Require Attention Direction Level Level Deteriorating Improving Very Weak Realism Results framework 33% → 32% 20% → 36% Moderate Implementation Ownership 57% → 51% 18% → 55% Source: IEG reviews of CAS Completion Reports as of December 1, 2010. The Strategic Alignment of Sector and Country Priorities 29 among a more diverse set of stakeholders. This is reflected in better results on ownership and responsiveness for IDA countries compared to IBRD in the CASCR reviews. Responsiveness and Effectiveness by Region Among the Bank’s Regions, Africa has the lowest share of country programs rated above-the-line overall—less than half—while programs in Latin America and the Caribbean were rated best, with 80 percent rated moderately satisfactory or better (see Appendix C). Two-thirds to three-fourths of country programs in the other four Regions were rated moderately satisfactory or better. The aggregate ratings for “responsiveness and effectiveness� are broadly similar to IEG’s ratings for country program outcomes in CASCR reviews in all Regions. Ratings were similar for programs in both IDA- and IBRD-eligible countries. Low ratings for realism and ownership and high ratings for flexibility are striking among nearly all the Regions. Insights from CASCR Reviews The quality of country programs—their responsiveness and effectiveness in meeting country needs—has been moderately satisfactory or better just under two-thirds of the time since the introduction of the matrix. Bank country programs have been stronger in four components—alignment with national priorities, flexibility, outcomes of projects, and AAA—and weak in the other four. To improve overall quality, and likely results, the other components need further attention. Realism has rarely been found satisfactory and has shown no improvement over the period. Careful consideration of partner capacity and likelihood of ownership by relevant levels of government and civil society during the design stage reduces the risks to the country program and increases the prospects for effective implementation. Moreover, selectivity in areas of intervention with an eye to the critical barriers to development can focus attention on the most strategic areas and deepen the Bank’s impact. The country program’s results framework is the area with the next lowest rating. It has shown modest improvement, likely related to the introduction of results- based CASs (piloted in 2003, and mainstreamed in January 2005), but it remains less than satisfactory in the vast majority of cases. Programs that begin with a results framework in mind would be better able to design a coherent approach that can realistically achieve its goals. Ownership has improved significantly but remains moderately satisfactory or better in just over half of programs in the second half of the review period. In fact, ratings of ownership are higher for IDA countries than for those eligible for IBRD, especially when excluding fragile and conflict-affected states (FCS). Ten of 14 (71 percent) non-FCS IDA countries were judged to have moderately satisfactory or better ownership in the more recent period, compared to 15 of 28 (54 percent) non-FCS IBRD-eligible countries. In order to improve further, 30 The Matrix System at Work country programs would have to broaden government commitment by reaching out to multiple levels of government—central ministries, relevant line ministries, and local government—in addition to civil society through strong outreach and consultations during program design. Portfolio implementation is moderately satisfactory or better in barely half of country programs; this rating has remained stagnant, even deteriorating slightly. Improved realism in design during project appraisal—including tailoring program design to local capacity and more realistic assessment of government commitment to the program—and more rigorous and candid supervision are essential for improving implementation performance. Alignment of Country and Sector Strategies This chapter set out to assess the extent to which strategic alignment—the first dimension of the conceptual framework has been achieved. More recent SSPs are better at internalizing country and regional experience than the first generation of SSPs, but they do not effectively influence country programs. The Regions contribute to preparation of SSPs but do not use them sufficiently. Integration of Country and Regional Experiences in Sector Strategies Lessons from country operational experience have increasingly found their way into sector strategies. Previous SSPs were seen to be anchor-driven, but in recent years there has been greater involvement and management by the Sector Boards, that is, by regional sector managers. For example, the recent health, nutrition, and population; transport; and social protection strategies are seen as more useful and are “owned� by regional managers and staff. This was also the experience for roughly half of the sector- and anchor-based staff and three-fifths of sector managers in the focus sectors, as shown in Table 2.6. Regional sector directors interviewed noted that the nature of SSPs has evolved and that recent strategies place more emphasis on consultations with operational staff in the Regions. Integration of Sector and Global Strategies in Country Strategies Sector strategies are more conceptual than operationally focused. Bank- wide sector strategies are often too general to provide specific guidance to country programs and operations beyond broad principles and guidelines. They need to be complemented by regional sector strategies to become more operational, a view expressed in interviews by managers of the water, environment, transport, and social protection sectors, and in the Africa, East Asia and Pacific, and South Asia Regions. Furthermore, there is a gulf between the timeframe, geographic scope, and audience of the two types of strategies. Sector strategies have very little effect on the Bank’s interaction with its clients. Desk reviews of CASs in the 14 focus countries found little trace of SSPs. Managers and staff confirmed this point. CASs and operations draw on The Strategic Alignment of Sector and Country Priorities 31 Table 2.6 Integration of Country and Regional Experience in Sector Strategies Survey question Sector staff Sector managers Percent indicating to a substantial or very large extent The sector strategy reflects relevant country 46 experience The sector strategy reflects relevant regional 52 experience The sector strategy reflects regional and country 60 experience (18 of 30) Source: IEG staff survey and IEG manager interviews. SSPs to a substantial or very large extent in the portfolios of two-fifths of sector managers (Table 2.7). Most SSPs have little effect at the country level because they do not have any “teeth.� SSPs have to be customized at the regional and country level to make them operationally relevant. Sector managers noted that since countries are very different, CASs are selective in their use of SSP guidance. Under the country model, selectivity is to be expected in the CAS, and not all sector priorities can, or indeed, should be reflected in each CAS. However, the influence of SSPs on CASs remains indirect and limited. In one-fourth of the operations overseen by sector staff, CASs draw on the relevant SSP to a substantial or very large extent. This contrasts with the perception of other staff. One-tenth of network anchor staff—those closest to the preparation of SSPs—perceived their SSPs to be adopted into CASs to a substantial or greater extent. On the other end, two-fifths of staff based in country units—those closest to the preparation of CASs—thought they had incorporated SSPs to a substantial or greater extent. One survey respondent from an anchor unit in SDN summarized the experience of the majority: “Overall, it seems that CASs are not very well aligned with sector strategies. While it is understandable that they cannot be 100 percent aligned, it seems that many regional vice presidencies ignore particularly the importance of the sector strategy in its CASs.� The effect of sector strategies on lending operations is not much better. Thirty percent of staff said the operations on which they work (or in the country whose portfolio they oversee) drew on relevant SSPs. As one survey respondent from an SMU working on environment put it, “Bank priorities are often determined by donors. Country operations are often determined by the client . . . Only by fluke [do] country operations meet Bank priorities.� 32 The Matrix System at Work Table 2.7 Alignment Among Country Strategies and Operations and Sector Strategies Survey question Operational staff Sector managers Country directors Percent indicating to a substantial or very large extent [The] country strategy(ies) draw on the sector strategy(ies) 27 [The] country strategy(ies) reflect GPG issues 36 [The] country strategy(ies) reflect regional issues 27 Country operations draw on the sector strategy(ies) 30 Country strategies and operations draw on sector priorities 43 45 Country strategies and operations reflect GPG issues and 41 36 regional issues Source: IEG staff survey and IEG manager interviews. Sector strategic thinking does have some influence on operations. Regional sector staff who provide inputs for the CAS serve as the vehicle for this alignment, but some of those interviewed suggested that their influence is more on the design of operations than the product mix. The Bank itself has acknowledged these tensions, most recently in the 2009 CAS Retrospective, which noted that “there remain tensions between strategic objectives of the Bank of a sectoral nature, such as support for infrastructure, gender, and climate change, and the need to align with the country’s own vision and priorities, which may or may not prioritize the Bank’s strategic objectives.� But instead of highlighting successful approaches for integrating corporate or sector priorities, it went on to discuss only processes that ensure CAS alignment with the country’s own development strategy and goals. The challenge in preparing a CAS is to bring together a series of tasks from many different sectors into a coherent, properly sequenced, and relevant package of support. But the fact that sector strategies tend to be prepared independently from each other, without buy-in from staff working in related sectors adds substantial complexity to the process, and results in “silo- ized� strategies as one sector director put it. For example, both the 2007 health, nutrition, and population strategy and the 2003 water supply and sanitation strategy emphasize the deep interrelation between the two sectors. But neither strategy indicates criteria for determining when intersectoral coordination or a multisector approach is desirable. Although the health SSP foresaw the CAS as the instrument for coordinating intersectoral action to improve health outcomes, this has not occurred consistently.8 The Strategic Alignment of Sector and Country Priorities 33 Aligning country strategies with sector strategies is further complicated by the fact that individual SSPs can sometimes act as lobbying documents for devoting more resources to that sector, as one sector director put it. Thus, the aggregation of these sector strategies does not provide good guidance for defining Bank priorities. Sector priorities emerge in country programs more often through the direct involvement of sector staff on the country teams than as a result of the SSPs. It is not surprising, then, that country directors adopt sector priorities into a CAS only when it already suits the client. SSPs are usually out of date when a CAS is being prepared; they can sometimes conflict with each other; they are relatively vague; and they are not backed by specific goals or budget. As one sector manager noted, the world would have been the same without the sector strategy in her sector. Three primary factors determine whether a sector strategy will be adopted into a CAS. First and foremost, those sector strategies perceived as reflecting more “real� corporate priorities and initiatives than others are prime candidates; in interviews sector and country managers are quick to distinguish among them. These strategies feature prominently in major speeches by the Bank’s president and are often associated with special mechanisms, such as senior management monitoring or special budget allocations. Corporate priorities, such as environment (in particular climate change) and governance, fall into this category and have acquired increasing importance in country operations. In others, such as gender, the degree of corporate priority is perceived to be lower, in spite of special funding and action plans. These are seen as externally driven and have less traction in the Bank. This is manifested in the data on thematic coding which reveals that the theme of gender (along with social development) was identified by 18 percent of lending operations approved in FY07-10 compared to 24 percent for environment and 79 percent for public sector governance. Second, the effect of sector strategies on CASs and country operations depends on the extent to which they attempt to break from previous Bank strategies and practices. When the strategy changes arose from operational knowledge and country experiences, such as the criticism of big dams, the changes were easier to implement. But major changes encounter more resistance because of entrenched practice among staff. As one sector director put it, “the last thing I want to do is sell some snake oil when I’m in a changing relationship with [a major partner] on its way to becoming a middle-income country.� Third, the impact of sector strategies on CASs and Bank operations is lower for middle-income countries than it is for lower-income countries. Bank strategies for middle-income countries often include specific references to the importance of flexibility in country programs and of responding to specific client demands. Interviews with regional sector managers emphasized this aspect, particularly in Latin America and the Caribbean, but this is increasingly apparent in East Asia and Pacific, Europe and Central Asia, the Middle East and North Africa, and South Asia as well. 34 The Matrix System at Work Incorporation of Global and Regional Issues in Country Strategies While alignment of country and sector strategies was one of the original aims of the 1997 reforms, GPGs and regional issues have emerged as additional priorities that are somewhat distinct from sector priorities. Some SSPs also include consideration of relevant global issues; more recently GPGs have taken on a life of their own leading to the preparation of corporate strategies, such as those on governance and anticorruption or climate change. Their integration into the country program poses new challenges. Global and regional issues are not easy to address under the Bank’s country-driven model. Public goods are undersupplied by markets because those who provide them do not capture all of the benefits but bear all the costs. Government intervention is justified to overcome such market failure. However, the problem is much more acute with GPGs, whose potential benefits are inherently cross-national or global, with no single agency to redress collective action problems. Some examples of GPGs are a stable global climate and the eradication of communicable diseases. The Bank’s country-based model comes under strain when global and country interests are seen to diverge, as noted in IEG’s 2008 Annual Review of Development Effectiveness and the Bank’s 2009 CAS Retrospective. Mirroring the international structure of sovereign states, the Bank’s system gives primacy to national decision making on the policies and programs it supports. As one interviewee put it, “there is always going to be a conflict because the countries don’t care about the Bank’s corporate priorities.� Given this dynamic, SSPs are the primary platform through which individual GPGs can be identified as corporate goals, and through which the Bank can formally encourage their provision. What is needed is a balance between what the client wants and the global perspective the Bank brings. When global and country interests diverge, additional incentives may be needed to increase client interest in GPGs. As with SSP priorities, GPGs were only integrated into CASs to a very limited extent in the 14 focus countries. GPGs were substantially present in strategies for only four of these countries, and their treatment in the strategy was all but nonexistent in six. In part, this may be because CAS templates do not have a specific provision for global programs or GPGs. This is consistent with earlier findings. IEG’s Annual Review of Development Effectiveness 2008 found consistent references to only a few, such as the Global Environment Facility (GEF). The 2009 CAS Retrospective notes, “there is scope to do more in ensuring the coherence between country-level interventions and global/regional objectives.� In the staff survey, barely one-third of staff indicated that country strategies reflect GPG issues to a substantial or very large extent. Staff from the East Asia and Pacific Region were outliers: for them, the figure was over half. Managers had a similar perspective. Several indicated that GPGs are mentioned but not truly integrated into country programs or operations. One suggested that they were conspicuous for their absence. Overall, however, more staff reported that integration of GPG issues in CASs was substantial (36 percent) The Strategic Alignment of Sector and Country Priorities 35 than integration of sector strategies (27 percent), while sector managers viewed them at par (Table 2.7). Integration of regional issues into CASs was similarly weak. Barely a quarter of staff respondents considered regional issues to be integrated to a substantial or very large extent. The figures for sector managers and country directors on both GPGs and regional issues were higher (Table 2.7) and were very similar to the figures for integration of sector strategies in CASs, roughly 40 percent. Middle-income countries, such as Brazil, India, and Mexico, sometimes view action on major GPG issues as a vehicle for global leadership and request support from the Bank on these issues. This reduces the tension between global and local demands. In these countries GPGs find their way into the country program more easily. In countries that do not see direct benefits from investment in GPGs, pressure from the Bank to take action is seen as an imposition, and GPGs are not seen as important for the national development agenda. This extends to Bank staff working in the Regions. At the same time it needs to be recognized that not all GPGs are relevant to all countries, nor can all countries make an appreciable contribution to GPG agendas. Mainstreaming of Corporate Priorities into Operations Corporate priorities—such as improving the quality of the environment and supporting improvements in governance and institutional capacity—typically cannot be achieved through assistance in a single sector. On the contrary, they must be mainstreamed into operations and analytical work in many sectors in order to be achieved. Task teams are required to assign sector and theme codes to all activities that directly serve an external client and are used for corporate reporting.9 In the Bank’s portfolio, each project can have a maximum of five sectors and five themes, with a percentage allocated to each sector and theme. The identification of sectors and themes is based on self-reporting by task teams in project appraisal documents and provides a proxy of the extent to which task teams consider sectors and themes of other Sector Boards relevant to their respective projects. IEG analyzed the distribution of sectors and themes by classifying them into affiliated and non-affiliated categories based on their linkage with particular Sector Boards as determined by the Bank (for a description of the methodology see Appendix B). Affiliated sectors and themes are those that are directly associated with the Sector Board sponsoring the project. Non-affiliated sectors and themes refer to those that are associated with other Sector Boards. Among the three corporate priorities that were analyzed, public sector governance (PSG) is the most prevalent in the portfolio, identified as a sector in 48 percent of all operations not mapped to the PSG Sector Board during FY07-10 and as a theme in 29 percent. Comparable figures for environment and social development and gender themes were 19 percent and 16 percent, respectively. Data from two periods—FY95-98 and FY07-10 show that the average allocation in the projects mapped to other Sector Boards declined for environment (from 9 percent to 7 36 The Matrix System at Work percent) and for social development and gender (from 7 percent to 4 percent), while for PSG it increased marginally from 9 percent to 10 percent as a theme, but decreased as a non-affiliated sector from 20 percent to 16 percent in the same time period (Appendix J). Efficiency and Effectiveness of Strategic Alignment This section assesses the efficiency and effectiveness of strategic alignment in terms of the efficiency of resource use for strategy development and the extent to which the Bank has been effective in achieving strategic alignment between country programs and operations, and Bank priorities and client needs. Costs of Country and Sector Strategies The Bank invests considerable resources in developing sector and corporate strategies. In aggregate, the Bank spends twice as much on sector strategies as it does on country strategies, and the average expended per SSP developed is 38 times higher than for a CAS. Actual expenditure on sector strategy development is substantial ($118 million over the FY02-10 period), just over half of which is spent by the Regions on sector strategy development, compared to approximately $62 million for CASs over the same period. In FY10 alone, the Bank reportedly spent $13.5 million on SSPs compared to $6.7 million on CASs.10 The funds expended on SSP development during the FY02-10 period, including regional costs, averaged $9.0 million per new strategy11 developed overall or $3.8 million if only including network anchor costs (Box 2.1). These are non- trivial sums, yet the country model militates against effective integration in country programs. There is an absence of corporate oversight of resource utilization for sector strategy development. While the aggregate amount charged to sector strategy development is clearly identifiable, obtaining precise numbers on how much individual SSPs cost is virtually impossible because of parallel expenditures on background papers as well as expenditures charged to internal orders all listed under the strategy development business process. Aggregate expenditures give a good picture of the orders of magnitude of staff time and resources being spent on, or at least charged to sector strategy development. The combination of high cost and lengthy preparation of SSPs, and the need to complement them with additional regional strategies, gives rise to questions about the effectiveness of SSPs as vehicles for setting strategic priorities and calls for a need to redefine their scope more realistically. Regional staff and managers (in CMUs and SMUs) consider the cost of preparing and frequently revising sector strategies too high and see it as a big tax on country programs and operations. Country unit and sector unit staffs also agree that the cost of SSPs substantially exceeds their benefits. There is a need for more rigorous oversight of resource use for SSPs, The Strategic Alignment of Sector and Country Priorities 37 Box 2.1 Sector Strategy Papers at What Cost? Actual Bank expenditure on the development of SSPs from FY02 to FY10a is $118 million according to the Bank’s records; a period in which it approved 11 sector and two corporate strategies, yielding an average of $9 million expended per strategy developed. Of the 13 SSPs, nine were from SDN sectors*—of which four were from the former Environmental and Socially Sustainable Development Network (ESSD), two were from the Infrastructure Network (INF), two were from FPD, one was from HDN, and one was from PREM.b Even if one excludes expenditures for sector strategy development from regional VPUs ($60.5 million) and corporate VPUs ($7.3 million), $50 million was expended by network anchors—of which $6.6 million is trust funded—either through direct costs allocated to the sector strategy development product line ($39 million) or through indirect costs ($11 million). Considering only the $50 million expended by network anchors and the 13 new SSPs developed during FY02-10, this equates to $3.8 million in total costs to network anchors per SSP developed. There appears to be a wide range of expenses subsumed under sector strategy development with no evidence of monitoring of associated ex- penditures or corporate governance to ensure efficiency of resource use. Source: World Bank database and OPCS. *In 2006, the ESSD and INF networks were merged into SDN. a. According to the Bank’s business process dictionary, development of sector strategies includes preparation of concept note, background papers, strategy development, and external/internal consultations. For the purposes of this analysis only the product line code for sector strategies and internal orders were considered, removing $3.7 million in economic sector work and $54,000 in knowledge products assigned to the sector strategy development business process. b. Strategies approved during FY02-10 period include “Private Sector Development Strategy—Directions for the World Bank Group,� “Information and Communication Technologies,� “Reaching the Rural Poor: An Updated Strategy for Rural Development,� “A Revised Forest Sector Strategy,� “Water Resources Sector Strategy Paper and Business Plan - Manage- ment Note and Revisions,� “Water Supply and Sanitation Sector Business Strategy,� “Empowering People by Transforming Institutions: Social Development in World Bank Operations,� “Financial Sector Strategy for the World Bank Group,� “Healthy Development: The World Bank Strategy for Health, Nutrition, and Population Results,� “Strengthening World Bank Group Engagement on Governance and Anticorruption,� “Safe, Clean, and Affordable Transport for Development. The World Bank Group’s Transport Business Strategy, 2008-12,� “Development and Climate Change: A Strategic Framework for the World Bank Group,� “Systems of Cities: Harnessing urbanization for growth and poverty alleviation.� clearer terms of reference specifying their purpose and audience, and better integration of corporate and sector priorities within country programs. Feedback From Clients Surveys of country clients confirm that responsiveness has improved, but considerable gaps remain between the Bank’s country programs and country needs. A 2008 multinational survey of opinion leaders reveals that across all Regions, three-fourths of respondents strongly agree that to address poverty more effectively, the World Bank Group needs to work in a different and more innovative way, with leaders from the Africa, East Asia and the Pacific, and South Asia Regions more insistent on the need for change. Results from surveys of leaders in client countries in 2009 and 2010 also critique responsiveness (Table 2.8). In the latest poll, one-third feel the Bank imposes technocratic solutions without regard to political realities and about 20 percent fault the Bank for not exploring alternative policy options. In the same year, nearly two-thirds of in-country stakeholders rated alignment of 38 The Matrix System at Work the Bank’s work with their country’s development priorities and the Bank’s effectiveness favorably. The Bank has been more effective in responding to client government demands but much less so in selecting strategic priorities that are realistic and have broader ownership. Alignment of the priorities of country and sector strategies is a relevant objective for maximizing the Bank’s development effectiveness, but the system designed for achieving this is not effective. Sector priorities are integrated into country strategies or operations through the contributions of sector staff on the country teams, not through, or because of, the sector and corporate strategies. Central to the question of alignment of sector strategies with country programs is the relative power of country directors and sector managers, an issue discussed later in chapter five. Notes The other four responsibilities are knowledge, quality, partnerships, and human 1 resources management for the sector. 2 For example, see businessdictionary.com. 3 Management has informed IEG that unlike the SSPs reviewed by IEG (the health, nutrition and population strategy (2007) and the water and sanitation strategy (2003), their most recent updates (health, nutrition and population strategy, 2009 and water and sanitation strategy, 2010) have elaborated on linkages with other sectoral and thematic areas. Derived from IEG reviews of sector strategy papers and IEG ­ 4 manager ­ interviews from the Agriculture and Rural Development (ARD); ­ Environment (ENV); Gender; Health, Nutrition, and Population (HNP); Public Sector Governance (PSG); Social Protection (SP); Transport; and Water sectors. Table 2.8 Greatest Weakness of the World Bank Country survey 2009 Country survey 2010 Survey question (N=2075) (N=2308) Percentage of respondents Imposing technocratic solutions without regard to political realities 44 33 Too bureaucratic in its operational policies and procedures 41 30 Not exploring alternative policy options 23 22 Source: World Bank Country Survey 2009 Review; Country Survey 2010 Review. Note: Responses to the 2009 and 2010 surveys are drawn from different pools of client countries and are not comparable. The Strategic Alignment of Sector and Country Priorities 39 5 The Pearson correlation coefficient is 0.68, and statistically significant at 95 percent confidence. Improvements may be related to the introduction of results-based CASs—required 6 to include a results framework and expected to be oriented to achieving specific and realistic outcomes—which were piloted in FY03-04 and required beginning January 2005. From “Country Survey 2010 Review,� World Bank. 7 See IEG 2009. Improving Effectiveness and Outcomes for the Poor in Health, Nutrition, 8 and Population: An Evaluation of World Bank Group Support Since 1997. Washington, DC.: World Bank An internal website describes the Bank’s sector and thematic coding system as 9 follows: “The purpose of Sector and Theme codes is to facilitate reporting of Bank activities by sectoral/thematic cut at corporate level. Sector codes are used to indicate which part of the economy is supported by the Bank intervention while theme codes are designed to capture Bank support to corporate priorities such as Millennium Development Goals or global public goods.� 10 Of the amount reported by IAD, IEG’s numbers show $8.5 million in FY10 for sector strategy development and $6 million for CASs. 11 In contrast, the 2009 CAS Retrospective found the average cost of a CAS to be $236,000 during the FY06-09 period. 40 The Matrix System at Work Chapter 3 The Promise of a Knowledge Bank Chapter Highlights • The objective of creating a global Knowledge Bank has not been achieved. • The matrix system has been highly inefficient and modestly effective in ensuring that innovation in the field is well captured across the institution and that country opera- ­ fficiently on global knowledge to meet client needs. tions are able to draw e • Most staff do not efficiently draw on existing knowledge generated inside and outside the Bank. • The Bank lacks the ability to efficiently retrieve and share the large volume of em- bedded knowledge generated during preparation and implementation of lending operations. ­ egional and network silos are even stronger. • Sector silos persist, but r • Actual inter-regional cross support accounts for less than one percent of regional staff time. • Cross support within the Regions is 15 percent of staff time, significantly higher than across VPUs but much lower than the 42 percent share of allocations to cross-sector work identified by the Bank’s coding system. • In fiscal 2010, expenditures for nonlending technical assistance, driven by trust funds, outstripped those for economic and sector work. “We have been in the business of researching and disseminating the lessons of development for a long time. But the revolution in information technology increases the potential value of these efforts by vastly extending their reach. To capture this potential, we need to invest in the necessary systems, in Washington and worldwide, that will enhance our ability to gather development information and experience, and share it with our clients. We need to become, in effect, the Knowledge Bank. “Let me stress one other point: the global knowledge partnership is not about machines. It is about people. The challenge is to harness the technology to link people together and to leverage its impact for development. That means both accumulating the right kind of knowledge, and helping our clients build the capacity to use it.� (James D. Wolfensohn, Speech at 1996 Annual Meetings) Introduction The Knowledge Bank The Knowledge Bank was envisaged both as the means to enhance the quality of World Bank lending and as an opportunity to grow a new stream of services to developing countries. One of the key drivers of the 1997 renewal strategy was the recognition that the Bank’s development programs were excessively driven by a culture of lending, with insufficient attention to client needs and the 42 The Matrix System at Work quality of results, which are crucial to development effectiveness (Wapenhans 1992, 1994; and Pincus and Winters 2002). This was accompanied by a growing acknowledgment that the Bank’s lending portfolio itself was under criticism from advocacy nongovernmental organizations (NGOs)1 and of diminishing importance due to the increasing flows of private capital to developing countries that offered alternatives to official aid. Under President Wolfensohn, with the 1996 announcement of the Knowledge Bank, there was an overt shift toward capacity building and knowledge creation, and with it a growing emphasis on institution building and market regulation, and later on governance reforms (Phillips 2009: 9). While World Bank annual commitments have been relatively steady, averaging $22 billion from 1990-95, foreign direct investment grew from $22 billion in 1990 to $95 billion in 1995,2 four times the size of Bank lending. The Knowledge Bank was seen as the solution to both forms of criticism, and the matrix system was to be the vehicle to create it. On the one hand, more careful attention to knowledge could enhance the quality and hence effectiveness of Bank-financed lending. On the other, in an increasingly competitive world, knowledge could offer an alternative stream of services to developing countries and a unique value proposition combining knowledge and finance. Specialized knowledge providers in the private sector—think tanks and universities—have already demonstrated the demand for knowledge services to inform public policy. With the introduction of the matrix system, the World Bank decided to emulate them more systematically. The relevant question is “how effective is the matrix system in enabling the Bank to distill and use country experience and global knowledge to meet client needs? To answer this question this chapter assesses to what extent: • is knowledge produced efficiently by Regions and network anchors? • do networks facilitate access and use of Bank and global knowledge? • is high quality expertise mobilized across Regions and sectors? The functioning of the Knowledge Bank is examined in this chapter by focusing on the capture and flow of knowledge. This complements the work of the Bank’s own report on the state of knowledge to examine the composition of knowledge products and the volume of expenditure on explicit knowledge products over time and includes findings from staff and managers on the supply and demand for knowledge. Knowledge flows are reviewed through an in-depth analysis of the volume and trends in cross support, as a proxy for the flow of tacit knowledge across Regions, networks, and sectors and findings on knowledge capture and dissemination. Following this assessment of the Bank’s ability to achieve these outputs we draw conclusions regarding the extent to which the related intermediate outcome—“Innovation in the field is well captured across the institution, and country operations draw efficiently on global knowledge to meet client needs�—has been achieved during FY1997-2010, the period reviewed by this evaluation. The Promise of a Knowledge Bank       43 Knowledge Needs to be Pooled and Shared Knowledge comes in three distinct forms—explicit knowledge, tacit knowledge, and embedded knowledge. Explicit knowledge consists of knowledge found in codified knowledge products that are produced by Bank staff, sometimes in collaboration with clients or partner organizations. Dissemination of explicit knowledge products is essential for them to be of operational use. Tacit knowledge is the stock of knowledge that Bank staff, clients, and partners gain from development experience. Of these, the most relevant to this report is the tacit knowledge among Bank staff. The principal mechanisms for transfer of tacit knowledge remain staff rotation and cross support to other organizational units, or personal networking. Embedded knowledge refers to knowledge within operational documents prepared as part of project appraisal, supervision, or assessment of lending operations. While embedded knowledge is often put to immediate operational use, and could be invaluable to similar operations elsewhere, without explicit arrangements for its capture and dissemination, this knowledge is unlikely to flow beyond an individual operation. Knowledge retrieval and transfer across the organization and externally could significantly enhance the value of the Bank’s knowledge products. The Knowledge Bank initiative, supported by the matrix and internal markets, was supposed to require a shift from an individualistic mode of working and storing knowledge toward a team-based mode (Phillips 2009: 71). Knowledge organizations need well-developed mechanisms to capture and share knowledge. The World Bank has the advantage of being staffed by highly qualified technical experts in comparison with most of its clients in developing countries. However, before the introduction of the matrix, the Bank did not explicitly commit to sharing of knowledge across the organization. Most technical staffs were housed in sector divisions within 20 country departments with minimal collaboration across country departments. The 1997 reform was designed to overcome these silos by grouping technical staff into Bank-wide sector families governed by Bank-wide Sector Boards, regardless of where they were located. The Knowledge Report (2011) describes three key roles for the Bank—producer, customizer, and connector—of knowledge. This chapter focuses more on knowledge flows—that is, the Bank’s role as a connector and customizer of knowledge, rather than on the quality of knowledge produced. Assessing the Bank’s Knowledge Work The magnitude of knowledge work in the Bank is very difficult to estimate. In FY2010, the Bank spent $606 million on “core knowledge products,� which includes knowledge for external clients, internal use, and research and data generated as a public good. Although how much the Bank spent on explicit knowledge products is known, tacit knowledge among clients and partners is inestimable. Nor has the Bank monitored the volume or quality of embedded knowledge in operational documents, except for a one-off review undertaken by the Middle East and North Africa Region, which estimated that in 2010 the Region spent about 10 percent of loan proceeds on knowledge products. 44 The Matrix System at Work Projecting from this, the Bank’s Knowledge Report estimates that knowledge spending embedded in lending was $2.5 billion (World Bank 2011: 11). However, given the wide variation in the nature of lending operations and knowledge products across the Regions, without further substantiation this estimate remains dubious. The volume of cross support across organizational units provides a proxy of the flow of tacit knowledge among Bank staff. The ultimate purpose of the Knowledge Bank is to ensure that the Bank uses cutting-edge knowledge, internal and global, to support client needs. An effective Knowledge Bank is predicated on maintaining an appropriate body of knowledge generated internally or externally and ensuring effective flows of knowledge. For this to be achieved, knowledge needs to be strategically relevant and customized to serve client needs. Effective knowledge flows require incentives and efficient systems for capture, dissemination and use of explicit, tacit, and embedded knowledge generated by the Bank, and global knowledge to operational teams and country clients. Bank management recognizes that the transformation into a Knowledge Bank has not yet been achieved. The 2010 management review of the matrix found that “anchors and Regions are increasingly disconnected,� “innovation in the field is not well captured across the institution,� and “not all country programs are able to draw on available knowledge.� IEG concurs with these findings. Management initiatives to address these weaknesses include targets for cross support, Global Expert Teams (GETs), Knowledge Platforms, a Global Fellows Program, and the “10 point program� toward enhanced Sector Practices.3 However, unless the underlying incentives inhibiting the flow of knowledge and expertise across regional and network boundaries are addressed, these initiatives will have limited effect. Investing in Knowledge Production The capture of knowledge from operations, analytical work, and technical assistance has been limited. Staff members believe that the Bank generally produces useful economic and sector work (ESW)—53 percent of sector- and anchor-based staff believe that their own sector produces operationally relevant knowledge to a substantial or very large extent. But in interviews, staff and managers repeatedly indicated that there was no mechanism or time to extract lessons from lending operations, and that this is a critical barrier to cross-regional and even cross-country knowledge sharing. The result is that much of the Bank’s knowledge is really embodied as tacit knowledge among individual staff members of the Bank. Most of this knowledge is not stored in easily searchable and retrievable forms resulting in considerable inefficiencies in the Bank’s knowledge work. Primacy of Lending Delivery Lending volume remains the most important metric of success. Incentives and accountability of country directors and sector managers is strongest for The Promise of a Knowledge Bank       45 lending operations and particularly lending delivery. Sector managers argue that in the current flat budget environment, knowledge and lending vie for scarce resources; lending always emerges as the winner (Box 3.1). The primacy of lending incentives is amply substantiated by the feedback from managers’ interviews (see Appendix E) and by the fact that success has been measured and reported primarily in terms of lending commitments and disbursements.4 The Bank’s Core Knowledge Products In FY11, the Bank spent $606 million, about 20 percent of its administrative and trust fund budget,5 on nine core knowledge products grouped into three categories (Figure 3.1).6 The first category, knowledge for external clients, accounted for the lion’s share ($444 million) and includes ESW, technical assistance, impact evaluation, and external client training. Most of the knowledge for external clients is produced by the regional sector units and network anchors, except for client training which is delivered mostly by the WBI. About 10 percent of client training is also provided by the network anchors and 8 percent by the Regions. The second category, internal knowledge ($99 million), includes knowledge products and new product development and is produced largely by the network anchors. The third category, knowledge as a public good ($63 million), includes the World Development Report, research, and global monitoring, and is produced by DEC. Core knowledge constituted 56 percent of operational expenditures (including trust funds) in FY10. About 60 percent was spent by the Regions, while the network anchors and DEC/WBI each spend about 20 percent of the total budget on core knowledge products.7 DEC and WBI lie outside the matrix structure and their knowledge activities appear to be less well connected to the Bank’s operational work. Within the Regions, the Middle East and North Africa and East Asia and the Pacific had the highest share of spending on core knowledge. In FY10, the six Regions together spent 24 percent of their expenditures (Bank budget plus BETFs) on core knowledge products. The Middle East and North Africa and East Asia and Box 3.1 Sector Manager Perspectives on Primacy of Lending “All that matters is lending; knowledge gets relegated to the tenth place.� “… in any competition between lending and knowledge-generation, lending wins.� “At the institutional level, knowledge management has not worked well. In the Re- gions, we are under too much pressure to deliver [lending]. We have no time to think or document our experience, although this was originally envisaged as part of the matrix re- ­ form. We have no resources to document and disseminate experience systematically. The main i nitiatives only.� constraint is pressure of lending delivery. Whatever is done is based on individual ­ Source: IEG manager interviews. 46 The Matrix System at Work Figure 3.1 The Production and Application of Knowledge at the World Bank Client Knowledge for Training Knowledge External Clients ($55m) Application Impact Evaluation ($11m) External External Partnerships Technical New Product Knowledge Assistance Development ($195m) ($18m) Economic and Internal Knowledge Sector Work Knowledge Generation ($184m) ($81m) World Development Global Report Research Monitoring ($8m) ($35m) ($19m) Source: Adapted from “The State of World Bank Knowledge Services: Knowledge for Development 2011� (World Bank 2011). Note: Volumes denote total FY10 Bank budget resources expended on core knowledge products. Pacific Regions spent 29 and 28 percent, while Africa and Latin America and the Caribbean had the lowest shares at 21 and 19 percent, respectively. A Trend Toward Just-in-Time Knowledge Nonlending technical assistance (NLTA) is the fastest growing segment of the knowledge portfolio. Knowledge for external clients made up almost three-fourths of all core knowledge expenditures throughout the FY02-10 period. As shown in Figure 3.2a, the composition of products within this category has changed. Spending on ESW has decreased since FY04, and in FY10 expenditures for NLTA outstripped that for ESW. This growth has been financed largely by trust funds. ESW continues to receive more Bank funds than NLTA, but the gap between the two products has decreased significantly over time (Figure 3.2b). A 2010 client survey indicates that technical and policy advice are highly appreciated by country clients.8 The growth in volume of NLTA could thus be seen as a positive trend as long as (a) the stock of analytical knowledge products is maintained, (b) NLTA selection is based on strategic planning rather than availability of donor funds, and (c) adequate quality assurance processes for NLTA are in place. In practice, the shrinking volume of ESW and the increasing reliance on trust funds raises questions about the sustainability of current trends and, as will be discussed in chapter five, quality assurance for NLTA is a mounting concern. With the exception of Latin America and the Caribbean, where trust funds are small, BETFs are driving knowledge products toward NLTA in all the Regions. The Promise of a Knowledge Bank       47 Figure 3.2 The Composition of Expenditures on Knowledge for External Clients Has Changed Substantially With Non-Lending Technical Assistance Overtaking Economic and Sector Work 3.2a. Financing from all sources 3.2b. Financing from bank budget only 200 (Constant 2010 US$ millions) 180 160 140 120 100 80 60 40 20 – FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Economic and Sector Work Non-lending External Client Training Impact Evaluation Technical Assistance Source: World Bank database. Dissemination and Use of Knowledge Dissemination of Knowledge The Bank, as noted, operates through a country-based model, with the country director having budget authority for all Bank activities in a partner country. This budget is allocated by the country director to all SMUs through a WPA that specifies the deliverables each SMU will be accountable for. The country model has increased responsiveness to client governments, but the incentives under the current matrix system and imbalance in power between country and sector units result in insufficient integration of sector and corporate issues and global knowledge in country programs. Under the country-based model, country directors are primarily accountable for country-level results in the near term and are therefore responsive to demand by in-country counterparts. There are very limited institutional incentives to finance strategic activities whose benefits would accrue to other countries (such as regional programs) or to the same country over the medium- to long- term. The view from sector managers was clear in this regard, and summarized by one: “Country directors have no incentive to give priority (and funding) to sector knowledge activities, and the sectors do not have their own 48 The Matrix System at Work funds to do so . . . The focus on individual countries increased greatly the responsiveness to the client, but three problems emerged: (i) loss of a regional perspective, as each director was concerned exclusively with his country; (ii) lack of strategic view for the country, as directors became too close to and dependent on the government; and (iii) neglect of generation and transmission of knowledge.� Correspondingly, the dissemination of knowledge is weak. Two-fifths of sector staff find the Bank-wide dissemination of internally generated knowledge and experience within their own sector effective to a substantial or very large extent (Table 3.1). And only one-fourth of staff surveyed report the dissemination of global knowledge from outside the Bank to be as effective. A frequent concern among staff was that the regional sector units are disconnected from each other, a fact that is complicated by the deep decentralization of sector staff in recent years. As one survey respondent from an SDN sector in Europe and Central Asia said, “with some technical staff decentralized, and little resources devoted to knowledge sharing among sector staff, it is difficult for operational staff to reflect experiences from other Regions or from outside the Bank to enhance [the] quality of the operations.� Sector managers and country directors were just as candid as staff regarding the dissemination of knowledge generated in their portfolios. While about half of the sector managers reported that lessons from their unit’s portfolio were provided to global, corporate, and sector strategies to a substantial or very large extent, a third provided knowledge to other Regions. Cross-regional sharing of lessons was rated even lower by country directors, with just two of 11 reporting that this occurred in their portfolio to a substantial or very large extent (Table 3.2). In most sectors, sector managers and regional sector staff were very critical of the usefulness of their anchors in disseminating operationally useful knowledge, and of Sector Boards for managing that knowledge. On Table 3.1 Dissemination of Knowledge Within the Bank Survey question To what extent was your sector family substantial Percent indicating to a ­ effective in . . . or very large extent . . . the production of operationally relevant knowledge 53 within the Bank? . . . the dissemination of internally generated knowledge 40 and experience across the Bank? . . . the dissemination within the Bank of operationally 25 relevant global knowledge produced outside the Bank? ­ Source: IEG staff survey. The Promise of a Knowledge Bank       49 Table 3.2 Most Knowledge Generated in a Unit is not Shared With Other Units Sector managers Country Survey question (n=30) directors (n=11) lessons from In the past five years, to what extent have ­ Share indicating to a substantial experience in your country/sector been ­ provided to . . . or very large extent Global, corporate and sector strategies 47% — Development Economics VPU 3% 27% World Bank Institute 20% 9% Network anchors 40% 19% Other regions 33% 18% Other units in your VPU — 60% Source: IEG manager interviews. a ten-point scale, sector managers gave their anchors an average rating of 5.0 regarding their effectiveness in providing knowledge and learning services to the Regions. A typical comment from a manager was, “The anchor does what it wants to do, with a cottage industry mentality and without reference to country needs or even discussion with sector people in the Regions.� Staff in the Regions indicated that the anchors “seem to have their own agenda� and are “too far away from operations.� The social protection sector was an exception and was generally viewed by sector managers as one of the best with regard to global knowledge management. On a ten-point scale, the Social Protection Sector Board received 9.0 for knowledge and learning services to the Regions. As one social protection manager put it, “Social Protection is one of the best sectors in global knowledge management. The Sector Board works well and has a business model that puts the Regions first. The sector anchor too is very client-focused and very supportive of the Regions.� Box 3.2 highlights some ingredients of success for the sector. Thematic groups, when properly funded, can serve as a useful platform for global knowledge sharing when compared to Sector Boards. The QAG Sector Board Assessment reported significant shortcomings in most of the Sector Boards. However, some of the thematic groups were found to be rather effective in sharing knowledge, an assessment also conveyed to IEG through feedback from staff and managers. Three main characteristics were offered for the more successful thematic groups in knowledge sharing. First, thematic groups cover narrower topics than a typical Sector Board, so they involve a smaller group of experts across the Bank interested in a particular topic. Second, they can be led or co-managed by staff in the Regions rather than exclusively by anchor staff; 50 The Matrix System at Work Box 3.2 Social Protection: A Model for Global Knowledge Management The Social Protection Sector Board is recognized by staff and managers as the best functioning Sector Board.a During IEG interviews, the sector anchor was widely praised for its work in provid- ing operationally relevant knowledge to the Regions. Interviewees attributed this to the service orientation of the anchor, and highlighted the following factors: • Since the 1997 reorganization, the management of the anchor unit has ensured that its top priority is to serve the Regions, and its research agenda is also driven by the Regions. • The anchor’s annual work plan and budget are discussed and approved by the Regions. The sector is not distracted by donor-driven initiatives due to a lack of sector-specific trust funds. Source: IEG manager interviews, IEG staff focus groups, 2005 QAG Sector Board Assessment. a. The Social Protection Sector Board was also identified as the best functioning one by the QAG Sector Board Assessment (2005). Staff feedback from the IEG matrix survey and focus groups, and managers’ interviews, reconfirmed this standing. this makes them more operationally focused. Third, the sole purpose of thematic groups and communities of practice is sharing knowledge among participants, unlike the multiple roles of Sector Boards. The GETs are not a substitute for thematic groups. Ten GETs, constituted by managers and recognized experts who serve as gurus in their respective technical areas, were launched as part of recent knowledge reforms. The GETs are aimed at issues of high corporate priority and are subsidized: the time of the experts is partly paid for from a central pool of funds. The GETs model is reported by managers and staff to be restrictive both in its membership and in limiting GETs members to a single mission per operation. Task teams who were able to obtain support from GETs valued its contribution, but many more reported being unable to obtain GETs support. The GETs provide high-quality advice in a top-down fashion to a limited number of priority operations and offer a partial solution to inadequate knowledge flows. In contrast, thematic groups are more inclusive mechanisms for staff to share knowledge and experience and are the primary vehicles through which sharing of tacit and embedded knowledge among regional staff occurs. Their inclusive nature and willingness to share knowledge was valued by task teams. Communities of practice (such as those on Financial Management Information Systems and Political Economy) are similar to thematic groups. Like thematic groups, membership is open to all interested staff who may access and contribute knowledge. Due to their openness, the thematic groups and communities of practice, when resourced, are able to capture and share knowledge across operational staff. GETS and thematic groups thus serve different purposes. While accountability and funding for the GETs has been established, rather than improving the governance of thematic groups most of them have been starved of resources and allowed to attenuate. While the incentives of the matrix system exacerbate the problem of knowledge dissemination, challenges go beyond just the matrix. Even The Promise of a Knowledge Bank       51 dissemination of country-specific analytical work—paid for by the country director’s budget to improve outcomes within the borders of that country, whose dissemination falls squarely within the interests of the country director—is substantially weaker than its relevance and quality. A 2009 QAG synthesis report of AAA assessments found most knowledge produced for clients to be strategically relevant and of high quality. However, the impact of analytical products has been more muted due to weaker dissemination and dialogue on the analytical findings and inadequate client ownership and involvement in the analytical activities. Results from 53 country AAA assessments undertaken from FY00-09 by QAG, covering 485 different tasks, show that in comparison with 87 percent moderately satisfactory or better on strategic relevance and 98 percent on internal quality, dialogue and dissemination was rated 72 percent and likely impact on government policy was rated 75 percent.9 Similarly, IEG found that only 55 percent of impact evaluations conducted by the Bank Group are disseminated to governments and clients (World Bank Group Impact Evaluations: Relevance and Effectiveness, 2012 forthcoming). Dissemination to clients was found to be a critical determinant of their influence on the policy dialogue and on the design of subsequent operations. Furthermore, of the 75 technical assistance activities in the QAG sample, 36 could not even be evaluated, indicating substantially weaker management oversight of technical assistance products and services. This suggests that incentives for knowledge work need to be revisited. Use of Knowledge Produced by Other Operational Units Knowledge is only useful if it is ultimately used for programs and operations in partner countries, whether or not the relevant program is financed by the Bank. The Bank must therefore go beyond the capture and dissemination of knowledge and actually use this knowledge. The use of analytical and advisory activities to underpin lending operations is somewhat limited. About half of the staff based in country units reported that their sector units do so to a substantial or very large extent. One field- based staff member in a country unit said: “AAA takes too long and is outdated by the time it is finalized. Often analytical work is not linked to any country operation and once finished is shelved and forgotten. Clients do not always see the value of AAA if it is not demand-driven and does not provide options and solutions to problems.� But the use of knowledge across organizational boundaries is even weaker. Table 3.3 shows that about one-quarter of the 262 respondents10 to the staff survey used analytical work produced by the anchor unit in their sector for operational work, and less than one-fifth used analysis from other Regions or other sectors. One focus group participant from a sector in SDN summarized the view of the group: 52 The Matrix System at Work Table 3.3 Few Staff Use Knowledge Produced Elsewhere Sector- and anchor-unit Survey question based staff In the preparation and supervision of your lending and substantial Percent indicating to a ­ non-lending operations, to what extent have you used ESW or very large extent or other analytic products produced by . . . . . . the anchor unit from your sector? 28 . . . sector units from other regions in your sector? 19 . . . sector units outside your sector? 17 . . . Development Economics VPU? 7 Source: IEG staff survey. “I was in operations for 13 years before moving to the anchor. Compared to the amount of knowledge generated, the amount that I was able to use was zero. The only time you could use knowledge was because of personal contacts through exchanges, or sometimes missions, or occasionally through thematic groups. But it was minimal.� Staff members report that incentives (lack of time and budget) and poor mechanisms to promote knowledge sharing prevent them from using the full range of knowledge available in their operational work. One focus group participant noted that once the time of the task team leader and the fiduciary and safeguards work are taken off the top, project budgets are only large enough for one or two technical people. Another asserted, “Knowledge is country-driven and country-retained.� Participants in multiple focus groups were universal in their assessment that “knowledge sharing is really personality driven and not institutionalized.� Sector managers and country directors provided a similar view regarding the use of knowledge products produced outside their own VPU, as shown in Table 3.4. Eleven of the 30 sector managers interviewed indicated that operations in their portfolio had used knowledge generated in other Regions to a substantial or very large extent, and the same number for other units in their Region. Country directors were slightly more positive regarding the use of knowledge produced elsewhere in their Region, but much less so regarding knowledge generated in other Regions. Knowledge Produced by DEC is Rarely Used by Operational Units Despite the recognition of the high quality of knowledge products generated by DEC, they are rarely used by operational staff. Seven percent of the 262 respondents to the staff survey reported using knowledge from DEC in their operational work, and 10 percent of sector managers reported that knowledge from DEC was used for the operations in their portfolios. A recent DEC paper (Ravallion The Promise of a Knowledge Bank       53 Table 3.4 Most Units Do Not Use a Substantial Amount of Knowledge Produced Elsewhere Sector managers Country directors Survey question (n=30) (n=10-11) SMs: To what extent have operations in your portfolio used knowledge products or analytical work produced by ... CDs: To what extent has your CMU received useful knowl- Percent indicating to a substantial edge from . . . or very large extent . . . Development Economics VPU? 10 18 . . . World Bank Institute? 17 0 . . . Network anchors? 0 . . . other regions? 37 18 . . . other units in your VPU? 37 50 Source: IEG staff survey. 2001) reports wide variability in familiarity with DEC’s research with less than 15 percent of staff from SDN sectors (primarily SDN) reporting familiarity compared to 60 percent for HDN and PREM staff, country-based staff being far less familiar with DEC research than staff in headquarters. Familiarity with DEC research is one step removed from using it in Bank operations. Those who benefited from actual collaboration with DEC were very positive about the high quality of input received from DEC staff, but the vast majority of operations did not benefit directly from DEC. DEC’s work is geared heavily toward knowledge as a public good. While this clearly is a useful contribution, the disconnect from the Bank’s country operations warrants some rethinking. At a minimum, DEC may want to invest in ways to improve internal dissemination of research findings in a form that is easily accessible by operational staff. Tacit Knowledge Does Not Flow There are two primary vehicles for the movement of tacit knowledge within the Bank, integral to making the matrix work effectively. The first is the periodic rotation of staff between departments and Regions. The second is the use of expertise from staff across units as temporary inputs for specific tasks, also known as cross support. Staff Rotation Staff mobility has been a slow vehicle for the flow of tacit knowledge. Under the Bank’s 3-5-7 policy in place since 2003, staff should remain in their assignments for three years for the sake of continuity, are encouraged to apply for rotation within five years, and should expect a rotation to a unit in another vice presidency [facilitated by their manager] within seven years. In practice, 54 The Matrix System at Work the policy was not implemented systematically and the internal labor market has not been efficient.11 In recognition that a large number of staff had spent more than seven years in their unit, a one-time mandatory rotation process was conducted for GH-level staff in the first quarter of FY09 and for GG staff in FY10. This resulted in a one-time level reduction in the share of staff working in the same unit for more than five years, as shown in Table 3.5, although no new policies or processes were introduced to address the underlying problem. Even after the mandatory rotation, in FY11, 29 percent of regional staff and 25 percent of network anchor staff had served in their VPU for more than five years and 17 percent of regional staff and 14 percent of network anchor staff had served for more than seven years. Overall, the share of staff working in a unit for more than seven years declined from 24 percent in FY09 to 19 percent in FY11, showing that even mandatory rotation was only partially successful in adhering to the seven year rule. An in-depth evaluation of HR policies and implementation—including recruitment, rotation, and exits— that could shed more light on some of these challenges to sharing knowledge was excluded from the scope of this evaluation. Staff transfers are only considered in terms of their impact on the flow of tacit knowledge. A natural time to capture tacit knowledge is when a long-serving staff member moves from a Region to a new VPU or retires, but the Bank lacks any mechanism and resources to “debrief� staff upon leaving a unit or to prepare staff moving to a new Region or country. While frequent staff Table 3.5 Proportion of Headquarters-Appointed GF-GI Staff in Unit for More Than Five Years FY09 Q1 FY10 Q1 FY11 Q1 # Percent # Percent # Percent AFR 231 42% 213 38% 189 32% EAP 93 37% 89 34% 69 25% ECA 136 43% 119 37% 93 29% LCR 121 37% 111 33% 101 29% MNA 57 31% 64 32% 53 27% SAR 74 32% 68 28% 56 24% Regional units 712 38% 664 35% 561 29% FPD 30 33% 30 29% 32 28% HDN 30 39% 23 25% 17 18% PRM 18 25% 17 20% 20 24% SDN 90 37% 85 32% 73 27% ­ nchor units Network a 168 35% 155 29% 142 25% Source: HR Analytics IBRD Quarterly Staffing Reports. Note: Staff graded GF–GI are professional staff up to the director level. The Promise of a Knowledge Bank       55 rotation can help to transfer operational lessons across Regions, it can also weaken institutional knowledge within any unit. Several staff and managers suggested that senior staff should be required to take a brief sabbatical or time-out between appointments to give them time to write a synopsis of the lessons they have accumulated from their time in the unit. Cross Support Cross support is defined by Bank management as the staff time of an expert or a specialist purchased from outside the responsible unit for a specific task. Budget data of cross support sold or purchased thus provides a “measurement of matrix functionality and organizational flexibility� from the knowledge perspective. Cross support is an indicator of the ability of anchors and Regions to make use of the tacit knowledge and expertise embodied in individuals across the Bank while making the knowledge and skills of their own staff available to others. A management review of FY08-10 data found inter-VPU cross support to be 3 percent of total staff time and resulted in the setting of higher targets for cross support. FY10 Bank-wide inter-VPU cross support sold accounted for 2.1 percent of staff time from regional VPUs and 8.3 percent from network anchors. However, estimated cross support as a proportion of total staff time results in underestimates of cross support by the technical sector staff who should be working across organizational boundaries, as opposed to support staff who cannot be expected to do so. Furthermore, eliminating silos and promoting cross- sector collaboration even within the Regions were deliberate aims of the 1997 matrix reform. The analysis was therefore expanded to include cross support across organizational units within VPUs. Variations in organizational structures and accounting practices across Regions posed significant methodological challenges. Accurate analysis of data covering all the Regions and networks is available only for FY10. For FY02- 09, inter-VPU cross support was analyzed separately for the four networks— FPD, HDN, PREM, and SDN—and for four of the six Regions (Africa, East Asia and the Pacific, Latin America and the Caribbean, and South Asia) for which data could be tabulated. Cross Support across VPUs IEG found cross support to be higher than the numbers in management’s review. Cross support across VPUs was about 5 percent of technical staff time in FY10.12 The figure was lower for staff in the Regions, at 3.3 percent of staff time for staff at level GF or above, than for staff in the network anchors at 12.5 percent. The Middle East and North Africa Region was the biggest seller among the Regions with 3.8 percent; Europe and Central Asia, with 2.9 percent, had the lowest inter-VPU cross support sales (see Appendix H for details). Demand for cross support varied considerably across the Regions. The Middle East and North Africa Region was also the biggest purchaser of cross support (10 percent), followed by South Asia (7.2 percent). Africa was at the regional 56 The Matrix System at Work Table 3.6 Total Cross-Support Sold by Regions and Network Anchors, FY10 Total cross support sold (% of seller’s Total cross-support sold (staff weeks) total time) Inter-VPU XS Intra-VPU XS Total XS Total time Inter-VPU XS Intra-VPU XS Total XS AFR 1351 4776 6127 41758 3.2% 11.4% 14.7% EAP 803 4249 5052 25150 3.2% 16.9% 20.1% LCR 703 3934 4637 20415 3.4% 19.3% 22.7% SAR 729 3847 4576 20648 3.5% 18.6% 22.2% 4 regions 3586 16806 20392 107971 3.3% 15.6% 18.9% Total cross support sold (% of seller’s Total cross support sold (staff weeks) time) Inter-VPU XS Intra-VPU XS Total XS Total time Inter-VPU XS Intra-VPU XS Total XS SDN 2663 2155 4818 21017 12.7% 10.3% 22.9% HDN 808 362 1170 5356 15.1% 6.8% 21.8% PREM 427 27 454 4476 9.5% 0.6% 10.1% FPD 686 639 1325 5958 11.5% 10.7% 22.2% All networks 4584 3183 7767 36807 12.5% 8.6% 21.1% 4 regions + 8170 19989 28159 144778 5.6% 13.8% 19.4% all networks Source: World Bank databases. Note: Europe and Central Asia and Middle East and North Africa underwent unit restructuring during FY10; reliable data on cross support within the Middle East and North Africa and Europe and Central Asia do not exist. average (6.5 percent), with the others somewhat lower. Among network VPUs, HDN, with 15.1 percent, was the biggest provider, followed by 12.7 percent inter- VPU cross support from SDN. Cross support from one network VPU to another was minimal (0.2 percent of staff time). Network anchors purchased an average of 7.8 percent of staff time, of which 5 percent was from the Regions. These findings indicate a slightly higher degree of cross support than recent findings by Bank management, and are more robust because they exclude nontechnical staff. Real inter-regional cross support, however, is less than one percent. While the aggregate volume of cross support sales across VPUs appears to be slightly higher than that reported by Bank management, in FY10 cross support across Regions was 1.5 percent of GF+ regional staff time. Even this amount masks actual collaboration across the Regions. Indeed, 0.6 percent of staff time was charged as cross support by staff who had transferred out of a Region within the past 12 months and were continuing work on their previous operations. The Promise of a Knowledge Bank       57 Correcting for this factor, actual inter-regional cross support in FY10 was 0.9 percent by staff who were not mapped to the purchasing unit in the previous fiscal year. The bulk of cross support from network anchors to regional sector units or vice versa has typically been within the boundaries of their respective networks. There was virtually no inter-VPU cross support from regional sector units to anchors of other networks during FY10. SDN and HDN anchor staff provided 9 percent and 10 percent of their time as cross support to HDN and SDN sector units, respectively, but very little to regional units of other networks (Appendix H, Tables H.4 and H.5). The share of staff time devoted to cross support has decreased from FY02 to FY10. Total inter-VPU cross support sold was 9,373 staff weeks in FY02, and 9,264 in FY10. Yet, as a percentage of staff time, inter-VPU cross support declined from 7.6 percent in FY02 to 5.2 percent in FY10. Cross-support from network VPUs to Regions also decreased over time. Total network anchor staff time increased from 24,537 in FY02 to 36,807 staff weeks in FY10 but the percentage of cross support from network VPUs to the Regions decreased from 23 percent in FY02 to 12.5 percent of network VPU staff time in FY10 (Figure 3.3) indicating that a growing proportion of anchor staff time is consumed by corporate or global actions. Cross Support within VPUs In FY10, cross support within VPUs (intra-VPU cross support) across sector units, between SMUs and CMUs, or from the regional functional nodes to SMUs and CMUs was almost three times as high as inter-VPU cross support. Intra-VPU cross support averaged 13.8 percent overall, 15.6 percent in the Regions, and 9 percent in the network anchors (Figure 3.4). Variations across networks and Regions were significant: cross support within VPUs amounted to 10 percent staff time in the SDN anchor but was barely 0.6 percent in the PREM anchor, indicating that silos are strong within the PREM anchor. Intra-regional cross support also appears to have declined in FY10 compared to previous years. Intra-regional cross support for the period FY02-10 is available on a comparable basis for four Regions—Africa, East Asia and the Pacific, Latin America and the Caribbean, and South Asia.13 In East Asia and the Pacific and South Asia, intra-regional cross support remained steady in FY10 (compared to the average for FY02-10) at 17 percent and 18.6 percent, respectively. In Africa, it dropped from an average of 14.8 percent to 11.4 percent in FY10, possibly due to a change in budget settlement rules for fiduciary and safeguards services, which are no longer considered cross support. In Latin America and the Caribbean, it dropped from the FY02-10 average of 21.5 percent to 19.3 percent in FY10, which is still the highest among the Regions. On aggregate, intra-VPU cross support for the four Regions declined from an average of 17.3 percent in FY02-10 to 15.6 percent in FY10. 58 The Matrix System at Work Figure 3.3 Cross-Support from Network VPUs to the Regions, FY02-FY10 Percentage cross-support sold 35% 30% 25% 22.8% 20% 15% 12.5% 10% 5% 0% FY02 FYY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Fiscal Year SDN HDN PRM FPD Total 4 Anchors Source: World Bank databases. Note: SDN = Sustainable Development, HDN = Human Development, PREM = Poverty Reduction and Economic Management, and FPD = Financial and Private Sector Development. Intra-VPU cross support consisted largely of sales across sector units within network boundaries and sales from central units to sector and country units. Cross-support sales by sector staff to other SMUs in their network averaged 8 percent over the period FY02-10, ranging from 10.1 percent in East Asia and the Pacific to 5.6 percent in Africa. Cross-support sales by regional function units Figure 3.4 Cross Support Under the Matrix System in FY10 0.9% Other Regions Other units 15.6% Regions in Region 18.3% total 0.6% Other VPUs 1.2% 11.4% 0.9% Other units 8.6% in Network Networks 21.1% total Other Network 0.2% Source: World Bank databases. Note: The system records an additional 0.6% inter-regional cross support that in practice is cross support to a staff member’s previous unit within a year of departure. The denominator for Regions ­ is total staff weeks in Regions; for networks, it is total staff weeks in networks. The Promise of a Knowledge Bank       59 to sector and country units averaged 4 percent across the four Regions, ranging from 3.8 percent in South Asia to 5.6 percent in Africa, but was exceptionally low (0.4 percent) in East Asia and the Pacific, which has the highest degree of decentralization (Appendix H, Table H.7). It is noteworthy that the cross support from the regional function units in Africa is the same as that across sectors within the Region, indicating a higher degree of centralized financial management and procurement services from the regional function unit. Efficiency and Effectiveness of the Knowledge Bank This chapter has assessed questions on the Bank’s production and dissemination of knowledge. Based on the findings, the evaluation draws conclusions in this section regarding the extent to which the expected intermediate outcome— “Innovation in the field is well captured across the institution and country operations draw efficiently on global knowledge to meet client needs�—has been achieved. During the period evaluated, the matrix system has been highly inefficient and modestly effective in achieving this outcome. Rather than functioning as a global organization, the evidence from the flow of virtual knowledge and technical expertise indicates that the Bank is at risk of evolving into a group of regional banks interconnected by fraying ties among the Regions and between the Regions and networks. Knowledge that is generated by the Bank or captured from operations implemented by clients is rarely used in future operations or in other contexts. Decentralization has created further impediments to the flows of knowledge and expertise, and until the underlying incentives and constraints inhibiting effective knowledge flows are addressed, further decentralization will inevitably aggravate these problems. The three forms of knowledge—explicit, tacit, and embedded—do not draw on each other efficiently, which reduces the Bank’s development effectiveness. Most staff, particularly those in the Regions and country offices, are unable to draw on existing knowledge generated inside and outside the Bank efficiently, and dialogue and dissemination is less effective, which reduces likely impact. The volume of expenditures by the Bank on explicit knowledge has been growing and reached $606 million in FY10. QAG assessments of AAA indicate that although knowledge produced for clients tends to be strategically relevant and of high quality, weak dissemination and dialogue limits their impact. To the extent that knowledge is disseminated, this is limited primarily to the client for whom it is produced. Knowledge products are not stored in an easily searchable and retrievable form and are rarely used by staff outside the units where they are produced. Notwithstanding the analytical quality of the Bank’s knowledge products, much of the Bank’s knowledge thus has a limited shelf life and limited use value. The only exception is knowledge produced by DEC, which is widely disseminated to a global audience, but Bank operations make very limited use of DEC’s knowledge. The Bank generates a large volume of knowledge embedded in its operations but lacks the ability to efficiently capture and share that knowledge. The Bank’s knowledge report (Knowledge for Development 2011: The State of World 60 The Matrix System at Work Bank Knowledge Services) estimates that the Bank spends $2.5 billion a year14 on embedded knowledge, four times the amount spent on core knowledge products. Embedded knowledge produced for individual operations suffers from the same production inefficiencies as explicit knowledge, and is mostly buried within the documentation of that project or program. Management does not have any system in place to capture, share, or even track embedded knowledge across the Bank. Given the volume of embedded knowledge, these inefficiencies need to be addressed urgently. The availability of trust funds has led to an increase in core knowledge production by 50 percent, but the Bank’s budget for knowledge barely increased. The growing reliance on trust funds generates inefficiencies and risks. While the share of the Bank’s gross administrative budget spent on core knowledge has remained steady at approximately 15 percent over the FY02-10 period, trust funds for core knowledge have more than doubled in real terms over the same period, growing from $107 million in FY02 to $245 million in FY10. This has resulted in growing reliance on external financing for the Bank’s core knowledge products, with external financing for core knowledge growing from 27 percent in FY02 to at least 40 percent in FY10 (Figure 3.5). Reliance on trust funds also gives rise to questions about relevance and quality since most BETFs are earmarked for special, sometimes supply- driven, purposes. IEG’s evaluation of trust funds (IEG 2011) also raised concern over the growing use of BETFs as they are not allocated within the Bank’s budgeting process which relies on the tensions between country and sector units to balance the needs of both sides of the matrix structure, leading to non-strategic allocation of resources. Staff report high transaction costs both in soliciting funds from multiple trust funds and in tracking and reporting separately on a myriad of different trust funds that have their own grant- Figure 3.5 BETFs Now Finance More Than 40 Percent of the Bank’s Core Knowledge Products 700 (Constant 2010 US$ millions) 600 500 27% 28% 28% 31% 30% 33% 31% 36% 40% BETFs 400 BB Actuals 300 200 73% 72% 72% 69% 70% 67% 69% 64% 60% 100 0 FY02 FYY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Source: World Bank database. The Promise of a Knowledge Bank       61 making procedures and reporting requirements. The priority given to lending induces even greater reliance on trust funds for AAA tasks since country directors feel that the sector units will be able to mobilize trust funds for AAA more easily. But this increases inefficiency and undermines the effectiveness of strategic planning for knowledge, sometimes leading to a distortion of Bank priorities and quality problems arising from looser oversight than for operations financed by the Bank’s budget. Trust funds play a significant role in the implementation of corporate priorities such as governance and gender, and GPGs, such as climate change, which may not otherwise be integrated into country programs. However, this undermines the effectiveness of strategic planning for knowledge and creates a disincentive to integrate these corporate priorities within country program budgets. This may be justifiable for those GPGs where the externalities are significantly greater than in-country benefits. Tacit knowledge is largely unmonitored, making it difficult to assess its efficiency and effectiveness. Tacit knowledge resides in the form of operational experience of Bank staff, clients, and partners and is much harder to monitor. While the magnitude of explicit knowledge is known—at least in terms of the number of knowledge products “delivered� and spending on knowledge products—“the Bank has yet to put a value on the tacit knowledge staff, partners, and clients have developed in the course of their work� (World Bank 2011: 20). The only portion of tacit knowledge that can be tracked and analyzed at present is the movement of staff and cross support, and even that is constrained by incomplete and inconsistent data across the Regions. Sector silos persist, but Region and network silos are even stronger. Overall, the flow of tacit knowledge through cross support has been rather limited, is declining in volume, and is constrained by structural boundaries. This is not what the matrix system aimed to achieve. First, there has been a decrease in inter-VPU cross support. Second, the percentage of cross support from network anchors to the Regions has halved during the same period as a percentage of network anchors’ staff time. Third, real inter-regional cross support is miniscule; the incentive of staff to prioritize operational services in their own Region creates strong regional silos. Fourth, to the extent that there is cross support from the network anchors to the Regions and from the Regions to the network anchors, those flows occur within network silos. There was virtually no cross support from SDN sector units to sector units of other networks, and not much cross support among sector units across other network boundaries. The communication tools for the Bank to act as a connector of knowledge are available, but not systematically used. Some sector groups (such as the HDN chief economist’s seminars) connect sector staff in the Regions through webinars, which are more easily accessible by staff in country offices. The governance and anticorruption team has also been in the forefront of using technology to reach out to regional staff. However, this is more the exception than the rule. Operational staff in the Regions, especially those in country 62 The Matrix System at Work offices, complain about the ineffectiveness of the Bank’s role as a connector of knowledge. The problems arising out of weak knowledge flows have become more acute with increasing decentralization of sector staff to country offices, 54 percent of technical staff mapped to the Regions now being country-based. The dissemination of operationally relevant global knowledge was considered to be substantial by only 25 percent of staff surveyed and internally generated knowledge by only 40 percent of staff. Managers interviewed were not more encouraging. Recognizing that personal networks are the primary mode of knowledge sharing, staff and managers alike lamented the loss of networking opportunities due to the scaling back of “sector weeks.� In the absence of other effective, and readily accessible, mechanisms for sharing of virtual knowledge, sector weeks are an effective forum for sharing new initiatives and cross-fertilization of operational knowledge, especially for country-based staff. They are also the most valued opportunity for country-based staff to network and demonstrate their ability to their sector colleagues, a prerequisite for mobility across the Bank. Innovation in the field is not well captured across the institution, and country operations do not draw efficiently on global knowledge to meet client needs. Given the vast amount of resources expended on the Bank’s knowledge work—$4 billion annually, of which $2.5 billion is embedded knowledge—knowledge that is not captured or disseminated has limited use, often to a single client. Innovations and knowledge generated at the country level, is rarely disseminated further. Despite the numerous seminars organized by the networks, the travel schedules, decentralized location of operational staff, and time differences with Washington reduce their reach. Seventy percent of task teams report that they rely primarily on their own resources and personal networks for knowledge. This is not an efficient use of resources given the constraints of the internal search engines and limited opportunities for collaborative networks to access knowledge, and is a far cry from the promise that the Knowledge Bank offered when the concept was introduced in 1996 and was a major justification for introducing the matrix system. Based on the Bank’s performance since the introduction of the matrix, both in efficiency and effectiveness, the intermediate outcome of the Bank’s knowledge work—capturing innovation across the institution, and drawing on global knowledge to meet client needs—has not been achieved. This shortcoming assumes even greater importance as feedback from country surveys indicates that clients value the Bank’s knowledge and policy advice even more than its financial assistance. Improving its knowledge portfolio is critical to the Bank’s future. The Bank has recently launched several initiatives to rectify these shortcomings. The creation of the KLC is an attempt to redress the weaknesses in the Bank’s knowledge agenda. The KLC has increased corporate oversight of knowledge activities and launched several initiatives, including GETs and The Promise of a Knowledge Bank       63 Knowledge Platforms. The restructuring of the FPD into Global Practice Teams is also aimed at improving knowledge flows. The KLC also oversaw production of the first-ever Knowledge Report, following which management has committed to actions to fill gaps in governance arrangements of individual knowledge products, establish a results framework for all core knowledge products, and establish mechanisms to promote greater connectivity and support management of knowledge as a portfolio. A more detailed implementation plan is under preparation. While these are steps in the right direction, the underlying incentives that effect demand for and use of knowledge will also need to change if these initiatives are to address the constraints that have hampered the Knowledge Bank. The Bank’s role in development knowledge is equally vital to the quality of its lending portfolio. In addition to knowledge services that have a direct bearing on policy advice and institutional development, there has always been a complementarity between knowledge and lending. IEG’s evaluation of the Bank’s use of knowledge to improve development effectiveness found that ESW and technical assistance was effective in shaping Bank lending and was statistically associated with better loan design (IEG 2008). The Bank’s ability to produce high-quality knowledge needs to be preserved to improve the quality of Bank lending, and its ability to access and customize cutting-edge global knowledge is essential to maintaining its relevance to the increasingly complex demands of its clients. Notes 1 A campaign against the multilateral banks, targeted most pointedly at the World Bank called “Fifty Years is Enough� was launched in 2004 (the 50th anniversary of the World Bank) by a group of NGOs advocating economic justice and environmental protection. Doug Hellinger subsequently wrote that despite its rhetoric, in the sixth decade the Bank has continued with business as usual, even backsliding in some key areas (see “Fifty Years was Enough,� The Development Gap, September 2004. http://www.developmentgap.org/ worldbank_imf/50Years/50_Years_Was_ Enough_Sept_2004.pdf) 2 Global Development Finance: External Debt of Developing Countries, World Bank 2011. 3 By the end of FY11, expenditure on strategic knowledge activities, overseen by the Knowledge and Learning Council was projected to be $10.1 million out of a budgeted amount of $11.2 million. Major expenditures were on the GETs ($4.7 million) and Knowledge Platforms ($1.9 million), with about $1 million each to unspecified “support to front line teams� and South-South innovation managed by WBI. (See The World Bank’s Budget: Trends and Recommendations for FY12. August 9, 2011 (R2011-0134/1, IDA/R2011-0206/1), page 21.) 4 In FY11, 14 years after the matrix was introduced, Bank management has introduced a Corporate Scorecard to broaden results measurements, including outcome level reporting on knowledge. 64 The Matrix System at Work 5 Includes the Bank’s budget plus non-BETF external financing (reimbursables and fee income). 6 The nine product groups were defined by the Knowledge Strategy Group in “Transforming the Bank’s Knowledge Agenda: A Framework for Action.� February 2010. The World Bank. 7 See “Towards Open Knowledge: The State of World Bank Knowledge Services.� Washington DC: World Bank, 2011, page 14. 8 In a country survey of 4,710 stakeholders, 22 percent said that technical advice and policy and economic advice was the greatest value (even more than financial resources) brought to their country by the World Bank, while another 22 percent said it was the second greatest value. See “Country Survey 2010 Review,� The World Bank. 9 “Synthesis of Past Assessments of Analytical and Advisory Activities (AAA): A QAG Review,� Quality Assurance Group, September 17, 2009. 10 The respondents to this question exclude staff from country units and network anchors, yielding a response rate of 35 percent. 11 Mobility of staff through transfers has proved to be very difficult, however, and had to be imposed through a mandatory rotation system called the 3-5-7 system which forced the managed rotation of staff who had been in one vice presidency for more than seven years. However, that one-off rotation has not been continued on an ongoing basis. 12 Variations in organizational structures and accounting practices across Regions posed significant methodological challenges. Accurate analysis of data covering all the Regions and networks is available only for FY10. For FY02-09, inter-VPU cross support was analyzed separately for the four networks-FPD, HDN, PRM, and SDN- and for four of the six Regions (Africa, East Asia and the Pacific, Latin America and the Caribbean, and South Asia) for which data could be tabulated. 13 Europe and Central Asia and the Middle East and North Africa underwent unit restructuring during FY10; reliable data on cross support within the Middle East and North Africa and Europe and Central Asia were unavailable. 14 This estimate of embedded knowledge has been extrapolated from a one-off estimate made by the Middle East and North Africa, the Region with the smallest lending portfolio and may be an overestimate. Even so, embedded knowledge is much larger than explicit core knowledge. References IEG (Independent Evaluation Group). 2008. Using Knowledge to Improve Development Effectiveness: An Evaluation of World Bank Economic and Sector Work and Technical Assistance, 2000–2006. Washington, DC: World Bank. The Promise of a Knowledge Bank       65 Pincus, Jonathan and Jeffrey Winters, eds., 2002. Reinventing the World Bank. Ithaca, NY: Cornell University Press. Phillips, David A. 2009. Reforming the World Bank: Twenty Years of Trial and Error. New York: New York University Press. Ravallion, Martin. 2001. “Knowledgeable Bankers? The Demand for Research in World Bank Operations.� Washington, DC: World Bank Development Economic Group, Policy Research Working Paper 5892. Wapenhans, Wili A. 1992. The Wapenhans Report. Portfolio Management Task Force. Washington, DC: World Bank. (Also published as Effective Implementation: Key to Development Effectiveness.) World Bank. 2011. “Knowledge for Development 2011: The State of World Bank Knowledge Services.� Washington, DC: World Bank. ———. 1994. “Efficiency and Effectiveness: Is the World Bank Group Well Prepared for the Task Ahead?� In Bretton Woods: Looking to the Future, ed. The Bretton Woods Commission, pp. C289-304. Washington, DC: Bretton Woods Commission. 66 The Matrix System at Work Chapter 4 Delivering Quality Services Chapter Highlights • Although outcomes of Bank-financed operations improved significantly over the past two decades, there has been a significant decline in outcome ratings since FY06. • Quality assurance systems are more effective for lending than for ESW and are weakest for NLTA. • The focus on reputational risk has crowded out attention to other aspects of quality of lending operations. • Accountability for quality is highest for task team leaders and nonexistent for sector directors. • Regional sector units, responsible for operational delivery, are receiving a decreasing share of regional budgets. • Messages from senior ­management, control over task budgets, and desires for promo- tion and salary increase are the primary incentives for meeting lending targets. • Job satisfaction and ­ recognition by peers are the ­ primary incentives for ­ “matrix ­ behaviors� designed to i­mprove development o ­ utcomes, including focusing on results, collaborating across sectors, and mobilizing Bank-wide expertise. • Disincentives arising out of budgetary and information management rules discourage cross-sector collaboration. Introduction The matrix system was designed to strengthen the Bank’s ability to deliver quality services that combine relevant local solutions with global knowledge. This was to be achieved by the combination of country priorities articulated by country management and global technical expertise brought to bear by sector management and the Bank-wide networks. The key question this chapter seeks to answer is “Are incentives and accountability mechanisms aligned in a manner that rewards high quality in client services and results?� To answer this question, the evaluation assesses the extent to which: • incentives and accountabilities are clearly defined to assure quality. • financing and budget incentives influence the quality of outputs and promote effective knowledge and technical services. • the matrix system fosters cross-sector collaboration and teamwork to achieve results. The evaluation has looked at both formal and informal incentives in recent years to assess the effectiveness of the matrix system. Improving quality in portfolio performance and encouraging greater cross-sector work were objectives of the 1997 reforms. This chapter briefly discusses the role of formal and informal incentives for quality and client orientation and their links to institutional arrangements. 68 The Matrix System at Work It then presents an overview of recent trends in portfolio outcomes to assess if there have been any changes in quality and in the patterns of sector and thematic distribution of the portfolio. The chapter then focuses on incentives, drawing on data from survey results and manager interviews as well as analysis of budgetary incentives. The chapter then presents a brief review of quality assurance mechanisms and accountabilities for results. Based on these findings, the evaluation draws conclusions regarding the efficiency and effectiveness of the matrix system in achieving the related intermediate outcome—“Incentives and accountability ensure teamwork and high quality country dialogue and client services.� Portfolio Trends in Quality Outcomes of Bank-financed operations as measured by year of exit from the portfolio improved significantly over the past two decades, rising from less than two-thirds satisfactory in the early 1990s to a peak of nearly 80 percent satisfactory in the mid-2000s, as shown in Figure 4.1a. Reporting outcomes by year of exit is a lagging indicator of quality at entry, reflecting performance of projects prepared an average of seven years earlier and affected in part by Bank supervision. Of the lending operations exiting the portfolio in FY01-10, 76 percent of operations were rated moderately satisfactory or better, compared to 68 percent for those exiting in FY1991-2000, a difference that is statistically significant. In more recent years, quality has become a concern. After the long upward trend in outcomes, the trend has turned decidedly negative. The share of Bank- supported operations exiting the portfolio in FY08-10 whose outcomes were moderately satisfactory or better was 74 percent, compared to 79 percent in FY05-07, a decline that is statistically significant at 95 percent confidence.1 This decline is more evident for investment projects, which are typically prepared six to seven years earlier, than for development policy operations. Indeed, partial results for FY102 were the worst in an individual year since FY99 (although year-to-year variations should be interpreted with caution). This suggests that the Bank’s quality assurance system deserves a closer look. For project outcomes by year of approval, the most significant improvement in the share of Bank-financed operations with moderately satisfactory or better outcomes occurred before the introduction of the matrix. The share of operations with outcomes rated moderately satisfactory or better was 74 percent for those approved in FY91-96, and 76 percent for those approved in FY1997-2005; the difference is not statistically significant. The slight gain for projects approved during the early years of the matrix (FY1998-2001) was lost by projects approved after FY01. A downward trend commenced after that peak (Figure 4.1b). Using a three-year moving average, the share of operations approved in FY01-03 rated moderately satisfactory or better was 79 percent, compared to 70 percent among operations approved during FY04-05, a drop that is statistically significant with 95 percent confidence. This trend needs to be monitored carefully. Delivering Quality Services 69 In addition to assessing the outcomes of operations after they have closed, IEG assesses Bank performance in ensuring quality at entry and the quality of supervision. Supervision is more crucial for investment lending projects, which are typically active for six to seven years, than for development policy operations which are typically closed within a few months of approval. For this reason, the analysis is restricted to investment lending projects. The quality of supervision and quality at entry have shown a declining trend since the introduction of the matrix. The quality of supervision peaked at 86 percent moderately satisfactory or better during FY97-99, falling to 78 percent during FY04-06, as shown in Figure 4.2, a difference that is statistically significant at 95 percent confidence. Bank performance in ensuring quality at entry peaked at 74 percent during FY96-98, falling to 67 percent during FY04- 06, but this decline is not statistically significant. Figure 4.1 Outcomes of Bank-Financed Operations by Fiscal Year 4.1a. By Fiscal Year of Exit 90% 3-year average Share moderately satisfactory 85% 80% 75% or better 70% 65% 60% Launch of matrix 55% system 50% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 4.1b. By Fiscal Year of Approval 90% Percent moderately satisfactory 85% 80% 75% or better 70% 65% 3�year average 60% Launch of matrix 55% system 50% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Source: IEG ICR review evaluations. 70 The Matrix System at Work Figure 4.2 Investment Lending Quality and Outcomes by Fiscal Year of Approval 90% Share of moderately satisfactory Quality of supervision or better (3�yr centered avg) 85% 80% 75% Outcomes 70% Introduction of matrix system 65% 60% 55% Quality at entry 50% 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 Source: IEG evaluations. Three indicators offer additional clues to the quality of the portfolio, based on the information made available in real time to task teams and managers. Without candid information on the likelihood an operation will meet its objectives, Bank managers and government implementing agencies cannot allocate attention and resources to where they are needed. The net disconnect compares self-ratings to ultimate ratings in independent evaluations: it is the difference between the percentage of projects rated as moderately unsatisfactory or worse by IEG and the percentage rated as such by the Regions in the final Implementation Status and Results report. The Bank has never met its target of 5 percent or lower. The disconnect has been at least 14 percent in every year since FY07. The realism index compares self-ratings to recent history: it is the ratio of the percentage of actual problem projects in the active portfolio and the percentage of operations with moderately unsatisfactory or worse outcomes as rated by IEG for the most recent completions (1,000 evaluations), on a rolling basis. The Bank has set a target of 70 percent for its realism, and despite improvement over the past decade, the target has only been met once (in FY09) in the past 10 years IEG ratings of the quality of monitoring and evaluation (M&E) in operations assess the extent to which M&E systems were appropriately designed, implemented, and utilized and serve as a measure of the real-time feedback available for project management. These ratings were fully introduced for operations exiting the portfolio beginning in FY05. A high quality rating would indicate an M&E system with a strong influence on project performance. M&E systems have been weak during the entire period for which data are available, with just 25 percent rated of substantial or high quality among operations closing in Delivering Quality Services 71 FY10, and never more than 36 percent reaching that threshold. The Bank has not identified a target for this rating, but compared to the other dimensions assessed by IEG highlighted above, these ratings are very low. One last indicator—the proactivity index—measures the extent to which the Bank has taken action when problems arise, and is defined as the proportion of projects rated as actual problem projects 12 months earlier that have been upgraded, restructured, suspended, closed, or partially (20 percent plus of commitments) or fully canceled. This index has followed a downward path for a decade, dropping to 60 percent in FY01. It has not met its target of 80 percent since FY08. Institutional Incentives for Quality Institutional Incentives for Client Orientation and Technical Excellence after 1997 The 1997 reforms differ qualitatively from the earlier ones in the budgetary arrangements and introduction of dual reporting relationships. The success of the matrix reform was predicated on the alignment between the new organizational arrangements and accompanying incentives, and on the effectiveness of the framework of formal and informal incentives. The new organizational arrangements consisted of (a) reducing the number of countries covered by each country unit (thus increasing client orientation), (b) creating networks (and Sector Boards) to link technical sector units at the regional and Bank-wide level (to increase global technical expertise available to country teams), and (c) creating network anchors (to facilitate exchanges of global technical knowledge among the sector units and with external partners). In addition, decentralization was intended to increase client orientation by facilitating more frequent contacts with clients. The formal incentives consisted of country budgets allocated by country units to the different sector units and the dual reporting relationships in the sector units. This, and not the creation of the networks and Sector Boards (or decentralization), is what makes the reforms of 1997 qualitatively different from the earlier ones and made the Bank a matrix-managed institution. New budgetary arrangements were introduced with country budgets allocated to sector units by country departments to enhance client orientation. Prior institutional arrangements left excessive room for supply-driven behavior by sector units because of their ability to deploy budget resources to countries without need for approval by the country units. Although this problem had been reduced after the reorganization of 1987 it had not completely disappeared. After 1997 the introduction of country budgets was intended to allow the country units (country director/CMU) to purchase services from the different sector units in line with the CAS and country needs. Sector units, on the other hand, had the skills (technical staff) but needed to finance their staff with budget allocations from different country units. The fluctuating demand among countries was expected to balance out, particularly with arrangements designed to make 72 The Matrix System at Work Bank-wide resources available to individual country teams through cross-sector teamwork and cross support. The formal incentive of dual reporting relationships in the sector units was envisaged to ensure technical excellence and strategic alignment between country and sector strategies. The founding documents envisaged that regional sector managers and sector directors would report to the regional vice president as well as to the network vice presidents. Sector staff members working in a given country were expected to report to their sector managers as well as to the country directors. In practice, formal reporting arrangements for sector staff were to their regional sector managers, while informal incentives, particularly for staff decentralized to country offices induced them to be more responsive to country directors. Country units, however, continued to operate in a line management environment: country directors report only to regional vice-presidents. Moreover, as discussed in chapter two, alignment between country and sector strategies has been weak. The matrix system was put in place in part to ensure delivery of quality services by Bank-wide networks of sector specialists. Quality enhancement reviews were introduced by QAG in FY00 to focus on quality by enhancing the role of sector specialists. They were subsequently transferred to the networks and Regions. Internal and External Factors Budgetary processes and informal incentives had an impact on the effectiveness of the matrix system. The new budgetary system quickly ran into difficulties because its detailed functioning principles had been left undefined and resulted in variations and sometimes inconsistencies across the Regions.3 In the first few years the transaction costs were very high. The uncertainty created for large numbers of technical staff members who were concerned about finding effective demand for their services and generated supply-driven behaviors to ensure new tasks. Sector managers felt obliged to find work for their staff rather than allow the labor market to determine demand for different staff members. Organizational culture also played a major role in undermining the incentives to use global knowledge through increased cooperation among sectors and across Regions: defensive “turf� behaviors developed quickly and persist today, particularly among sector units with overlapping mandates. The poor performance of Sector Boards is partly attributable to such behaviors, as well as to the excessively large span of control of sector managers. External effects also had an impact on the matrix system. Just when the matrix system was being introduced, the onset of the Asian financial crisis created an unanticipated increase in demand for quick-disbursing loans, which distracted management’s attention away from the Knowledge Bank that was to offset the lending culture at precisely the same time that new incentives and processes were to be put in place to support the matrix structure. The growing importance of GPGs and the expanded external role of the networks in Delivering Quality Services 73 the new international aid architecture and growing corporate demands, which are unrelated to the internal matrix management of country and sector units, also had an impact on the matrix by diverting the attention of the network anchors. Incentives Arising from the Metrics of Success The matrix system was expected to reduce the emphasis on lending volume as the metric of success, but has not succeeding in doing so. The long- standing “lending culture� continues to privilege Board approval of lending operations and disbursement ratios over results in client countries. The Bank monitors and reports regularly on lending commitments, and supervision reports track disbursements closely, sending a signal that volume rather than quality is the measure of success. On knowledge products, the Bank tracks the number of ESW and technical assistance deliveries and the resources expended on them. It was only in FY11 that the Bank introduced a Corporate Scorecard and a message from the president announced that the World Bank was embarking on “the next phase of the results agenda� (President’s End-of- Year Message, July 1, 2011). Even so, that commitment to results was followed immediately by a list of the volume of commitments and investments by each arm of the World Bank Group. The priority given to lending (while not new) and the numeric count of knowledge products rather than results generate incentives for management and staff that can lead to suboptimal outcomes. Managing for Quality under the Matrix System Quality under the Matrix System The matrix system was expected to simplify review processes with clear responsibilities for quality. The design of the matrix system had envisaged dual responsibility: the articulation of client demand was the responsibility of country management, while primary responsibility for quality rested with the managers of sector units. Quality was to be ensured through a simplified, two-stage review process for project appraisals, supplemented by independent reviews by QAG to provide third-party feedback to senior management on the quality of the portfolio. In practice, additional formal and informal quality control functions have been added to the “simplified review process.� Sector managers interviewed for this evaluation consider these multiple roles, incentives, and accountabilities to be “absolutely inconsistent with the official [matrix] arrangements.� The matrix systems’ aim to improve quality has been undermined by the growing number of quality assurance layers in different units— sector units, regional OS units, and CMUs—with a growing emphasis on operational risks that could damage the Bank’s reputation, with relatively less attention to other aspects of quality and development impact. CMUs were originally expected to rely on quality assurance systems established at the regional level and the sector level. Over time, they have hired operational 74 The Matrix System at Work advisers and other operational staff in the CMUs, adding another layer of clearance and quality control. The country directors’ role in quality assurance receives mixed appreciation. While some sector managers who were interviewed view their role and the additional layers of control they have introduced as helpful, others find them both misplaced and redundant. “Two sets of eyes (in the CMU and the SMU) looking at the portfolio are better than just one. It is good to have the CMU also look over quality.� But also, “I am accountable for quality but the CMUs have also started doing technical reviews through their own quality assurance systems because accountabilities are not clear and they want to protect themselves.�4 Moreover, the quality improvements from these additional layers often consist of safeguarding against risk or presentational changes in project documents rather than substantive changes in project design. Regional functions units have grown substantially as they have strengthened their fiduciary and safeguards teams. For example, fiduciary staff in the Regions grew from 53 in FY02 to 407 in FY10. The growth during the first two years can be attributed to remapping of country unit staff to OS units reaching 158 staff in FY05 but the subsequent increase reflects heightened attention to fiduciary areas. A few high-profile investigations of corruption and safeguards have resulted in widespread defensive behavior, leading many country directors and some sector units to add staff to undertake additional quality reviews focused on operational and reputational issues. Feedback from staff and managers confirms that the Bank’s quality control systems are more concerned with reputational risks than other aspects of quality. Staff survey results reinforce the findings from managers’ interviews that the quality assurance system is imbalanced. Based on their operational experience, the majority of the 342 staff responding to the survey reported that the quality assurance system ensures high quality for fiduciary aspects (59 percent) and safeguards aspects (54 percent), while one-third report effective use of peer reviews (34 percent) and quality enhancement reviews (33 percent), and high quality of technical design (36 percent). Roughly one-fifth reported that the quality assurance system provided independent feedback on quality, and that there is effective supervision for timely and frank feedback on projects. Current incentives and accountability mechanisms lead to task team leaders having the greatest accountability for lending operations and sector directors having the least accountability. Sector directors were identified as primarily accountable for lending or nonlending activities beyond business development in only one of the four Regions that have formal accountability matrices (see Appendix G), while two Regions do not even have formal accountability matrices. Country directors were accountable for concept review meetings, and in all but East Asia and the Pacific, decision meetings. Sector managers and country directors were identified as the managers most accountable for lending operations. A third of the 342 respondents Delivering Quality Services 75 to the staff survey reported that sector managers and country directors were accountable for lending quality to a substantial or very large extent, another third considering them accountable to some extent. Their accountability is reported to be significantly lower for AAA quality. At the same time, the span of control of regional sector managers—managing 37 staff on average, compared to seven to 15 in comparator external organizations and 19 in the International Finance Corporation (IFC)—is unrealistic, making it impossible for them to provide quality control for all lending and nonlending operations produced in their unit. They are instead forced to rely on sub- regional or country-based “shadow managers,� introducing an additional layer of non-technical review, as they are not always experts within the sector. Information from managers’ interviews validates these findings. Sector managers typically described accountability for quality as high for the sector manager and the task team leader, none for the sector director, and substantial for the country director. Across all the Regions, sector directors’ accountability for lending operations was rated as substantial (or very large) by only 13 percent (four of 30) of sector managers and 9 percent (one of 11) of country directors interviewed. Sector directors are deemed accountable only for cases where technical assistance has substantial potential impact on country dialogue. In comparison, sector managers are reported to have much greater accountability for lending, which was rated substantial by 90 percent of the sector managers and 55 percent of country directors interviewed. The staff survey yielded similar results. Quality assurance systems are described as being clearer and more effective for lending than for ESW, while accountability for NLTA is even weaker. Review processes have been in place for lending operations since before the matrix. A two-stage review process was introduced for ESW following the 2001 evaluation of the Strategic Compact, but until FY11 no Bank-wide requirement was in place for review of NLTA quality. Managers and staff report that performance of the current quality assurance system is relatively effective in ensuring fiduciary and safeguards quality, but less so for assuring the quality of lending operations, even lower for the quality of ESW, and lowest for NLTA. Feedback from country directors and sector managers is fairly consistent. Managers find that the current incentive structure results in attention to processes and reputational risk more than operational and AAA quality. Many argued that, as one respondent put it, “too much attention on safeguards and fiduciary takes away from time spent ensuring technical quality of lending operations,� and that “the quality assurance system, such as it is, provides some assurance for fiduciary/safeguard issues and ESW reports, but slight [assurance] for quality of lending and none for NLTA.� The implications of weak accountability of NLTA are especially serious for institutional interventions, such as those in the public sector governance portfolio. There is little opportunity for the quality assurance system to materially improve many pieces of ESW. For nearly 40 percent of ESW that have decision meetings—in which a study is put up for institutional scrutiny, and 76 The Matrix System at Work is commented on by peer reviewers—those meetings occur less than 30 days before delivery to the client, as shown in Figure 4.3. It is implausible that substantive peer review critiques of a sophisticated piece of analytical work can be adequately addressed by such a large share of ESW in such a short period of time. Furthermore, decision meetings are held within 10 days of delivery to the client for more than 15 percent of these studies. Some studies are even shared with the client the day of the meeting. The June fiscal year deadline in effect “cuts off� the period in which a task team can make significant improvements before sharing analytical work with a client, causing much of the problem. Budget rules create considerable uncertainty for task teams: allocations for a product in a given fiscal year must be disbursed that year, and there is no guarantee additional resources will be made available in subsequent years. A very large share of ESW products—over 40 percent—is delivered in June, the final month of the fiscal year. Among those products delivered within 30 days of the decision meeting, more than 60 percent over FY00-FY11, and more than 70 percent in FY11, were delivered in June. In focus groups and interviews, staff indicated that review meetings for products facing a fiscal year deadline are perfunctory—in practice everyone in the room is aware of the fiscal year crunch, and no one has an incentive to slow down the process unless deficiencies are extreme. Peer reviewers have also been known to self-censor their comments so that they might be addressed within the time available. The Role of Network Anchors and Sector Boards on Quality Although the matrix system had envisaged that sector units would perform the primary role of ensuring technical excellence and quality services, the network Figure 4.3 The Quality Assurance System Has Little Time to Influence Many Studies 45% Share of ESWs with a decision meeting 40% 35% 30% Less than 30 days between dicision meeting and delivery to client, 3-year 25% centered average 20% 15% 10% Less than 10 days between dicision 5% meeting and delivery to client, 3-year centered average 0% FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 FY11 Source: World Bank databases. Delivering Quality Services 77 anchors and Sector Boards have played a minimal role in quality assurance. Anchor units review corporate documents and sometimes contribute to high- risk projects, but for the most part quality assurance has been left to regional sector units that may invite individuals from anchor units to be peer reviewers. Quality enhancement reviews (QERs)—meetings focused on quality held during preparation and chaired by the sector manager—were introduced 10 years ago and provide a structural role for anchor units but they have become optional, are often not timely and are not being systematically used, except in HDN and in two of the six Regions (East Asia and the Pacific and Africa). The majority of staff report that anchor units have little influence over the quality of regional operations. Thirteen percent of the 342 respondents to the staff survey (including anchor staff) reported that anchor units substantially enhanced the quality of their operations, and 22 percent reported that anchor units provided direct operation support to Regions to a substantial or large extent. The Role of Regional Operational Services Units The regional OS units are the regional nodes of the OPCS network and have dual reporting to the regional vice presidents and OPCS. The regional OS units typically have four functions, with a fifth added in some Regions: (a) quality review of all lending operations in the Region during project preparation, with intermittent review during supervision; (b) quality control: clearances and signoff for fiduciary and safeguards issues in all operations; (c) direct operational services on financial management and procurement to all task teams; and (d) coordinating knowledge and learning services for staff in the Regions. In most of the Regions they are also responsible for (e) coordinating trust funds. The existence of responsibility both for fiduciary and safeguards quality control, and provision of operational services for fiduciary areas, and quality reviews of all operations within the same organizational unit creates multiple conflicts of interest. Both of these functions—quality control and operational support for fiduciary areas (financial management and procurement)—currently reside within the regional OS departments and contribute to fiduciary and safeguards receiving excessive attention compared to other aspects of quality. The separation of quality control for safeguards from operational services on environmental and social issues is evidence that management is aware of the risk of conflict of interest. The housing of financial management and procurement in the OS departments also results in their separation from the upstream dimensions of public expenditure management that is handled by the public sector governance unit. The additional responsibility for quality review—in addition to operational support and quality control—creates yet another conflict of interest. The incentives of OS reviewers are heavily tilted toward fiduciary and safeguards issues, which are managed by their own unit. It is telling that staff and managers complain 78 The Matrix System at Work about the excessive focus on fiduciary and safeguard matters and relative inattention to other aspects of quality in regional review processes. Quality Assurance Systems Need to be Strengthened In line with the recently observed decline in development outcomes, most managers believe the quality of the portfolio has not improved since the introduction of the matrix system. Half of the sector managers and country directors/managers interviewed for this evaluation said that quality of lending and AAA is not significantly better than during the pre-matrix period, possibly reflecting recent pressures on quality. Among the remaining managers interviewed, more think quality has worsened, particularly for AAA, rather than improved. This may be a reflection of the growing volume of NLTA, which has had weaker quality review procedures than ESW. The majority of staff whose Bank careers began prior to the matrix system also do not report a noticeable improvement in the quality of supervision and AAA, while 39 percent feel it has worsened. While these perceived changes may not be due to the matrix reforms, at the very least it indicates that quality assurance systems are under strain and have not prevented quality deterioration, particularly in recent years. Available tools for quality enhancement are not being used effectively by sector units. The two main tools available to sector units are independent peer reviews and QERs. Sector managers value QERs much more than peer reviews for improving quality. Peer reviews in regional operations are not well managed. For the most part, task team leaders select their own peer reviewers, sometimes even from their own unit. In contrast, peer reviewers in DEC and IEG are selected by managers to ensure high quality and global expertise. In regional operations, due to delivery deadlines, peer reviewers are aware that substantive criticisms at the decision meeting or at meetings of the Operations Committee—where operations receive internal approval to move forward with appraisal and to request Board approval—are neither welcomed nor likely to be addressed. QERs have the potential to be more effective because they are considered a nonthreatening venue to obtain high-quality technical advice. But they are now considered optional, they are often undertaken too late, and their use is uneven across Regions and sectors. More systematic and timely use of QERs would significantly enhance quality of lending operations. The potential of the matrix is not being fully exploited to enhance operational quality. In most Regions, the concept review meetings—where teams are authorized to prepare an operation—and decision meetings for investment projects and ESW are chaired by the country director. Development Policy Loan (DPL) review meetings are held at the Regional Operations Committee or Bank-wide Operations Committee. Only in East Asia and the Pacific have both sides of the regional matrix been involved: the country director chaired the concept review meeting to ensure strategic relevance and client demand for the lending or nonlending product, and the sector director chaired the decision meeting to ensure quality. This good practice could easily be replicated in other Regions but ironically, to harmonize procedures across Regions, it has been done away with, as of July 2011. Delivering Quality Services 79 Staff Incentives The System of Formal Incentives The primary set of formal staff incentives at the Bank—performance management, rewards and recognition, and career progression—are not aligned with the Bank’s primary objective, to achieve development results in client countries. While these formal incentives are not the only means by which staff are motivated, they do have some influence and serve as an important benchmark regarding which behaviors the institution values. This evaluation does not directly assess the HR system of the Bank—that is a task for future evaluations—but rather identifies areas where incentive systems set by HR policies and practices affect the functioning of the matrix system. The Bank’s performance management system is not tailored to the unique circumstances of the Bank’s work, but rather it is based entirely on individual results agreements between staff members and their managers. In the absence of any specified guidance on linking individual efforts to higher-order organizational objectives—such as sharing knowledge across organizational units, collaborating across sectors, focusing on development outcomes of lending operations, and focusing on dissemination, use, and impact of AAA— managers are free to orient performance assessments around the narrower goals of individual units. Moreover, assessments do not hold staff accountable for longer-term results of work from previous years. By contrast, indicators for assessment of performance in IFC are based on objective criteria and are linked to IFC’s core objectives. Rewards are based on individual and team-based performance, project delivery, longer-term project and portfolio outcomes, and knowledge sharing. IFC also has a system of more rapid promotions enabled by a higher number of grade levels for professionals and staff than at the World Bank (see Box 5.3 and Appendix L). Opportunities for promotion are infrequent, and rewards are inadequate. The grouping of professional grades into large bands means that some of the best performing staff may only be promoted once or twice over the course of their career. Other forms of recognition, such as annual bonuses for strong contributions and awards for specific good work, do not fill this “acknowledgment gap.� Annual salary increases are linked to performance, but the difference in salary adjustment between a typical performer and the best performer amounts to as little as two or three percentage points, minimizing its efficacy as an incentive. Team awards for specific pieces of good work are distributed on criteria aligned with strategic objectives—a focus on development results, collaboration, responsiveness, and innovation—but these awards are rare and inadequate. Behaviors of Staff in Response to Formal and Informal Incentives Staff receive substantial encouragement to meet country-specific needs and lending targets from their managers, but less encouragement to work across sectors or use expertise across organizational boundaries. A primary informal incentive is direct encouragement from managers. Over three-fifths 80 The Matrix System at Work of anchor and sector unit staff received substantial or greater encouragement from managers to give priority to meeting lending targets, and a majority received such encouragement to adapt knowledge to country needs, as shown in Table 4.1. By contrast, they received substantially less encouragement to adopt “matrix behaviors,� such as collaborating across sectors and mobilizing Bank-wide expertise. Country unit staff believe sector unit staff have even less incentive to engage in these matrix behaviors. As a senior sector unit staff member in a Region summarized, “the matrix environment has so many mixed messages and incentives, that the only thing that everyone can agree on is the importance of the ‘deliveries’ of lending products.� Traditional organizational incentives—messages from senior management, desire for promotion and salary increase, and control over budgets—were in general the strongest incentives for sector- and anchor-based staff to meet lending targets and the weakest incentives for matrix-related behaviors. Messages from senior management and desire for promotion or a salary increase were weak incentives for collaborating across sectors (25 percent and 28 percent assessing this as having substantial influence). Mobilizing Bank-wide technical expertise (19 percent, 20 percent) and adapting knowledge to country needs (19 percent, 20 percent) were also weak incentives. Factors unrelated to the Bank’s management and performance systems—job satisfaction and recognition by peers—were far more influential in this regard (Table 4.2). One Matrix Behavior: Cross-Sector and Cross-Thematic Work Given the time constraints faced by staff and limited incentives to collaborate across sectors relative to other goals, to what extent does this “matrix behavior� occur in practice? Recent IEG evaluations of health, nutrition, and population (2010), social safety nets (2011), agriculture and agribusiness (2011), and climate change (2009) flagged weak cross-sector collaboration as both systemic and as an important barrier to good development outcomes. What do the data show? Table 4.1 Percentage of Anchor and Sector Staff Indicating They Received Substantial Encouragement to Carry Out the Following Behaviors Prioritize Use M&E to Collaborate Mobilize Bank- Adapt knowledge lending Focus on improve Encouragement to: across sectors wide expertise to country needs targets results quality From sector 40 41 54 61 53 36 management From country 35 25 52 67 41 27 management Country staff view of 16 18 25 48 32 14 sector staff incentives Source: IEG Matrix Evaluation Staff Survey. Delivering Quality Services 81 An expectation of cross-sector and cross-thematic work is typical, but declining, in the design of operations. On average, 45 percent of each operation approved by the Bank between FY1995 and FY2010 was flagged as devoted to a sector not affiliated with the Sector Board that governed the operation.5 The share for investment lending was lower—at 41 percent—and relatively flat until FY07, after which it began a steep decline to 33 percent for the three-year average over FY08-10, as shown in Figure 4.4. For non-affiliated themes, the share was 56 percent over the period for all operations and 55 percent for investment loans. The share devoted to non-affiliated themes among investment loans rose modestly to 61 percent in FY01, after which it began a steady decline, dropping to 46 percent for the three-year average over FY08-10. Appendix J provides more details. DPLs are designed as more cross-sectoral and cross-thematic than investment loans. The average proportion of development policy operations devoted to non-affiliated sectors and themes rose steadily over the period, exceeding 70 percent for both for the three-year average over FY08-10. But even for DPLs, the policy focus inherent in the operation makes it easier to indicate linkages to multiple sectors without actual collaboration across sector units, as demonstrated in IEG’s evaluation of Poverty Reduction Support Credits (IEG 2010). But the Bank has not been able to mobilize the relevant expertise to deliver these cross-sector and cross-thematic operations: actual collaboration across sectors by staff working in different units is far below expectations. One of the objectives of matrix reform was to break the silos of sector staff trapped in divisions within separate country departments. The creation of Table 4.2 Percentage of Respondents Indicating That Incentives Influence Their Behavior to a Substantial Extent Col- Adapt Give prior- Use laborate Mobilize knowledge ity to meet- Focus M&E to across Bank-wide to country ing lending on improve sectors expertise needs targets results quality Messages from senior management 25 19 19 47 34 22 Desire for performance review, 28 20 20 44 39 24 Incentives promotion, salary increase Control over task budgets 45 40 29 47 37 27 Access to earmarked trust funds 34 33 33 28 26 20 Cross sector career advancement 32 21 19 32 31 22 Recognition by peers 42 43 46 38 54 37 Job satisfaction 71 60 62 45 70 51 Source: IEG Matrix Evaluation Staff Survey. 82 The Matrix System at Work Bank-wide Sector Boards and regional SMUs helped to break those silos but re- created them at the regional and network levels. In FY10, total cross support— work by a staff member of one organizational unit for another unit—was 19 percent, of which about 14 percent was within VPUs6 (mostly across sectors) and 5 percent across VPUs (inter-VPU). Collaboration across sectors within the Regions is much better than elsewhere. But even the combined inter- and intra- regional cross support is far below what could be expected from the cross-sector and cross-thematic design of operations, as encapsulated by the assignment of 50 percent and 64 percent of the portfolio to non-dedicated sectors and themes, respectively. If task teams themselves recognize the cross-sectoral nature of their projects, what disincentives come in the way of actual collaboration across sector units? The Bank’s budget processes have not been effective in promoting cross- sector collaboration and teamwork under the matrix system. The inability to share tasks and budgets across organizational units discourages this type of cross-sector collaboration. Creating cross-sectoral teams for multisectoral tasks has proved particularly difficult. Interviews with staff and sector managers and feedback from the focus groups suggest that the inability to share responsibility and budget for tasks across organizational units is a major hindrance, a finding similar to that in IEG’s social safety net evaluation (IEG 2011). Even when organizational units seek to collaborate with each other, budget rules do not permit co-management of a task or subdivision of budgets. One unit must be declared the managing unit with a single task team leader accountable for the task, while staff members from other units have to charge their time on a transactional basis. Sector managers try to retain their most valued staff, which tend to be in greatest demand, to work on activities for which the manager will be held accountable. One sector manager captured the Figure 4.4 Average Share of Investment Loans Allocated to Non-affiliated Sectors and Themes 70% Avg. allocation of investment lending to non-af�liated themes (3-year centered average) 60% 50% 40% Avg. allocation of operations to non-af�liated sectors 30% (3-year centered average) 20% 95 96 97 98 99 00 01 02 03 04 05 06 07 08 09 10 FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY FY Source: World Bank databases. Delivering Quality Services 83 perspective of many: “Business brought in by a sector stays in the sector… There isn’t even a point in the system when the possibility of cross-sectoral or cross-regional teamwork can be considered.� Conclusions on Efficiency and Effectiveness of Quality Assurance Systems In this section, the evaluation draws conclusions regarding achievement of the intermediate outcome. The efficiency and effectiveness of the matrix system in achieving this outcome has been suboptimal, and the matrix objectives of improved quality, collaboration, and teamwork have not been achieved. Systems for accountability and quality control are more developed for lending operations than for knowledge products and are weak for NLTA, the fastest growing segment among nonlending services. Despite the intent to move away from a lending bank, the widespread recognition of lending approvals as the primary indicator of success by managers and staff alike is incontestable. This is reinforced both through informal incentives and in recognition and career advancement rewards. Feedback from staff and managers alike indicates that the Bank’s quality control systems focus more on fiduciary and safeguard risks with relatively less attention to other aspects of quality. The selection of peer reviewers is not well managed and end-of-year pressures undermine the potential value of peer review feedback, as there is hardly enough time to address substantial comments.. Quality enhancement reviews are a more effective mechanism to address quality, but they are optional and are often not timely. The checks and balances created by a mechanism, independent of line management, to take periodic stock of the robustness of the Bank’s quality assurance system, disappeared with the dissolution of QAG in 2010. An alternate mechanism has not yet been put in place. Quality assurance systems are better for lending than ESW and weakest for NLTA. NLTA therefore has widely varying quality. QAG reviews of AAA are the main source of evidence on AAA quality used in the Knowledge Report. With QAG disbanded, the Bank has lost its sole mechanism for autonomous review of portfolio quality. Staff lack incentives to focus on quality since outcomes of operations and knowledge work are de-linked from the Bank’s current performance management system. Financing and budgetary incentives also militate against high-quality client services and results. The regional sector units, responsible for operational delivery, are receiving a decreasing share of regional budgets. The squeeze on SMU budgets adversely effects quality, especially in the context of escalating lending targets and the growth of more demanding, middle-income countries in recent years. It creates a perverse incentive to mobilize trust funds but creates inefficiencies through an unwanted burden on sector technical staff both for fund-raising and for multiple, sometimes duplicative reporting for 84 The Matrix System at Work different trust funds. A larger share of trust funds finance Bank instruments with weaker quality assurance systems—particularly NLTA—compounding the quality challenge for knowledge and technical services. Finally, the Bank has not been effective in promoting the kind of matrix behavior—cross-sector collaboration and teamwork—that were expected to improve quality. Financial and budgetary incentives play a significant role in discouraging cross support and cross-sector collaboration. The inability to share budgets and task team leadership across organizational units discourages collaboration across sector units. Budget rules create a cross support penalty and budget settlement rules vary across the Regions, creating further uncertainty. Delayed budget settlement for cross support and differential norms for staff costs compared to consultant costs also discourage cross support. Notes 1 This calculation differs from that in IEG’s Annual Report 2011: Results and Performance of the World Bank Group because it includes the latest evaluations for operations exiting the portfolio in FY10. Inclusion of these operations brings down the share that was moderately satisfactory or better from 76 percent (as reported in the earlier evaluation) to 74 percent for the FY08-10 period, and makes the difference from the FY05-07 period statistically significant. 2 So far, 153 operations have been evaluated. 3 See the Assessment of the Strategic Compact (World Bank 2001c). 4 Findings from sector manager interviews. See Appendix E. 5 Sectors and themes were identified as affiliated with a Sector Board using classifications prepared by OPCS. IEG made one adjustment, considering those sectors and themes identified as affiliated with the Legal Sector Board (to which no projects are mapped) as affiliated with the PSG Sector Board. 6 Although cross support data are incomplete, in FY10 total cross support, including inter-VPU and intra-VPU cross support could be computed for all the networks and four of the Regions. For each of these, that is the networks and the four Regions, total cross support came to 19 percent. References Independent Evaluation Group (IEG). 2010. Poverty Reduction Support Credits: An Evaluation of World Bank Support. Washington, DC: World Bank. World Bank. 2001. “Assessment of the Strategic Compact.� Matrix Management Strategy and Resource Management Vice Presidency, March 13, 2001. Washington, DC: World Bank. Delivering Quality Services 85 Chapter 5 Institutional Issues and Organizational Structure Chapter Highlights • Incentives of regional sector managers toward regional priorities severely hamper the effectiveness of Sector Boards and weaken their essential corporate functions. • There is overlap and fragmentation among some Sector Boards but no mechanism for periodic review of Sector Board and network configuration to keep pace with the Bank’s technical practices. undermines the effectiveness of the • The large span of control of regional managers ­ ­matrix system. merging two previous ­ • The creation of SDN by ­ exacerbated the networks has considerably ­ organizational challenges for the Bank under the matrix system. • Financial incentives and budgetary rules constrain the operation of the matrix system. • Over the FY02-10 period, sector units have received a declining share of the Bank’s bud- get. • The Bank spends 47 percent of its budget on “client services� when trust funds and resource mobilization are excluded. • Expenditure on country ser­ v ices is 32 percent of the Bank’s budget, or one in three dol- lars, for direct costs of lending and non-lending products delivered to country clients. • The Bank currently lacks a mechanism for transparent assessment of operational qual- ity independent of line management. The overall conceptual rationale of the matrix system remains valid and even more urgent than ever, given the changes in the external environment currently facing the Bank. Client needs have diversified and the growth of regional and corporate priorities and GPG is creating tension with the country model. The Bank’s ability to renew itself and function as a truly global Bank is critical to its success. However, several organizational and institutional impediments need to be addressed for the objectives of the matrix system to be realized. The 1997 reforms (of which the matrix system was a major part) have led to greater client responsiveness, but the sector side of the matrix did not succeed in creating global technical practices with world-class knowledge systems. This perception needs to be qualified by the fact that the very strength of the country units and, by extension, the Regions, created an imbalance and undermined the centrality of the sector model in the Bank-wide matrix. At the regional level, the imbalance in the relationship between country and sector units created suboptimal tensions giving primacy to short-term client demands over strategic country needs. Rather than functioning as a global institution, the Bank is at risk of evolving into six regional banks. Regional variations in structure and processes compound the problems arising from insufficient inter-regional flows of knowledge and expertise. Sector configuration has varied across Regions and 88 The Matrix System at Work over time in terms of the sector groupings within SMUs or sector departments (for example, the bundling of FPD with PREM in some Regions but not others; environment with water in one, urban with water in others, and so on) and the cost centers, some recording budgets at the SMU level, others at the sector department level. The role of the country sector coordinators—who coordinate network activities in each CMU—also varies across Regions and networks. And norms and procedures for quality assurance differ significantly. In East Asia and the Pacific, until July 2011, country directors chaired concept review meetings and sector directors chaired the decision meeting, while in the other Regions, sector directors have had no formal role in the quality assurance process. Regions also vary in the use of QERs, some (such as South Asia and Africa) relying on them more systematically for quality, others infrequently. Some of the inputs and rules put in place to manage the matrix system need revamping to enhance the relevance and effectiveness of the matrix. The current distortions cannot be resolved by tinkering at the margins if the objectives of the matrix system are to be achieved. Major changes are needed, explicitly targeting the key problems among existing structures that are barely coping with operational demands. Addressing these organizational constraints may require “lateral thinking� (De Bono 1967)1 to find creative solutions lying beyond the range of current assumptions and practices. In this chapter we focus on a few of the reform inputs—listed in the horizontal row at the top of the evaluation framework (see chapter 1)—where organizational effectiveness issues that merit further examination have emerged from the evaluation’s findings. The chapter begins with a summary of the key organizational and institutional impediments—the Sector Board model; and span of control issues and the effects of network restructuring that affect the regional and Bank-wide matrix systems; financial and budgetary issues; and quality assurance arrangements. These cover four of the six boxes in the row of reform inputs in the evaluation framework. Decentralization and HR processes require much more in-depth analysis and have been left out of this evaluation The Sector Board Model Effectiveness of Sector Boards Sector Boards, which were one of the main organizational innovations under the matrix system, have not been able to deliver on most core functions effectively. Sector Boards were created as a governing body to manage a professional association of all technical staff working on similar technical products. In addition to governance of the professional association, Sector Boards were responsible for strategy formulation, knowledge management, quality support to Regions, HR management, and external partnerships. The design of Sector Boards appears to have been flawed due to the inadequate division of labor between the sector managers  and sector directors in Sector Board governance and quality. Accountability for quality is high for sector managers, Institutional Issues and Organizational Structure     89 but far less for sector directors. Regional sector managers are overburdened with a wide range of regional responsibilities and have little incentive, time, or resources to perform the multifarious corporate roles expected of them. On the other side, the sector anchors have taken on additional corporate and global tasks that diverted attention from their ability to support frontline operations in the Regions. The documentary and survey evidence, as well as the management self- evaluation, point to the following major weaknesses (some of which were discussed in greater depth in earlier chapters), which are also reiterated by the majority of the regional sector managers: • Strategy formulation has been costly and has added insufficient value to country strategies and country programs. As discussed in chapter two, SSPs have had little direct impact on country strategies and programs. Although SSPs increasingly reflect country and regional experiences, the relationship is one-sided, and they have little direct influence on the CAS despite the high costs incurred to prepare them. • Sector Boards have been ineffective in ensuring knowledge flows across the Bank and operational support to the Regions. Sector managers are very critical of the knowledge function of anchors and Sector Boards, and of the operational support received from them. None of the country directors and managers interviewed received substantial knowledge or operational support for their respective portfolios from the anchor units.2 Both the quantity and quality of operational support—the principal means of transferring tacit knowledge—were inadequate. • Quality is one of the five core functions of Sector Boards, but it takes a backseat. As a corporate entity, the Sector Boards have little influence over the quality of regional operations. Quality control in regional operations is exercised by regional sector managers; however, they are overwhelmed by span of control challenges. As pointed out in the previous chapter, in regional quality assurance processes, attention to other aspects of quality ends up taking a backseat to risk management concerns. • HR management has been poor and is heavily criticized by staff and regional sector managers. In the World Bank’s Staff Survey (2009) HR policies and procedures received the lowest rating (33 percent favorable) among all policies and procedures on which the survey solicited feedback. Regional sector managers, too, find Sector Boards to be ineffective (20 percent favorable) in balancing hiring priorities among regional and Bank-wide needs, and unable to deploy staff across the Bank to serve regional needs.3 • However, the role of anchor units in managing partnerships and corporate priorities is valued. Sector anchors are viewed by staff and regional managers as somewhat effective on corporate priorities (such as climate change) and in providing technical support on highly specialized topics (such as agribusiness, results-based financing, and national health accounts). SDN and HDN have been very successful in mobilizing trust funds. However, this is a double-edged sword as trust fund management diverts attention from the core functions of Sector Boards. Indeed, the Sector Board found 90 The Matrix System at Work to be the most effective—Social Protection—attributes some of its success to the fact that it has not been distracted by trust funds and has been able to link its work more directly to regional operational priorities. Effect of Regional Incentives on the Sector Boards The incentives of regional sector managers toward regional priorities severely hamper the effectiveness of Sector Boards. The regional sector managers receive their entire budget from the regional CMUs and are accountable for delivery of regional WPAs, in the context of a growing portfolio and shrinking share of regional budgets. Anchor units do not provide any resources to regional sector managers or to regional sector directors. It should be no surprise that their primary loyalties and accountabilities are with their Region, and that they serve as advocates and defenders of the Region’s interests rather than as members of a collective Bank team. Within the Regions, sector managers are responsible for mentoring and managing technical staff, overseeing quality in project preparation and supervision, and ensuring quality of knowledge products and services provided by their unit, with little time or incentive to focus on responsibilities for the Bank-wide Sector Board. In practice, the functioning of the board defaults to the anchor unit in most cases, which was originally conceived simply as the secretariat to the Sector Board. Most regional sector managers are highly critical of anchor units. Despite their own inability to contribute to Sector Board governance meaningfully, sector managers are highly critical of anchor units. As one sector manager put it, “The anchor meets its own needs and corporate demands, but not the Regions’ needs ... knowledge generated by the Regions is uploaded and disseminated, but mainly through—and depending on—the thematic groups.� In 2011, the FPD vice presidency and the FPD departments in the Regions were restructured in an attempt to address the perception that “the network anchor and the six Regions were not working in full alignment.� Among other measures, FPD introduced six global practice groups, each focused on one of the six areas identified as strategic priorities in the sector. All staff members in the FPD network—whether in the anchor or the Regions—have been mapped to one or more global practice. And each of six dual directors now has both a regional and functional responsibility: they are accountable to the FPD network vice president for quality across all Regions in one of the six global practices, and they are accountable to a regional vice president for project implementation and risk mitigation across all global practices in one of the six Regions. Their pay is divided equally between the Region and the network VPUs, and they have the support of one manager in the Region, as well as one to three service line managers for the practice group. This design was intended to address matrix-related issues on strategy, knowledge, and quality and accountability. Dual accountability of sector directors was expected to make them “think corporate� rather than just regionally or topically, and thus allow them to both better integrate corporate Institutional Issues and Organizational Structure     91 strategic priorities into operational work and build on-the-ground realities into strategic work. New relationships were expected to increase the sharing of knowledge and expertise across Regions and between the Regions and network anchors by creating natural communities of practice among staff centered on the strategic priorities of the institution. And the new system was expected to increase attention to quality by making sector directors accountable for quality on a global practice Bank-wide. Box 5.1 outlines a prospective assessment of the FPD model based on the matrix issues identified in this evaluation. Recent efforts by management to rejuvenate the Sector Boards do not address the underlying constraints that impede them. The Matrix Leadership Team has agreed on a 10-point program4 to enhance sector practices which states that henceforth sector managers will also report to both sides of the matrix. The program consists of 10 steps that Sector Boards will take to brand their sectors as “Enhanced Sector Practices.� But the underlying incentives and structural constraints, including span of control of sector managers that have prevented efficient functioning of the Sector Boards have not yet been addressed. Sector Board Architecture The current configuration of Sector Boards has not evolved to match the Bank’s operational priorities. On the one hand, there is overlap and fragmentation, sometimes leading to competition, among some Sector Boards. On the other hand, new operational areas, such as disaster management, climate change, and governance and anticorruption have emerged as global technical practices that do not have Sector Boards. Given the evolving nature of country programs, the rationale for continuing the same configuration of Sector Boards originally established in 1997 is hard to justify. The assessments by QAG5 and the 2004 Organizational Effectiveness Task Force identified several instances of overlap or fragmentation in the structure of Sector Boards, some of which fell across network boundaries. The review of country and sector programs reveals that the situation has not changed. Examples include Financial Management and Procurement with Public Sector Governance, Poverty with Social Protection, Social Development with Gender and Public Sector Governance, and Water with Urban Development. Overlap and fragmentation of sector agendas across Sector Boards was reported by a majority of sector and anchor staff and brought up frequently by sector managers. As operational priorities evolve, it becomes more feasible and, indeed, necessary for organizational structures to be fine-tuned to keep pace with those changes. No mechanism exists for periodic review and adjustment of Sector Board mandates and network configuration to keep pace with the Bank’s technical practices. Such adjustment has been undertaken periodically among country units as country programs have evolved. However, the Sector Boards have, until now, been immutable as sector boundaries have been assumed to coincide with professional families based on technical competencies. On 92 The Matrix System at Work Box 5.1 The New FPD Model: Creating “Global Technical Practices� FPD set ambitious goals for its restructuring, including more cross support, improved outcomes, and improved client satisfaction. Still in its infancy, the restructuring cannot be evaluated for its outcomes, but there are some hints given the degree to which the new system introduces new incentives and removes disincentives for achieving its objectives. Strategy: The introduction of dual directors responsible for operations in one Region and one Global Technical Practice (GTP) will likely help to integrate the Bank’s strategic sector priorities, where appropriate, into country operations. At a minimum, the directors now have an incentive to integrate the GTP for which they are responsible into the operations of the Region for which they are responsible and, to a lesser degree, to reciprocate on other practices. Knowledge: The mere setting of a goal for cross support and mapping of staff to more specific global practices is unlikely to dramatically increase the use of cross support unless cross support penalties are removed. FPD is addressing this by switching to pay-as-you-go cross support which addresses the disincentive caused by delays in budget adjustment. Sharing of knowledge across VPUs could benefit from the orientation of teams around global practices through informal networking. Experience in other sectors suggests that for a thematic group to be useful, it must (i) be financed well-enough to provide a minimum threshold of knowledge management functions; and (ii) focus on a topic specific enough to attract a group of motivated staff to yield a strong network. Accountability for quality: The directors’ new accountability for quality for one GTP Bank-wide may heighten their attention to quality. This will incentivize directors to insist on more robust quality assurance systems for all operations in their GTP. Sustainability: In several ways, the FPD pilot is quite similar to the original vision of the matrix system, with staff assigned to Bank-wide Sector Boards and expected to work across Regions as needed. These teams would ensure that current knowledge is integrated into country operations. The FPD pilot holds a similar vision with a few adjustments: GTPs are more specific than Sector Boards; directors serve in a dual role; and some incentives have been modified. While these changes have the potential to enhance the effectiveness of the FPD Network, it remains to be seen if the initial excitement over the new structure will be internalized by FPD staff and can be sustained. Replicability: The new structure for FPD is designed as a pilot exercise. Its applicability to other networks may be constrained by differences in the relative share of anchor staff and in overall staff size. Almost half of FPD staff are in the Network Anchor—giving it more flexibility for sharing knowledge—compared to less than a quarter in PREM and HDN and a third in SDN. PREM and HDN are roughly 60 and 80 percent bigger than FPD. SDN staff is more than the combined strength of the other three and the breadth of its technical sectors is too vast for the FPD model to be feasible under its current structure. Source: “Frequently Asked Questions, FPD Global Practices Pilot,� interviews with FPD staff. Institutional Issues and Organizational Structure     93 the other hand, emerging multisector technical practices—such as disaster management or climate change—have drawn on staffs from several sectors without a formal governance structure to support them, suggesting that the Sector Board design is not essential for creating new technical practices, as long as fragmentation and competition can be overcome. Currently sector managers, and not sector directors, are responsible for overseeing all of the functions of the Sector Boards. Under the current structure, each Sector Board serves as the governing body for that sector and is responsible for strategy, knowledge, quality, partnerships, and staffing. As noted earlier, the regional incentives and work pressures on sector managers prevent them from doing justice to their corporate responsibilities related to Sector Boards. At the same time, regional sector directors neither serve on the boards nor are accountable for any of these five functions.6 The original design of the matrix had envisaged making the sector units responsible for the quality of global technical practices without clarifying the division of responsibilities among sector directors and sector managers. While all five roles may be essential for sector coherence, it does not follow that all the accountabilities should lie with the Sector Board and sector managers and none with the sector directors. However, given the range of sectors overseen by current sector directors, especially in SDN, any readjustment would have to be accompanied by an examination of the configuration of networks and network size. Span of Control in the Regions As does every large organization, the Bank must meet the three major criteria of an efficient structure: non-overlap, non-fragmentation, and a manageable span of control (Oakerson 1989).7 Non-overlap implies that two or more units should not have the same authority to act in the same circumstances; non-fragmentation allows aligning authority with responsibility; and span of control calls for manageable organizational size. There are inevitable tradeoffs among the three criteria. In particular, small unit sizes tends to lead to inter-unit coordination problems and risks of incoherent policy, and large unit sizes entail intra-unit coordination problems and risks of inefficient and unaccountable implementation. The challenge of finding the right balance among the three criteria is particularly complex in a matrix organization because such balance must be found on both the “technical� and the “client� sides of the matrix. Span of Control of Sector Managers The unrealistic span of control and multitude of responsibilities of sector managers in the Regions makes it unlikely that they can either devote sufficient attention to quality assurance and people management or contribute effectively to Sector Board functions. A 2009 study by Bank management reported that the span of control8 in the Bank was 23 staff per manager, which was already much higher than comparator organizations, 94 The Matrix System at Work which were in the range of 7-15 staff per manager. The study found that sector managers had the most acute span of control problem with an average of 34 staff per manager in 2009, which also affects their ability to contribute effectively to Sector Board functions. This evaluation found that at the end of FY11, the average staff per regional sector manager was 37,9 more than three times the average for comparator organizations, and compared to 19 at IFC (Appendix L). Although the number of staff to be managed is only one aspect of the workload, this alone renders it highly unlikely that regional sector managers can participate meaningfully in their Bank-wide role on Sector Boards. As one former sector manager said, “The overload of sector managers and many staff is ridiculous. The anchor can only facilitate participation in network activities if the parties are free and interested in contributing. Time, more than incentives, is the main constraint.� The current model of one sector manager per sector or group of sectors in each region is unsustainable. Span of control issues are most acute in Africa and SDN. Among the Regions, the span of control challenge is most acute in Africa at 54 staff per sector manager, compared to the Middle East and North Africa, which has the smallest span at 23 staff per manager, even this being much higher than any of the comparator organizations. Among the networks, SDN has an average of 42 while the remaining networks range from 30–35 staff per regional sector manager; including the network anchor, SDN has 47 staff per sector manager. These network averages mask significant variations with 15 sector units having more than 50 technical staff as direct reports, the highest being 90. To cope with this, many units are functioning with shadow managers who have been delegated tasks but are not always paid10 or officially acknowledged as managers. Other units are relying on alternative models that have moved away from the notion of one sector manager per Region for each sector. In Africa, PREM has three geographic sector managers covering several sectors and one for public sector governance region-wide. In East Asia and the Pacific, SDN has decentralized sector managers to country offices, but they cover all the SDN sectors for individual country groupings. In both cases the regional structure meets specific regional and network needs but assumes minimal contribution by operational sector managers to their Sector Boards. The two headquarters- based sector managers who represent the Region on all eight of the Sector Boards in East Asia and the Pacific have no operational role. The reliance on shadow managers and country sector leaders or coordinators, particularly in Africa and SDN, puts additional stress on its ability to deliver quality services. Span of Control of Sector Directors Span of control is also an issue for sector directors, especially in SDN. SDN directors have the heaviest burden in terms of staff per sector director. The average number of regional staff under sector directors ranges from 41 in FPD to 258 for SDN, with PREM and HDN having slightly more than 100. Even adjusting for the fact that FPD has moved to a model of dual directors, each overseeing one Region and one Practice Group, the total number of staff per FPD director is 93, while the average for SDN, including anchor directors and staff, is 210. The excessive burden on SDN sector directors is a consequence of Institutional Issues and Organizational Structure     95 the merger of two vice presidencies with eight Sector Boards between them. The merger did not just increase the number of staff under each sector director, it also vastly increased the substantive scope of work to eight sectors. Since sector directors cannot have technical competence in every one of their sectors, and sector managers who have the specific technical competence are too overburdened, there is neither an adequate mechanism for technical oversight nor accountability for quality (also see chapter four). These findings suggest that at least for the sectors currently under SDN, the span of control of sector directors is also an issue. The Effects of the SDN Merger on the Matrix System Subsequent changes added stress to the matrix design. The rigidity in Sector Board structures proved to be an impediment to the flexibility needed to address issues arising out of emerging operational priorities and subsequent organizational restructuring adding to the stress arising from inadequate knowledge flows and accountability for quality. In 2006, the two smallest sectors—Financial Sector and Private Sector Development networks—were merged into the FPD network, and the two largest—ESSD and Infrastructure networks were merged into the SDN network. The FPD network recognized the constraints on knowledge flows and quality and was restructured at the end of FY 11 (Box 5.1). The two Sector Boards have been replaced by 6 Global Technical Practices designed to address these issues. While it is too soon to evaluate the effects of the reorganization, or to assess its replicability across networks, FPD’s efforts to recognize and confront these issues are to be commended. Bank-Wide Level The creation of SDN has considerably exacerbated the organizational challenges facing the Bank under the matrix system. At the Bank-wide level, the benefits of the merger are mixed. Infrastructure lending has increased, but accountability for quality has not. Regardless of the significant increase in infrastructure lending—which started after the Infrastructure Action Plan in 2003, even before the SDN merger—from the perspective of organizational effectiveness and quality assurance in country operations, the SDN merger has been suboptimal. The decision to merge the ESSD and INF networks was motivated in part to integrate environmental concerns within infrastructure sectors, which generate the highest environmental impacts. There is some evidence from the work on climate change that this has had positive effects although it is still too soon to evaluate outcomes on the ground. However, this rationale for merging environment does not extend to the agriculture sector. Nor was there clarity initially on the benefits of merging social development with SDN rather than with other networks, such as HDN or PREM, which have overlapping mandates.11 96 The Matrix System at Work The merger of the two networks was also induced in part by the possibility of cost savings, but there is little evidence that this has occurred and organizational complexity has increased. SDN attempted to reorganize itself after the merger was announced. In the network VPU, the three infrastructure departments were merged into two departments,12 but staff from former ESSD and urban sectors resisted a further merger. More substantial cost savings might have been realized if Sector Boards under the two previous networks were merged. However, even among the INF sectors, despite the reduction in number of departments, the number of Sector Boards remained unchanged. SDN has eight Sector Boards,13 compared to three in HDN and four in PREM. Within the Regions, some sector units were merged but this was done unevenly across Regions and span of control increased to 47 staff per manager in SDN (Table 5.1). SDN has adapted to budget constraints by becoming even more dependent on trust funds than other network anchors. SDN is the largest network, its staff numbers exceeding that of the FPD, HDN, and PREM networks combined. The matrix system directly affects 78 percent of the Bank’s staff. In FY 11, of the 11,402 total Bank staff, 6,235 were located in the six regions (Appendix M, Table M.1) and 1,630 in the anchors of four networks (FPD, HDN, PREM, SDN) plus OPCS14 (Table 5.2). The size differential between SDN and other networks remains true among network anchor staff as well as when the combined strength of anchor and regional staff of the networks are compared. The networks also differ in the share of staff located in the network anchors. Half of FPD’s technical staff is housed in the network anchor compared to a third for SDN, and less than a fourth for PREM, HDN, and OPCS (Appendix M, Table M.4). A larger share of anchor staff implies less likelihood of regional capture and greater potential for cross support and work across Regions, Table 5.1 Selected Indicators for Networks VPU FPD HDN PREM SDN Total regional and anchor staff (as of June 2011) 565 904 827 2308 Share of anchor staff/total staff 48% 23% 21% 30% Span of control (total staff/sector manager) 28 40 34 47 Scope of control (Sector Boards/Global Practices) 6* 3 4 8 Network Anchor Expenditures (FY10 BB + BETF actuals) (in US$ million) 68 64 47 208 Share of BETFs in Anchor Expenditures 39% 43% 32% 52% Source: World Bank databases. Note: Data exclude financial management and procurement sector boards which fall under the OPCS Network. *FPD reorganized itself into 6 global practices at the end of FY11. FPD staff data include only IBRD head counts. Institutional Issues and Organizational Structure     97 Table 5.2 Staff Head Counts of Network Anchors and Selected Central VPUs, FY11 VPU ETT ETC GA-GD GE+ Total FPD 3 16 45 248 312 HDN 4 36 35 143 218 PREM 3 30 27 123 183 SDN 12 74 118 553 757 OPCS 6 10 28 116 160 Total NVPUs 28 166 253 1183 1630 DEC 2 39 37 187 265 WBI 5 21 38 152 216 Source: World Bank databases (snapshot of June 2011). Note: Head counts of staff mapped to the units exclude staff on special assignments (SPAS). Includes IBRD FPD staff head counts only. ETT = extended term temporary; ETC = extended term consultant provided the incentives that currently inhibit cross  support and operational support services to the Regions can be addressed. At the Bank-wide level, the creation of SDN led to a separation of quality assurance for environmental and social safeguards from the units with technical expertise in environment and social development to OPCS. The former Quality Assurance and Compliance Unit (QACU) was moved from ESSD to OPCS to avoid any perception of conflict of interest that might have occurred if QACU were under the same vice president who oversaw three-quarters of the Bank’s investment lending portfolio. Regional Level The creation of SDN considerably aggravated the span of control issue for sector managers. In the Africa, South Asia, Middle East and North Africa, and East Asia and the Pacific Regions, the number of sector managers was reduced and multiple sectors combined within larger SMUs. The decentralized Bank also resulted in the proliferation of regional variations, making it impossible to rethink Sector Board structures—Africa combined Urban and Water sectors; South Asia initially combined Environment, Social Development, and Water sectors and subsequently hived off Social Development; the Middle East and North Africa combined Urban and Social Development in one SMU, and Environment and Agriculture and Rural Development in another; East Asia and the Pacific—the most decentralized Region—created the most radical structure described in paragraph below. Under SDN, some sector managers have been assigned responsibility for multiple sectors, including for sectors in which they lacked technical 98 The Matrix System at Work expertise. While in some instances this has been caused by having responsibility for more than one sector, there have also been cases where sector managers were appointed to oversee sectors that were beyond their previous sector expertise, substantially limiting their ability to play any role in quality assurance for lending and knowledge. Board approval for lending operations and mentoring staff in project management took precedence over accountability for quality. Integration of environmental and social safeguards into infrastructure sectors has improved, but this may be crowding out attention to other environmental and social issues. Some SDN managers (and some country directors) reported in interviews that the creation of SDN has led to better integration of environmental and social issues within infrastructure projects. On the other hand, there is considerable evidence from the Regions that the SDN merger has led to much greater demand for safeguards work, crowding out attention to other aspects of environmental and social sustainability. By bringing environmental and social safeguards specialists under the same roof as infrastructure staff, the Bank changed incentives and increased attention to “do-no-harm� issues. In Europe and Central Asia, 15 of the 18 country-based environmental specialists work on safeguards. In Africa, there has been attrition of senior (GG-GH level) social development staff, with GH-level replacements brought in dedicated exclusively to social safeguards. As discussed in IEG’s safeguards evaluation15 (2010), this has had the unintended consequence of a separation of the work on safeguards from the work on environmental and social sustainability. The dichotomy between safeguards and non-safeguard risks is a false one, arising out of flaws in the Bank’s safeguards framework. IEG’s evaluation of safeguards drew attention to the inadequate coverage of social risks in the Bank’s current suite of safeguard policies (IEG 2010: 92). While the SDN merger appears to have generated increased demand for safeguards work, social risks such as labor and community impacts, not covered by the safeguard policies, receive inadequate attention.16 Feedback from focus groups, staff surveys, and manager interviews reveals that environmental and social development staff perceive the SDN merger as a takeover by infrastructure, rather than a genuine integration of their agendas. As one focus group participant from an infrastructure sector pointed out, “they are starting to think like us.� Staff and managers alike are critical of the excessive attention to safeguards and fiduciary issues relative to the inadequate attention to other aspects of quality and AAA. The SDN merger exacerbated this tendency. From an organizational perspective, the creation of the SDN Network is an imperfect solution to the perception that network boundaries are insurmountable. If structural merger was the only solution, SDN would need to be merged with the PREM Network to ensure the integration of poverty objectives (which represent the Bank’s overarching mission) into infrastructure sectors . These are precisely the kind of barriers that the matrix system was designed to overcome but has as yet been unable to do so because of the lack of a supportive incentive framework. Institutional Issues and Organizational Structure     99 Country Level The creation of SDN has added another layer to an already complex management structure, with coordination of the portfolio, rather than quality, as the primary objective. Most Regions appointed SDN sector leaders or coordinators at the country level—adopting the Latin America and the Caribbean model—to help country directors cope with the large number of SDN staff in each country team. Latin America and the Caribbean Region has used a model of “sector leaders� for each sector department since 1997, but responsibilities of SDN sector leaders substantially increased after the merger. Latin America and the Caribbean’s sector leaders are de facto shadow managers, co-located with country directors and providing a key link between the country departments and sector departments. They are responsible for work program coordination, quality assurance, portfolio monitoring, and performance feedback on country team members to sector managers. After the creation of SDN, the sector leader model was retained in Latin America and the Caribbean, but the sector leaders for ESSD and INF were merged at the country level, increasing their span of control and reducing their ability to provide quality assurance across all eight SDN sectors evenly. East Asia and the Pacific created a matrix within a matrix within a matrix (Box 5.2). EAP combined the former ESSD sectors under one sector manager and all the infrastructure sectors under another sector manager in headquarters, while merging all eight sectors under one manager or sector leader in each CMU. In the three largest country offices—China, Indonesia, and Vietnam— SDN staff are managed by a field-based sector manager who oversees the SDN portfolio. In the three smaller CMUs, SDN staff are managed by a field-based “country sector leader� who acts as a de facto manager. These managers manage all country-based staff and the country portfolio for all eight SDN sectors, while the sector managers in headquarters represent the Region on three and five Sector Boards, respectively, with no associated operational role. East Asia and the Pacific thus has a third-level matrix within the Region. This structure came under withering criticism from the Region’s SDN staff in an FY11 survey17 because of its diffused accountability and inadequate quality assurance from the field-based managers. SDN staff in that Region felt that the broad range of sectors under their country-based managers has reduced them to a “postbox� coordinator, rather than an effective line manager. Europe and Central Asia, South Asia, and Africa have structures that lie between the Latin America and the Caribbean and the East Asia and the Pacific model, with sector leaders or coordinators designated for each CMU. The primary function of the sector leader or sector coordinator appears to be coordination of the work program and liaison with the country director. The range of sectors they are responsible for is too large for them to be able to provide meaningful quality assurance and they have not been delegated any 100 The Matrix System at Work Box 5.2 Views on Matrix Issues of East Asia and the Pacific Staff Based in Sustainable Development Network Units In an FY11 surveya EASIN (East Asia and the Pacific infrastructure) and EASER (East Asia and the Pacific social, environment, and rural development) staff expressed concerns on issues ranging from internal clearance processes and reporting procedures, the proliferation of management layers, decentralization of sector management, and the balance of power between headquarters- based and field-based managers, and extra efforts to sustain networks of decentralized staff. “The department’s structure with managers in the field has resulted in people without knowledge of particular sectors leading the dialogue in those sectors.� … “Having managers responsible for multiple sectors does not work.� “The [department’s] structure is completely dysfunctional. The managers in the field add no value and are mere bean counters/transaction processors. We need to go back to sector-based management with a fair mix of managers in the field and in DC.� “Do not have both [field and DC-based sector managers], it is redundant. Having Sector Leaders and Sector Coordinators simply adds to this complexity. The solution to challenges with multiple layers is not to add more layers!� “In [East Asia and the Pacific], the Country Offices have too much control over resources and decisions. It is difficult to engage unless one is based in the field office.� … “Better coordination between sector managers in the field and HQ. Budget has to be managed by the two sector managers, not only the ones in the field.� “Budget for full attendance of sector weeks and for travel from HQ to the field and vice versa is essential for such a convoluted structure to function. Skype, VCs etc. can maintain relationships, but they are not enough. There has to be face to face contact and relationships.� Source: EASIN and EASER FY11 Pulse Surveys. a. The 2011 Pulse Survey was undertaken by management as a follow-up to the World Bank Staff Survey, 2009. responsibility for staff management. Discussion of regional variations can be found in Appendix G. The SDN merger also has a distortionary effect at the country level due to the disproportionate size and resources of SDN relative to the other networks. As one country director put it, “SDN now basically owns my program, and I have to spend more time in protecting the smaller sectors than in anything else.� The overwhelming majority of staff interviewed (with the exception of senior SDN managers), both in SDN and in other networks, feel very strongly that the costs of SDN merger significantly outweigh its benefits. Budgetary Incentives as a Driver of Matrix Behaviors The Bank’s matrix-related budgeting process involves allocations to each regional and network VPU, which are then negotiated and allocated within each VPU according to strategic priorities and work program agreements. In the Regions, Institutional Issues and Organizational Structure     101 the regional vice presidents hold back some money for regional functions, including the chief economist’s office (responsible for regional studies), and operational services and quality control functions undertaken by the OS unit. The remaining budget is allocated to CMUs, who negotiate work programs with SMUs, reserving some funds for their own administrative expenses. Once work program agreements are finalized, the fiscal year budget is released to SMUs for product delivery to country clients, with limited oversight and monitoring by the CMUs and regional vice president. Bank-executed trust funds for individual products are mobilized directly by task team leaders rather than through the normal budget allocation process. IFC recently underwent an organizational change—termed IFC 2013. The matrix design and process adopted recently by IFC offers relevant lessons for the World Bank (see Appendix H for a full description of IFC 2013). The most significant variation from the Bank’s matrix system lies in the manner in which budgets and deliverables are negotiated among the three entities responsible for regional product delivery rather than providing complete budget authority to the country units. A second major difference is the refinement of the performance management system that links recognition and rewards in a transparent manner to contribute to corporate goals and to the longer-term outcomes of operations in which each staff member has been involved. IFC has also introduced metrics to monitor implementation effectiveness to identify problems early and allow further refinement of their matrix system. The effectiveness of IFC 2013 (Box 5.3) will depend on IFC’s ability to make necessary adjustments in response to the feedback from the organizational monitoring metrics. In the interim, one can learn from the significant challenges the IFC 2013 change initiative has faced in its short lifespan and how it is addressing them. Budgetary arrangements under the matrix system were singled out by staff as well as country and sector managers as a major impediment to the Bank’s efficiency. The imbalance in budgetary authority between country units and sector units leads to a short-term focus driven by lending delivery and fiscal year deadlines rather than strategic selectivity that balances short-term needs with business development to prepare for future risks or policy opportunities. The matrix was supposed to reduce the focus on lending in favor of a better balance with quality and knowledge, but it has not done so. Overall Resources for Frontline Delivery Greater transparency on overheads would increase accountability and encourage efficiencies. Out of concern for the decline in frontline resources prior to the introduction of the matrix, management had committed, as part of the Strategic Compact, to an increase in frontline expenditures to 60 percent of the budget (Phillips 2009: 38). The first indicator on organizational effectiveness and modernization in the corporate scorecard issued at the end of FY11 (World Bank for Results 2011) states that 61 percent of the World Bank’s resources (Bank budget plus BETFs) are spent on “client services�. However, this 102 The Matrix System at Work Box 5.3 The International Finance Corporation’s 2013 Change Initiative IFC introduced organizational changes in 2011 intended to strengthen development impact, financial sustainability, client satisfaction, talent management, and people development, and to use global knowledge in a decentralized environment. The new system adopts more specific development goals for driving financially sustainable growth and maximizing development impact where IFC has a comparative advantage; strengthens alignment of the performance management system to the corporate strategy and budget cycle; and restructures the investment departments to clarify roles and accountabilities. These changes were supported by a communication strategy, milestones, and a monitoring system. The revised organizational structure divides IFC into: • Global Industry Groups made up of Global Industry Departments responsible for managing global clients and partnerships, developing sector strategies, linking advisory and investment services, and developing global knowledge networks. • Regional Client Service Groups composed of Regional Departments and Regional Industry Departments. Regional Departments are responsible for managing country strategies, business development, and client relationship. Regional Industry Departments are responsible for portfolio and transaction management, providing functional specialization, and improving responsiveness, efficiency, and risk management. The new business process and incentives system include: • A revised budgeting process with vice presidents more involved in the planning phase to determine budget envelopes that are later allocated at the department level. Technical staff are mapped to global or regional industry units and the budget is negotiated and allocated directly to each of the three departments (Global Industry, Regional, and Regional Industry) based on an agreed investment program and staff requirements from each department. • Staff incentives aligned with strategic goals. Managers’ objectives are adjusted to focus on development and financial results, client orientation, people management and development, and contribution to corporate initiatives with achievement of both quantifiable and qualitative development objectives included in performance measurement and compensation. In addition, IFC will continue its system of performance rewards linked to project outcomes and more frequent promotions enabled by a higher number of grade levels for staff. Source: IFC Road Map FY11-13- Maximizing Impact, Unlocking Our Potential (IFC 2010), FY11 Business Plan and Budget (IFC 2010); and IFC’s FY12 Business Plan & Budget (IFC 2011), IFC 2013 Change Initiative - an update at an informal meeting with the Board of Directors (IFC 2010). definition of client services includes trust fund resources (9 percent), as well as resource mobilization and management (5 percent), primarily for trust fund management and treasury functions.18 Adjusting for these two, the amount spent on client services drops to 47 percent of the Bank’s budget. “Client services� covers a broad range of other expenditure items with differing levels of separation from the end product delivery to the client. These include country services (making up 32 percent of Bank budget), sector and global services (12 percent), other operational support (2 percent), and administrative support to products and programs (1 percent) s. Expenditure on country services is 32 Institutional Issues and Organizational Structure     103 percent of the Bank’s budget, or one in three dollars, for direct costs of lending and non-lending products delivered to country clients. Budget Expenditures Within the Regions Over the FY02-10 period, the regional sector units are receiving a decreasing share of regional resources, largely due to the growth of regional functions. While the Regions continue to utilize two-thirds of overall Bank resources, within the Regions, the units responsible for operational delivery and quality are under increasing budget pressure. Based on end-of-year expenditures, the share of regional Bank budget resources expended by regional sector units, which are directly responsible for product delivery, declined by 5 percent, receiving a 61 percent share ($724 million in 2010 constant dollars) in FY10 compared to a 66 percent share ($681 million) in FY02 (see Figure 5.1).19 Under the matrix system all the technical staff, including those based in country offices, are mapped to and paid for by the sector units, and these charges include their overheads (sustaining costs). CMUs share of regional Bank budget, decreased slightly from a 23 percent share ($231 million) in FY02 to a 21 percent share ($246 million) in FY10 but with decentralization of sector staff, paid for by the sector units, that side of the matrix is relatively well endowed. The remainder of the Regions’ Bank budget expenditures went to regional functions, which grew from $117 million in FY02 to $213 million in FY10 (see Appendix L). Regional variation shows sector units in Africa and the Middle East and North Africa and CMUs in Latin America and the Caribbean and South Asia receiving the lowest share of regional Bank budgets in FY10 (see Appendix L). The trend of growth in regional functions and declining share of SMU budgets is found in all Regions. There thus appears to be a strong association between budget authority and the share of resources withheld, first at the regional vice presidency level for regional functions, then at the CMU level. The growth in regional functions has been driven by regional core operations services. Growth was found in the majority of individual units located within regional functions. However, in FY10 the regional OS units made up approximately 45 percent of the Bank’s budget toward regional functions and explained almost half (48 percent) of the growth over the period. These OS units are the regional counterparts of OPCS and provide procurement, financial management, quality assurance, and other operational support to the Regions. Unsurprisingly, the largest growth within the OS units has been in the numbers of fiduciary staff, initially due to the remapping of fiduciary staff from country units to the OS units, but which grew by more than 150 percent from FY05-10, even after the remapping. Reliance on Trust Funds In FY10, a total of $1.4 billion (57 percent) of the total Bank budget expenditures of $2.46 billion was administered by the Regions and networks. When trust funds are included, the share of funds managed through matrix arrangements 104 The Matrix System at Work Figure 5.1 Regional Sector Units Are Delivering More With a Smaller Share of Regional Bank Budgets as Overheads Grow 1,200 (Constant 2010 US$ millions) 15% 17% 17% 18% 1,000 14% 16% 16% 11% 14% 21% 21% 21% 21% 21% 21% Regional 800 23% 22% 22% Functions BB Actuals 600 CMU BB Actuals 400 66% 64% 65% 63% 64% 62% 62% 62% 61% SMU BB Actuals 200 - FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Fiscal Year Source: World Bank databases. increases to 59 percent since the Regions and networks together absorb 85 percent of trust funds. The Bank has become more reliant on external financing through trust funds, the allocation of which falls outside the Bank’s annual budgeting process under the matrix system. External financing supports a growing proportion of the Bank’s resources–from an annual average of 19 percent in FY02-04 to 26 percent in FY08-10.20 From FY02-04 to FY08-10 the share of budget resources from BETFs has increased from 10 percent to 16 percent in the Regions, 32 to 44 percent in network anchors, and 23 to 26 percent in DEC and WBI combined. Trust funds have become critical to the Bank’s core business and are, in effect, a way of maintaining the appearance of a flat budget ceiling while allowing sector staff and managers to cope with the constraints in the Bank’s budgeting process under the matrix system to fill funding gaps. Network anchors have become more reliant on trust funds than the Regions. In FY08-10 the Regions taken together received 49 percent of trust funds and 49 percent of overall Bank budget funding. Network VPUs received a greater share of trust funds (36 percent) than their 9 percent share of overall Bank budget (Figure 5.2). BETFs financed an average of 16 percent of regional budgets and 44 percent of network anchor budgets, with more than half of the SDN anchor’s expenditures coming from trust funds as it manages over 70 percent of all network anchor trust funds. The five largest trust funds managed by the network anchors—which alone amount to almost 50 percent of network trust funds—are the Water and Sanitation Program, Energy Sector Management Assistance Program, Carbon Fund, and Global Facility for Disaster Reduction and Recovery in SDN, and the Consultative Group to Assist the Poorest in FPD. Institutional Issues and Organizational Structure     105 Access to trust funds gives the network VPUs a degree of autonomy to finance sector and corporate priorities such as gender and governance, and global priorities such as climate change whose externalities may extend beyond individual  client countries and which may not otherwise be integrated into country programs. However, trust funds also create contribute to the decline noted earlier in cross support provided by the network VPUs to the Regions.21 Budget Incentives for Cross Support Budget rules hinder staff from working across organizational boundaries. Current arrangements require each unit to estimate the amount of cross support it expects to sell and purchase in the coming year. Even though the Bank’s time recording system captures actual cross support across organizational units in real time, deviations from the beginning-of-year estimate for inter-VPU cross support are only reconciled through adjustments in the budget for next fiscal year, or in limited circumstances twice per fiscal year. Each VPU is free to set internal rules for intra-VPU budget settlements, but in some cases, cross support within the same sector department may not involve budget settlements at all. This non-payment or substantial delay between the provision of cross support and the payment for that work penalizes a unit for providing unanticipated cross support. On the demand side, a disincentive arises from the significant differential between the cost of staff and the cost of consultants due to Figure 5.2 Network Anchors and WBI Receive a Much Higher Proportion of Trust Funds (BETF) Than Bank Budget FY08-10 Budget Actuals FY08-10 BETFs ($2,375 milllion annual average in ($446 million annual average in 2010 constant US$) 2010 constant US$) 1% 8% 1% 3% 0% 1% 2% 9% Regions 13% Network Anchors DEC/WBI Corporate 49% 49% Finance 8% Administration Other Operations 6% Other 36% 5% 9% Source: World Bank databases. Note: Other included Central Accounts and Other Unit expenditures. Other Operations includes OPCS and QLP. 106 The Matrix System at Work the markup charged to purchasing units for sustaining costs of staff.22 Any initiative to jump-start cross support must address these penalties. Quality Assurance and Conflicts of Interest The Bank currently lacks a mechanism for transparent assessment of operational quality in real time, independent of line management. Effective organizations ensure the quality of product delivery and operational support processes by undertaking additional due diligence for quality control, although both tend to be under line management. In addition, separate mechanisms for quality assurance, financial audits, and independent evaluations provide independent feedback to senior management, governing bodies, and external constituents on quality. These systems for quality assurance, audits, and evaluation are not considered credible unless they are administered by third parties. While the Bank has an Internal Audit Department and an Independent Evaluation Group, no entity is responsible for undertaking third-party quality assurance of operations during implementation since QAG was dissolved in June 2010. Upon its dissolution, QAG’s budget and remaining staff were transferred to the OPCS vice presidency which retains primary responsibility for those functions, including portfolio reviews. Management has responded that the funds were used in part for completing an ongoing QAG review; that review has not been made public.23 As of the end of FY11, no alternate mechanism had been put in place to generate independent or arm’s-length data on lending and AAA quality24 and provide feedback to senior management. In effect, an information gap has been created in the Bank’s quality assurance system. Because of its heavy corporate responsibilities and support functions for regional operations, OPCS itself cannot play an independent quality assurance role. OPCS performs a crucial corporate role for the Bank. Its core corporate functions are (a) formulating operational policies and guidance to Bank staff for lending and nonlending operations, (b) overseeing country services across the Bank, (c) acting as the secretariat to the managing directors and the Operations Committee, and (d) consolidating results and results reporting to the Board of Directors. In addition to those four vital roles, OPCS also provides a home for additional functions, including (e) the financial management and procurement network anchors, which provide operational services to task teams, (f) quality control for fiduciary and safeguards for all lending operations, (g) results reporting, (h) the Internal Reforms Secretariat, and (i) the fragile and conflict affected countries unit. Because it houses the financial management and procurement Sector Boards, OPCS states that it is both inside and outside the matrix. As the anchor of country services, it prepares the CAS and DPL Retrospectives, and as the home of the Results Secretariat, it is responsible for the production of the Results Report and the Corporate Scorecard initiated in FY11. Its counterpart Operational Services units in the Regions provide fiduciary and safeguards clearances and undertake quality Institutional Issues and Organizational Structure     107 control for all operations; they also act as the secretariat to the Regional Operations Committee. The additional functions beyond the core corporate functions of OPCS can give rise to conflict of interest. The quality control function for safeguards was moved to OPCS from the vice presidency that houses the environmental and social development sectors when it merged with infrastructure sectors to avoid potential conflicts of interest by ensuring that the units providing clearances do not provide operational support to task teams. However, the same firewall to avoid conflict of interest has not been established for fiduciary areas—financial management and procurement—where the OS units provide both operational support and clearances. Senior management currently lacks a mechanism to receive transparent feedback on operational quality independent of line management. Given the corporate role of OPCS, its role in quality control, and its role in providing guidance to staff and oversight of specialized areas, including CASs, DPLs, and new lending instruments, OPCS and its regional OS units play a critical role in the Bank’s delivery of services to country clients. As such, it is unreasonable to expect that OPCS or its regional OS units will be able to undertake quality assurance to assess the quality and robustness of the Bank’s quality control systems with any degree of independence. Key Findings and Conclusions Imbalance in the regional matrix. At first glance it appears that one of the two objectives of the Bank’s matrix system—enhancing client responsiveness—has been achieved but that the sector side of the matrix did not succeed in creating global technical practices with world-class knowledge systems, delivering quality services. The very strength of the country units created an imbalance in the regional matrices. The imbalanced relationship between country and sector units created suboptimal tensions giving priority to short-term client demands over longer-term strategic needs. This is reflected in the weaker results on realism, broader country ownership, and implementation in country programs. Imbalances also exist in the Bank-wide matrix. The very strength of the regional matrices weakened the networks and consequently the Bank-wide matrix. The design of Sector Boards appears to have been resulted in the overburdening of regional sector managers, who have little incentive, time, or resources to perform the multifarious corporate roles expected of them. A more realistic sharing of responsibilities and accountabilities between the sector directors and sector managers might have enabled a more realistic design of the sector units. Further, the sector anchors took on additional corporate and global tasks that diverted attention from their ability to support frontline operations in the Regions. Subsequent changes added stress to the matrix design. The matrix system has not been able to adapt to issues arising out of span of control or subsequent 108 The Matrix System at Work organizational restructuring. The merger of SDN forced the Regions to adopt second-best or, in some instances, third-best solutions that varied across the Regions. The broadening of the sector scope of responsibility overburdened sector units limiting their ability to ensure knowledge flows and quality services. And the checks and balances created by the existence of a quality assurance system, independent of line management, disappeared. Finally, the matrix system also suffered from the absence of a shared set of organizational objectives with clearly identified indicators to monitor progress. At the end of FY11, management introduced a Corporate Scorecard to track results. A complete assessment of the scorecard is beyond the scope of this evaluation. The viability of the instrument will depend on the realism and transparency of the indicators and the alignment between the indicators of organizational effectiveness and modernization (Tier 4 of the Corporate Scorecard), the Bank’s development effectiveness, and country results supported by the Bank. Notes 1 According to the Oxford English Dictionary, lateral thinking is “a way of thinking which seeks the solution to intractable problems through unorthodox methods, or elements which would normally be ignored by logical thinking.� Edward de Bono invented this concept to mean thinking beyond the range of existing perceptions, concepts, and boundaries. See The Use of Lateral Thinking. London: Jonathan Cape, 1967. Also see New Think: The Use of Lateral Thinking in the Generation of New Ideas. New York: Basic Books, 1969. 2 See findings from country director and sector manager interviews in Appendix E. 3 See findings from sector manager interviews in Appendix E. 4 The program is summarized in a document issued by the Reform Secretariat, “10 Point Program towards Enhanced Sector Practices.� April 26, 2011. 5 See “QAG Assessment of Sector Boards� (2004). 6 The new structure of Global Technical Practices being piloted by FPD intends to change this pattern by making the directors accountable for several of these functions. 7 Oakerson, Ronald J. 1989. “Governance Structures for Enhancing Accountability and Responsiveness,� in James L. Perry, ed., Handbook of Public Administration. San Francisco: Jossey-Bass. 8 The span of control is defined as the number of people who can be effectively managed by one manager. 9 This number excludes OPCS units. Institutional Issues and Organizational Structure     109 10 In the Africa Region, SDN has appointed sector leaders as de facto deputies to help manage their larger sector units. But this role is still not part of the Bank’s formal managerial structure. 11 The first vice president of SDN said in her welcoming speech to ESSD staff, she did not know why social development too should be in SDN but she went along with the recommendation of the acting vice president of ESSD. It was only subsequent to the formation of SDN that social sustainability emerged as a business line under the social development Sector Board. 12 The former Department for Transport and Urban Development was eliminated. The Transport sector moved to the Infrastructure Department (TWI); Urban Development joined the Finance and Economics Department (FEU). 13 The eight Sector Boards do not include climate change or disaster management, which have large teams financed by two of the largest Bank-executed trust funds. 14 OPCS is included under network VPUs as they house the two Sector Boards for financial management and procurement. Although some OPCS functions fall outside the matrix, the bulk of their staff, especially in the Regions are part of the matrix. An additional 481 staff were in DEC and WBI, which have important roles in the Bank’s knowledge services but are considered outside the matrix. 15 See Safeguards and Sustainability Policies in a Changing World: An Evaluation of World Bank Group Experience (IEG 2010). 16 Subsequent to IEG’s evaluation of the safeguard policies, the World Bank has initiated a review of the safeguards framework with a view to updating it and addressing policy gaps. Management has indicated that the review will be completed in 2013. 17 The Pulse Survey was conducted in FY11 by Bank management for some networks and Regions on demand. 18 Resource Mobilization and Management includes trust fund administration, trust fund program management, project finance and guarantee operations, resource mobilization, banking and asset liability management, capital markets, and asset management. 19 Throughout the report, budget figures are reported from the Bank’s cost reporting system as assigned to the fund center (holder of budget). Because they are not disaggregated in the Bank’s reporting system, Bank budget actuals listed here include both actual expenditures from the Bank’s budget as agreed by the Board of the World Bank and non-BETF external financing (reimbursables and fee income). All budget figures are reported in 2010 constant U.S. dollar millions, converted using the deflator obtained from the Bank’s corporate finance unit. 20 Including 11 percent for Bank-executed trust funds and 8 percent for reimbursables and fee income. Figures for this calculation are taken from the Bank’s expenditure report, which is the only available source of comprehensive figures on reimbursables and fee income. 21 Management reports that “based on cumulative BETF disbursements in FY07-11, about 60 percent of network BETFs were used to support their work on global partnerships, followed by about 30 percent used in knowledge activities.� 22 These budget disincentives were specifically identified by the FPD network, which is undertaking a pilot experiment to address them. See Budget Report to the Board for FY12. 110 The Matrix System at Work 23 The portfolio reviews are issued quarterly by compiling existing data but these are not resource intensive. Management is also reported to have transferred a fifth of the funds ($1 million) directly to the Regions for quality-related work. 24 The report Towards Open Knowledge: The State of World Bank Knowledge Services, Reform Secretariat, OPCS 2011, prepared by Bank management, drew heavily on QAG AAA findings. Management’s FY11 Results Report also drew heavily on QAG findings. However, results from the second round of Quality Assessment of Lending Portfolio have not been publicly released. References de Bono, Edward. 1967. The Use of Lateral Thinking. London: Jonathan Cape, 1967. Independent Evaluation Group (IEG). 2010c. Safeguards and Sustainability Policies in a Changing World: An Independent Evaluation of World Bank Group Experience. Washington, DC: World Bank. International Finance Corporation, 2010. IFC Road Map FY11–13—Maximizing Impact, Unlocking Our Potential, Washington, DC: International Finance Corporation. Oakerson, Ronald J. 1989. “Governance Structures for Enhancing Accountability and Responsiveness.� In Handbook of Public Administration, ed. James L. Perry. San Francisco: Jossey-Bass. Phillips, David A. 2009. Reforming the World Bank: Twenty Years of Trial and Error. New York: New York University Press. Institutional Issues and Organizational Structure     111 Chapter 6 Conclusions and Recommendations Despite the many positive aspects of the 1997 reorganization and improvement in client-responsiveness and in country-level outcomes during more recent years, the general conclusion of the evaluation is that the creative tensions and balances sought by the reorganization—between country and sectors, and between the Regions and networks—have not been achieved. Nor has the Bank succeeded in elevating its knowledge work to capture innovations from the field and sharing Bank and global knowledge across the institution. The matrix system, which was predicated on collaboration and teamwork among Regions and sectors, left intact or strengthened the existing incentives and underlying organizational culture. Coupled with the drive for decentralization, the matrix has strengthened regional operations while weakening linkages to the network anchors, and has served to maintain organizational silos and competition among the Regions and networks. In addition, a critical risk factor is looming: the erosion of the Bank’s global primacy in development knowledge, partly caused by short-term lending pressures and the concomitant disincentive to the internal transmission of knowledge, and more so by the recognition that the Bank is not the primary producer of development knowledge. The Bank’s Knowledge Report recognizes that the Bank’s ability to connect its staff and country clients to global knowledge and customize that knowledge to make it relevant to country conditions is as critical to the preservation of the Bank’s operational relevance as is its ability to produce relevant knowledge. Unless the Bank significantly modifies the means and incentives to ensure efficient sharing of Bank and external knowledge through better connectivity with global knowledge, the institution risks losing global relevance on both the knowledge and the lending side. Such a prospect cannot be averted by addressing only one part of the Bank’s interdependent system. There is widespread recognition that the benefits of the matrix system have in significant ways not been realized and have led to some substantial costs. It does not follow, however, that the matrix system should or indeed can be done away with. Changes in the external environment indicate that the matrix system is even more relevant today than when it was introduced. Client needs have diversified, with greater differentiation among countries, even within the Regions; the growth of GPGs and corporate priorities is creating tensions with the country model; demand for cutting-edge knowledge is growing both to enhance the quality of lending and as a growing business line for policy and program advice to clients; new global practices have emerged to meet needs such as information, communication and technology, and disaster management. The Bank’s ability to renew itself and function as a truly global Bank is critical to its success. Some modification of the balance of power or authority, revision of operating processes and incentives, and adjustments to the organizational structure is needed to recalibrate incentives, modify procedures to ensure greater balance across elements of the matrix, and in some instances revisit structural impediments to operation of the matrix. The wide diversity among the Bank’s country clients and the multiple challenges they face require the application of 114 The Matrix System at Work selectivity and customization from the wide range of sector and global practices that the Bank can offer. There is thus a need to match client and country needs with the products and sector practices that the Bank can produce or mobilize, inevitably leading to some form of a matrix system. As a public institution, the Bank is limited in its ability to implement market- based practices and processes that are commonly applied, with success, in private firms to implement matrix systems. Yet even as a public institution, the Bank can learn from the experience of the private sector in implementing matrix structures and systems. Research and experience from the private sector (Gottlieb 2007, Galbraith 2009) shows that effectively operating matrix structures requires internal consistency among five key factors: (a) alignment and tracking of goals; (b) clear responsibilities and expectations; (c) a collaborative culture and leadership; (d) mature resource management processes; and (e) a robust performance management framework. The second key lesson from matrix organizations is that organizational design is not just about structure. Incentives and processes have to be aligned with structures, and all three have to be aligned with strategic objectives for matrix organizations to work. These key factors and the need for internal consistency remain fully valid for the Bank, while recognizing the Bank’s unique challenges as a public institution with multiple objectives and global reach. The World Bank moved quickly to revamp its organizational structure. However, as the evidence from this assessment reveals, changes in incentives and processes were inadequate and not fully aligned with the strategic goals of the 1997 reforms. The incentive created by transferring budget authority to country directors increased responsiveness to the client but failed to create the conditions for quality services based on global technical excellence. The quality of Bank knowledge services is viewed by many as having deteriorated since the matrix was introduced. Budgetary processes and individual incentives militated against flows of knowledge and attention to quality. Organizational silos re-emerged along different boundaries. The very strength of the regional matrix undermined the effectiveness of the Bank-wide matrix. And the links between the center and Regions was further weakened by decentralization and the emergence of global and corporate priorities. The absence of alignment among the Bank’s corporate goals and the interests of the Regions and networks did not ring any alarm bells because no metrics were put in place to track or reward achievement of strategic goals other than through lending volume and numeric counts of knowledge products. IFC revised its matrix structure in 2010. IFC has built on its previous systems and includes individual incentives, and performance assessments aligned with the organization’s goals. Its compensation program is aligned with strategic objectives and includes achievement of quantifiable development goals. IFC has also proposed a set of monitoring metrics to track organizational results and identify problem areas more rapidly. While it is far too early to evaluate its implementation, these design features are worth considering as part of subsequent efforts to strengthen the Bank’s organizational effectiveness. Conclusions and Recommendations 115 It is to the credit of the staff of the World Bank that people in the institution, at all levels, have largely adapted to the matrix system and have found ways to reduce its transaction costs. The improvement in responsiveness to the client must be preserved while expanding ownership to other country stakeholders, and the efficiency gains made by several regions and some sectors should be protected, and their flexibility preserved. The evaluation considered options ranging from a complete organizational restructuring to minimalist changes. Persuaded by the finding that attempts to replace the entire system would entail substantial disruption and incur large transition costs, and the evidence on client responsiveness and initial gains on quality, the evaluation does not recommend replacement of the matrix system. Nonetheless, the constraints and institutional deficiencies in the matrix system are sufficiently out of sync with current client needs and emerging operational challenges that a significant effort to address these systemic constraints in a holistic manner is in order. Bank management is cognizant of deficiencies in the matrix system and in FY10 established a Matrix Leadership Team, comprised of operational vice presidents to reinvigorate the system. In parallel, management established a Knowledge and Learning Council, headed by a managing director (MD), to improve the Bank’s ability to deliver knowledge, enhance learning opportunities for Bank staff, and link country practitioners to global sources of knowledge to increase development impact. An Internal Reform Secretariat supports these entities, and in FY11, an MD was appointed to lead the Bank’s modernization efforts. This has led to the identification of key constraints, such as the span of control of sector managers and, in the past year, to several initiatives to upgrade the quality of the Bank’s knowledge work. However, the measures proposed have yet to address the underlying incentives, operational processes, or structural constraints that have impeded the matrix system. The general IEG recommendation is to make substantial interrelated changes in the matrix system to realize its potential and reduce its downside. The objectives of the system cannot be achieved by further tinkering at the margins or cosmetic relabeling, but instead requires major adjustments to remove unnecessary constraints to its implementation and the establishment of the clear rules and incentives required by its logic. The Bank-wide matrix— between the network anchors and the Regions—has become increasingly weak and needs substantial strengthening, with some rebalancing needed within the Regions. In order to do so, the responsibilities and accountabilities of sector directors and managers also need to be adjusted to ensure a balance between corporate sector responsibilities and regional needs. And the unfinished task of designing incentives and operational processes that support operational needs must be addressed urgently. The overarching theme of the adjustments recommended by IEG is to restore the creative tensions among the Regions and networks and revitalize the corporate citizenship of its staff, while allowing operational flexibility and remaining responsive to external demands for Bank global leadership and support to the 116 The Matrix System at Work development challenges faced by its external clients. The main adjustments are described below. Recommendations The findings of this evaluation point to separate actions on incentives, processes, and structures. These are described generally below. Incentives Institutional incentives: Under guidance of the Knowledge and Learning Council, the Bank should strengthen the incentive system for effective sharing and use of internal and external knowledge across the Bank, ensuring resources to enable documentation and sharing of key lessons from country experience; financing of thematic groups and communities of practice for capture and sharing of knowledge; providing opportunities and incentives for field-based staff to access knowledge, including through sector weeks, webinars, training, short secondments, and cross-country work; developing and implementing monitoring systems to track use of knowledge from relevant operations and AAA from other regions, and from DEC; and assessing AAA effectiveness and monitoring its use by clients. Staff incentives: Management should revamp the system of formal incentives for individuals to ensure that middle and senior management and staff are assessed on performance, including (a) on sharing of knowledge and collaboration across sectors, and (b) the quality and results of operations for which they are responsible; and reward performance that goes beyond Board approval for lending operations, while being more transparent and better linking career advancement decisions to quality and achievement of development objectives. Processes Redefine the purpose and frequency of SSPs as a more modest five-yearly report on the “state-of-the-sector� and provide independent funding for strategic ESW and business development. Sector reports should communicate emerging good practice and operational experience in different country contexts to staff and external partners. To ensure that country strategies are responsive to medium-term sector priorities, funding for strategic ESW and business development by regional sector units should be independent of CMU budgets and WPAs (either from a more substantial pool with the chief economist, or at the sector department level). Quality: Strengthen mechanisms for review of quality and establish a quality assurance system independent of line management. The focus on quality should be enhanced by (a) strengthening the peer review process and QERs for technical quality, (b) ensuring dual accountability of country and Conclusions and Recommendations 117 sector directors for quality, and (c) establishing a quality assurance mechanism for feedback on quality independent of line management. Budget: To remove constraints on cross sector and cross Region collaboration, management should eliminate cross support penalties and introduce flexibility in budgetary rules to allow sharing of tasks and budgets across organizational units. Structure Bank-wide matrix: To enhance matrix effectiveness, strengthen the links among the Regions and between the Regions and anchors, and restructure Sector Boards making sector directors accountable for Bank-wide linkages and sector governance. This would need to address: fragmentation of sectors by reconfiguring Sector Boards to regroup those with related or overlapping mandates under the same networks; the range of responsibilities, particularly in SDN, by adjusting network size (number of staff) and scope (number of sectors); inadequate accountability to networks by introducing co-financing of regional sector directors by both networks and Regions; and inadequate division of responsibilities and accountabilities for sector governance between sector directors (for Bank-wide tasks) and sector managers (for regional tasks). Regional matrix: Rebalance the relationship between country and sector units and regional functions to ensure appropriate attention to lending, knowledge, and quality. This would need to address: span of control of regional sector managers by relieving them of responsibility for Bank-wide sector governance; the number of regional sector managers based on workload (number of staff and operations) in sectors and Regions, recognizing actual costs of shadow managers; and the declining regional sector unit budget shares with a consideration of restoring them to FY02 levels. Network rebalancing: Address key organizational constraints in the Sustainable Development Network. This includes: (a) excessive span of control of SDN sector managers; (b) dissimilarity in structures across the Regions; (c) unrealistic range of sector responsibilities of sector directors and country sector leaders/coordinators; and (d) removing the separation between safeguards and non-safeguards work by paying equal attention to both aspects of environmental and social sustainability. References Galbraith, Jay R. 2009. Designing Matrix Organizations That Actually Work: How IBM, Procter & Gamble, and Others Design for Success. San Francisco: Jossey-Bass. Gottlieb, Marvin R. 2007. The Matrix Organization Reloaded: Adventures in Team and Project Management. Westport, Conn.: Praeger Publishers. 118 The Matrix System at Work Appendix A Matrix Management and Organization of the World Bank in Historical Perspective This Appendix discusses what a matrix organization is, the history of organizational arrangements at the World Bank, and the role of formal and informal incentives for technical quality and client orientation. Matrix Organizations and the World Bank Matrix organizations are characterized by their managerial arrangements, which are assigned to functional and operational units. Individuals in matrix organizations are grouped by functions—for instance engineers in a construction company or sector experts at a development bank—and report to a functional manager. They are then identified for particular tasks delivered to a client—to carry through the example, the developer of an office building or a government implementing a project or reform program—which is overseen by an operational manager. Individuals thus have dual reporting lines, accountable to both the functional and operational managers. The purpose of this arrangement is to allow for better information sharing across task boundaries, better customization of services to the needs of clients, and better management of career progression in particular subject areas. The World Bank is organized geographically. The Bank’s country units are operational, serving as the face of the institution to its clients. The Bank’s sector units are functional, and manage sector-based tasks to be delivered to clients. To avoid confusion with other uses of terms “operations� and “functions� at the Bank, the evaluation sometimes refers to country units as client groups rather than operational units and to sector units as product groups rather than functional units. The World Bank’s current organization consists of several layers of matrices. The “regional matrix� refers to the relationship of country director- controlled country management units (CMUs), which are responsible for managing relationships with country counterparts and packaging a coherent group of products and services that will contribute to country-level development results; and sector manager-controlled sector management units (SMUs), which consist of technical experts who manage individual lending and non-lending tasks. CMUs are client groups and SMUs are product groups. Each of the six regional vice presidential units (VPUs) is responsible for the functioning of its own regional matrix. Staff based in SMUs are accountable both to their own sector manager who controls their performance evaluation and compensation, and to the country director who typically controls the final sign-off on all operational work and allocates budget across sectors. Staff in the CMUs are only accountable to the country director. Country directors hold the budget and negotiate a series of work program agreements (WPAs)—contracts to deliver knowledge and lending services for an agreed upon budget—with sector managers. A representation of the regional matrix is shown in Figure A.1. The “Bank-wide matrix� refers to the relationship between operational regional VPUs (which include CMUs, SMUs, and central regional units, and the matrix through which they interrelate) responsible for delivery of services to clients 120 The Matrix System at Work Figure A.1 The Regional Matrix F U L C R U M : Control Budget and Strategy Regional vice president and CMU1 CMU2 CMU3 CMUn regional functions (country director) (country director) (country director) … (country director) SMU1 WPA WPA (sector manager) WPA Sector Department SMU2 between WPA (sector director) (sector manager) SMU2 & CMU2 SMU3 WPA WPA WPA (sector manager) SMU4 WPA (sector manager) Control Staff WPA Sector SMU5 between Department SMU5 & CMU3 WPA (sector director) (sector manager) WPA SMU6 between WPA (sector manager) SMU6 & CMU1 … SMUn-1 WPA WPA Sector (sector manager) Department (sector director) SMUn WPA WPA (sector manager) Note: WPA=work program agreement; SMU=sector management unit; CMU=country management unit. and functional networks or sector families that develop knowledge, interface with external partners, and manage products in the Regions. These networks are primarily composed of the staff assigned to SMUs in each of the six geographic Regions. In addition, network anchors are housed in separate network VPUs and intended to serve as a hub for individual sectors, helping to manage knowledge in the sector and provide support to operations managed by the Regions. Staff from the network anchors are also part of these networks. Staff in the regional SMUs are therefore linked to both network and regional VPUs. Regional VPUs are governed by the regional vice president, while sector families are governed by a Sector Board primarily consisting Appendix A 121 of a sector manager from each of the Regions and a director located in the network anchor. For individual staff, there is no dual reporting in the big matrix. SMU and CMU staff ultimately report to the regional vice president; network anchor staff report to the network vice president. Moreover, there is no equivalent of a WPA, nor requirement that staff provide direct support to tasks managed in one of the other VPUs (whether network or Region). A representation of the Bank-wide matrix is depicted in Figure A.2. Organizational Arrangements before 1997 The existence of geographically organized country units and technical sector units does not necessitate a matrix management system. Like other organizations in the public and private sectors with a diverse client base and Figure A.2 The Bank-wide Matrix FULCRUM: President and Region1 Region2 Region3 Region6 managing directors (RVP) (RVP) (RVP) … (RVP) Anchor1 Sector Board managed by regional and network representatives SB1 Network1 Anchor2 (NVP) Regional VPU, including SMUs, CMUs, Anchor3 and regional functions Anchor4,5 SB4 Network2 SB4 (NVP) Anchor6 … Network4 Anchorn (NVP) Note: SB=Sector Board; NVP=network vice president; RVP=regional vice president. 122 The Matrix System at Work a broad range of products, the Bank has operated with an internal structure that included client groups (programs departments, country departments, and CMUs) and product groups (technical departments, sector divisions, central vice presidencies) for several decades. The tensions between the two groups have been managed through a combination of institutional arrangements and reporting relationships, as well as informal incentives. Prior to the 1997 reforms, the management arrangements remained hierarchical, with staff members and middle managers reporting directly to a single line supervisor. In the much smaller Bank of the 1960s and early 1970s, the project groups were central (Bank-wide) projects departments, dealing with client groups that formed what we now know as regional VPUs. As the Bank expanded rapidly in the 1970s under President Robert Strange McNamara, the central projects departments were gradually spun off into regional projects departments (by the mid 1970s there were still a combination of regional and central project departments), so that the search for an equilibrium (and creative tension) between client and product groups was shifted to the Regions and the conflict resolution role was placed on the regional vice president. In 1987 a different mechanism was introduced, as highlighted in [Box A.1] to regulate the relationship between country and sector units, in response to an assessment that that client orientation was weak, and sector units operated in a supply-driven manner. Thereafter, country and sector units operated together under a single director with technical staff split into sector divisions under the country director. While the reform in 1987 strengthened the country focus, technical staff worked primarily in country department silos and the technical quality of the Bank’s project work deteriorated.1 The 1987 reform was aimed at transforming the Bank from one with an exclusive focus on “projects� toward one that viewed projects in the context of country programs. The 20 Country Departments were composed of subregional clusters of countries, with technical and program management skills combined in each under Division Chiefs to carry out the work program. Highly specialized, one-of-a kind skills were housed in four Technical Departments to service the Country Departments.2 However, the relationship between country and sector units remained hierarchical and perpetuated silos across Country Departments, affecting knowledge flows and operational quality. The 1997 Reform By the mid-90s, pressure for change was acute. First, tensions between CMUs and SMUs continued and despite increasing focus at the country rather than global level, client orientation was lower than desired. Second, technical quality of the SMUs became a growing concern: subdivision of technical staff into country departments was seen to have led to a loss of global technical leadership. Third, changes in the external environment created an imperative for reinventing the Bank. The share of total capital flows coming from development Appendix A 123 Box A.1 A Thumbnail Sketch of the 1987 Reorganization After its initial contributions to postwar reconstruction, until the 1970s, the Bank financed mainly physical infrastructure and progressively expanded to other sectors, but always with a project focus. Its internal organization and rules were consistent with that focus. Most substantive decisions on choosing what project to support and in what country, as well as the project design, supervision, and implementation, were entrusted to specialists in technical departments. The economic development literature of the late 1950s and 1960s made evident that a project cannot achieve its development objectives if the “strategy of development� is defective and the surrounding economic policy environment is deficient or distortionary (Hirschman, 1958). Next there was recognition of the fungibility of financial resources (De Vries, 1967). A project financed by an aid donor and prepared under strict technical standards, in fact, released the government’s own resources (or aid from a less demanding source) to finance another project X which may be neither technically nor economically sound (Schiavo-Campo and Singer, 1970). In reality, the donor enabled project X, of which it knew little, while reassuring itself that its finan- cial support met the highest technical and economic requirements. And, if the donor-financed project was not a government priority, neither the donor-financed project nor the aid was likely to be effective. By the 1970s, it became necessary to consider the economic profile of all public investment, as well as the soundness of the policy framework that conditioned the projects’ rate of return in na- tional economic terms. The Bank’s approach began to evolve from an exclusive focus on isolated projects to the probable development impact of the entire public investment portfolio and, later, of public expenditure as a whole. In the meantime, under the McNamara presidency, the World Bank’s resources expanded several fold, leading to a scarcity of “bankable� projects, increasing doubts about the quality of lending, and a view of the institution as a “hobbled giant� (Please, 1984) constrained by its project-dominated culture and organizational structure. This led to a gradual shift toward “program� lending, linked to the soundness of the public investment pro- gram and the economic policy framework of the recipient government. By 1986, program lending and budget support grew significantly in the Bank’s portfolio, and the previous organizational architecture built around technical departments was no longer tenable. The 1987 reorganization aimed at a new balance between country focus and technical quality, by creating “country departments� composed of several countries in a subregion and compris- ing all sector divisions as well as a “country operations� division charged with formulating the Bank “strategy� of assistance for each recipient country. The country departments were led by a department director allocating budgets to the sector divisions for their projects (presumably consistent with the strategy for the individual client countries). To complement the arrange- ment, a “technical department� was created in each Region to provide specialist support as and when demanded by the country departments. External circumstances and the understanding of development changed dramatically in the sub- sequent decade—from the seismic repercussions of the disappearance of the Soviet Union, the informatics revolution, the demands of globalization, the new institutional economics, the rec- ognition of the governance-development nexus, and the increasing importance of private financ- ing of development. Source: IEG. 124 The Matrix System at Work assistance declined from 50 percent to 20 percent between 1990 and 1997, while new sources of knowledge services emerged in partner countries, including universities, think tanks, and the private sector in the developing world. The 1997 reforms, as detailed above, tried to address all three of these challenges through three main changes: (a) a new set of organizational arrangements; (b) increased decentralization; and (c) matrix management arrangements which included dual reporting in the SMUs (but not the CMUs) and budgets authority with the CMUs. Thus, the introduction of matrix management was a major element of the 1997 changes, but not the only change. The new matrix system was layered on top of existing operational realities: rapid changes in the Bank’s external environment, such as renewed demand for lending in the wake of the Asian financial crises and the growing significance of global public goods; longstanding conventions in organizational culture, including the widespread use of lending commitments as the dominant metric for success; and competing objectives by varying stakeholders, for instance those of donors, nongovernmental organizations (NGOs), and country partners. While these underlying realities are unrelated to the internal matrix management of client and SMUs, they represent a part of the operating environment in which the matrix system exists, and thus affect its effectiveness directly. Incentives for Client Orientation and Technical Excellence after 1997 Formal incentives. All the changes introduced in 1997 were expected to operate by affecting formal incentives and behavior, in a way that improved client orientation and technical excellence. The new organizational arrangements consisted of (i) reducing the number of countries covered by each CMU, (ii) creating networks (and Sector Boards) to link SMUs at the regional and Bank- wide level (to increase the level of global technical excellence available to client teams) and (iii) creating network anchors (to facilitate exchanges of global technical knowledge among the Bank SMUs and with external partners). In addition, decentralization was intended to increase client orientation by facilitating the contacts with clients. The two sets of changes above were in line with (and an evolution of) the successive institutional arrangements during the previous two decades, with similar objectives and similar instruments. What made the package of reforms substantially different was the introduction of matrix management with dual reporting relationships and country budgets allocated by CMUs to the different SMUs. This, and not the creation of the networks and Sector Boards, or decentralization, is what makes the reforms of 1997 qualitatively different from those that came earlier. The main formal incentive introduced by the matrix system to affect client orientation was the country budget and allocation. All prior institutional arrangements were thought to have left excessive room for supply-driven behavior Appendix A 125 by SMUs because of their ability to deploy budget resources to countries without need for approval by the client teams. Although this problem had been reduced after 1987, it had not completely disappeared. Thus, after 1997, the introduction of country budgets was intended to allow the client groups (country director, CMU) to purchase services from the different product groups (sector manager, SMU) in line with the country assistance strategy and country needs. Product groups, on the other hand, had the skills (technical staff) but needed to finance their staff with budget allocations from different client groups. The fluctuating demand among countries was expected to balance out, particularly with arrangements designed to make Bank-wide resources available to individual country teams through the Bank-wide networks and Sector Boards. The other major matrix management incentive, to guarantee technical excellence and strategic alignment between country and sector strategies was the dual reporting relationships in the product teams. Technical staff working in a given country report to their sector managers as well as to the country directors, and regional sector managers and sector directors report to the regional vice presidents as well as to the network vice presidents. Client teams, however, continued to operate in a line management environment: country directors report only to regional vice presidents (but their performance reviews have inputs from regional sector managers and directors). The above set of formal incentives continues to date, 14 years after introduction, although with some adjustments in the institutional arrangements (e.g. change in the number of networks) and in the matrix management (particularly budget mechanisms). Informal incentives: organizational culture and external effects. Formal incentives are only part of the mechanisms that determine behavior, and are often affected, enhanced, or even overwhelmed by informal incentives. In the case of the reforms adopted in 1997, the two main types of informal incentives were the Bank’s organizational culture and the external effects. Changing external conditions have been particularly important in the 14-year period since the matrix system was introduced. The external conditions facing the Bank changed suddenly, the day after implementation of the reforms formally began, subjecting the new system to unanticipated stress that led to behaviors that were inconsistent with matrix operations. On July 2, 1997, the devaluation of the Thai baht ushered a major regional (East Asian) crisis that later spread to other countries such as Brazil and Russia. With it, there was a major drying up of private capital flows to developing countries, and a sudden increase in demand for Bank financing. The sudden increase in demand, and in the approval of quick disbursement loans, took management attention away from the support of the new matrix instruments and arrangements. Some of the responses to the crisis were totally contrary to the principles of the reforms, particularly the leading roles of country directors and CMUs. The 126 The Matrix System at Work most important of these responses was the establishment of the special financial operations unit (SFO) which was given budget and, effectively, autonomy from the CMUs in which it operated. This created considerable resentment between the affected CMUs and the SFO and, arguably, cast doubt on the support from senior management for the principles of the new matrix arrangements. At the same time, it was also criticized by CMUs that did not have access to the SFO resources, as source of budget misallocation. Thus, it alienated both groups of client teams and weakened the foundations of the new budgetary principles. The new budgetary system also quickly ran into difficulties because its detailed functioning principles had been left undefined, and in the first few years, the transaction costs became very high, and because the uncertainty associated with the large number of technical staff members (and their managers) who were concerned about finding effective demand for their services led to a new source of supply-driven behaviors and to efforts to ensure new tasks (to “fill the dance cards�). During the early 2000s, as the global economy recovered and private capital flows gradually rematerialized, the external conditions turned again closer to those that had led to the 1997 reforms: demand for Bank lending decreased, particularly from upper middle-income countries, including major repayments and reductions in country exposure. During this period, budgetary mechanisms were improved, as were some of the coordination mechanisms. Faced with the decline in demand, however, senior management encouraged lending increases, particularly in infrastructure, sometimes at the expense of developing cutting edge technical skills and knowledge generation and dissemination. In 2008, a new change in external conditions, the global financial crisis, led again to large increases in demand and in lending, mainly through development policy loans (DPLs), which have required leadership by CMUs, with only limited involvement of the SMUs (except for the Poverty Reduction and Economic Management Network, PREM). This may have reduced again the effectiveness of “matrix behaviors� sought by the 1997 reforms. Organizational culture (including the “loan approval culture�) also played a major factor in counteracting the incentives to use global knowledge through increased cooperation among sectors and across Regions. The 1992 Wapenhans report identified a focus on lending volume rather than implementation and supervision, and thus development outcomes. Its primary recommendation was “to focus attention at all levels more fully on the central objective of achieving sustainable development impact� (pg. 24). While the Bank has place more emphasis on “results� in recent years, staff and managers have confirmed that the lending culture still reigns. Given the primacy of lending, devoting attention to �matrix behaviors�—behaviors such as utilizing knowledge produced across the Bank, mobilizing expertise from across the Bank, and carefully considering the complex interactions of the various sectors—is seen to detract from, rather than contribute to successful working. Appendix A 127 The Bank and the New Aid Architecture In addition to client responsiveness and technical excellence, the reforms of 1997 were intended to address new efforts to increase coordination among donors and align donor systems with country systems to improve aid effectiveness. The United Nations adopted the Millennium Development Goals in September 2000. The central units (network anchors) became the external face of the Bank on issues relating to their sectors, adding an important external responsibility to their internal Bank function as aggregators and disseminators of knowledge. This aspect of the network anchors has greatly expanded over time, in line with the evolving international aid architecture and the emergence of new global partners. This expansion has gone well beyond what was expected in 1997, and has coincided with the emergence of global public goods and cross- national challenges as new features of Bank assistance. The Bank’s action in these areas developed in a reactive manner in response to external demands and expectations, and often without a clear results framework linking objectives, instruments, and expected results. Notes 1 The Wapenhans Report (1992) and first Quality Assurance Group (QAG) reviews, which cover older projects approved prior to 1997, indicated that about one third of Bank projects were not likely to have satisfactory outcomes. 2 In 1991, the Technical Departments were reduced in size—one third of the staff went to the country departments, one third to Central Vice Presidencies, and one third stayed in the Technical Departments within the Regions. References De Vries, Barend A. 1967. “External Aid for Plans or Projects: An Alternative Approach.� Economic Journal. Vol. 77, No. 307. Hirschman, Albert O. 1958. The Strategy of Economic Development. New Haven, Conn.: Yale University Press Please, Stanley. 1984. The Hobbled Giant: Essays on the World Bank. Boulder, Colo.: Westview Press. Schiavo-Campo, S. and Singer, H.W. 1970. Perspectives of Economic Development. Boston, Houghton Mifflin Company. Wapenhans, Wili A. 1992. The Wapenhans Report. Portfolio Management Task Force. Washington, DC: World Bank. (Also published as Effective Implementation: Key to Development Effectiveness.) 128 The Matrix System at Work Appendix B Evaluation Methodology This evaluation’s findings, conclusions, and recommendations are supported by evidence from a wide range of available documents and data sources, including relevant Bank management assessments of the matrix, portfolio data from the Bank’s administrative database, human resources data, and budget data. The evaluation also used matrix-relevant findings of previous Independent Evaluation Group (IEG) assessments, including country assistance evaluations (CAEs), country assistance strategy completion report (CASCR) reviews, and sector and thematic evaluations. Data Sources Literature Review IEG’s review of the literature on organizational reform, in general, and matrix- relevant Bank reforms, in particular, drew on the assessment of Sector Boards by the Bank’s former quality assurance group (QAG), the organizational effectiveness task force (OETF) report, and the 2010 Bank management review of the matrix, as well as on research and analytical pieces on both private sector matrix reform experience and the Bank’s various internal reform processes. Country- and Sector-Level Desk Reviews Country and sector-level bank documents were analyzed to better understand cross-regional and cross-sectoral variations in the effectiveness of the matrix and link these, to the extent possible, to regional and sector-specific incentives and performance indexes. Sector-level analysis covered eight sectors and cross- sector themes which have been evaluated by IEG since 2005. These reviews covered agriculture; environment; gender; health, nutrition and population; public sector governance; social safety nets; transport; and water. The 14 focus countries in the country-level review include the 10 East Asia and Pacific, Europe and Central Asia, Latin America and the Caribbean, Middle East and North Africa, and South Asia countries for which recent CAEs are available, and four Africa Region countries randomly selected among the eight with recent CAE reports. The 14 countries (Madagascar, Mali, Nigeria, Uganda, Cambodia, Indonesia, Georgia, Ukraine Honduras, Peru, Egypt, Yemen, Bangladesh and Nepal) include large and small, and low- and middle-income countries. Secondary Analysis of CASCR Reviews, Bank Staff Surveys, and Client Surveys IEG’s comparative secondary analysis of the 96 CASCR reviews undertaken for country programs designed after the introduction of the matrix system provided the means to assess the responsiveness and effectiveness of country programs. Subdividing the universe into two equal cohorts of programs helped identify the program elements along which performance has improved overtime, and those along which it has stagnated or deteriorated. (See Appendix C for more information on CASCR reviews secondary analysis methodology). 130 The Matrix System at Work Relevant Bank documents were surveyed for information regarding incentives and accountability mechanisms to ensure technical quality. The evaluation looked specifically into (i) managers’ documented roles and responsibilities for strategy development and implementation, staff management, client relations, and operational quality; and (ii) the role of regional and central mechanisms for quality assurance. Reviews of relevant questions and answers from five rounds of Bank-wide staff surveys as well as client and country surveys conducted between 2002 and 2010 supplemented the analyses. Budget Data Analysis Budget data Bank-wide, as well as for regional and network VPUs, over the FY02-10 period was analyzed to better understand and assess budget distribution across organizational units, trends in actual expenditures, and the impact of budgetary incentives on the working of the matrix. The budget review describes the evolution of the distribution of the Bank’s administrative budget, reimbursable and fee income received from clients, and Bank-executed trust funds. It also details matrix-relevant trends in expenditures on the various product lines as well as appraisal and supervision. Cross-Support Data Analysis The analysis of FY02-10 Bank-wide data on cross-support (i.e. time experts spend working outside their responsible units) provided evidence on the extent to which anchors and regional SMUs were able to mobilize technical expertise across the Bank and the degree of flexibility for technical staff to work across Regions and sectors. Data analyzed included data on cross- support from Regions to anchors, from anchors to Regions, across Regions and across sectors within Regions. Carryover work by recently transferred staff was separated from actual cross-support. Portfolio Data Analysis The degree of cross-sector and cross-thematic elements built into project design was approximated via analysis of trends in the sector and thematic composition of the portfolio (that is, the cumulative share of projects addressing a particular sector or theme) for Bank lending operations approved between FY95 and FY10. The review compared trends in the sector and thematic allocation of each of the major lending instruments, investment lending and development policy operations. Data specific to three corporate priorities (public sector governance, environment, and social development and gender) was also analyzed to determine the degree to which they have been mainstreamed in the Bank’s portfolio before and after the introduction of the matrix. Staffing Data Analysis Appendix B      131 Descriptive data on the staff strength of regional, network, and selected central VPUs was extracted from the Bank’s administrative database. The proportion of sector staff mapped to anchors as opposed to regional SMUs was also established. Data on decentralization, mobility/time in management unit, and span of control were obtained from human resources (HR) analytics to compute results for spans of control of managers and directors. Counts and shares of net staff in VPUs for more than three years, three to five years, five to seven years and seven years or more were used as indicators of staff mobility. IEG Staff Survey IEG surveyed 1,054 randomly selected technical Bank staff who had worked at the Bank for at least two years, a one-third sample of the universe. The Internet-based survey was tailored to three different self-identified staff groupings (those who worked primarily in a CMU, an SMU or an anchor unit for the past five years) and was fielded between February and March 2011. Staff were asked to rate the performance of the matrix along the three key dimensions of interest of the evaluation: (a) strategic alignment between country, sector, and corporate strategies; (b) knowledge production and dissemination; and (c) incentives and accountability for quality. Informed views on the strengths of the matrix compared to the pre-1997 system were also sought. Forty-one percent (435/1,054) of staff responded. Manager Interviews Structured face-to-face interviews were held with over 60 Bank managers. Interview format and questions varied slightly, depending on the areas of responsibility of the interviewees. Managers interviewed included 31 Bank sector managers and 13 country managers/directors for the eight focus sectors and themes and 14 focus countries. Managers, directors, and vice presidents were interviewed between March 2011 and June 2011. Focus Group Discussions Three two-hour long formal focus group discussions with 21 anchor and regional senior staff were held between July and September 2011. The main purpose of the discussions was (a) to gather in-depth feedback on the effectiveness of the matrix, (b) to better understand cross-regional and cross- sectoral differences in the working of the matrix, and (c) to test findings from the staff survey and manager interviews as well as conclusions derived from budget, staff, and portfolio analyses against focus group participants’ views. Limitations of the Evaluation Attribution of impacts solely to the matrix is difficult because of the overlapping effects of other elements of the renewal strategy and subsequent developments (i.e. decentralization of country management and sector staff, 132 The Matrix System at Work creation and then elimination of the independent quality review function served by QAG, integration of budget and human resource management processes, and the merger of networks in 2006, among others). While focusing primarily on matrix issues, the evaluation has acknowledged and dealt with these collinear effects to the extent that they have impacted the effectiveness of the matrix system. Appendix B      133 Appendix C Responsiveness and Effectiveness of the Bank’s Country Programs Are Bank country programs responsive and effective in meeting country needs? For the last eight years, the Bank has been issuing Country Assistance Strategy Completion Reports (CASCRs)—self-assessments of country program achievements at the completion of the Country Assistance Strategy (CAS) period—and the Independent Evaluation Group (IEG) has conducted CASCR reviews to validate their accuracy and draw additional lessons from the experience of CAS implementation. These reviews show that outcomes, measured as achievement of objectives by CAS pillars, were satisfactory less than two- thirds of the time. As the number of completed CASCR reviews now approaches 100, it is possible to take a closer look at specific factors that underlie the aggregate results. IEG reviewed 96 CASCR reviews—all those available as of December 2010 and which reviewed CASs prepared after the introduction of the matrix system— covering country programs in 80 countries. IEG developed an aggregate measure of the two goals of the 1997 reform decomposing the quality of country programs into eight component variables—four that reflect responsiveness and four representing effectiveness. Responsiveness encompasses the most basic measure—alignment with a country’s national priorities—and includes also a measure of ownership of the program by the government and other stakeholders including civil society, realism of the objectives and planned instruments, and flexibility to adapt to changing country circumstances. Effectiveness is measured by the quality of the results framework of the CAS, the quality of portfolio implementation of the program, outcomes of projects and operations, and outcomes of analytical and advisory activities (AAA). Each component was rated separately on a four-point scale —Unsatisfactory, Moderately Unsatisfactory, Moderately Satisfactory, or Satisfactory—using only evidence presented in IEG’s CASCR reviews. An aggregate rating was arrived at for each country program, applying equal weight to each of the eight components. Ratings for each dimension and in aggregate were compared across Regions, income categories, and two time periods (CAS periods beginning between FY98 and FY03 and those beginning between FY04 and FY08). How does responsiveness and effectiveness vary by Region? What drives Bank performance in each of the eight components? What are the primary deficiencies in areas of poor performance? This appendix explores the underlying factors driving the ratings for alignment with national priorities, ownership, realism, and flexibility on the responsiveness side, and results frameworks, quality of portfolio implementation, outcomes of projects and operations, and outcomes of AAA on the effectiveness side. Responsiveness and Effectiveness by Region Among the Bank’s regions, Africa has the lowest share of country programs rated above-the-line overall—less than half—while programs in Latin America and the Caribbean were rated best, with 80 percent rated moderately satisfactory or better (Table C.1). Two-thirds to three-fourths of country programs in 136 The Matrix System at Work the other four Regions were moderately satisfactory or better. The aggregate ratings for “responsiveness and effectiveness� are broadly similar to IEG’s ratings for country program outcomes in CASCR reviews. Overall ratings were similar for programs in both International Development Association (IDA)- and International Bank for Reconstruction and Development (IBRD)-eligible countries. The low ratings for realism and ownership and high ratings for flexibility are striking among nearly all Regions, but in particular in Middle East and North Africa. There were three programs in which ratings for ownership or realism were unsatisfactory but flexibility was considered satisfactory. In Djibouti (FY06-08) and Yemen (FY06-09), this was in part due to rapid responses to emerging crises, such as elevated food prices and avian flu, while the underlying program continued to lack ownership and realism. In Egypt (FY02-04), the Bank changed its program to better align with changing government priorities. Alignment with National Priorities The most basic assessment of responsiveness is the degree to which the program’s objectives are appropriate for the country’s development and consistent with the country’s own development strategy. This measure does not rise to the level of “relevance,� which would also include whether the program is likely to produce an impact that is significant for the partner country’s development. Percentage of Country Programs Rated Moderately Satisfactory Table C.1 or Higher by Region Results Imple- Project AAA Ag- CAS Align- Owner- Real- Flex- frame- men- out- out- gregate outcome ment ship ism ibility work tation come come rating rating AFR (n=29) 97 38 34 50 14 48 55 55 45 52 EAP (n=7) 100 43 29 57 29 71 83 100 71 67 ECA (n=25) 92 40 40 77 48 64 64 88 76 71 LCR (n=20) 100 35 40 74 20 45 85 85 80 75 MNA (n=9) 100 22 0 71 44 56 89 67 67 44 SAR (n=6) 100 33 17 83 17 50 83 50 67 67 Over all (n=96) 97 36 32 67 28 54 71 74 66 63 Source: IEG reviews of CASCRs as of December 1, 2010. Note: For flexibility in Africa, n=24; Europe and Central Asia, n=22; Latin America and the Caribbean, n=19; the Middle East and North Africa, n=7; for project outcome in East Asia and the Pacific, n=6. The entire universe of evaluations was used in this analysis, but nonetheless there were few observations in several Regions. While the results remain valid, they may fluctuate substantially as new observations are added. CAS outcome rating is the rating of the outcomes of country programs as assessed in CAS Completion Report Reviews. Appendix C 137 In many cases, the definition of national priorities (as well as CAS objectives) is broad and so a strong degree of overlap is to be expected. Indeed, 93 of the 96 country programs were deemed to be moderately satisfactory or better in aligning program objectives with national priorities. In all of the three cases where country programs fell short—Uzbekistan (FY02-05), Armenia (FY02-04), and Sao Tome and Principe (FY01-05)—programs were designed in the early phase of the matrix era. In Uzbekistan, the program directly pursued goals that were different from those of the government. In the other two, the programs were deemed not to be appropriate for the partner’s development. In Armenia, there was limited attention to the impact of growth on jobs and inequality, an omission consistent with later concern by the Bank’s Board regarding the impact of the program on poverty. And in Sao Tome and Principe, the Bank’s strategy did not align itself adequately to the country’s development situation. Ownership One-third of country programs—35 of the 96 that were rated—had moderately satisfactory or satisfactory country ownership, defined as including support from inside and outside of government, as well as at all levels of government where appropriate, as shown in Figure C.1. There was a large improvement in ownership over the period, with 55 percent of programs initiated between FY04 and FY08 rated moderately satisfactory or better, compared to 18 percent initiated between FY98 and FY03. Figure C.1 Most Programs Had Weak Government Ownership from the Outset 50 45 Satisfactory Unsatisfactory 40 ownership (34%) ownership (66%) Number of Programs 35 30 25 20 15 10 5 0 Broad Sustained Minor Govt support Insuf�cient Weak govt consultation govt differences weakened consultation support from commitment outset Category of Program 138 The Matrix System at Work In roughly half of the satisfactory programs, there was broad consultation that underpinned the design of the CAS, and government commitment was well- established. For example, in Armenia (FY05-08), there was deep consensus building in the design and implementation of the program, and collaboration with private and public bodies in project design, implementation, and monitoring and evaluation. In Maldives (FY00-07), CAS preparation engaged civil society, the private sector, nongovernmental organizations, and donors; projects showed strong political leadership and ownership. And in Morocco (FY06-08), there was a wide-ranging consultation process with key internal and external stakeholders, and during implementation the Bank maintained a continuous dialogue with the government and donors. In the other satisfactory programs, there was no evidence of a consultative process presented in CASCR reviews. In practice, however, there was a sustained commitment to the program by government throughout the CAS period. This was the case in Mexico (FY05-08), where the program was very well-customized and driven by emerging demands by the client, and in Madagascar (FY04-08), where the government showed strong backing of reforms to support the private sector and a willingness to adopt new approaches suggested by the Bank. For the other 61 programs, country ownership was rated as less than satisfactory. In an overwhelming majority of cases, government commitment to the Bank’s reform agenda embedded in its program appeared weak from the outset. For example, in Cameroon (FY04-08), political economy factors held the program back. The Bank was seen as imposing its own agenda, and while there were consultations, the Bank could have worked to cultivate ownership. In Bangladesh (FY01-04), several areas of reform did not have government support, particularly the governance reform agenda, but that support may have been attainable with more consistent policy dialogue. In a few cases, government support weakened or government priorities changed during the period. This was the case in Mauritania (FY03-07), where government interest in reform waned rapidly after the country reached the completion point under the Heavily Indebted Poor Countries Initiative. In the remaining CASs, the key problem was that the Bank’s consultation with stakeholders was insufficient to ensure broad-based support. For instance, in Brazil (FY 04- 07) the CAS preparation was highly participatory, but some components of the program—particularly lending to municipalities, which ultimately led to conflict with the Ministry of Finance—was not sufficiently discussed with all the relevant players. And in El Salvador (FY05-09), the program should have reached beyond the executive branch in order to mitigate political gridlock. Realism Realism was also rated low. Only about a third of country programs—31 of the 96 that were rated—demonstrated moderately satisfactory or satisfactory realism in country program design. There was no change in the realism ratings Appendix C 139 over time, with 32 percent rated moderately satisfactory or better in the second half of the period of review, compared to 33 percent in the first half. Four programs were explicitly singled out as fully realistic in their CASCR reviews. The program in Guatemala (FY05-08) fell into this category. There, the Bank carefully calibrated its program to take into account risky political conditions, using single-tranche programmatic loans that would allow it to adjust the program as needed. Twenty-seven other programs were also moderately satisfactory or better in terms of their realism. Most had only minor shortcomings with regard to their estimation of national capacity, estimation of government commitment, identification of appropriate lending targets and outcome objectives, and selectivity and prioritization. For example, in Rwanda (FY03-08) the lending and non-lending program were considered realistic, with slightly ambitious overall objectives. There were some minor project delays due to gaps in government capacity. The satisfactory program categories are summarized in Figure C.2. In the two-thirds of programs that were not realistically designed, the major culprits were poor assessments of national capacity constraints, absorptive capacity and robustness of government commitment, the desire for lending , and/or Bank capacity constraints (as manifested in a lack of selectivity). Indeed, many programs made misjudgments along several of these dimensions, with over half considered unrealistic along at least two, as shown in Figure C.3. An overestimation of government capacity was the most common shortcoming, occurring in half of the unsatisfactory programs, and a third of programs overall. For example, in Yemen (FY03-05) the program included complex investments and AAA for which the country did not have sufficient implementation capacity, and in Ukraine (FY01-03 and FY04-07), a large number of lending operations were planned despite past experience with long-standing implementation constraints and limits to institutional capacity. The next most common reason for low realism was over-ambitiousness in lending targets. This was evident in Romania (FY06-09) where lending targets were high despite the government’s moderate interest in remaining engaged with the Bank, and the Russian Federation (FY03-06) where the Bank proposed a large lending program in the face of slow disbursement of existing project loans and a general reluctance to borrow. An over-estimation of government commitment was evident in about a third of unsatisfactory country programs. In Mauritius (FY02-06), the strategy was bold, but similar to the previous strategy, which was unable to influence policy reform. The government declined to take on two of the three public expenditure reform loans, an action that may have been predicted given the knowledge that Mauritius often cancels requests for donor financing. In Ethiopia (FY03-07), the Bank overestimated the government’s commitment to private sector development, and so made little progress on key issues in that area. 140 The Matrix System at Work Figure C.2 Most Programs Were Not Realistically Designed 80 Realistic (32%) Unrealistic (68%) 70 Number of Programs 60 50 40 30 20 10 0 Well-matched De facto realistic Minor Unrealistic to capacity shortcomings Category of Program Figure C.3 An Overestimation of Capacity Was the Most Frequent Shortcoming of Unrealistic Programs 35 30 Number of Programs 25 20 15 10 5 0 Overestimated Overambitious Overestimated Lacking in At least two of country lending govt selectivity these capacity targets commitment Characteristics of 65 Unrealistic Programs Appendix C 141 Lack of selectivity also limited a program’s realism. For example, the program in Ukraine (FY04-07) was set up to support all seven of the country’s development pillars, and in Cambodia (FY00-03), the program was too broad, placing heavy demands on limited field-based staff and spreading resources too thin. Figure C.3 summarizes the frequencies of each of these characteristics among the 65 country programs with low realism ratings. Flexibility In 57 programs—roughly two-thirds of those where a judgment could be made— flexibility in responding to changes in government demand or changes in country conditions was moderately satisfactory or better. Ratings for flexibility improved over the period, with 73 percent rated moderately satisfactory in the second half of the period, compared to 60 percent in the first half. In Morocco (FY06-08), the Bank responded rapidly to changes in government commitment, priorities, or capacity, scaling up lending and AAA accordingly. In El Salvador (FY05-09), the Bank used a CAS Progress Report to recast its program in response to political gridlock in the first half of the period, scaling back the lending program and associated objectives. The program was still able to make an impact through AAA, project preparation, and the use of alternative funding sources. Other programs responded well to changes in country conditions, including in Sri Lanka (FY03-07) where the Poverty Reduction Strategy Paper (PRSP) was discarded when the new government pursued military operations and the Figure C.4 Most Programs Responded Flexibly to Changing Conditions 25 Satisfactory flexibility (66%) Unsatisfactory flexibility (34%) 20 Number of Programs 15 10 5 0 Responded to Responded to Responded to No program Response was changes in changes in both revision too small or govt demand, country although one too slow commitment conditions was needed Category of Program 142 The Matrix System at Work country was besieged by the 2004 tsunami. A CAS Progress Report narrowed the scope of objectives in the existing pillars and added a new pillar focused on tsunami reconstruction. The 29 programs that were rated moderately unsatisfactory or worse for flexibility fall into three categories. First, in about half of unsatisfactory programs an important change in conditions occurred, but the program was not adjusted in any significant way. This happened in Argentina (FY06-09) where the CAS anticipated a Progress Report at the program’s midpoint, but none was prepared even in the face of the global financial crisis. In addition to major strategic adjustments to the lending program, the Progress Report would have provided an opportunity to rethink the size and composition of the AAA program. Similarly in Benin (FY04-06), the Bank did not revise its program in response to deteriorating external events. Moreover, while risks to the program were well identified in advance, the mitigation measures focused on long-term policies such as diversifying the economy and strengthening incentives. In the event, an infusion of quick-disbursing assistance would have been necessary to appropriately cope with multiple shocks. In other programs where flexibility was rated moderately unsatisfactory or worse, a response to changing conditions came either with significant delay, or was too minor to have a meaningful impact. For example in The Gambia (FY03- 07), it emerged quickly that the government lacked the necessary commitment to implement the program. In response, the Bank dropped most of the high case projects from the pipeline, but replaced them with several others. It also introduced portfolio monitoring improvements, but these actions were far from timely. And in Romania (FY06-09), the Bank was slow to recognize and act upon the need to overhaul its strategy when the government felt little need for a Bank program after European Union accession, despite a plan at the outset to carry out a full review in that scenario. Quality of Results Frameworks The quality of results frameworks is the first component related to program effectiveness. Just over one-quarter (27 of 96) of country programs were deemed to be moderately satisfactory or better along this dimension. Ratings for results frameworks improved over the period, from 20 percent moderately satisfactory or better in the first half to 36 percent in the second half, but the share rated above-the-line in the second half of the period remains very small. The most common deficiency was that the program did not link well Bank instruments and activities to intended outcomes, which was the case in 40 percent of the programs rated moderately unsatisfactory or worse. For instance, in Mauritania (FY03-07), objectives were defined in terms of countrywide goals from the PRSP without being translated into achievable, program-level outcomes. And in India, (FY05-08), there was inadequate linkage between the strategy, program design, and desired outcomes; outcomes were largely built Appendix C 143 up from the existing portfolio designed under a previous strategy rather than built down from the new strategic design. Another common shortcoming was the lack of an outcome orientation, evident in about 30 percent of programs with results frameworks rated moderately unsatisfactory or worse, with indicators focused on inputs, processes, and outputs rather than outcomes. Programs have improved on this indicator over time—three-quarters of the programs with this shortcoming were designed before the rollout of results-based CASs. Finally, in just over a quarter of cases, the programs’ core objectives were not well-enough defined to be reasonably tracked during implementation. Quality of Portfolio Implementation Over half (54 percent) of country programs were rated moderately satisfactory or satisfactory with regard to the quality of the implementation of their lending and non-lending portfolios. There was a slight decline over the period, with 57 percent rated moderately satisfactory or better in the first half, compared to 51 percent in the second half. Of those rated moderately unsatisfactory or worse, there were three common barriers to high-quality implementation. First, in about half of these cases, there were significant delays in implementation. These were frequently due to challenges in procurement and issues related to the capacity of the implementing agency. There was deterioration in portfolio quality in about half of these cases as well, as measured by the share of projects and commitments at risk. Finally, slow disbursement plagued about one-third of the moderately satisfactory or worse programs. Outcomes of Projects and Operations The rating for project outcomes was based on the average rating of independent assessments of projects that closed during the CAS cycle, as reported in the CASCR review. This rating therefore is focused on final outcomes of Bank operations rather than implementation progress; however, it is backward looking at the portfolio of operations that have closed during the period. By contrast, the measure of portfolio implementation was focused on progress in implementing the concurrent lending and non-lending portfolio. The rating for both project outcomes was high—71 percent of country programs were rated moderately satisfactory or better in this regard. This high rating was stable over the period. Outcomes of Analytical and Advisory Activities (AAA) When available, CASCR Reviews use ratings of AAA from assessments of the now defunct Quality Assurance Group in their discussion of knowledge products, 144 The Matrix System at Work but they do not otherwise explicitly rate the outcomes of the AAA program. Nonetheless, even when no rating is specified, all CASCR Reviews make some judgment regarding the usefulness or effectiveness of individual pieces of AAA or of the AAA program in its entirety. This exercise did not review any AAA directly; instead, it considers statements in CASCR Reviews that AAA affected the design of Bank operations or impacted the government’s policies as indications that the Bank’s AAA work had positive outcomes. The rating for outcomes of AAA in the Bank’s country programs was moderately satisfactory or better in 74 percent of cases. In contrast to the ratings for outcomes of projects and operations which were stable over the period, AAA outcomes showed significant improvement rising from 64 percent of country programs moderately satisfactory or better in the first half of the period to 81 percent in the second half. Appendix C 145 Appendix D Experience in the Trenches: Results from the Staff Survey The Independent Evaluation Group (IEG) conducted a staff survey on the effectiveness the Bank’s matrix system. The survey was fielded from February to March, 2011 and was sent to 1,054 Bank staff, of whom 435 responded, a rate of 41 percent. Results are statistically significant with 95 percent confidence at +/– 5 percent. The survey asked about the impact of the matrix system on (i) strategic alignment among country, sector, and corporate strategies; (ii) knowledge production and dissemination; and (iii) incentives and accountability for quality. The key messages and results from the survey are highlighted below. There are two types of survey questions: some are grounded in experience; others are opinions. For many questions, respondents were asked to provide hard evidence regarding very specific items that the matrix system was intended to affect based on their own experiences in the last five years. For example, they were asked to what extent they had used economic and sector work produced by sector units outside their own sector. Answers to these questions are akin to those gathered in a household survey. For other questions, respondents were asked to share their opinion of how the Bank functions. For example, they were asked whether the benefits of some aspects of the matrix system outweigh the costs. While still based on staff’s own recent experiences, answers to these questions are akin to those gathered in an opinion survey. The analysis below makes an implicit distinction between the two, referring to responses to the earlier type of question with more certainty than those to the latter type of question. The survey was lightly tailored to three different self-identified staff groupings: those who primarily worked in a country management unit, a sector management unit, and a network anchor unit over the past five years. Figures reported below refer to questions asked of all these groups (or extremely close variants) unless otherwise noted. Further, “substantial� refers to a respondent agreeing that the statement is true either to a “substantial extent� or to a “very large extent�, which is here considered a positive, or “above-the-line,� response (other options for most questions were “some extent� and “slightly or not at all�). Strategic Alignment Strategy: Sector strategies appear to reflect experiences in the regions to a limited degree, but few staff perceive operations and Country Assistance Strategies (CASs) as drawing on sector strategies or reflecting regional or Global Public Good (GPG) issues. Sector strategies are seen to reflect relevant country experiences to a limited degree. Roughly half (46 percent)1 of sector and anchor staff reported this to be true to a substantial or very large extent in their sector, with the sentiment stronger among field-based staff. Fifty-six percent of staff mapped to the Human Development Network (HDN) sectors held this view to a substantial or greater extent at the high end, compared to 29 percent of 148 The Matrix System at Work staff mapped to sectors in the former Environmentally and Socially Sustainable Development Network (ESSD).2 Country strategies and country operations are not reported to draw on sector strategies substantially. Less than a third of respondents (30 percent) indicated that the country operations on which they worked reflected the relevant sector strategy to a substantial or very large extent and fewer respondents (27 percent) said that country strategies do so. Operations and CASs, respectively, were seen to draw on sector strategies the most in the Africa Region, where 44 percent and 35 percent of staff indicated this was the case. The relationship was weakest in Regions with more middle-income countries: in the Europe and Central Asia Region, 25 percent and 32 percent of staff indicated that operations and country strategies, respectively, drew on relevant sector strategies; in the Latin American and Caribbean Region, the figures were 27 percent and 26 percent. Regional and GPG issues were not reported to be substantially reflected in country strategies. Just over a third of respondents (36 percent) indicated that the country strategies with which they are familiar reflect GPG issues, with wide variation among Regions (54 percent of East Asia and the Pacific Region staff vs. 23 percent of South Asia Region staff). And about a quarter of respondents (27 percent) indicated that the country strategies with which they are familiar reflect regional issues, including only 12 percent each among staff in the Middle East and North Africa Region and the South Asia Region. Knowledge Production and Dissemination Knowledge: Staff are roughly equally divided over whether the Bank generates relevant knowledge to a substantial extent. A minority view dissemination efforts as effective, particularly for knowledge generated outside the Bank. The sectors generate relevant knowledge for about half of staff.3 A little more than half of respondents indicated that the sectors have been effective to a substantial or higher extent in producing relevant knowledge (53 percent). Staff mapped to HDN sectors hold the most positive views (64 percent) among network groups in this regard, and those mapped to the Finance and Private Sector Development Network (FPD) were least positive (41  percent). Staff in Europe and Central Asia (61 percent) were most positive among regions and those in the Middle East and North Africa (34 percent) were most negative. Sectors fail to adequately disseminate internal and global knowledge, particularly from the view of staff in country units. Forty percent of staff reported that sectors have been substantially effective in dissemination of internally generated knowledge and experiences across the Bank, and a quarter of staff indicated that the sectors successfully disseminated relevant global knowledge. On both Bank-generated and external knowledge, staff in Africa were most positive (48 percent, 33 percent). On Bank-generated knowledge, staff in East Asia and the Pacific Appendix D 149 were by far the most negative (18 percent). There was limited variation among staff mapped to different Sector Boards, but staff mapped to the Poverty Reduction and Economic Management Network (PREM) sectors were least positive regarding dissemination of Bank-generated knowledge (32 percent), and HDN-mapped staff were least positive regarding dissemination of external knowledge (20percent). Only half of country management unit (CMU) staff indicated that the sectors are substantially effective in using economic and sector work (ESW) in country operations. Knowledge: The flow of knowledge and expertise across Regions and sectors is seen as extremely limited, with staff from East Asia and the Pacific and the Middle East and North Africa utilizing the least knowledge and expertise from anchors, other Regions, and other sectors within the Region. Sector unit staff use very little explicit knowledge produced by sector units in different Regions or the network anchor. About one-quarter (28 percent) of sector unit staff reported using ESW produced by the anchor in operations to a substantial or greater extent; and less than a fifth (19 percent) reported using ESW produced by same-sector units in other Regions. The flow of technical support has been equally weak. Twenty-nine percent of sector unit staff have received substantial sector-specific technical support from their anchor, and only 11 percent have received such support from same- sector units in other Regions. The flow of knowledge across sectors at the operational level was seen as even rarer. Seventeen percent of sector unit staff reported using ESW produced by other units in their Region to a substantial extent, and an equal share reported receiving substantial sector-specific technical support from the same source. East Asia and the Pacific and the Middle East and North Africa were consistently the two Regions who least utilized analytical work or technical support from the anchors, other Regions, and other sectors within the Region. For example, 8 percent of East Asia and the Pacific staff and 11 percent of Middle East and North Africa staff considered their use of ESW in their sector produced by other Regions to be substantial; the figures are 8 percent and 4 percent for technical support from the same sources. Regional sector staff mapped to PREM were most likely to report using analytical work produced by other units, while those mapped to FPD were most likely to utilize technical support. But even these numbers were low. Twenty-four percent of regional sector staff mapped to PREM substantially used ESW produced by PREM units in other Regions; 29 percent substantially used ESW produced by other sectors within their Region, and 11 percent substantially used ESW produced by the Development Economics Vice Presidential Unit (DEC). For regional sector staff mapped to FPD, 150 The Matrix System at Work 45 percent received substantial technical support from the FPD anchor, 21 percent from FPD units in other Regions, and 30 percent from other sectors in their own Region. Anchor units were not successful in generating operationally useful knowledge products and disseminating knowledge across the Bank to a very large or substantial extent, according to anchor and sector staff. Around a third of sector and anchor unit respondents indicated that anchor units in their sector generated useful knowledge products (37 percent), facilitated access to global knowledge (32 percent), and disseminated knowledge across the Bank (31 percent) to a substantial or very large extent. Fewer sector and anchor unit staff (11 percent) indicated that the anchor units added substantial value in developing new lending products. And as already reported, about a quarter (28 percent) of sector staff used ESW produced by their own anchor in operations to a substantial or greater extent. DEC was the least utilized source of knowledge—both knowledge products and technical support—in operational work. Seven percent of sector-based staff used ESW produced by DEC to a substantial extent and 6 percent received substantial technical support from DEC in operational work. Quality, Incentives, and Accountability Quality: The quality assurance system for lending under the matrix ensures higher quality of fiduciary and safeguards aspects than of technical design. The Bank pays substantial attention to fiduciary duties and safeguards aspects of projects. A clear majority of staff indicate that the quality assurance system for lending substantially ensures a high quality of fiduciary aspects (59 percent) and safeguard aspects (54 percent). There was little variation among Regions on this question, but fewer than half of staff mapped to HDN and ESSD sectors reported the quality of both these aspects as being substantially ensured. But the quality assurance system pays less attention to the technical design and other quality aspects of operations. In the experience of about a third of staff did the quality assurance system substantially ensure effective use of peer reviews (34 percent) and quality enhancement reviews (QERs) (33 percent), and for 36 percent did it substantially ensure high quality of technical design. The staff responses also indicated a lack of effective mechanism to provide independent feedback on quality and effective supervision for timely feedback. Nineteen percent of respondents indicated that the current system provides a substantial mechanism for independent feedback on quality and in the experience of 20 percent there was substantially effective supervision for timely and frank feedback on projects. There was little variation among Regions along these dimensions, although staff in Africa (43 percent, 41 percent) and East Asia and the Pacific (41 percent, 38 percent) report being less dissatisfied with the peer review and QER processes than in other Regions, and staff in Appendix D 151 Table D.1 Percentage of Anchor and Sector Staff Indicating They Received Substantial Encouragement to Carry Out the Following Behaviors Adapt Mobilize knowledge Give prior- Collaborate Bank-wide to country ity to lending Focus on Use M&E to Encouragement to: across sectors expertise needs targets results improve quality From sector 40 41 54 61 53 36 management From country 35 25 52 67 41 27 management Country staff view of 16 18 25 48 32 14 sector staff incentives South Asia report being least satisfied with the extent to which the system ensures high quality of technical design (26 percent). Anchor units were not effective in improving the quality of regional operations and providing direct operational support to Regions. Thirteen percent of staff (including anchor staff) indicated that anchor units substantially enhanced the quality of regional operations, and for 22  percent of staff, anchors were reported to provide direct operational support to Regions to a substantial or very large extent. Incentives: Sector- and anchor unit staff received substantial encouragement to meet country-specific needs, but less encouragement to work across sectors or utilize expertise across organizational boundaries. Sector and anchor unit staff reported receiving substantial encouragement from sector and country management to meet lending targets and adapt knowledge to country needs, but much less so to collaborate across sectors and mobilize Bank- wide technical expertise. Over three-fifths of anchor- and sector-based staff received substantial or greater encouragement from managers to give priority to meeting lending targets, and a majority received such encouragement to adapt knowledge to country needs. But only two-fifths received encouragement from sector management to collaborate across sectors or to mobilize Bank-wide expertise (and between a quarter and a third received this encouragement from country management). Sector and anchor unit staff received more encouragement from sector management than country management to focus on results and outcomes. Roughly 10 percent more sector and anchor unit staff received substantial encouragement from sector management to focus on results and to use monitoring and evaluation (M&E) to improve quality than received such support from country management. 152 The Matrix System at Work Table D.2 Percentage of Country Unit Staff Indicating Incentives to Adopt the Following Behaviors are Substantial Respond with flex- Adapt Use Incentives Promote Ensure ibility to Collabo- Mobilize knowl- Give M&E to enable staff ownership realism of changing rate Bank- edge to ­priority Focus im- in country of country country circum- across wide country to ­lending on prove units to: program programs stances sectors expertise needs targets results quality Country unit 48 40 36 33 25 32 51 30 64 staff Table D.3 Percentage of Sector/Anchor Unit Staff Indicating That Incentives Influence Various Behaviors to a Substantial Extent Behaviors Give Adapt priority to Collabo- Mobilize ­knowledge meeting Focus Use M&E rate across Bank-wide to country lending on to improve sectors expertise needs targets results quality Top behavior Messages from senior 25 19 19 47 34 22 Lending management targets Control over task budgets 45 40 29 47 37 27 Lending targets Access to earmarked TFs 34 33 33 28 26 20 Cross-sector Incentives collaboration Desire for OPE, promotion, 28 20 20 44 39 24 Lending salary targets Cross sector career 32 21 19 32 31 22 Lending/ advancement cross-sector Recognition by peers 42 43 46 38 54 37 Focus on results Job satisfaction 71 60 62 45 70 51 Cross-sector collaboration Top incentive Job Job satis- Job satis- Sr. mgmt/ Job Job satis- satisfaction faction faction budget satis- faction control faction Appendix D 153 Country unit staff believe sector unit staff have even less incentive to engage in matrix-related behaviors. Sixteen percent of country unit staff believes sector staff is substantially incentivized to collaborate across sectors and 18 percent believe sector staff is substantially incentivized to mobilize Bank-wide expertise. Incentives: A majority of country unit staff indicate they are not substantially incentivized under the matrix to focus on responsiveness in country programs. Incentives under the matrix enable fewer than half of country unit staff, in their view, to substantially promote ownership or ensure realism of country programs, or to respond flexibly to changing circumstances. Views on other matrix-related behaviors are more negative than for sector and anchor unit staff, as shown in Table D.2 below. Incentives: Traditional organizational incentives—messages from senior management, desire for promotion and salary increase, and control over budgets—were the strongest incentives for sector and anchor unit staff to meet lending targets and the weakest incentives for matrix-related behaviors. The top-reported incentives for meeting lending targets were traditional incentives from hierarchical organizations, as shown in Table D.3. Nearly half of sector and anchor unit staff reported messages from senior management and control over task budgets as substantially important incentives influencing their focus on meeting lending targets, the highest of any incentives. These traditional incentives were reported as least influential in affecting matrix-related behavior of staff, which was driven by job satisfaction and recognition by peers. Messages from senior management and desire for promotion were the weakest incentives for collaborating across sectors (with substantial influence for 25 percent and 28 percent, respectively), mobilizing Bank-wide technical expertise (19 percent, 20 percent), and adapting knowledge to country needs (19 percent, 20 percent). Non-organizational factors were far more influential in this regard: job satisfaction was substantially influential for 71 percent of sector and anchor staff in driving collaboration across sectors, for 60 percent in driving mobilization of Bank-wide technical expertise, and for 62 percent in driving adaptation of knowledge to country needs. The next most common incentive for these behaviors was peer recognition. Accountability: Only the Task Team Leader (TTL) is seen by a majority to be substantially accountable for the results of lending operations; No individual is seen as substantially accountable for ESW. Staff indicated a greater degree of accountability of TTLs for results of lending operations than of sector managers and country directors/manager: TTLs were accountable to a substantial or very large extent according to 52 percent of staff, while 35 percent and 34 percent of staff attributed substantial accountability to sector managers and country directors/managers, respectively. Staff graded at the H-level were least likely to attribute accountability to managers and directors. 154 The Matrix System at Work Table D.4 Most Staff Believe the Matrix is Working in a Manner That is Moderately Effective or Better All things considered, do you believe that the matrix is operating in a manner that is . . . Share of staff Highly effective 1% Effective 19% Moderately effective 37% Moderately ineffective 19% Ineffective 15% Highly ineffective 9% Few staff think anyone is substantially accountable for ESW. Thirty-six percent of staff attributed substantial accountability to TTLs for dissemination of ESW products, compared to 22 percent to sector managers and 17 percent to country directors/managers for impacts of ESW on projects or policy. Accountability for results of lending operations are strongest in East Asia and the Pacific, where more than two-thirds (69 percent) consider the TTL substantially accountable for results of lending and half consider sector managers and country directors/managers accountable. Accountability is weakest in Latin America and the Caribbean, where the same figures are 49 percent, 20 percent, and 25 percent respectively (only staff from the Middle East and North Africa judge TTLs to be less accountable for lending with 35 percent considering them substantially so). Similarly, staff in Latin America and the Caribbean rarely view anyone substantially accountable for dissemination or impact of ESW—the analogous figures for TTLs, sector managers, and country directors/managers are 28 percent, 10 percent, and 10 percent. Staff based in Washington are less likely to view sector managers or country directors/managers as accountable for results. About a quarter (27 percent) of headquarter-based staff consider sector managers substantially accountable for results of lending operations, compared to 47 percent of field-based staff. The figures for country director accountability are 28 percent and 44 percent. And only 10 percent of headquarter-based staff view the country director as substantially accountable for the impact of ESW on policy or projects, compared to 27 percent of field-based staff. Overall Assessment Overall: A majority considered the matrix moderately effective or better overall, with staff based in the field, in Africa, and mapped to HDN sectors the most positive. The matrix system was considered to have worked moderately effectively or better by a majority of staff (57 percent). But still, as shown in Table D.4, Appendix D 155 nearly a quarter of staff consider the matrix system to be ineffective or highly ineffective. Staff in Africa and mapped to HDN sectors were most positive about the effectiveness of the matrix system, while those mapped to FPD were least positive. Nearly three-quarters of staff mapped to HDN sectors (73 percent) judged the matrix system as working in a manner that is moderately effective of better, compared to only 38 percent of those mapped to FPD. Nearly two- thirds (65 percent) of staff in Africa held the same opinion. Staff based in the field viewed the effectiveness of the matrix system far more positively than those based in headquarters. Nearly three-quarters (73 percent) of field-based staff indicated the matrix was working in a moderately effective or better manner, compared to less than half of those based in headquarters (47 percent). Indeed, 37 percent of field-based staff deemed the matrix system fully effective or highly effective, higher than any other subcategory with a sample size over 10. Overall: A clear majority of staff indicate that costs exceed benefits under the matrix system with regard to dual reporting, dual accountability, and budgetary arrangements. Costs of budgetary arrangements are seen to outweigh benefits by more than three-quarters of respondents, while the same figures for dual accountability for quality and dual reporting to SMUs and CMUs is 61 percent and 58 percent, respectively. For budgetary arrangements, the most common response— indicated by one-third of respondents—was that costs significantly outweigh benefits, the worst possible assessment. Least pleased with budgetary arrangements were staff based in anchor units (for 80 percent costs outweighed Table D.5 Effect of the Matrix on the Bank’s Work According to Staff Who Joined the Bank Prior to 1997 Improved Worsened Responsiveness to country needs 52% 20% Borrower ownership 45% 11% Implementing agency ownership 14% 11% Civil society ownership 36% 7% Quality of implementation 24% 32% Quality of supervision 19% 39% Quality of ESW 24% 29% Impact of ESW 18% 31% Quality of outcomes of Bank support 26% 29% Links between GPGs and country programs 31% 29% 156 The Matrix System at Work benefits, compared to 68 percent in country units), those in East Asia and the Pacific (87 percent), and those mapped to FPD (93 percent) and former ESSD sectors (89 percent). In no subgroup did a majority believe the benefits outweighed the costs of budgetary arrangements. Overall: Staff who joined the Bank prior to 1997—a subsample of about 85 respondents—considered the matrix to have improved responsiveness to country needs, but worsened overall organizational effectiveness, as well as the quality of supervision and the impact of ESW. The matrix system was viewed as helping to improve Bank responsiveness to country needs and borrower ownership by staff that began employment at the Bank prior to the introduction of the matrix. In 52 percent of responses, staff indicated that the matrix system improved Bank’s responsiveness to country needs compared with the organization of the Bank before 1997, and 45 percent of staff indicated that the current system enhanced borrower ownership of country programs. On average, these same long-time staff indicated that the matrix either did not affect or worsened the quality of supervision, the quality of outcomes, and the quality and impact of ESW. Around 43 percent of long-time staff indicated that the quality of supervision did not change and 32 percent think it worsened. The quality and impact of ESW did not change according to 47 percent and 51 percent of long-time staff, respectively, while 29 percent think that the quality worsened and 31 percent think the impact worsened. And they reported that the matrix system worsened the Bank’s overall organizational effectiveness. Around 49 percent of long-time staff think that the current system worsened Bank’s overall organizational effectiveness. Appendix D 157 STAFF SURVEY RESULTS EVALUATION STRATEGY Based on your experience in the sector(s) where you worked for the most time in the past 5 years, to what Q1 extent did… Slightly Sub- Very Don’t or not at Some stantial large Above know all extent extent extent N the line a The sector strategy reflects relevant country 6% 12% 39% 36% 8% 353 46% experience? [excludes country unit staff] b The sector strategy reflects relevant regional 7% 10% 35% 42% 7% 352 52% experience? [excludes country unit staff] c Country strategies draw on the sector strat- 6% 27% 42% 22% 3% 404 27% egy? [for country unit staff: The country strategy draw on sector strategies?] d Country operations draw on sector strategy? 7% 29% 37% 25% 3% 405 30% [for country unit staff: Country operations draw on sector strategies?] e Country operations draw on the regional 19% 24% 38% 19% 1% 395 25% sector strategy (if any)? [for country unit staff: country operations draw on regional sector strategies?] f Country strategies reflect global public good 11% 19% 39% 29% 3% 405 36% issues (e.g. environmental commons, commu- nicable disease, etc.)? g Country Strategies reflect regional issues? 8% 21% 45% 22% 3% 406 27% Q2 To what extent do the operations that your currently work on… [excludes country unit staff] Slightly Sub- Very Don’t or not at Some stantial large Above know all extent extent extent N the line a Draw on global experience from outside the 1% 17% 35% 37% 10% 353 48% Bank? b Draw on country experience from other 2% 9% 38% 42% 9% 353 52% countries in the region? c Draw on country experience from other 2% 18% 42% 33% 5% 352 39% regions? d Integrate World Bank programs to address 6% 28% 40% 23% 4% 351 28% global public goods? e Integrate regional issues? 3% 25% 44% 25% 3% 351 28% (Continued) 158 The Matrix System at Work KNOWLEDGE Q3 In the past 5 years, to what extent was your sector family effective in… [for country unit staff: ... to what extent have the sectors been effective in …] Slightly Sub- Very Don’t or not at Some stantial large Above know all extent extent extent N the line a The production of operationally relevant 1% 10% 36% 39% 13% 383 53% knowledge within the Bank? b The dissemination of internally generated 1% 11% 48% 30% 10% 382 40% knowledge and experience across the Bank? c The dissemination within the Bank of opera- 3% 30% 42% 20% 5% 381 25% tionally relevant global knowledge produced outside the Bank? d The use of economic and sector work in 2% 21% 27% 44% 6% 52 51% country operations? [asked of country unit staff only] Q4 In the past 5 years, in the preparation and supervision of your lending and non lending operations to what extent have you used economic and sector work or other analytic products produced by… [excludes country unit and anchor unit staff] Slightly Very Don’t or not at Some Substan- large Above know all extent tial extent extent N the line a The anchor unit from your sector? 2% 33% 38% 20% 8% 262 28% b Sector units from other regions in your 1% 35% 46% 17% 2% 261 19% sector? c Sector units outside your sector? 2% 41% 41% 15% 2% 262 17% d DEC (Development Economics VPU)? 9% 59% 25% 6% 1% 260 7% Q5 In the past 5 years, to what extent have you received sector specific technical support from… [excludes coun- try unit and anchor unit staff] Slightly Very Don’t or not at Some Substan- large Above know all extent tial extent extent N the line a The anchor unit in your sector? 0% 34% 37% 24% 5% 262 29% Other units in your region (e.g. regional b 5% 46% 32% 15% 2% 260 17% hubs)? The sector units in your sector from other c 2% 46% 41% 11% 0% 260 11% regions? DEC (Development Economic Vice d 10% 65% 20% 5% 0% 259 6% Presidency)? OPCS (Operations Policy and Country e 2% 45% 33% 15% 4% 259 20% Services VPU)? (Continued) Appendix D 159 QUALITY, INCENTIVES AND ACCOUNTABILITY Q6 Based on your experience in the last 5 years, to what extent do you feel the current matrix system en- sures adequate attention to results of lending operations and impacts of economic and sector work (ESW) through... Slightly Very Don’t or not at Some Substan- large Above know all extent tial extent extent N the line a Accountability of task team leaders for 6% 21% 24% 35% 14% 342 52% results of lending operations? b Accountability of sector managers for results 6% 29% 33% 27% 6% 342 35% of lending operations? c Accountability of country directors/manag- 7% 28% 34% 26% 6% 341 34% ers for results of lending operations? d Accountability of task team leaders for dis- 5% 27% 33% 30% 4% 341 36% semination of ESW? e Accountability of sector managers for im- 6% 42% 32% 19% 2% 340 22% pacts of ESW on projects or policy? f Accountability of country directors/manag- 8% 42% 34% 13% 2% 340 17% ers for impacts of ESWon projects or policy? Q7 Based on your experience in the last 5 years, to what extent do you feel that the current quality assurance system for lending operations ensures…? Slightly Very Don’t or not at Some Substan- large Above know all extent tial extent extent N the line a High quality of technical design? 5% 23% 39% 30% 4% 342 36% b High quality of fiduciary aspects? 6% 9% 29% 49% 6% 342 59% c High quality of safeguard aspects? 7% 9% 34% 42% 8% 342 54% d Effective use of peer reviews? 3% 21% 43% 30% 2% 342 34% Effective use of quality enhancement 6% 19% 44% 28% 2% 342 33% e reviews? f An effective mechanism to provide indepen- 5% 36% 39% 18% 2% 340 20% dent feedback on quality to task teams? g Effective supervision to ensure frank and 5% 34% 41% 18% 2% 342 21% timely feedback on projects? (Continued) 160 The Matrix System at Work Q8 Based on your experience in the last 5 years, to what extent do you feel the anchor unit in your sector has added value by… [for country unit staff: ... to what extent do you feel the sector anchor units have added value by …] Slightly Very Don’t or not at Some substantial large Above know all extent extent extent N the line a Generating useful knowledge products? 2% 20% 42% 30% 6% 342 37% b Developing new lending products? 8% 60% 22% 7% 3% 341 11% c Disseminating knowledge across the Bank? 2% 16% 51% 25% 7% 342 32% d Facilitating access to global knowledge? 2% 26% 42% 24% 6% 340 31% e Training and mentoring sector staff? 4% 50% 30% 12% 4% 342 16% f Enhancing quality of regional operations? 5% 53% 30% 11% 2% 341 13% Mobilizing trust funds for regional 8% 33% 29% 22% 8% 341 33% g operations? Providing direct operational support to 5% 41% 34% 18% 3% 341 22% h regions? Strengthening monitoring and evaluation of 6% 59% 25% 9% 2% 339 11% i country operations? j Other( please rate here and specify below) 70% 10% 9% 7% 3% 98 34% Q9 Based on your experience in the last 5 years, to what extent did the incentives under the matrix system en- able you to adopt the following behaviors in your operational work? [asked of country unit staff only] Slightly Very Don’t or not at Some Substan- large Above know all extent tial extent extent N the line a Promote ownership of country programs 2% 24% 27% 40% 7% 45 48% b Ensure realism of country programs 2% 30% 30% 34% 5% 44 40% Respond with flexibility to changing country c 0% 21% 43% 32% 5% 44 36% conditions Promote greater cross- sector collaboration, d 0% 29% 38% 29% 5% 42 33% when appropriate e Mobilize Bankwide technical expertise 0% 36% 39% 23% 2% 44 25% f Adapt knowledge to country needs 0% 25% 43% 30% 2% 44 32% g Give priority to meeting lending targets 11% 18% 25% 34% 11% 44 51% h Focus on results and outcomes 0% 18% 52% 27% 2% 44 30% Use M&E to improve quality of country i 0% 5% 32% 48% 16% 44 64% operations Q10 Based on your experience in the last 5 years, to what extent do you feel that the incentives under the matrix system enabled sector staff to adopt the following behaviors in their operational work? [asked of country unit staff only] Slightly Very Don’t or not at Some Substan- large Above know all extent tial extent extent N the line a Collaborate across sectors 2% 42% 40% 11% 4% 45 16% b Mobilize Bank-wide technical expertise 2% 36% 44% 16% 2% 45 18% c Adapt knowledge to country needs 2% 33% 40% 22% 2% 45 25% d Give priority to meeting lending targets 7% 22% 27% 38% 7% 45 48% e Focus on results and outcomes 2% 29% 38% 29% 2% 45 32% Use M&E to improve quality of country f 5% 36% 46% 11% 2% 44 14% operations (Continued) Appendix D 161 Q11 Based on your experience in the last 5 years, do you feel you received encouragement from sector manage- ment to adopt the following behaviors in your operational work? [excludes country unit staff] Slightly Very Don’t or not at Some substantial large Above know all extent extent extent N the line a Collaborate across sectors 2% 23% 37% 30% 9% 297 40% b Mobilize Bank-wide technical expertise 2% 21% 37% 32% 8% 297 41% c Adapt knowledge to country needs 2% 13% 32% 40% 13% 297 54% d Give priority to meeting lending targets 10% 12% 24% 35% 19% 295 61% e Focus on results and outcomes 1% 14% 33% 39% 13% 296 53% Use M&E to improve quality of country f 4% 24% 37% 26% 9% 294 36% operations Q12 Based on your experience in the last 5 years, do you feel you received encouragement from country manage- ment to adopt the following behaviors in your operational work? [excludes country unit staff] Slightly Very Don’t or not at Some Substan- large Above know all extent tial extent extent N the line a Collaborate across sectors 7% 26% 34% 25% 8% 297 35% b Mobilize Bank-wide technical expertise 7% 31% 38% 19% 5% 297 25% c Adapt knowledge to country needs 6% 16% 30% 36% 13% 297 52% d Give priority to meeting lending targets 12% 13% 17% 35% 25% 296 67% e Focus on results and outcomes 7% 20% 35% 29% 10% 294 41% Use M&E to improve quality of country f 10% 30% 37% 19% 6% 291 27% operations Q13 Based on your experience in the last 5 years, to what degree did the following incentives influence your abil- ity to collaborate across sectors? [excludes country unit staff] Very Nega- Not at Some Substan- large Above tively all extent tial extent extent N the line a Messages from Senior Management 1% 37% 37% 22% 3% 297 25% b Control over task budgets 10% 24% 25% 28% 13% 294 45% c Access to earmarked trust funds (e.g. for 2% 30% 35% 26% 7% 292 34% environment, gender, or governance) d Desire for positive performance assessment 1% 38% 34% 21% 6% 296 28% (OPE), promotion, salary increase e Career advancement opportunities across 1% 40% 28% 24% 8% 294 32% sectors f Recognition for quality work, including 1% 22% 36% 30% 12% 295 42% publications, by peers inside and outside the Bank g Job satisfaction 1% 7% 22% 40% 30% 295 71% (Continued) 162 The Matrix System at Work Q14 Based on your experience in the last 5 years, to what degree did the following incentives influence your abil- ity to use Bank-wide technical expertise? [excludes country unit staff] Very Nega- Not at Some Substan- large Above tively all extent tial extent extent N the line a Messages from Senior Management 1% 44% 35% 18% 1% 294 19% b Control over task budgets 7% 26% 30% 27% 11% 294 40% c Access to earmarked trust funds (e.g. for 2% 30% 36% 24% 9% 295 33% environment, gender, or governance) d Desire for positive performance assessment 1% 44% 35% 16% 4% 295 20% (OPE), promotion, salary increase e Career advancement opportunities across 0% 50% 29% 18% 3% 292 21% sectors f Recognition for quality work, including 0% 23% 34% 33% 10% 293 43% publications, by peers inside and outside the Bank g Job satisfaction 0% 13% 28% 41% 18% 292 60% Q15 Based on your experience in the last 5 years, to what degree did the following incentives influence your abil- ity to adapt Bank-wide and global knowledge to country needs? [excludes country unit staff] Very Nega- Not at Some Substan- large Above tively all extent tial extent extent N the line a Messages from Senior Management 1% 42% 38% 16% 3% 296 19 b Control over task budgets 4% 30% 38% 21% 7% 294 29 c Access to earmarked trust funds (e.g. for 1% 29% 37% 26% 6% 292 33 environment, gender, or governance) d Desire for positive performance assessment 1% 38% 12% 17% 3% 293 20 (OPE), promotion, salary increase e Career advancement opportunities across 1% 45% 35% 16% 3% 290 19 sectors f Recognition for quality work, including 0% 21% 32% 37% 9% 293 46 publications, by peers inside and outside the Bank g Job satisfaction 1% 8% 29% 41% 21% 292 62 Q16 Based on your experience in the last 5 years, to what degree did the following incentives influence your abil- ity to give priority to meeting lending targets? [excludes country unit staff] Very Nega- Not at Some Substan- large Above tively all extent tial extent extent N the line a Messages from Senior Management 2% 23% 29% 31% 15% 295 47% b Control over task budgets 4% 23% 28% 35% 11% 296 47% c Access to earmarked trust funds (e.g. for 1% 43% 29% 20% 7% 294 28% environment, gender, or governance) d Desire for positive performance assessment 1% 28% 28% 32% 12% 294 44% (OPE), promotion, salary increase e Career advancement opportunities across 1% 38% 29% 24% 9% 294 32% sectors f Recognition for quality work, including publi- 1% 31% 31% 30% 8% 295 38% cations, by peers inside and outside the Bank g Job satisfaction 2% 25% 30% 30% 14% 292 45% (Continued) Appendix D 163 Q17 Based on your experience in the last 5 years, to what degree did the following incentives influence your abil- ity to focus on results and outcomes? [excludes country unit staff] Very Nega- Not at Some Substan- large Above tively all extent tial extent extent N the line a Messages from Senior Management 2% 27% 38% 24% 9% 294 34% b Control over task budgets 4% 28% 32% 27% 9% 293 37% c Access to earmarked trust funds (e.g. for 1% 38% 36% 19% 7% 291 26% environment, gender, or governance) d Desire for positive performance assessment 1% 24% 36% 29% 10% 295 39% (OPE), promotion, salary increase e Career advancement opportunities across 0% 39% 30% 23% 8% 290 31% sectors f Recognition for quality work, including 0% 20% 26% 37% 17% 294 54% publications, by peers inside and outside the Bank g Job satisfaction 0% 9% 22% 40% 30% 294 70% Q18 Based on your experience in the last 5 years, to what degree did the following incentives influence your abil- ­ ountry operations through M&E ? [excludes country unit staff] ity to improve quality of c Very Nega- Not at Some Substan- large Above tively all extent tial extent extent N the line a Messages from Senior Management 2% 37% 40% 16% 5% 296 22% b Control over task budgets 6% 35% 33% 19% 7% 295 27% c Access to earmarked trust funds (e.g. for 1% 43% 36% 14% 6% 295 20% environment, gender, or governance) d Desire for positive performance assessment 1% 40% 36% 18% 6% 292 24% (OPE), promotion, salary increase e Career advancement opportunities across 1% 48% 29% 16% 6% 292 22% sectors f Recognition for quality work, including 1% 31% 32% 28% 8% 291 37% publications, by peers inside and outside the Bank g Job satisfaction 1% 18% 30% 35% 16% 290 51% Q19 Do you see any overlap of responsibilities Yes No N in practice between the mandate of your sector and that of other sectors? [excludes 82% 18% 276 country unit staff] Q20 Do you see any fragmentation of func- Yes No N tions across the responsibilities of your sector and that of other sectors? [excludes 55% 45% 276 country unit staff] (Continued) 164 The Matrix System at Work OVERALL ASSESSMENT Q21 Based on your own experience in the last 5 years, do you feel that the benefits of the current system out- weigh the transaction costs in terms of… Costs Benefits Benefits some- Costs sig- some- what sig- nificantly what out- nificantly outweigh weigh weigh outweigh Don’t Above costs costs benefits benefits know N the line a Dual reporting to sector and country units, 9% 33% 27% 21% 10% 340 46% where applicable? b Dual accountability of country director/ 7% 32% 26% 25% 10% 342 44% manager and sector manager for quality of operations? c Budgetary arrangements? 3% 21% 29% 33% 14% 338 28% d Other transaction costs? (other transactions 1% 6% 17% 30% 46% 229 13% were identified) Q22 All things considered do you Moder- Moder- High- believe that the matrix system is Highly ately ately ly working in a manner that is: inef- Ineffec- ineffec- effec- Effec- Above fective tive tive tive Effective tive N the line 9% 15% 19% 37% 19% 1% 340 57% Q23 Overall, comparing the performance of the current matrix organization with the organization of the Bank before 1997, how has the matrix organization affected… [asked of staff who joined the Bank prior to 1997] Don’t Im- No Worsened N Im- Worsened Know proved change proved a Bank responsiveness to country needs? 7% 48% 26% 19% 85 52% 20% b Borrower ownership of country programs? 10% 40% 40% 10% 82 45% 11% c Ownership by implementing agencies? 12% 12% 66% 10% 82 14% 11% d Ownership by civil society? 12% 32% 50% 6% 82 36% 7% e The quality of implementation? 9% 22% 40% 29% 82 24% 32% f The quality of supervision? 7% 17% 40% 36% 81 19% 39% g The quality of economic and sector work? 7% 22% 44% 27% 82 24% 29% h The impact of economic sector work? 10% 16% 46% 28% 82 18% 31% i The quality of outcomes of Bank support? 10% 24% 41% 26% 81 26% 29% j The links between global public goods and 15% 27% 34% 24% 82 31% 29% country programs? k The Bank’s overall organizational 7% 21% 27% 45% 82 22% 49% effectiveness? Notes 1 A related question asking whether sector strategies reflect relevant regional experiences had a slightly more positive response, with 52 percent agreeing that it did so to a substantial or higher extent. 2 ESSD was merged with the Infrastructure Network in 2006 to create the Sustainable Development Network. 3 Sector and anchor staff were asked to consider the effectiveness of their own sector. Country staff were asked to consider the effectiveness of “the sectors.� Appendix D 165 Appendix E Key Findings from Sector Manager Interviews and Country Director/ Manager Interviews This annex presents the findings from face-to-face semi-structured interviews of 31 Bank sector managers (covering the six administrative Regions and the evaluation’s seven sectors of focus1) and 13 country directors and country managers for the evaluation’s 14 focus countries. Sector managers (six from Africa, five from East Asia and the Pacific, four from Europe and Central Asia, five from Latin America and the Caribbean, four from the Middle East and North Africa and seven from South Asia sector units) and country directors/ managers were interviewed between March 2011 and June 2011. They were asked to rate the performance of the matrix along the three key dimensions of interest of the evaluation: strategic alignment of country, sector, and corporate strategies; knowledge production and dissemination; and incentives and accountability for quality. Managers were also encouraged to provide substantive comments and specific examples illustrating their rating and viewpoint in general. Regional and sector-specific differences are highlighted to the extent possible.2 Strategic Alignment To What Extent Have Sector Strategies Been Informed by Country and Regional Experiences? Bank-wide sector strategies are informed by country and regional experience. An overwhelming majority of sector managers agree that recent sector strategies (and action plans) reflect country and regional experience. Ninety-seven percent (29/30)3indicated that lessons from country and regional experience have some influence on Bank-wide sector strategies. Sixty percent (18/30) of sector managers believe that that influence is substantial (i.e. substantial or very large). The sector strategy papers (SSPs) they nominally referred to (e.g. the 2003 Agriculture and Rural Development Strategy and the Agriculture Action Plan 2010–2012, the 2003 Water Resources Strategy, the 2007 HNP strategy, and the 2008 Transport Strategy) have been drafted with extensive input from regional staff and borrowed heavily from existing and up-to-date regional sector strategies. The 2007 Governance and Anti- Corruption (GAC) strategy is the outlier and was criticized for “lacking conceptual clarity, selectivity, and operational utility� in addition to not being “grounded in country realities.� Older SSPs (e.g. the 2001 Environment Strategy) were designed with limited consultation and are less likely to reflect country experience and regional staff’s various contributions. How Effectively Do Country Strategies and Programs Draw on Sector Strategies and Global Best Practice? Bank-wide sector strategies and priorities have limited influence on Country Assistance Strategies (CASs) and operations. Overall, sector managers were more positive about the uploading of country and regional experience 168 The Matrix System at Work into Bank-wide strategies than they were about SSPs’ influence on country programs and operations. Forty-three percent (13/30) of sector managers and 45 percent (5/11) of country directors/managers rated SSPs’ influence on CASs and operations as substantial. SSPs per se have been described as too “broad� and “internally driven� to have direct bearing on operations. Yet, the sector priorities they champion are considered somewhat relevant to country programs and operations. As managers put it: “While the specific provisions of […] sector strategies do not really translate into country programs and operations, the main strategic directions usually do.� “SSPs identify conceptual directions but countries are very different so we are selective while ensuring consistency with the broader framework.� Integration of Bank-wide sector strategies into country programs is mediated by country directors. Sector managers argued most emphatically that: “Country directors could not give a hoot about Sector Strategy Papers.� “Sector strategies are pieces of paper without influence: the bottom line is that the country director meets with the client, finds out what the client wants and formulates the program—without any reference to sector strategies.� “In practice the Bank has always accommodated the countries, and the strategic shifts have been driven by changes in countries’ ideology.� All in all, Bank sector strategies are less likely to feature prominently in middle- income countries’ CASs. The balance is liable to be tipped strongly to the country (as opposed to the sector) side in middle-income country-dominated Regions such as Europe and Central Asia and Latin America and the Caribbean. None of the sector managers from Europe and Central Asia (0/4) described the influence of Bank-wide sector strategies on CASs and operations as substantial. Country directors and managers from Europe and Central Asia and Latin America and the Caribbean were equally skeptical of the influence of SSPs on CASs. There are also sector-specific divides. Transport sector managers all rated Bank-wide sector strategies’ impact on CASs as substantial. As one put it, “Alignment of the CASs with the sector strategy is partly guaranteed by the large demand by country directors for transport projects (large projects, with good delivery track record). At the same time, however, the composition of the transport sector program in individual countries may lead to tensions […] because clients (and country directors) and sector staff may want different types of transport projects.� Three out of four Social Protection and Agriculture and Rural Development managers believe sector strategies’ influence on operations is less than substantial. Appendix E 169 How Effectively Are Corporate Strategic Priorities and Global Public Goods (GPGs) Issues Reflected in Country Strategies and Programs? Sector managers hardly ever discussed GPGs in isolation from corporate priorities. Most argued that GPGs issues are integrated into country programs either: (i) when a GPG becomes a corporate priority and garners higher visibility and closer management attention, country directors are then keener to follow the GPG’s agenda; or (ii) when the client espouses a GPG issue. Once again Europe and Central Asia sector managers were the most skeptical of the relevance of GPGs. As one maintained, “In Europe and Central Asia, global or regional public goods have been conspicuous for their absence.� Other managers argued that GPGs are not relevant to most countries and that there is, in general, little incentive for country directors to dedicate resources to GPGs. According to a South Asia sector manager, “Most country programs and operations are determined by country demand and countries are understandably interested in their own problems, not global problems. There are cases, however, where the Bank corporate priorities were picked up by the country—e.g., in Brazil, the large metro areas of Sao Paulo and Rio became interested in the climate change agenda. The real connection of Bank corporate priorities to country programs is permissive and facilitative. When a country becomes interested in a particular global or regional issue—or cross-sector theme—the fact that it is a priority for the World Bank means that one can respond much better and with more resources. Another instance is found in disaster risk management. Before it became a corporate interest, response was always post-crisis, but more recently the Bank has been able to work with clients on preparedness—e.g., the Caribbean Disaster Risk facility, and the Cyclone Preparedness project in India.� Only 36 percent (4/11) of country directors and managers rated the influence of GPGs and corporate priorities on country strategies and operations as substantial (or very large). Knowledge The Role of Anchors and Sector Boards Sector managers are typically critical of the knowledge function of their anchors and Sector Boards and the operational support received from them. None of the country directors and managers reported receiving substantial knowledge or operational support from anchors. High on the list of criticisms is the anchors’ failure to provide adequate operational support to the Regions. Both quantity and quality of operational support are reported to be inadequate. Yet, anchors are somewhat effective on corporate priorities and to some extent on selected highly specialized 170 The Matrix System at Work topics (e.g. agribusiness, climate change, resource-based financing, national health accounts, pharmaceuticals). Sector managers were asked to rate the effectiveness of their anchor on a variety of activities on a scale of 1 to 10. The average score for knowledge and learning services to the Regions is 5.0; the score for operational support to the Regions is 4.3. Middle East and North Africa and Europe and Central Asia managers’ average scores for operational support are the lowest, both at 3.3. Anchor performance was rated the highest (7.5/10) on addressing corporate demands on the sector, not a knowledge activity. Sector managers voiced their dissatisfaction with anchor performance rather forcefully. One said, “The anchor does what it wants to do, with a cottage industry mentality and without reference to country needs or even discussion with sector people in the Regions.� Another argued that, “[t]he role of the anchor is mixed to low, as most relevant knowledge is generated by the Regions—linkages are very weak and the anchor is essentially a self-standing entity. The anchor meets its own needs and corporate demands, but not the Regions’ needs. The Sector Boards are inactive in the knowledge area, although they have some involvement in strategic staffing. Some of the knowledge generated by the Regions is uploaded and disseminated, but mainly through—and depending on—the thematic groups.� Social Protection sector managers on the other hand, are rather satisfied with both anchor and Sector Board knowledge performance. Social Protection managers’ scores were higher than average across the board. Their average scores for anchor effectiveness in knowledge and learning services to the Regions and operational support to the Regions are 9.0 and 7.8. As one Social Protection manager put it, “Social Protection is one of the best sectors in global knowledge management. The Sector Board works well and has a business model that puts the Regions first. The anchor too is very client-focused and very supportive of the Regions.� Support from the World Bank Institute (WBI) and the Development ­Economics Vice Presidential Unit (DEC) Support from WBI was generally rated higher than support from DEC, except by Agriculture and Rural Development managers. Collaboration with both was criticized and described as people-based rather than a result of adequate organizational arrangements by both sector managers and country directors. Seventeen percent (5/30) of sector managers and none of the country directors ranked support from WBI to their unit as substantial; 10 percent (3/30) of sector managers and 18 percent (2/11) of country directors said the same of knowledge from DEC. Sector managers from Africa were the most positive about DEC support. One of them indicated that, “[his] sector unit does have some interaction with DEC, both through DEC work on trade growth (in Africa) and on transport costs and the sub-regional corridors. With WBI the Region is planning an urban transport course.� Appendix E 171 A few managers mentioned collaboration with DEC and WBI on specialized topics/methodologies (e.g. community-driven development (CDD), public- private partnership (PPP) issues, good governance, climate change and carbon finance, impact evaluations, etc). Yet, regional units are likely to be knowledge net providers rather than consumers in most cases. Work with DEC has been on the poverty-environment nexus, health-pollution links, impact evaluations, and trade growth, transport costs and Africa sub-regional corridors. Sector units collaborated with WBI on governance issues, CDD, climate change and carbon finance, South-South initiatives, conditional cash transfers, and PPP issues. The Role of the Regional Vice Presidents (RVPs) and Country Directors Sector managers have described RVPs and country directors as being focused on lending and short-term delivery, and unwilling to internalize learning costs. “Country directors are short-sighted and will not give money for AAA [analytical and advisory activities], policy dialogue, much less business development which is needed to build a program of work.� “CMU budget priorities are clear: (i) supervision; (ii) new lending.� “The contracting relationship with country directors means that long-term, strategic analytical work and strategic dialog that is not demanded by the client is not financed by the country director.� “Country directors have no incentive to give any priority (and funding) to sector knowledge activities, and the sectors do not have their own funds to do so. Hence the reliance on trust funds, which carries its own problems and risks. The focus on individual countries increased greatly the responsiveness to the client, but three problems emerged: (i) loss of a regional perspective, as each director was concerned exclusively with his country; (ii) lack of strategic view for the country, as directors became too close to and dependent on the government; and (iii) neglect of generation and transmission of knowledge.� Cross-Regional and Sectoral Collaboration According to managers, cross-regional collaboration generally occurs on a bilateral basis and is mainly on cutting-edge and corporate issues (e.g. on climate change, on issues related to the fee-based assistance to Russia, gender, and transparency in extractive industries). Creating cross-sectoral teams for multi-sector tasks has proved particularly difficult. Intra-regional inter-sectoral work is more frequent. Main obstacles cited are budgetary arrangements, SAP coding and sector managers’ tendency to retain in- demand staff. The global expert teams (GETs) and active thematic groups came in handy to those who were able to rely on their support. Thirty-three percent (9/30) of sector managers indicated that their unit provided a substantial amount of knowledge to units in other Regions. Seventy-five percent (3/5) and 172 The Matrix System at Work 50 percent (2/4) of Latin America and the Caribbean and Middle East and North Africa sector managers reported substantial knowledge flows to other Regions. “The budgetary and staffing arrangements have produced large transaction costs and no benefits. Their rigidity has caused managers to have to find ways to get around the arrangements and encourages hoarding of staff. E.g., for a decade the budget of each sector in each country has remained the same despite major changes in lending and country demands.� “Within a Region, some cross-support does take place, as well as exchange of knowledge, but almost never between Regions—the barriers are too strong-“ “Business brought in by a sector stays in the sector … There isn’t even a point in the system when the possibility of cross-sectoral or cross-regional teamwork can be considered … Rather than use colleagues from another Region, sector managers would rather hire consultants; managers guard their staff.� “The country model has worked well bringing us closer to the client. But some countries are a world unto themselves leading to missed opportunities in knowledge sharing. While we encourage cross-support, budget rules make it more difficult because we cannot have joint codes with units we are collaborating with.� Ultimately, sector managers argue that in the current flat budget environment, knowledge and lending vie for scarce resources; lending always emerges as the winner. As they put it, “All that matters is lending; knowledge gets relegated to the 10th place.� “At the institutional level, knowledge management has not worked well. In the Regions we are under too much pressure to deliver [lending]. We have no time to think or document our experience, although this was originally envisaged as part of the matrix reform. We have no resources to document and disseminate experience systematically. The main constraint is pressure of lending delivery. Whatever is done is based on individual initiatives only.� “Some improvement might be achieved by setting aside a share of the budget and administering it centrally for sectors to compete based on clear rules for access, but in any competition between lending and knowledge generation, lending wins.� “The acquisition of [trust funds] and allocation of staff time is always secondary to the need to deliver on lending operations. Moreover, in middle- income country-dominated Regions like Europe and Central Asia and Latin America and the Caribbean, the budget is declining—leaving very little for knowledge activities of any kind. If the Bank is serious about knowledge, it must resource it separately, adequately and with appropriate recognition to the staff involved—rather than rely so heavily on lending as the metric of success.� “The rhetoric about the ‘knowledge Bank’ is wishful thinking devoid of reality. All the pressure is for lending, and within lending for new lending rather than Appendix E 173 supervision. If the Bank really wants to encourage knowledge, it should put muscle and money behind it, both of which are now not there.� Incentives and Accountability for Quality Relative roles, incentives, and accountability are considered “absolutely inconsistent with the official arrangements.� An example cited more than once is the example of clearances processes and their tendency to go “up and down different levels of authority.� Sector managers typically described accountability for quality as being high for the sector manager and the task team leader (TTL), none for the sector director and substantial for the country director. Sector directors are deemed accountable only for cases where technical assistance has substantial potential impact on country dialogue. Accountability systems are described as clearer and more effective for lending than ESW. Accountability for non-lending technical assistance (NLTA) is weak across the board. Ninety-three percent (28/30) of sector managers and all country directors/managers believe that the current incentives and accountability mechanisms ensure the accountability of TTLs for lending operations to a substantial extent. Sector managers’ accountability for lending was rated substantial in 90 percent of cases by sector managers and in 55 percent of cases by country directors. Sector directors’ accountability for lending operations was described as substantial (or very large) by only 13 percent (4/30) of sector managers and 9 percent (1/11) of country directors. Accountability of country directors and managers was rated higher in Europe and Central Asia, the Middle East and North Africa and South Asia than in other Regions. A South Asia sector manager insisted that “[d]despite the theoretical matrix assumption of sector accountability for quality, if a major lending operation goes wrong, it is the country director who gets the heat since s/he’s the one interacting directly with the both the client and the RVP (or the president, in case of a very large country or a high-profile problem). This is the only logical and functional aspect of the matrix since in effect the country director calls all the shots, s/he also bears the de facto responsibility.� A Europe and Central Asia sector manager further argued that, “TTLs carry by far the highest responsibility for lending quality, with significant involvement of country directors and none at all by sector directors. For AAA, accountability is shared between [the] TTL and sector manager, with virtually no input from the sector director and country director.� Performance of the current quality assurance system was ranked relatively low for quality of lending operations, lower for ESW quality and even lower for NLTA. The system is however judged effective in ensuring fiduciary/ safeguards quality. All in all, managers feel that the attention is on processes and reputational risk rather than operational and AAA quality. Many argued that “too much attention on safeguards and fiduciary takes away from time spent ensuring technical quality of lending operations,� and that “ the quality assurance system, such as it is, provides some assurance for fiduciary/safeguard issues and reports, but slight for quality of lending and none for NLTA.� 174 The Matrix System at Work Overall, 37 percent (11/30) of sector managers and 55 percent (6/11) country directors/managers rated the current system’s effectiveness in ensuring quality of lending operations as substantial; the percentage drops to 40 percent of sector managers and 18 percent of country directors for  ESW and 23 percent of sector managers and 9 percent of country directors for NLTA quality. Eighty-three percent (28/30) of sector managers and 91 percent (10/11) of country directors rated the systems effectiveness in ensuring fiduciary and safeguards quality as substantial (or very large). As an Africa country manager put it: “Much of the pain is self-inflicted. Reviewing is a self-perpetuating industry that feeds itself (…). Extending a closing date can take months, and minor disbursement issues can become a nightmare. Everyone is tied up in knots and nobody takes decisions based on good judgment.� Appreciation of regional functions’ and the country director’s role in quality assurance is mixed. While some sector managers view their role and the additional layers of control they have introduced as helpful, others find them both inappropriate (being symptomatic of distrust) and redundant. “Two sets of eyes (in the CMU and the SMU) looking at the portfolio are better than just one. It is good to have the CMU also look over quality.� “I am accountable for quality but the CMUs have also started doing technical reviews through their own [quality assurance] systems because accountabilities are not clear and they want to protect themselves.� “QK [the OS unit in the Region] has no impact and should be abolished— too many non-technical people providing feedback but can only make grammatical corrections; occasional helpful feedback on results framework.� Other Institutional Issues Sector Boards and Human Resources Management The effectiveness of Sector Boards as corporate bodies for managing human resources of the sectors is perceived as rather low. Sector managers feel they are unable to secure the technical skills needed to serve current operational priorities; Sector Boards are ineffective in balancing hiring priorities for a particular position and Bank-wide needs; and they are not able to deploy staff across the Bank to serve regional needs. Only about 20 percent of managers believe Sector Boards’ role is substantial in those three areas. SDN Integration Regarding Sustainable Development Network (SDN) integration, dissatisfaction with the merger is more often than not associated with the increase in the span of control of both sector managers and sector directors. Having to handle technical issues in more areas than before also raises quality issues. Size is mentioned in relation to span of control and interaction/competition with other networks and sectors at regional Appendix E 175 and country levels. Yet, the integration is credited by some former Infrastructure Network (INF) and agriculture and rural development sector managers with creating synergies among former ESSD and INF network sectors. Social and Environment sector managers disagree and feel the merger has been more of a “take-over of ESSD by INF�. “SDN is too big and crowds out PREM and HDN which are equally relevant to our clients.� “SDN size is a problem both inside and outside the department. We have too many operations therefore quality slips.� “SDN is much too big and unwieldy. Aside from problems of organization, the main result of the merger has been to dilute the focus on the individual sectors and thus to some extent the quality of their interventions. It is just not possible for a sector director to be on top of the issues in five very different sectors.� “The SDN director has too much power and an unmanageable workload driven by infrastructure, energy, and agriculture while the other sectors receive less attention. (…) The Bank should reorganize the networks based on operational synergies and balance in size—rethink PREM poverty group, social protection, and parts of social development as a global practice group—and SDN integration must be redone.� Managers were asked to rate the performance of the matrix system in comparison with the pre-1997 organization with regard to Bank responsiveness to country needs, technical excellence in lending and AAA, and Bank overall organizational effectiveness. They described the Bank as more responsive to country needs. Yet, many doubt quality and overall organizational effectiveness have improved overtime. Nineteen sector managers described the Bank as more responsive to country needs, 11 are unsure and one sees no significant change in the Bank’s responsiveness. Eighty-eight percent of country directors/managers believe Bank responsiveness to country needs has improved. Five sector managers believe quality of lending operations has improved, six that it has worsened, eight see no significant change, 11 are unsure. Three sector managers said that AAA quality has improved, four said the exact opposite, i.e. that it has worsened. Country directors see no improvements either in lending or in AAA quality. Five out of 30 sector managers see an improvement in organizational effectiveness; 8 out of 30 sector managers believe effectiveness has decreased. All in all, 57 percent (17 out of 30) of sector managers and 73 percent (8 out of 11) country directors believe the matrix system is moderately effective or better. (Table E.1.) Notes 1 The seven Bank sectors of focus are Agriculture and Rural Development; Health, Nutrition, and Population (HNP); Public Sector Governance; Transport; Water; Environment; and Social Protection. 176 The Matrix System at Work Table E.1 Country Directors and Managers’ Response to the Quantitative Section of the Questionnaire # Q. 1-6 Q. 7 Benefits-Costs Q. 8 Pre- and Post-Matrix Q. 9 Matrix Effectiveness % Benefits (Sig or Somewhat) Outweigh Costs Benefits Significantly Outweigh Costs Costs Significantly Outweigh Benefits Benefits Somewhat Outweigh Costs Costs Somewhat Outweigh Benefits % Very Large/Substantial Interviewee/Questions Moderately Ineffective No Significant Change Don’t Know/Not Sure Slightly or Not at All Moderately Effective Substantial Extent Very Large Extent Highly Ineffective Highly Effective Not Applicable % No Change Some Extent % Worsened % Improved % Effective Don’t Know Ineffective Worsened Improved Effective Not Sure CAS/Oper draw on SSPs 1.a 2 3 5 1 0 0 45 CAS/Oper draw on global 1.b 1 5 4 1 0 0 55 good practice CAS/Oper reflect GPG/ 1.c 2 2 7 0 0 0 36 regional issues Knowledge from DEC 2.a 1 1 3 6 0 0 18 Knowledge from WBI 2.b 0 0 2 9 0 0 0 Knowledge from Anchors 2.c 0 0 8 3 0 0 0 Knowledge from other 2.d 1 1 5 4 0 0 18 regions Knowledge from other 2.e 4 1 5 0 0 1 50 units in VPU To DEC 3.a 0 3 2 6 0 0 27 To WBI 3.b 0 1 1 9 0 0 9 To Anchors 3.c 0 1 7 3 0 0 9 To other regions 3.d 1 1 5 4 0 0 18 To other units in VPU 3.e 2 4 4 0 0 1 60 Expertise for 4.a 1 3 4 3 0 0 36 multisector tasks To augment regional 4.b 1 1 5 4 0 0 18 capacity advice across Specialized ­ 4.c 1 2 6 2 0 0 27 Bank TTLs for lending quality 5.a 4 7 0 0 0 0 100 SMs for lending quality 5.1 0 6 4 1 0 0 55 (Continued) Appendix E 177 Table E.1 Country Directors and Managers’ Response to the Quantitative Section of the Questionnaire (cont’d) Q. 9 Matrix Q. 1-6 Q. 7 Benefits-Costs Q. 8 Pre- and Post-Matrix Effectiveness % Benefits (Sig or Somewhat) Outweigh Benefits Significantly Outweigh Costs Costs Significantly Outweigh Benefits Benefits Somewhat Outweigh Costs Costs Somewhat Outweigh Benefits % Very Large/Substantial Interviewee/Questions Moderately Ineffective No Significant Change Don’t Know/Not Sure Slightly or Not at All Moderately Effective Substantial Extent Very Large Extent Highly Ineffective Highly Effective Not Applicable % No Change Some Extent % Worsened % Improved % Effective Don’t Know Ineffective Worsened Improved Effective Not Sure Costs SDs for lending quality 5.c 0 1 1 9 0 0 9 CDs for lending quality 5.d 4 4 3 0 0 0 73 TTLs for AAA quality 5.e 4 5 2 0 0 0 82 SMs for AAA quality 5.f 1 3 3 4 0 0 36 SDs for AAA quality 5.g 0 0 1 10 0 0 0 CDs for AAA quality 5.h 3 3 5 0 0 0 55 QA for technical quality 6.a 0 6 4 1 0 0 55 QA for fiduciary/ 6.b 5 5 1 0 0 0 91 safeguards QA of ESW 6.c 0 2 8 1 0 0 18 QA of NLTA 6.d 1 0 4 6 0 0 9 Dual reporting by staff 7.a 2 4 1 0 4 85.0 Budgetary arrangements 7.b 0 3 5 3 0 27.3 Other transaction costs 7.c 0 1 2 2 6 20.0 Responsiveness to 8.a country needs Technical quality in 8.b lending Responsiveness to 8.a 7 1 0 3 88 13 0 country needs Technical quality in 8.b 0 3 5 3 0 38 63 lending Technical quality in AAA 8.c 0 6 1 4 0 86 14 Bank’s overall org. 8.d 2 2 3 4 29 29 43 effectiveness Other 8.e 1 0 0 10 100 0 0 Effectiveness of matrix 9 0 2 6 2 1 0 73 system 10 178 The Matrix System at Work Table E.2 Sector Managers’ Response to the Quantitative Section of the Questionnaire Q # 11Matrix # Other Questions Q#7 Q #9 Benefits-Costs Q#10 Pre-post Matrix effectiveness % Benefits (Sig or Somewhat) Outweigh Costs Benefits Significantly Outweigh Costs Costs Significantly Outweigh Benefits Benefits Somewhat Outweigh Costs Costs Somewhat Outweigh Benefits Average Anchor Effectiveness % Very Large/Substantial Moderately Ineffective No Significant Change Don’t Know/Not Sure Moderately Effective Slightly or Not at All Substantial Extent Highly Ineffective Very Large Extent Highly Effective Not Applicable % No Change Some Extent % Worsened % Improved Don’t Know Ineffective % Effective Worsened Improved Effective Not Sure SSPs reflect 1.a 5 13 11 1 0 0 60 country/regional experience CAS/Oper draw on 1.b 2 11 13 4 0 0 43 sector priorities CAS/Oper reflect 1.c 3 9 11 6 1 0 41 GPG/regional issues Knowledge from DEC 2.a 1 2 12 15 0 0 10 Knowledge from WBI 2.b 1 4 11 14 0 0 17 Knowledge from 2.c 3 8 11 8 0 0 37 other regions Knowledge from 2.d 2 9 16 3 0 0 37 other units in VPU From SMU to Global, 3.a 4 10 15 1 0 0 47 Corporate, SSPs To DEC 3.b 0 1 8 21 0 0 3 To WBI 3.c 1 5 12 12 0 0 20 To Anchors 3.d 4 8 14 4 0 0 40 To other regions 3.e 2 7 14 4 0 3 33 Expertise for 4.a 2 6 13 9 0 0 27 multisector tasks To augment regional 4.b 0 5 11 13 0 1 17 capacity Specialized advice 4.c 1 6 18 5 0 0 23 across Bank TTLs for lending 5.a 10 18 2 0 0 0 93 quality SMs for lending 5.b 8 19 3 0 0 0 90 quality SDs for lending 5.c 0 4 13 13 0 0 13 quality CDs for lending 5.d 4 9 15 2 0 0 43 quality TTLs for AAA quality 5.e 10 18 2 0 0 0 93 SMs for AAA quality 5.f 7 18 5 0 0 0 83 SDs for AAA quality 5.g 0 3 15 12 0 0 10 CDs for AAA quality 5.h 1 7 15 7 0 0 27 Appendix E 179 Table E.2 Sector Managers’ Response to the Quantitative Section of the Questionnaire Q #9 Q #11 Matrix # Other Questions Q#7 Benefits-Costs Q#10 Pre-post Matrix effectiveness % Benefits (Sig or Somewhat) Outweigh Costs Benefits Significantly Outweigh Costs Costs Significantly Outweigh Benefits Benefits Somewhat Outweigh Costs Costs Somewhat Outweigh Benefits Average Anchor Effectiveness % Very Large/Substantial Moderately Ineffective No Significant Change Don’t Know/Not Sure Moderately Effective Slightly or Not at All Substantial Extent Highly Ineffective Very Large Extent Highly Effective Not Applicable % No Change Some Extent % Worsened % Improved Don’t Know Ineffective % Effective Worsened Improved Effective Not Sure CDs for AAA quality 5.h 1 7 15 7 0 0 27 QA for technical quality 6.a 4 7 16 3 0 0 37 QA for fiduciary/safeguards 6.b 8 16 4 1 0 1 83 QA of ESW 6.c 3 9 16 2 0 0 40 QA of NLTA 6.d 1 6 13 10 0 0 23 K&L services to regions 7.a 5.03 Operational support to 7.b 4.33 regions Mobilizing resources 7.c 5.03 Promoting & integrating 7.d 4.87 GPGs Addressing corporate 7.e 7.47 demands Managing global 7.f 6.50 partnerships SB ensures technical skills 8.a 1 5 14 10 0 0 20 in Bank SB balances particular & 8.b 0 6 13 11 0 0 20 Bank needs SB deploys staff across the 8.c 1 4 9 16 0 0 17 Bank effectively Dual reporting by staff 9.a 4 13 8 1 4 65 Budgetary arrangements 9.b 3 2 13 10 2 18 Responsiveness to country 10a 18 0 1 11 95 0 5 needs Technical quality in lending 10.b 5 6 8 11 26 32 42 Technical quality in AAA 10.c 3 4 12 11 16 21 63 Bank’s overall org. 10.d 5 8 4 13 29 47 24 effectiveness Other 10.e 3 0 0 27 100 0 0 Effectiveness of matrix 11 1 4 12 10 2 1 57 system 180 The Matrix System at Work 2 Managers (and directors) were asked to respond with specific reference to their experience. Yet, managers have worked in more than one Region and their perspective is therefore colored by their experience in their previous Vice Presidential Units (VPUs) as well as the Region they are currently mapped to. 3 One of the sector managers did not rate the performance of the matrix, but provided useful comments and was therefore not dropped from the sample. Appendix E 181 Appendix F Focus Groups Discussions: The Key Findings This annex summarizes the key findings from three hour-long focus group discussions held with seven Sustainable Development Network (SDN), one Operations Policy and Country Services (OPCS), seven Poverty Reduction and Economic Management Network (PREM), and six Human Development Network (HDN) anchor and regional senior staff members on July 19, August 23, and September 7, 2011. The main purpose of the discussions was (i) to get in- depth feedback on the effectiveness of the matrix, (ii) to better understand cross-regional and cross-sectoral differences in the working of the matrix, and (iii) to test findings from the staff survey and manager interviews as well as conclusions derived from budget, staff, and portfolio analyses against focus group participants’ views. Participants either responded to the staff survey and volunteered for focus groups or were otherwise identified as likely to contribute significantly to the deliberations. Staff were asked to describe and provide their opinions on the adequacy of knowledge production, cross-regional and cross-sectoral knowledge transfers, regional and Bank-wide incentives and accountability mechanisms for quality, as well as the interplay between sector and country strategies. SDN staff were also asked to elaborate on the benefits and costs of the SDN merger. Facilitators probed for additional information where necessary. Knowledge Production and Dissemination “Knowledge is country-driven and country-retained. Many of us work on specific countries or geographic focus. We don’t even know what is happening in our unit let alone in other Regions.� “For the most part I am able to use knowledge [produced] both in my unit and [in units] outside the VPU [VicePresidential Unit], but mostly through personal contacts.� “There are formal instruments that are helpful for (…) specific activities. One is the Global Expert Team [GET].� “I have used GETs. If you ask a question to the thematic group, you get an answer pretty quickly. So the formal mechanisms are there. The real handicap is the technology, the search engine, which they have been working on but has been abysmal until now, as well as the coding of products.� “People are not putting things in the correct repositories. That makes it very difficult to do a search and find something.� “Nobody has the time to read a 60-page paper or 600-page book.� “In terms of producing knowledge (…) [the anchor] is sort of caught in the same dynamic, not asking the right questions. I am just not convinced that it helps me in my day-to-day work.� “I think there is a lot of generation and sharing of knowledge in the network. [Yet], what is generated or shared often reflects personal interest, not what is in demand in the country offices.� 184 The Matrix System at Work “When you are sitting in a field office, it is extremely difficult to stop and write up what you have done. I don’t see the average task manager being able to add that to his or her work program.� “The purpose of the network was to disseminate global knowledge. Now they are utilized for corporate priorities and donor priorities. The mandate they were supposed to have is not the mandate they actually have. They cannot generate knowledge without input from the Regions. The Regions have become suppliers of knowledge to the anchor; it should be the opposite.� As the selected quotes above suggest, focus group participants described knowledge sharing across Regions and sectors as weak and heavily dependent on informal networks, and formal systems for knowledge sharing as useful but inadequate. Technology-driven media, although much improved, still miss the mark. Staff find GETs particularly useful but wish the teams had greater capacity to respond to requests. Thematic groups (and communities of practice) have also generally been described as proactive and effective when led by a motivated core in spite of the budget constraints they face. The reliance on informal channels for cross-regional and cross-sectoral knowledge transfers makes the need for face-to-face contacts and networking opportunities for both headquarter- and country office-based staff all the more necessary. Fora of communication staff mentioned included sector and training weeks. Participants understand that such events are costly but wish effort was made to organize them more frequently. As one put it, “there is value in getting people together, but there is also a cost, which has to be taken into account to determine in which way and how often to bring people together.� Given the importance of informal networks, staff argued that anchors should do their best to facilitate their development. Focus group participants expressed concern with the relative isolation of country office-based staff and locally recruited staff in particular. Most strongly believe that “if you are based in a country office, you are on your own.� All agreed that only a small fraction of the knowledge produced is actually relevant to and used in project design. Some suggested the sheer amount of information to process was to blame; others argued that knowledge produced tends to be supply-driven, rather than demand-driven, and rarely originates directly from the field. There was broad consensus that knowledge production in the field is underfunded and that the need to rely on trust funds leads to a non-strategic whole as well as major inefficiencies. Strategy According to the focus groups participants, Bank-wide strategies play a limited to nonexistent role in analytical and advisory activities (AAA) and lending operations. One noted that he was unaware of the existence of his sector’s strategy for years, yet prepared projects nonetheless. Another argued that since strategy papers reflect emerging Bank knowledge at the time they are drafted, Appendix F 185 their prescriptions simply cannot find their way into country programs either explicitly or implicitly. Incentives and Accountability for Quality Regarding lending operations, participants argued that more should be done at the concept note stage to address technical issues instead of focusing on the quality of documents and the neatness of procedures. Technical issues should in any case be addressed before the operational committee (OC) or decision meeting stage since “once things are at the OC stage, (…) very little can be done.� Quality assurance for large economic and sector work (ESW) was described as strong and “rigorous at the concept note stage and at delivery.� There are, however, no formal quality assurance systems for just-in-time notes and non-lending technical assistance (with the possible exception of those in place in Europe and Central Asia). Quality therefore varies widely: some products’ quality rivals that of large ESW reports; others are produced in under a week, largely copying and pasting text from other sources. Senior staff insisted strengthening peer review processes should be the basis for improvements to quality assurance for small ESW reports and non-lending technical assistance. As is, peer reviewers are often selected by team leaders and they do not expect their suggestions (and criticisms in particular) to be taken on board. Also, according to focus group participants, concentration of budget decisions with the country director sometimes conflicts with strategic selection and quality of work. Activities can be initiated in an effort to increase lending volumes and in the name of client responsiveness even if the Bank is not well- suited to the task or it creates excessive risks. Delivery is expected to come sooner than later and quality suffers from the self-imposed time constraints. Business development and longer-view/strategic AAA, which can underpin future operations are clearly under-resourced. “HR is the one function that gives Sector Boards some cache.� Most participants were unaware of the mandate of the Sector Boards beyond human resources and a few suggested that Sector Boards play a role in quality assurance. The Regional Matrix “Where it works well, it works well, and maybe one of the reasons is personal chemistry.� “When sector managers and country directors get along, the system works well.� “In the Region, it is SMUs [Sector Management Units] versus CMUs [Country Management Units]. It is not one team but two who are competing. If you have the money, you have a lot of power. That does have to be rebalanced.� 186 The Matrix System at Work Regional matrices’ functionality is seen as highly dependent on the personal relationships between sector managers and country directors. Staff argued that power and resource balances between CMUs and SMUs should be rethought. The balance of resources should still be tilted to the CMUs to ensure responsiveness, but SMUs should have direct access to funds for business development and strategic AAA with no direct bearing on current operations. In addition, participants suggested that work program agreements (WPAs) should be negotiated around block grants which allow sector units flexibility in allocating resources to operations, AAA, and business development as appropriate so long as they deliver the agreed upon products. SDN Merger “The merger was awful for environment and social staff. We [infrastructure] dominate them. Now we have 60 to 70 percent of country programs.� “If we have to compete with the energy sectors who bring a lot of business, we don’t have much of a chance. There is no business left. Solid waste has moved to urban, climate change to energy.� Environment and social development specialists think their operations now suffer from comparison with infrastructure projects, particularly when lending volume is a high priority; infrastructure specialists do not disagree. Another drawback mentioned is the managerial cost that the merger has generated. The main benefit from the merger is the ability to “work­ cross-sectorally more easily.� Organizational Impediments “There aren’t that many organizational models and they have pros and cons. The matrix has the potential to work. We need to make sure to identify the specific problems. It may be a problem of mandates and incentives, and we don’t pay sufficient attention to incentives across the board.� “The current structure does not create any major impediments, it is more a question of incentives. For instance, incentives for managers to facilitate and promote rather than constrain or forbid cross support; incentives in the hub to go out to the country teams and be more effective in getting and sharing information; incentives for the TTL [task team leader] to share knowledge with the hub.� “On peer review for non-lending, you get there as a peer reviewer, and every person in the room wants the project approved by June 28. If it slips, the TTL loses the budget. Multi-year budgeting is not possible, so you have artificial deadlines.� “The pressure on lending is even greater. [The] rule used to be in Europe and Central Asia that there are no review meetings in the last quarter. If review is serious, there has to be time to change the product.� All in all, focus group participants argued that organizational impediments are not of a structural nature; incentives are more problematic. Yet, budget Appendix F 187 arrangements make it to the top of the list of organizational constraints. Innovation and strategic knowledge generation should not be subject to the fiscal year deadlines (or so they argue), and AAA and lending operations’ quality should not suffer for fear of losing budget allocations. 188 The Matrix System at Work Appendix G Comparison of the Six Regional Matrices Background The Bank’s six administrative Regions vary considerably in terms of the nature of the region they serve, including the number and needs of client countries, the distance from Washington, their time zones, and the capacity and cost of potential local employees, as highlighted in Table G.1.1 In addition, given the allocation system and demand by clients, the scale and composition of the Bank’s program varies substantially by administrative Region, as shown in Table G.2. Structure Not surprisingly, the Regions have adapted their structures in an attempt to reconcile the demands of the matrix system with the various requirements of doing business in each operating environment. This section explores the structural differences among the Regions that have followed, summarized in Table G.3. Size of Country Units The matrix reforms of 1997 anticipated what was referred to as a “naked country director�, meaning the country director would only keep a skeletal Table G.1 The World’s Regions Vary Dramatically in Nature AFR EAP ECA LCR MNA SAR # of countries (non- 47 (36) 22 (12) 23 (23) 30 (20) 12 (12) 8 (6) small states) % middle income 21% 55% 83% 87% 83% 25% (IBRD-eligible) Typical flight time from 20 hours 24 hours 12 hours 8 hours 15 hours 24 hours Washington Typical time zone differ- 6 hours 12 hours 8 hours 1 hour 7 hours 9 hours ence from Washington Table G.2 The Scale and Composition of the Bank’s Programs Vary by Administrative Region FY08-10 avg. AFR EAP ECA LCR MNA SAR Annual lending $8.2 billion $6.7 billion $8.1 billion $10.9 billion $2.3 billion $7.0 billion Annual # of projects 96 47 49 64 19 42 Average loan $85 $143 $165 $170 $121 $167 size million million million million million million 190 The Matrix System at Work Table G.3 The Regions Have Experimented with Different “Small Matrix� Arrangements AFR EAP ECA LCR MNA SAR 6 CDs, 15 Country units 15 CDs, 25 CMs 6 CDs, 6CMs 7 CDs, 18 CMs 4 CDs, 8 CMs 5 CDs, 1 CM CMs Sector unit Vary by dept; In SDN sectors, SM Piloted program Region- Region-wide Region-wide structure FPD & PREM (or SL functioning in SDN sectors wide SMs SMs cover SMs cover 1 or 2 cover entire as SM) co-located where SM was cover 1 or 2 1 or sectors network of with CD covers responsible for sectors 2 sectors sectors by sub- all SD sectors, sector Region- region; others while 2 SMs in HQ wide as well as are region-wide cover corporate all SDN sectors by 1 or 2 sectors responsibilities in one country grouping; officially in place, but faded Sector/country In HD and 27 coordinators CSCs co-located Sector No sector HD and PREM CSCs unit interface SD, SL is co- support the 6 SDN with CD; Leader department- co-located with located with CD; SMs and SLs in field 25–40% of time; co-located wide SLs CD; 3 SDN CSCs includes QA in in specific sectors; responsibility with CD or CSCs; for India, 2 for HD, excludes QA 7 SLs in HQ support excludes QA for each of some SMUs Nepal, 1 for each in SD the 2 SDN SMs in HQ HDN, SDN, maintain of others; 20% of FPD, and point person time; includes QA quality; to liaise 100% with some of time; CMUs includes QA 12 of 15 CDs in All 5 CDs in Country Director All 6 CDs in field 4 of 7 CDs in field 3 of 6 CDs All 6 CDs in field field? field decentralization in field Share of 58% (45%) 71% (48%) 49% (17%) 33% (14%) 36% (21%) 64% (35%) technical staff decentralized (share of IRS) staff, outsourcing tasks to the sectors. This vision appears to have been over- ambitious—country directors have gradually built up their teams to support portfolio monitoring, quality control, strategy development, [and in a few cases even task team leadership for operations]. But the size and composition of Country Management Units (CMUs) varies by Region. Structure of Sector Units There are three basic structures –and one hybrid—for sector units with which the Regions have experimented. The first is the most basic, and currently used by the Latin America and Caribbean, South Asia, and Middle East and North Africa Regions. In this model, there is a single sector manager (SM) that covers one sector (or two when the sectors are small and there is a reasonable overlap in scope) for the entire Region. Sector managers report to one of three or four sector directors. Appendix G 191 The second model is used in East Asia and the Pacific Region. While units for most of the sectors follow the model as above, the Region maintains a sector manager (or sector leader that serves as a de facto manager) for all of the seven Sustainable Development Network (SDN) sectors2 as a group in each country grouping, six in total, co-located in the field with the country director. In addition, there are two sector managers based in Washington who cover the former infrastructure and Environment and Socially Sustainable Development Network (ESSD) sectors respectively. These headquarters-based sector managers fulfill corporate duties, such as sitting on the various Sector Boards, as well as managing staff based in Washington. In essence, the staff in the SDN sectors sit within three matrices—the “big matrix� of the networks and Regions; the “small matrix� of the Country Units and Sector Units; and the “very small matrix� of the field-based sector managers and headquarter-based sector managers. The rationale for this structure is that the Region is highly decentralized with multiple large offices, and so it makes sense to co-locate the manager with her staff. Moreover, travel and time zone differences make it difficult to communicate efficiently if the manager remained based in Washington. A hybrid of the first two models was launched in the Europe and Central Asia Region, where SDN sector managers are responsible both for an individual sector across all countries in the Region, as well as coordinating all SDN sectors in one country grouping (although not based in the field). While this responsibility technically continues, it was not found to be manageable in practice. The third model has been tested in the Africa Region. In this model, the Region was split up into several sub-regions, with a sector manager covering several sectors rather than one. This resulted in several sector managers for the same sectors. In most sectors, this sub-regional structure was seen to result in a lack of vision and consistency within individual sectors and was ultimately abandoned. Today it is maintained in the Poverty Reduction and Economic Management Network (PREM), where there are four sub-region-based sector managers that cover all PREM sectors, and the Financial and Private Sector Development Network (FPD) where the Region is split by two sector managers. Decentralization The degree of decentralization varies significantly by Region. East Asia and the Pacific and South Asia, Regions far from headquarters, are highly decentralized with roughly two-thirds of staff based in the field. Among the two, East Asia and Pacific is more tilted toward internationally recruited staff (IRS), with half of all IRS staff based in the field. South Asia, perhaps with its legacy of high capacity but lower cost local staff, is more tilted toward locally recruited staff in the field. In these Regions, all country directors are based in the field. South Asia is beginning to move sector managers to the field; East Asia and the Pacific already locates many sector managers in the field. Among the least decentralized Regions are Latin America and the Caribbean, Europe and Central Asia, and the Middle East and North Africa. Most of the 192 The Matrix System at Work countries in Latin America and the Caribbean are within a time zone of Washington and so communication with headquarters is not an issue. Moreover, travel to the Region is shorter and cheaper than for any other Region. And even though capacity is high among potential staff in the Region, labor costs are also very high. In Europe and Central Asia, just a quarter of projects are managed from the field, and barely half of country directors—four of seven—are field- based. In these Regions, roughly one-fifth of IRS are located in country offices. Interface between Sector and Country Units None of the several structures adopted by the Regions for sector units has addressed one item of concern: the large span of control for sector managers. Most sector managers have a wide span of control of staff and countries—as much as 90 staff for the sector manager of Agriculture and Rural Development in Africa. In addition, sector managers can have either a wide span of control of countries which makes it difficult to lead the policy dialogue in each of them, or a wide span of control of sectors which makes it difficult to exercise quality control over the work of the unit. Regions have adapted to address some of these span-of-control issues, although none has increased the number of sector managers. Early on, Latin America and the Caribbean created sector leaders for each network of sectors in each country grouping who serve several purposes. First, they act as shadow managers by mentoring junior staff and fulfilling quality assurance functions; this serves to address the problem of a large span of control of direct reports. Second, sector leaders are responsible for coordinating the various operations of the covered sectors in their country grouping, developing future business, and serving as the face of the sectors for government counterparts. The time of sector leaders is fully funded for these functions. In more recent years, other Regions have adopted a similar structure, but without the same level of budget and responsibility. Europe and Central Asia and South Asia have assigned country sector coordinators (CSCs) to oversee network coordination in each country grouping. In Europe and Central Asia, 25–40 percent of CSC time is devoted to the task; in South Asia 20 percent. But South Asia expects CSCs to provide a quality assurance function whereas Europe and Central Asia does not. This may not be manageable given the wide range of sectors and the limited amount of time allocated to the assignment (although South Asia does have three CSCs in India for the SDN sectors). Arrangements for Quality Assurance Accountability In 2009, President Robert Zoellick requested that the Regions clarify the accountabilities for various activities and decisions. Four Regions—Africa, East Asia and the Pacific, Latin America and the Caribbean, and the Middle East and Appendix G 193 North Africa— completed comparable “accountability matrices� that formally identified the roles and responsibilities of management within each Region. The matrices are very similar across Regions; an aggregation of the four is depicted in Table G.4. For decisions on strategy, country directors are accountable for the country strategy (although in Latin America and the Caribbean, the vice president holds primary responsibility). Sector directors are responsible for sector strategies, which are developed Bank-wide, except in the Middle East and North Africa, where this falls to the sector manager. For decisions on operational work during the preparation stage, there is modest variation among Regions. In three Regions, the accountability at the three main decision points during lending preparation is the same. At the beginning of the process, country directors hold primary accountability for the concept review meeting; part way through, sector managers are accountable for the (sometimes optional) quality enhancement review which is intended to focus on quality; and towards the end of preparations, country directors are accountable at the decision meeting stage (except in high risk and very large operations, where the vice president holds accountability), at which point the project team is authorized to send an operation for full appraisal. The Regions follow a Table G.4 Formal Accountability is Concentrated on the Sector Manager Sector manager Sector director Country Strategy Country strategy 333 Bank-wide sector strategy 3 333 Lending preparation Business development ½ 33 ½ Concept review 3333 Quality enhancement review 3333 Decision meeting 3 333 Negotiations/Board 3½ 33½ Lending supervision Implementation supervision 3333 Approval of ISRs 3333 Approval of mid-term reviews 3 Non-lending operations Concept review 3333 Decision meeting ½ 3 33½ Delivery to client 3 33 Note: 3 indicates accountability in one of the four Regions for which data is available. 194 The Matrix System at Work similar pattern for non-lending operations, with country directors accountable at both the concept and decision stage. East Asia is the exception to this pattern. There, sector directors are accountable at the decision meeting stage, the theory being that country directors are best placed to judge the responsiveness of a proposed operation at the concept stage, and the sector director is best placed to judge its technical quality and likelihood of success at the decision meeting stage. During the supervision stage, all four Regions assign the sector manager accountability. This includes approval of implementation supervision and results reports and mid-term reviews (only two Regions list this in their matrices). There are two striking patterns in Table G.4. First, sector directors formally hold very little accountability. Aside from the rather vague activity of business development, only East Asia and the Pacific assigns accountability to sector directors for operational work. In that Region, sector directors chair the decision meeting for lending and non- lending operations. Second, sector managers are assigned enormous responsibility. In addition to their large span of control of staff, countries, and operations, they oversee the staff who prepare lending and non-lending operations. While they are not accountable for the results of concept review and decision meetings, sector managers are generally considered responsible for the technical quality of these operations. And as the table shows, they are primarily accountable for each of the key responsibilities during supervision. Quality Review Processes The processes for quality review vary modestly among the Regions, but there is one constant theme: there are multiple layers of review. Operations under preparation are typically reviewed by at least three sources. First, the sector manager, or her designee (for example a sector leader or country sector coordinator), provides comments on the technical quality of an operation. In some departments, there is also a quality review unit (for example in the SDN department in Europe and Central Asia) or point person who reviews one or more aspects of the operation. Second, most country departments review operations for quality, typically led by an operations adviser to the country director or senior operations officer. Third, central quality units, discussed below, provide support to task teams and provide comments at the review meetings. Review meetings also include comments from peer reviewers. Except in a few departments, peer reviewers are identified by the project team rather than the sector manager, causing concern that their views are not fully independent. The Quality Assurance Group (QAG) was introduced concurrently with the matrix reforms to provide arms-length independent review of the quality of operations during preparation and supervision. Assessments were agreed upon by a panel of reviewers, but were not ready in time for review meetings. Since QAG was moved into Operations Policy and Country Services (OPCS) in FY10, there is no review of the quality of operations independent of regional management. Appendix G 195 Table G.5 Comparison of Regional Quality Units AFR EAP ECA LCR MNA SAR Annual projects 96 47 49 64 19 42 (FY08-10 avg.) Annual lending $8.2 billion $6.7 billion $8.1 billion $10.9 billion $2.3 billion $7.0 billion (FY08-10 avg.) Quality team budget $8.3 million NA ($658k) NA ($1.1 million) $4.2 million $3.3 million $2.3 million (3 year avg.) Quality team total 26 5 11 9 5 5 staffing (current) GH-level staff 4 2 3 4 2 1 Quality team staffing Manager, 3 GH, 2 GH, 1 GG, 1 GF, Manager, 2 GH, Manager, 3 2 GH, 2 GG, Manager, 2 GG, distribution 10 GG, 7 GF, 1 1 GE 5 GG, 2 GE GH, 1 GF, 1 1 GE 2 GE (current) GE, 4 JPA (+ TFC) ETC (+ TFC, 1 GF, 1 GE) Quality Units Separate units within Regions to review the quality of analytical and advisory activities (AAA) and operations under preparation or supervision were not anticipated during the matrix reforms. Today, each of the six Regions maintains a development effectiveness unit, although they vary substantially in size, mandate, and influence as shown in Table G.5. By far the largest of the quality units is in Africa, which has a manager and 25 technical staff,3 including three besides the manager at level GH. In addition to reviewing operations for quality, the department provides support in designing results frameworks and plans for monitoring and evaluation. At the other end of the spectrum is South Asia, with a manager and four professional staff, none of which are level GH. Notes 1 They are Africa (AFR), East Asia and the Pacific (EAP), Europe and Central Asia (ECA), Latin America and the Caribbean (LCR), Middle East and North Africa (MNA), and South Asia (SAR). 2 There can be great variation in these categories among countries within a Bank Region as well. 3 This includes the former Infrastructure (INF) sectors (Energy and Mining, Transport, Urban Development, Water) and the former ESSD sectors (Agriculture and Rural Development, Environment, and Social Development). 4 This includes staff at the GE level and above, as well as JPAs. 196 The Matrix System at Work Appendix H Working across Organizational Boundaries: The Use of Cross Support This annex presents the key findings of the Independent Evaluation Group’s (IEG’s) analysis of Bank-wide data on cross support including data on cross support from Regions to network anchors, from network anchors to Regions, across Regions and across sectors within the Regions. Levels of cross support—percentage of time experts work outside their responsible units—are used as indicators of anchors’ and Regions’ ability to make use of the technical expertise embodied in individuals across the Bank while making their own staff’s knowledge and skills available to others. IEG’s analysis of cross support data therefore serves the twin purpose of providing evidence of anchors’ and Regional Sector Management Unit’s (SMUs’) success or failure to mobilize and disseminate country-specific experience and cross-country technical knowledge across the Bank; and accounting—in part—for the degree of flexibility for technical staff to work across Regions and sectors. The analysis used data extracted from the Business Warehouse’s Resource Management Budget Accounting module as of December 2010. It covers the period FY02-10: Pre-SAP data was unavailable and FY01 data has been excluded because it was reported on the basis of different budget rules. Cross support is calculated as the amount or percentage of GF+ staff time, i.e. time staff graded GF and above spent working outside their responsible unit both during FY10 and on average over the period FY02-FY10. Main Findings In FY10, inter-Vice Presidential Unit (VPU) cross support sold (i.e. cross support across VPUs) was about 5 percent of GF+ staff time overall, with network anchors selling four times as much cross support as Regions (in terms of percentage of staff time): Regions sold 3 percent cross support and network anchors 12 percent. Intra-VPU cross support (i.e. cross support within VPUs) was much higher (13.8 percent overall, 15.6 percent in the Regions and 9 percent in the network anchors). The time trend in inter-VPU cross support sold (in staff weeks) over the period FY02-10 has by and large been flat. There was, however, a decrease in inter-VPU cross support as a percentage of seller’s time due to the increase in technical staff. Cross support from anchors to Regions, in particular, has decreased over time. FY10 cross support across Regions was 1.5 percent, 0.6 percent in reality being continuity support, i.e. time staff spent supporting the regional VPU in which they worked in FY09. There has been limited cross support across network boundaries i.e. cross support from anchor to anchor, from regional sector units to anchors of other networks, or between sector units of different networks within each 198 The Matrix System at Work Region (e.g. FY10 cross support from network anchors to other anchors accounted for only 0.2 percent of anchors’ staff time). Key Definitions Cross Support Cross support is defined as “staff time of an expert or specialist purchased from outside the responsible unit for a specific task.� Such tasks include joint missions, analytical and advisory activities (AAA) pieces or projects, peer reviews, and multi-sectoral work (e.g. project-level fiduciary and safeguard services, Public Expenditure Reviews (PERs), Poverty Reduction Strategy Paper (PRSP) process and reviews, Poverty Assessments, etc). Cross support is short-term by nature (typically a few staff weeks) and should not be confused with staff movement/rotation which refers to longer term arrangements and involves staff’s administrative transfer from one unit to another without budget transfer. Staff from one unit can also spend time working in/for another unit under a Work Program Agreement (WPA). In this case both budget and task management responsibilityare transferred to the receiving unit. Types of cross support There are two broad categories of cross support: inter-VPU cross support and intra- VPU cross support. Inter-VPU cross support refers to cross support from one VPU to another (e.g. from one Region to a network anchor or vice versa); while intra-VPU cross support is cross support within a VPU (see examples in Table H.1). Cross support is either prepaid or pay-as-you-go depending on the selling unit- purchasing unit pair and on the amount of cross support purchased. Inter-VPU prepaid cross support is negotiated and paid for at the start of each fiscal year. The Corporate Planning and Analysis Department (CFRPA) calculates the actual amount of cross support utilized during the preceding calendar year. This amount constitutes the ‘benchmark’ prepayment for the following fiscal year. Inter-VPU pay-as-you-go cross support payments, which cover deviations from the benchmark, are made by budget transfers. There are two settlements every Table H.1 Selected Case Scenarios for Intra-VPU Cross Support VPU Receiving Unit/Buyer Providing Unit/Seller Examples of Tasks Region Sector Management Units Regional Core Services unit, or Fiduciary and Safeguard work, multi-sector (SMUs) other SMU in the same region d iagnostic work tasks, core ­ Region Country Management Units SMUs in the same region CMU managed work including country monitor- (CMUs) ing, and Environment and social fiduciary work Appendix H 199 year. The first is in December and is based on end-November data. The second is in June and is based on the December to June data. Each VPU is free to set internal rules and procedures for intra-VPU cross support planning and budget settlements. Summary of Management’s Findings on Cross Support Management ’s Approach and Findings Recent analyses of FY08, FY09, and FY10 cross support data by management, and the Reform Secretariat (OPCRS) in particular, focused on inter-VPU cross support.1 Key findings were: • FY10 Bank-wide inter-VPU cross support sold accounted for 3.2 percent of total staff time: 2.1 percent of staff time for regional VPUs, and 8.3 percent for network anchors. • Cross support has mainly been provided by operational staff graded GG and GH. • Inter-VPU cross support has been consistently low and steady over the period FY08-FY10. • FY10 intra-VPU cross support was about 7 to 10 percent within R ­ egions. Management figures suggest that FY10 intra-VPU cross support within Regions was three to five times higher than inter-VPU cross support sold by Regions. Also, safeguard, fiduciary services and multi-sectoral initiatives are likely to generate significant amounts of intra-regional (i.e. intra-VPU) cross support. Given the importance of such activities in a post-QAG Bank environment, looking into the amount and nature of intra-VPU cross support is clearly warranted. Methodology for IEG Analysis Data and Data Sources IEG’s analysis used data extracted from the Business Warehouse’s resource management budget accounting module as of December 2010. It covers the period FY02 to FY10: Pre-SAP data was unavailable and FY01 data has been excluded because it reported on the basis of different budget rules. Rather than calculating cross support as a percentage of total staff time, IEG elected to restrict its analysis to technical staff graded GF and above. Although part of the motivation for the analysis lies in the potential for comparisons across Regions and discussions of time trends, a strong note of caution should be issued in view of the following: • Cross support sold, whether in staff weeks or as a percentage of total staff time, is a function of the size of the selling unit. Hence, ­ reorganizations and mergers of units have an impact on the amount of cross support sold. 200 The Matrix System at Work • Intra-VPU cross support budget settlement rules vary from one Region to the other; so do unit sizes and organizational structures. Though the reviewers attempted to control for those aspects (e.g. by adopting consistent categorization of Regions’ various units), disparities of a structural nature are likely to remain. Key Findings Inter-VPU Cross Support In FY10, inter-VPU cross support sold was about 5 percent of GF+ staff time overall: 3.3 percent from Regions and 12.5 percent from network anchors. As Table H.2. shows, the biggest seller (in terms of percentage time sold) among Regions was the Middle East and North Africa, with 3.8 percent; Europe and Table H.2 Inter-Vice Presidential Unit Cross Support (XS) Sold, Seller’s Perspective, FY10 Inter-VPU cross support sold (staff weeks) Inter-VPU cross support (% of seller’s total time) To Total To To Network To other Total XS seller To Network To other Regions Anchors units sold time Regions Anchors units Total XS sold AFR 424 644 283 1351 41758 1.0% 1.5% 0.7% 3.2% EAP 346 239 218 803 25150 1.4% 1.0% 0.9% 3.2% ECA 360 208 93 661 22472 1.6% 0.9% 0.4% 2.9% LCR 290 322 91 703 20415 1.4% 1.6% 0.4% 3.4% MNA 245 118 71 434 11511 2.1% 1.0% 0.6% 3.8% SAR 450 216 62 729 20648 2.2% 1.0% 0.3% 3.5% Regions 2115 1747 818 4681 141954 1.5% 1.2% 0.6% 3.3% Inter-VPU cross support sold (staff weeks) Inter-VPU cross support (% of seller’s total time) to Net- Total To To work To other seller To Network To other Regions Anchors units Total XS time Regions Anchors units Total XS SDN 2428 40 195 2663 21017 11.6% 0.2% 0.9% 12.7% HDN 745 4 60 808 5356 13.9% 0.1% 1.1% 15.1% PREM 369 12 45 427 4476 8.3% 0.3% 1.0% 9.5% FPD 646 22 17 686 5958 10.8% 0.4% 0.3% 11.5% Networks 4189 78 318 4584 36807 11.4% 0.2% 0.9% 12.5% TOTAL 6304 1825 1136 9265 178760 3.5% 1.0% 0.6% 5.2% Regions + networks Source: World Bank Database. Note: Other units include DEC, WBI, ISG… Appendix H 201 Central Asia, with 2.9 percent had the lowest staff time sold outside the VPU. Among network VPUs, the Human Development Network (HDN), with 15.1 percent, was the biggest provider of inter-VPU cross-support followed by 12.7 percent from the Sustainable Development Network (SDN) for inter- VPU cross support. Anchors mainly sold cross support to Regions. Out of the 12.5 percent of staff time sold by network anchors, 11.4 percent was sold to Regions: HDN2 sold 14 percent of time to the Regions; the lowest sales to Regions (i.e. 8.3 percent of staff time) were by Poverty Reduction and Economic Management Network (PREM) staff. Cross support from anchors to anchors amounted for only 0.2 percent of anchors’ staff time. HDN had the lowest percentage of staff in other VPU’s time sold to other anchors (i.e. 0.1 percent); the Finance and Private Sector Development Network (FPD) sold 0.4 percent of its time to other network anchors. From the buyers’ perspective in FY10, Regions purchased 6.5 percent of time worked within their VPUs: 1.4 percent from other Regions, 2.9 percent from network anchors, and 2.2 percent from other VPUs. Inter-VPU cross support represented 10 percent of time worked in the Middle East and North Africa and 7.2 and 6.5 percent staff time worked in South Asia and Africa, respectively (see Table H.3). Network anchors bought, on average, 7.8 percent of staff time worked within their unit. Of those 7.8 percent, 5 percent were purchased from Regions, 0.2 percent from other network anchors, and 2.6 percent from other VPUs. The HDN anchor bought 9.7 percent cross support from Regions; only 1.6 percent of time worked within PREM was provided by Regions. There was on the whole virtually no inter-VPU cross support from regional sector units to anchors of other networks during FY10. Similarly, the bulk of cross support from network anchors to regional sector units has typically been within the boundaries of their respective networks (see Table H.4 and Table H.5). SDN regional sector units sold 1.6 percent staff time to the SDN anchor and no cross support at all to anchors of other networks. HDN sector units sold 2.2 percent of staff time overall to anchors: 2.1 percent to the HDN anchor and 0.1 to the SDN anchor. PREM and FPD combined regional sector units sold 1.2 percent of staff time: 0.7 percent to the FPD anchor, 0.1 to the PREM anchor and 0.3 to the SDN anchor. The time trend in inter-VPU cross support sold (in staff weeks) over the period FY02-FY10 has, by and large, been flat. There was, however, a decrease in inter-VPU cross support as a percentage of seller’s time: Total inter-VPU cross support sold in FY02 was 9,373 staff weeks; it was 9,264 in FY10. Yet, due to the increase in technical staff, inter-VPU cross support was 7.6 percent staff time in FY02; it was 5.2 percent in FY10. 202 The Matrix System at Work Table H.3 Inter-Vice Presidential Unit Cross Support, Buyer’s Perspective, FY10 Inter-VPU cross support (% of buyer’s own Inter-VPU cross support bought (staff weeks) managed + cross support time ) From From From From network other Total XS Total time From network From other Total XS regions anchors units bought (a) regions anchors units bought AFR 631 1254 920 2805 43213 1.5% 2.9% 2.1% 6.5% EAP 319 636 490 1445 25792 1.2% 2.5% 1.9% 5.6% ECA 328 640 342 1310 23121 1.4% 2.8% 1.5% 5.7% LCR 246 566 343 1155 20867 1.2% 2.7% 1.6% 5.5% MNA 320 527 386 1233 12310 2.6% 4.3% 3.1% 10.0% SAR 271 566 710 1547 21466 1.3% 2.6% 3.3% 7.2% Regions 2115 4189 3191 9495 146769 1.4% 2.9% 2.2% 6.5% Inter-VPU cross support (% of buyer’s own managed Inter-VPU cross support bough (staff weeks) + cross support time) From From From From network other Total XS Total time From network From other Total XS regions anchors units bought (a) regions anchors units bought SDN 951 32 390 1373 19727 4.8% 0.2% 2.0% 7.0% HDN 506 9 175 690 5237 9.7% 0.2% 3.3% 13.2% PREM 71 4 235 310 4358 1.6% 0.1% 5.4% 7.1% FPD 220 33 94 347 5618 3.9% 0.6% 1.7% 6.2% Networks 1748 78 894 2720 34940 5.0% 0.2% 2.6% 7.8% TOTAL 3863 4267 4085 12215 181709 2.1% 2.3% 2.2% 6.7% Regions + networks Source: World Bank Database. Note: (a) Total time=time worked within the VPU i.e. own managed plus cross support to VPU. Cross support from anchors to Regions has decreased overtime. Total anchor staff time went from 24,537 in FY02 to 36,807 staff weeks in FY10. Yet, as shown in Figure H.1., the growth of the anchors has not come with an increase in the percentage of staff time dedicated to support the Regions. There has instead been an 18 percent decrease in total cross support (in staff weeks). The percentage of cross support from anchors to the Regions has also decreased (from 23 percent anchors’ staff time in FY02 to 12.5 percent overall in FY10). Appendix H 203 Table H.4 Inter-Vice Presidential Unit Cross Support from Regional Sector Departments to Network Anchors, FY10 Cross support sold (% total regional sector Cross support sold to Network Anchors (staff weeks) units’ time) Sub- Total regional Sub- To To To total ­sector units total To SDN HDN PREM FPD XS time To SDN To HDN To PREM To FPD XS SDN in the 791 18 18 5 832 50334 1.6% 0.0% 0.0% 0.0% 1.7% Regions HDN in the 20 469 1 3 493 22571 0.1% 2.1% 0.0% 0.0% 2.2% Regions PREM and FPD 84 8 38 207 337 28130 0.3% 0.0% 0.1% 0.7% 1.2% in the Regions TOTAL 895 495 57 215 1662 101035 0.9% 0.5% 0.1% 0.2% 1.6% Source: World Bank Database. Table H.5 Inter-Vice Presidential Unit Cross Support from Network Anchors to Regional Sector Units, FY10 Cross support sold to regional sector units (staff Cross support sold (% total Network Anchor’s weeks) time) From From From From Sub-total From From From From Sub-total SDN HDN PREM FPD XS SDN HDN PREM FPD XS SDN in the 1895 62 42 90 2089 9.0% 1.2% 0.9% 1.5% 5.7% Regions HDN in the 105 532 14 12 663 0.5% 9.9% 0.3% 0.2% 1.8% Regions PREM and FPD 356 103 271 531 1261 1.7% 1.9% 6.1% 8.9% 3.4% in the Regions TOTAL 2356 697 327 633 4013 11.2% 13.0% 7.3% 10.6% 10.9% Source: World Bank Database. Intra-VPU Cross Support In FY10, intra-VPU cross support sold (estimated in four Regions and network anchors for which comparable data was available) was 13.8 percent of staff time overall: 15.6 percent for Regions and 8.6 percent for network 204 The Matrix System at Work Figure H.1 Cross Support from the Anchors to the Regions, FY02-FY10 35.0% % cross support sold 30.0% 25.0% SDN 20.0% HDN 15.0% PRM 10.0% FPD 5.0% 0.0% 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 Fiscal Year Source: World Bank Database. anchors. Variations across networks and Regions were significant: intra-VPU cross support amounted to 10 percent staff time in the SDN anchor but barely reached 0.6 percent staff time in the PREM anchor; intra-VPU cross support was about 11.5 percent in the Africa Region, yet was as high as 19.3 percent in Latin America and the Caribbean. Total cross support sold, i.e. inter-VPU plus intra-VPU cross support (for four Regions and network anchors) amounted to 19.4 percent of staff time: 18.9 percent for Regions and 21.1 percent for network anchors as shown in Table H.6. Variations across Regions and networks were again considerable: total staff time dedicated to cross support was about 22 percent for South Asia and Latin America and the Caribbean, but barely reached 15 percent for Africa;3 SDN sold 23 percent of staff time to other units of the Bank, while PREM sold 10 percent of staff time. Intra-Regional Cross Support A closer look at the composition of intra-regional cross support only stresses cross-regional differences. Table H.7 shows the distribution of intra-regional cross support over the period FY02-10. Africa, East Asia and the Pacific, Latin America and the Caribbean and South Asia combined sold on average of about 17 percent intra-regional cross support over the period FY02-10: intra-regional cross support was 21.5 percent in Latin America and the Caribbean, about 19 percent in South Asia, 15 percent in the Africa Region and 17 percent in East Asia and the Pacific over the period FY02-10. Most intra-regional cross support sold (i.e. 8 percent of regional staff time) was cross support from sector units to other sector units in the Region. Inter- Appendix H 205 Table H.6 Total Cross Support Sold by Regions and Network Anchors, FY10 Total cross support sold (staff weeks) Total cross support sold (% of seller’s total time) Inter-VPU XS Intra-VPU XS Total XS Total time Inter-VPU XS Intra-VPU XS Total XS AFR 1351 4776 6127 41758 3.2% 11.4% 14.7% EAP 803 4249 5052 25150 3.2% 16.9% 20.1% LCR 703 3934 4637 20415 3.4% 19.3% 22.7% SAR 729 3847 4576 20648 3.5% 18.6% 22.2% 4 Regions 3586 16806 20392 107971 3.3% 15.6% 18.9% Total cross support sold (staff weeks) Total cross support sold (% of seller’s time) Inter-VPU XS Intra-VPU XS Total XS Total time Inter-VPU XS Intra-VPU XS Total XS SDN 2663 2155 4818 21017 12.7% 10.3% 22.9% HDN 808 362 1170 5356 15.1% 6.8% 21.8% PREM 427 27 454 4476 9.5% 0.6% 10.1% FPD 686 639 1325 5958 11.5% 10.7% 22.2% Networks 4584 3183 7767 36807 12.50% 8.6% 21.1% TOTAL 4 8170 19989 28159 144778 5.6% 13.8% 19.4% Regions + Networks Source: World Bank Database. Note: Europe and Central Asia and the Middle East and North Africa underwent unit restructuring during FY10; reliable data on cross support within the Middle East and North Africa and Europe and Central Asia was unavailable. In East Asia and the sectoral cross support could not be entirely isolated from Pacific, because of the unique structure of SDN sector units, inter-­ cross support across countries. sectoral cross support was 10.1 percent in East Asia and the Pacific, 9.5 percent in Latin America and the Caribbean and 5.6 percent in Africa. Yet, inter-sectoral cross support was most significant within network boundaries. There was, in particular, hardly any inter-sectoral cross support between SDN sector units and sector units of other networks. The Social, Environment and Water sector unit,4 and the Governance and Public Sector unit have been the biggest providers of inter-sectoral cross support within the South Asia Region over the period FY02-10: they sold respectively about 42 percent and 15.6 percent staff time to other sector units in the Region. The Social, Environment and Water sector unit sold 36 out of 42 percent cross support to other SDN sector units: 15 percent to the Agriculture and Rural Development sector unit, 14 percent to the Energy sector unit, 5 percent to the Transport unit and 3 percent to the Urban and 206 The Matrix System at Work Table H.7 Intra-Regional Cross Support Sold, Average FY02-10 Cross support sold (staff weeks) - Average FY02-10 Cross support sold by sector units Cross support sold by country units central Sector units Sector Country Country units to to other Sector units to units to Country units to sec- sector units units to regional Sub- other units to regional Sub- tor and in the region country central total XS country sector central total country Total (a) units units sold units units units XS sold units XS time AFR 2001 592 152 2745 305 146 57 508 2013 35681 EAP 2070 541 59 2670 543 168 15 726 85 20505 LCR 1730 247 323 2300 243 316 24 583 1025 18180 SAR 1538 271 86 1895 594 70 19 683 659 17296 Regions 7339 1651 620 9610 1685 700 115 2500 3782 91662 FY02-10 Cross support sold (% of seller’s time) - Average FY02-10 Cross support sold by sector units Cross support sold by country units Central Sector Country Country Country units to Sector units Sector units to units to units units to sec- Total to other sec- units to regional Sub- other to regional Sub- tor and intra- tor units in country central total XS country sector central total country VPU the region units units sold units units units XS sold units XS XS AFR 5.6% 1.7% 0.4% 7.7% 0.9% 0.4% 0.2% 1.4% 5.6% 14.8% EAP 10.1% 2.6% 0.3% 13.0% 2.6% 0.8% 0.1% 3.5% 0.4% 17.0% LCR 9.5% 1.4% 1.8% 12.7% 1.3% 1.7% 0.1% 3.2% 5.6% 21.5% SAR 8.9% 1.6% 0.5% 11.0% 3.4% 0.4% 0.1% 3.9% 3.8% 18.7% Regions 8.0% 1.8% 0.7% 10.5% 1.8% 0.8% 0.1% 2.7% 4.1% 17.3% FY02-10 Source: World Bank Database. Note: Cross support from sector units to other sector units in the Region does not include staff weeks from sector depart- ments’ front offices to sector units and from sector units to sector departments’ front offices. Water sector unit. It only sold 3.5 percent cross support to HDN sector units. The Governance and Public Sector unit sold 14.1 out of 15.6 percent inter- sectoral cross support to the Economic Policy and Poverty sector unit and 1 percent to HDN. A similar pattern is observed in Latin America and the Caribbean, where the Agriculture and Rural Development sector unit sold 19 out of 20 percent of inter-sectoral cross support to other SDN sector units and 14.5 percent to the Environment sector unit alone. The Latin America and the Caribbean Social Appendix H 207 Development unit sold 42 out of 50 percent of inter-sectoral cross support to other SDN sector units, (12 percent to the Agriculture and Rural Development unit and 15.5 to the Environment sector unit), yet only 1.7 percent of its staff time to the Poverty sector unit and 1.8 to the Social Protection sector unit. The Post-Conflict and Social Development, the Environment, and the Water Resource Management sector units were the three biggest sellers of inter- sectoral cross support in the Africa Region over the period FY02-10. They sold 33, 22, and 15 percent staff time on average over the period FY02-10. The Post–Conflict and Social Development sector unit sold 27 out of 33 percent of staff time to other SDN units, 1 percent cross support to Africa PREM sector units and 4 percent cross support to the HD sector units. The Water Resource Management sector unit sold about 15 percent inter-sectoral cross support overall, all of which were sold to SDN sector units. There may be fewer constraints for inter-sectoral cross support between PREM and HD sector units. The Latin America and the Caribbean Social Protection sector unit sold 14 percent staff time to other HDN sector units, 7 percent to PREM sector units, yet only 0.1 percent to the Social Development sector unit and 0.3 percent to SDN sector units overall. The Africa Public Sector Reform sector unit sold 12.2 percent inter-sectoral cross support overall: about 6 percent to other PREM sector units, 3 percent to HD sector units and 2.5 percent to SDN sector units There were substantial variations in levels of cross support from regional central units (i.e. units from the regional functions node including Resource Management, Procurement, the Office of the Chief Economist, Operations Services and regional program units) to sector and country units over the period: central units provided a maximum of 5.6 percent cross support in Latin America and the Caribbean and Africa, yet only 0.4 percent cross support in East Asia and the Pacific. Conclusions Three major conclusions can be drawn from this analysis. • First, although levels of inter-VPU cross support calculated as amounts and percentages of GF+ staff time are significantly higher than man- agement figures, they are on the whole quite low: “Real� inter-regional cross support was 0.9 percent in FY10, cross support across network anchors was 0.2 percent, and cross support from anchors to Regions 11.4 percent. • Second, intra-VPU cross support was on average significantly higher than inter-VPU cross support throughout the period FY02-10, mostly as a result of cross support from sector units to other sector units in the Regions (i.e. intra-regional inter-sectoral cross support). Yet, there was virtually no cross support from SDN sector units to sector units of 208 The Matrix System at Work other networks, and not much cross support from HDN sector units to PREM and FPD sector units (and vice versa). This suggests the exis- tence of strong barriers to cross support outside network boundaries. • Finally, although the time trend in inter-VPU cross support sold (in staff weeks) over the period FY02-10 has by and large been flat, there has been a decrease in inter-VPU cross support sold as a percentage of sellers’ time as the Bank has staffed up. Cross support from anchors to Regions, in particular, has decreased over time. Notes 1 See “Transforming the Bank’s Knowledge Agenda—A Framework for Action�, (World Bank 2010) discussed with the Executive Directors at the informal meeting held on March 9, 2010, and “Current State of Cross Support,� a presentation to staff as part of the Reform Secretariat December 2010 cross support retreat. 2 Although there is in practice more than one anchor unit within each network VPU, we will hereafter refer to the network anchor as a whole instead of each Sector Board’s anchor unit within their respective network VPUs. 3 The lower level of intra-VPU cross support for the Africa Region can in part be explained by the change in budget settlement rules for fiduciary and safeguard services effective FY10. They are no longer considered cross support but are now part of the WPA. This led to a drop in cross support from regional central units to SMUs and CMUs from 7.1 percent in FY09 to 1.8 in FY10. 4 South Asia’s Social Development sector unit was separated from the Social, Environment and Water sector unit during the period, but in this analysis we consider this group a single unit to maintain consistency. Appendix H 209 Appendix I Budgetary Incentives As part of the Independent Evaluation Group’s (IEG’s) assessment of the Bank’s matrix system, the Bank’s budget was reviewed to better understand the impacts of budgetary incentives on the effectiveness of the matrix. Information was primarily obtained from the Bank’s budget data and the analysis was informed by interviews with sector managers, country management teams, regional operations teams, chief administrative officers, and resource management staff. This section describes the components of the Bank’s budget and contextual background before highlighting the key findings further below. Bank budget resources are made up of three sources of finance: 1) the Bank’s administrative budget which is approved by the Board, 2) reimbursable and fee income which is received from clients for services rendered such as externally financed outputs, fee based services, and trust fund management fees, and 3) bank-executed trust funds (BETFs) which are funds received from donors to be used alongside the Bank’s administrative budget but earmarked for specific activities.1,2 It should be noted that only the Bank’s administrative budget is negotiated to the unit level with reimbursables, fee income, and BETFs to varying degrees considered, but not incorporated into the budget negotiation process and therefore their amounts are largely an artifact as opposed to a strategic allocation of the full funds available to conduct Bank activities. Unless otherwise noted, all figures in this appendix refer to total sources of finance and are listed in 2010 constant US dollars.3 As shown in Figure I.1, the Bank’s matrix related budgeting process involves allocations to each regional and network Vice Presidential Unit (VPU) which are then negotiated and allocated within each VPU according to strategic priorities and work program agreements. For the regions, the budgeting process flows from the regional vice presidents to the country units (country directors, country management units, CMUs) with some funds held back for regional functions such as the chief economist which is responsible for regional studies and regional Operations Policy and Country Services (OPCS) functions. The country units, in representing client needs and ensuring demand for services, purchase services from sector units (SMUs) through a work program agreement (WPA) while holding some funds back for administrative expenses and contingency. Once WPAs are in place for the fiscal year, the budget is released to sector units for completing product delivery to country clients with country units and regional functions providing limited oversight and monitoring of the agreed upon outputs. BETFs on the other hand are typically allocated directly for individual products and mobilized directly by associated task leaders, rather than through the Bank’s strategic allocation process, which relies on the tensions between country and sector units to balance the needs of both sides of the matrix structure. Budgetary arrangements under the matrix system were singled out by staff as well as country and sector managers as a major impediment to the Bank’s efficiency. The imbalance in budgetary authority between country units and sector units leads to a short-term focus driven by lending delivery and fiscal year deadlines, rather than strategic selectivity that balances short- 212 The Matrix System at Work Figure I.1 Matrix-Related Bank Budget Allocation and BETF Mobilization Flow Quite Differently (FY10 Figures)a,b RVP Totals (BB $1,183m) (BETF $257m) Reg Functions CD/CMUs SD/SMUs (BB $213m) (BB $246m) (BB $724m) (BETF $9m) (BETF $20m) (BETF $227m) BB Total BETF Total (BB $2,462m) (BETF $508m) Regional OPCS (BB $95m) (BETF $5m) To Other Units Chief To Other Units (BB $1,069m) Economist (BETF $74m) Sector Depts Other NVP Totals (BB $210m) (BETF $177m) Source: World Bank databases. a Solid lines highlight the Bank’s budget allocation system whereas the dotted lines indicate unofficial and potential trust fund sources of financing typically ­mobilized directly rather than through a systematic allocation process. b Figures in italics are amounts held within that unit as opposed to total amounts flowing through a given unit. term needs with business development to prepare for future risks or policy opportunities. The matrix was supposed to reduce the focus on lending in favor of a better balance with quality and knowledge, but it has not done so. Budget Expenditures Across the Matrix Structure Overall Resources for Frontline Delivery Greater transparency on overheads would increase accountability and encourage efficiencies. Out of concern for the decline in frontline resources prior to the introduction of the matrix, management had committed, as part of the Strategic Compact, to an increase in frontline expenditures to 60 percent of the budget (Phillips 2009: 38). The first indicator on organizational effectiveness and modernization in the corporate scorecard issued at the end Appendix I 213 of FY11 (World Bank for Results 2011) states that 61 percent of the World Bank’s resources (Bank budget plus BETFs) are spent on “client services.� However, the 61 percent includes trust fund resources (9 percent), as well as resource mobilization and management (5 percent), primarily for trust fund management and treasury functions. Adjusting for these two, the amount spent on client services drops to 47 percent of the Bank’s budget. “Client services� covers a broad range of other expenditure items with differing levels of separation from the end product delivery to the client. These include country services (making up 32 percent of Bank budget), sector and global services (12 percent), other operational support (2 percent), and administrative support to products and programs (1 percent). Frontline expenditure on country services is 32 percent of the Bank’s budget, or one in three dollars, for direct costs of lending and non-lending products delivered to country clients. Budget Expenditures Within the Regions Over the FY02-10 period, sector units have received a declining share of the Bank’s budget, largely due to the growth of regional functions. When considering regional budgets, the share of Bank budget expended by regional sector units declined by 5 percent, receiving only a 61 percent share (US$ 724 million) of the Bank’s budget in FY10 compared to a 66 percent share (US$ 681 million) in FY02. While not to the same extent, CMUs also lost in share of regional Bank budget decreasing from a 23 percent share (US$ 231 million) in FY02 to a 21 percent share (US$ 246 million) in FY10. Under the matrix system, all the technical staff, including those based in country offices, are mapped to and paid for by the sector units, and these charges include their overheads (sustaining costs). That being the case, it is unclear why CMUs hold back one-fifth of regional budgets. The remainder of the Regions’ Bank budget expenditures was expended by regional functions which almost doubled in size from US$ 117 million in FY02 to US$ 213 million in FY10, or a growth of 7 percent in terms of budget share from 11 to 18 percent share (see Figure I.2 and Table I.1 below). Regional variation shows sector units in Africa and the Middle East and North Africa and CMUs in Latin America and the Caribbean and South Asia receiving the lowest share of regional Bank budgets in FY10. SMUs in Africa and the Middle East and North Africa received 58 and 59 percent of regional Bank budgets respectively and Europe and Central Asia, Latin America and the Caribbean, and South Asia received the highest at 64, 64, and 63 percent, respectively. CMUs in Latin America and the Caribbean and South Asia received the lowest share of regional Bank budget in FY10 at 17 and 18 percent respectively while the Middle East and North Africa received the highest share at 25 percent. Regional functions in Europe and Central Asia, East Asia and the Pacific, and the Middle East and North Africa received the lowest share of Bank budget in FY10 at 14, 16, and 16 percent, respectively, while Africa and South Asia received the highest share of Bank budget at 21 and 19 percent respectively. The trend of growth in regional functions and declining share of SMU budgets is found in all the Regions. There thus appears to be a strong 214 The Matrix System at Work Figure I.2 Regional Sector Units Are Delivering More With Less Bank Budget as Overheads Grow 1,200 (Constant 2010 US$ millions) 15% 17% 17% 18% 1,000 14% 16% 16% 11% 14% 21% 21% 21% 22% 21% 21% 21% 800 23% 22% Regional Functions BB Actuals 600 CMU BB Actuals 400 SMU BB 66% 64% 65% 63% 64% 62% 62% 62% 61% Actuals 200 0 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Source: World Bank databases. Table I.1 SMUs and CMUs Are Receiving a Decreasing Share of Bank Budget Funds Managed by the Matrix System (Constant 2010 US$ Millions) % Increase (FY02 to Share of Fund Center FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 FY10) Matrix Anchor 169 184 195 196 193 188 199 207 210 25% 15% RVP Total 1,029 1,037 1,105 1,123 1,132 1,124 1,156 1,157 1,183 15% 85%   Regional Functions 117 145 157 177 169 184 201 202 213 82% 13%  CMU 231 225 231 236 243 240 240 243 246 6% 18%  SMU 681 667 718 710 720 701 715 712 724 6% 54% Matrix Total 1,198 1,221 1,301 1,319 1,325 1,312 1,356 1,364 1,393 16% 100% % SMU of Matrix Total 57% 55% 55% 54% 54% 53% 53% 52% 52% -5% Source: World Bank databases. association between budget authority and the share of resources withheld, first at the regional vice presidency level for regional functions, then at the CMU level. Regional sector units have also received a decreasing share of the Bank’s budget compared to the growth found in regional and network budgets. When the budgets of organizational units that fall directly within the ambit of the matrix system are compared, regional functions and network anchors come out as winners, while the sector units are the losers. The share of Bank budget Appendix I 215 going to the sector units in the Regions had declined from 57 to 52 percent of the budget administered by organizational units inside the matrix (see Table I.1). If one were to include Bank-executed trust funds, the story changes only slightly. As lending and grant portfolios continue to grow, trust funds assist SMUs in managing their portfolio in a flat budget environment. Despite new lending and grant commitments more than doubling in the wake of the dual food and financial crisis (from approximately US$ 20.5 billion in FY08 to US$ 48 billion in FY10 in current US dollars), the Bank’s administrative budget has remained almost flat in 2010 constant terms.4 To fill the remaining gap in funding, SMUs may attempt to mobilize trust funds. However, even when trust funds are considered SMUs have still lost approximately three percent in share of regional budget resource expenditures from FY02-10 (see Figure I.3 and Table I.2). Sector managers are almost entirely reliant on budget allocations from each country director and are accountable for all project- and program- level results in their sector(s), across all countries in their Region. Several country directors were under the impression that staff salaries were paid off- the-top. In reality this is true only for CMU staff, managers, and for the staffs of the financial management, procurement, and legal units. The budget for technical sector staff comes from the WPAs, and all staff members have to account for and charge their entire time and travel costs to product codes in the WPA. Sector managers have very limited variable budget (usually $50,000 a year) to finance their own travel or activities at their own discretion. This compels them to seek trust funds to finance any additional activities that are a priority for their sector but have not been financed by the country directors. Table I.2 SMUs Are Still Receiving a Decreasing Share of Budget Resources (Bank Budget Plus Bank-Executed Trust Fund) Managed by the Matrix System (Constant 2010 US$) % Increase (FY02 to Share of Fund Center FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 FY10) Matrix Anchor 239 274 291 296 302 315 346 370 387 62% 20% RVP Total 1,143 1,152 1,239 1,279 1,287 1,301 1,348 1,371 1,440 26% 80%   Regional Functions 122 150 166 189 177 191 209 211 222 83% 11%  CMUs 236 229 236 245 254 256 257 261 266 13% 16%  SMUs 785 773 837 844 856 853 881 900 951 21% 53% Matrix Total 1,383 1,425 1,530 1,575 1,589 1,616 1,693 1,741 1,827 32% 100% % SMU of Matrix 57% 54% 55% 54% 54% 53% 52% 52% 52% -5% Source: World Bank databases. 216 The Matrix System at Work Figure I.3 With the Inclusion of Bank-Executed Trust Funds Sector Management Units Are Still Being Squeezed 1,600 (Constant 2010 US$ millions) 1,400 15% 16% 15% 1,200 13% 15% 14% 15% 10% 13% 19% 1,000 20% 20% 19% 19% 19% 19% 21% 20% Regional 800 Functions Total 600 CMUs Total 66% 65% 65% 66% 66% 400 69% 67% 68% 66% SMUs Total 200 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Source: World Bank databases. The Growth of Regional Functions The growth in regional functions has been driven by regional core operations services. Growth was found across the majority of individual units located within regional functions. However, in FY10 the regional core operations services units made up approximately 45 percent of the Bank’s budget towards regional functions and explained almost half (48 percent) of the growth over the period. These units are essentially an extension of Operations Policy and Country Services (OPCS) and provide procurement, financial management, quality assurance, and other operational support to the Regions although they, like all units within regional functions are not directly linked to product delivery. Spending on Project Appraisal and Supervision As a smaller share of the Bank’s budget is going to the front lines, the Bank is spending significantly less to develop projects on a commitment dollar basis with earlier findings indicating that technical quality does not receive adequate attention. Bank budget spending on project appraisal increased on a per operation basis from an average of current US$ 358,000 for all operations in FY02-04 to US$ 378,000 in FY08-10, an increase of 6 percent, potentially indicating the Bank is now spending more on project appraisal.5 However, over the same period the average new commitment amount has almost doubled from current US$ 81 million to US$ 145 million overall. This resulted in appraisal expenditures per US$ 1,000 commitment dollars for all operations decreasing by more than one-third (35 percent) from a current US$ 4.44 in FY02-04 to US$ 2.87 in FY08-10. On the other hand, the Bank is spending more on supervision and implementation support. As was found with project appraisal, per project Appendix I 217 spending on project supervision has also increased from a per operation average of current US$ 76,000 for all operations in FY02-04 to US$ 104,000 in FY08- 10, an increase of 37 percent. Additionally, as the average net commitment of an operation has increased from current US$ 71 million in FY02-04 to US$ 93 million in FY08-10. Project supervision on a commitment dollar basis has also increased from US$ 1.08 per US$ 1,000 of open commitments in FY02-04 to US$ 1.14 in FY08-10, an increase of 6 percent, indicating the Bank is spending more on supervision. Both the appraisal and supervision trends for overall operations hold true for investment lending (see Table I.3. below) and support views expressed in sector manager and country director interviews that there is less focus on quality up front and that recent high-profile cases in which fiduciary or safeguards oversight has come into question has resulted in additional project supervision of those aspects.6 However, findings on quality of supervision indicate increased budget alone does not suffice to address quality issues. Impact of External Financing The Bank has become more reliant on external financing, the allocation of which falls outside the Bank’s annual budgeting process under the matrix system, and which depends on the tensions between country and sector units Table I.3 Bank Spending on Appraisal Is Down on a Per Commitment Dollar ­ upervision Spending Has Risen on All Accountsa Basis While S FY02–04 Averages FY08–10 Average Average cost per US$s per US$1000 of Average cost per US$s per US$1000 of Lending Instrument project (US$1000s) commitmentb project (US$1000s) commitment Project Appraisal DPLs $388 $2.37 $287 $1.41 ILs $352 $5.80 $409 $3.73 All projects $358 $4.44 $378 $2.87 Project Supervision DPLs $78 $0.44 $53 $0.28 ILs $76 $1.19 $108 $1.29 All projects $76 $1.08 $104 $1.14 Source: World Bank databases. a  Figures represent Bank budget and International Bank for Reconstruction and Development (IBRD)/International Deveo- pment Association (IDA) lending and grants. Complete data not available for trust funds. b  Commitment figures refer to new commitments for project appraisal and net open commitments throughout a given year for supervision. 218 The Matrix System at Work to balance the needs of both sides of the matrix structure. External financing supports a growing proportion of Bank resources (Bank  budget and BETFs), growing from an annual average of 19 percent in FY02-04 (11 percent for BETFs and 8 percent for reimbursables and fee income) to an annual average of 26 percent in FY08-10 (16 percent for BETFs and 10 percent for reimbursables and fee income).7 In the regions, the BETF share of Bank resources has increased from 10 to 16 percent while the network anchors share of BETFs has increased even more from 32 to 44 percent over the same time period. From FY02-04 to FY08- 10, BETF share of budget resources for research VPUs DEC and WBI together grew from 23 to 26 percent- not including non-BETF external financing whose funds are comingled with the Bank’s budget at levels below Bank totals (see Figure I.4). A further breakdown into the different VPUs highlights the reliance on trust funds for the individual regions and network anchors with the East Asia and the Pacific and Africa Regions and the SDN anchor relying most on trust funds to deliver on their respective mandates (see Figure I.5). The significant and growing reliance on trust funds brings into question the strategic intent of a large portion of client-facing Bank resources which are “selected mainly by donors, rather than through a systematic programming process.�8 IEG’s trust fund evaluation raises concern over how the funding of issues is determined and who is driving the Bank’s agenda; however, noting the impossibility of determining the degree to which this may take place. While it is not the intent of this evaluation to assess the allocation process of trust funds, there is cause for concern over their impact on matrix effectiveness given such a large proportion of funds are allocated outside its budgeting process and therefore outside the confines of the associated incentive mechanisms developed specifically for the matrix. Network anchors have become more reliant on trust funds than the Regions. In FY08-10, the Regions received the same proportion of overall BETF funding as they did overall Bank budget funding (49 percent). Network anchors by contrast received four times as much BETF funding (36 percent) as their proportion of overall Bank budget (9 percent). This means that during the FY08-10 period, network anchors generated 79 cents for each US dollar in Bank budget versus only 19 cents in the Regions (see Figure I.6a and b). The growth in BETF financing of the network anchors has been channeled through a few trust funds. The top five BETF programs disbursing through network anchors are the water and sanitation program, energy sector management assistance program, carbon fund, global facility for disaster reduction and recovery, and consultative group to assist the poorest (CGAP). Together, these programs make up almost half of FY10 BETF anchor disbursements. Across the anchors, SDN controls over 70 percent of BETFs including four of the top five aforementioned BETF programs (CGAP is disbursed through FPD). Access to trust funds gives the network VPUs a degree of autonomy to finance sector and corporate priorities such as gender and governance and global priorities such as climate change, whose externalities may extend beyond individual client Appendix I 219 Figure I.4 Trust Funds Represent a Substantial and Growing Share of Budget Resources for Anchors, Regions, and Research VPUs 1,400 BETFs 16% (Constant 2010 US$ millions) 1,200 10% 1,000 2% 800 Bank Budget 1% 600 84% 90% 400 32% 98% 44% 99% 200 23% 26% 68% 56% 77% 74% 0 o/w Regions o/w Network Anchors o/w DEC/WBI o/w Othera FY02–04 Avg FY08–10 Avg Source: World Bank databases. a  “Other�here includes Corporate, Finance, Administration and Other Operations (OPCS and QLP), Central Accounts, and Other Units. ­ countries and which may not otherwise be integrated into country programs. However, trust funds contribute to the decline in cross support provided by network VPUs to the Regions.9 Figure I.5 Regions Africa and East Asia and the Pacific and Anchor Sustainable Development NetworkHave the Most Reliance on Trust Funds (FY08-10 Avgs) 450 (Constant 2010 US$ millions) BETFs 400 20% BB Actuals 350 300 250 7% 6% 200 28% 15% 80% 13% 150 52% 100 93% 94% 39% 72% 85% 37% 25% 50 87% 48% 61% 63% 75% 0 EAP AFR SAR MNA LCR ECA SDN FPD HDN PREM Source: World Bank databases. 220 The Matrix System at Work Figure I.6a and b Network Anchors Receive a Much Higher Proportion of BETFs Than Bank Budget* FY08-10 Budget Actuals FY08-10 BETFs ($2,375 milllion annual average in ($446 million annual average in 2010 constant US$) 2010 constant US$) 1% 8% 1% 3% 0% 1% 2% 9% Regions 13% Network Anchors DEC/WBI Corporate 49% 49% Finance 8% Administration Other Operations 6% Other 36% 5% 9% Source: World Bank databases. * Other included Central Accounts and Other Unit expenditures. Other Operations includes OPCS and QLP. Overall, network anchors spend more on external partnerships and resource mobilization combined, than on core knowledge products. Over the FY08- 10 period, network anchors spent an average of 38 percent of budget resources on external partnerships10 (annual average of US$ 115 million) and resource mobilization11 (annual average of US$ 23 million) compared to only 35 percent on core knowledge products (annual average of US$ 128 million). Additionally, of the anchors’ core knowledge12, barely over half represents products for external clients meaning that only 19 percent of network anchors budget resources support Bank clients. This supports views expressed in sector manager interviews that a big part of the anchors’ role is to support the corporate center and external partnerships instead of supporting country client needs.13 Budget Incentives for Cross Support Budget rules hinder staff from working across organizational boundaries. Current arrangements require units to estimate the amount of cross support they expect to sell and purchase in the coming year. Even though the Bank’s time recording system captures actual cross support across organizational units in real time, deviations from the beginning-of-year estimate for inter-VPU cross support are only reconciled through adjustments in budget for next fiscal year, or in limited circumstances twice per fiscal year. Appendix I 221 Each VPU is free to set internal rules for intra-VPU budget settlements, but in some cases cross support within the same sector department may not involve budget settlements at all. This non-payment or substantial delay between the provision of cross support and payment for that work penalizes a unit for providing unanticipated cross support. On the demand side, a disincentive arises from the significant differential between the cost of staff and cost of consultants due to the markup charged to purchasing units for sustaining costs of staff.14 Any initiative to jump-start cross support will need to address these penalties. Management acknowledges that the current system of cross support is complex and transaction-intensive and favors a more simplified and transparent mechanism; however, given ongoing matrix and knowledge reforms, and the FPD pilot, institution-wide reform is likely to wait at least until FY13. The two binding constraints of cross support that the FPD pilot is trying to address are (a) the perceived high cost of staff compared to consultants, and (b) the time lag between the period when cross support is provided and when budget settlement for cross support takes place (see The World Bank’s Budget: Trends and Recommendations for FY12, August 2011).15 Given the widespread recognition of the dampening effect of these budget incentives it is not clear why management wants to wait for the FPD pilot to provide lessons. Notes 1 Trust Fund Support for Development: An Evaluation of the World Bank’s Trust Fund Portfolio (IEG 2011). 2 This analysis does not include other types of trust funds including recipient- executed trust funds (RETFs) which are executed in accordance with a grant agreement between the Bank and the recipient and financial intermediary funds (FIFs) which transfer funds to third-party recipients based on instructions from a governing entity and which the Bank does not provide operational oversight. 3 Throughout this appendix budget figures are reported from the Bank’s cost reporting system as assigned to the fund center (budget holder). As they are not disaggregated in the Bank’s reporting system, Bank budget figures include both actual expenditures from the Bank’s budget as agreed by the Bank’s Board and non-BETF external financing (reimbursables and fee income). All budget figures are reported in 2010 constant US dollars (millions) using the Bank’s corporate finance unit’s deflator. 4 The Bank’s gross administrative budget grew at a compound annual growth rate of three percent with the commitment to return to flat budget spending reaffirmed in FY11 (The World Bank’s Budget: Trends and Recommendations for FY11, World Bank 2010). 5 Figures for Bank budget and IBRD/IDA lending and grants only. 6 Findings from sector manager interviews. See Appendix E. 222 The Matrix System at Work 7 Figures for this calculation are taken from the Bank’s expenditure report as it provides comprehensive figures on reimbursables and fee income. 8   Trust Fund Support for Development: An Evaluation of the World Bank’s Trust Fund Portfolio (IEG 2011) 9   Findings from cross support analysis. See Appendix H. 10 Includes maintaining external partnerships, outreach/partnerships, global advocacy, partnerships. 11 Includes trust fund administration, trust fund program management, project finance and guarantee operations, and IDA replenishment activities. 12 Core knowledge products are defined by the knowledge strategy group as products for external clients (economic sector work, non-lending technical assistance, external training, and impact evaluation), products for internal use (knowledge products, external and internal knowledge management, and new product development), and research, global monitoring, and data (research services, global monitoring and data, and world development report). 13 Findings from sector manager interviews. See Appendix E 14 These budget disincentives were specifically identified by the FPD network, which is undertaking a pilot experiment to address them. See Budget Report to the Board for FY12. 15 See The World Bank’s Budget: Trends and Recommendations for FY12. August 9, 2011, page 39. Reference Phillips, David A. 2009. Reforming the World Bank: Twenty Years of Trial and Error. New York: New York University Press. Appendix I 223 Appendix J Working across Sectors and Themes The Bank’s thematic and sector coding system forms the basis for analy- sis of and standard reporting on the nature of Bank interventions. Sector codes roughly mirror the United Nations classification of economic sectors and help identify the sector(s) of the country clients’ economy Bank projects are designed to support. Thematic codes are meant to capture Bank support to corporate priorities such as Millennium Development Goals and global public goods. Each project is mapped to a single administrative sector board, which is then responsible for quality control and management of the product. Some sec- tors, themes, and Sector Boards are identified by the same or similar names. For example, gender is both a Sector Board and a theme; water is a Sector Board, water supply is a sector, and water resources management is a theme. At least one sector and one thematic code is assigned to all client-oriented Bank activities (mainly lending operations, economic and sector work, and technical assistance) by task teams at the appraisal stage. A total of five sec- tor and five thematic codes can be identified for a single activity. Each theme and sector code is coupled with a percentage indicative of the degree of focus of the Bank intervention on the designated sector or theme. The percentages reflect, by design, the allocation of Bank activity, both across economic sec- tors and toward the set of goals and objectives of the activity. For Investment Lending (IL), the percentages are derived from the project description and the cost table in the project appraisal document—an example for a $350 million Agriculture and Rural Development project in India (P089985) approved in 2010 is highlighted in Table J.1. For Development Policy Lending (DPL), the percent- ages assigned to each sector/theme must reflect the frequency to which the sector or theme occurs in the policy matrix. Projects with a variety of sector and thematic codes (including sector or thematic codes rarely associated with the Sector Board to which the proj- ect is mapped) do not necessarily involve cross support or collaboration of specialists from different sectors. The sector or thematic allocation does, Table J.1 Example of Sector and Thematic Coding for an Agriculture and Rural Development Project Sectors % Themes % Agriculture, fishing, and forestry 54 Social protection and risk management 50 (Agriculture and rural development) (Social protection) Water, sanitation and flood protection 32 Environment and natural resources management 50 (Water) (Environment) Public administration, law, and justice 14 (Public sector governance) Source: World Bank databases. Note: Sector Boards with which the sectors and themes are “affiliated�, as defined below, are noted in parentheses. 226 The Matrix System at Work however, indicate the extent to which task teams consider sectors and themes affiliated with other Sector Boards relevant to their respective projects. This annex presents the key findings of an analysis of time trends in the allocation of sector and thematic codes to Bank lending operations approved during the period FY95-10. The 76 current and nine expired sectors and the 82 current and three expired themes were each identified as affiliated with one or more Sector Boards using classifications prepared by OPCS. The Independent Evaluation Group (IEG) made one adjustment, considering those sectors and themes identified as affiliated with the Legal Sector Board (to which no operations are mapped) as affiliated with the public sector governance Sector Board. Some Sector Boards have are not affiliated with any sectors; other Sector Boards are not affiliated with any themes. Affiliated sectors and themes are defined as those associated with the Sector Board to which the project is mapped. Non-affiliated sectors and themes are sectors and themes affiliated with other Sector Boards. In the example above, 46 percent of the operation is allocated to a non-affiliated sector, and 100 percent is allocated to non-affiliated themes. This review compares trends in thematic allocations for each of the major lending instruments—investment lending (IL) and development policy lending (DPL) operations. Data specific to three corporate priorities—public sector governance (PSG), environment (ENV), and gender— is also analyzed to assess the degree to which they have been mainstreamed in the Bank’s portfolio before and after the introduction of the matrix system. ENV and gender were selected because they were identified in the 1997 as priorities for mainstreaming into the portfolio. PSG was selected because of the Bank’s focus on governance and anti-corruption issues in recent years. Results are compared in terms of the average percentage of operations assigned to a particular affiliated or non-affiliated sector or theme for projects approved during four time periods—fiscal years 1995-98, 1999-02, 2003-06, and 2007- 10. The first time period covers projects prepared before the introduction of the matrix system; looking at the others helps show patterns during different phases of implementation of the matrix system.1 Key Trends An expectation of cross-sector and cross-thematic work is typical, but declining, in the design of operations. On average, 45 percent of each operation approved by the Bank between FY1995-2010 was flagged as devoted to a sector unaffiliated with the Sector Board that governed the operation, and 56 percent was assigned to an unaffiliated theme. This share grew modestly upon introduction of the matrix system, to a peak of 47 percent for unaffiliated sectors and 61 percent for unaffiliated themes as shown in Table J.2, but fell substantially thereafter to 42 percent for unaffiliated sectors and 52 percent for unaffiliated themes. Appendix J 227 Table J.2 Bank-wide Allocation of Operations to Non-Affiliated Sectors and Themes Average percentage allocated to Average percentage allocated to Total projects approved non-affiliated sectors non-affiliated themes FY95-98 1184 46 58 FY99-02 1145 47 61 FY03-06 1299 47 56 FY07-10 1652 42 52 FY95-10 5280 45 56 Source: World Bank databases. The share of investment loans assigned to non-affiliated sectors was lower than the overall total—at 41 percent—and relatively flat until FY07, after which it began a steep decline to 33 percent for the three year average over FY08- 10, as shown in Figure J.1. The share devoted to unaffiliated themes was 55 percent for investment loans, about the same as the full portfolio. But this share among rose modestly to 61 percent for the three year average over FY00- 02, after which it began a steady decline, dropping to 46 percent for the three year average over FY08-10. The modest initial increase in the average share of a project devoted to themes not affiliated with the Sector Board to which the project is mapped coincided with the introduction of the matrix system, and may have reflected Figure J.1 Average Share of Investment Loans Allocated to Non-affiliated Sectors and Themes 70% Avg. allocation of investment lending to unaf�liated themes (3-year centered average) 60% 50% 40% Avg. allocation of operations to unaf�liated sectors 30% (3-year centered average) 20% FY95 FY96 FY97 FY98 FY99 FY00 FY01 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 Source: World Bank databases. 228 The Matrix System at Work initial enthusiasm for the new system. But this boost to cross-thematic work was short-lived, and as time passed, less and less attention was devoted to addressing these non-affiliated corporate priorities. The sharp drop in the as- signment of non-affiliated sectors to investment projects coincided with, but cannot be attributed to, the rapid scaling up of lending volume in response to the global financial crisis and followed shortly after the creation of the Sustainable Development Network (SDN), merging the two largest networks at the time in 2006. The scaling up of lending could plausibly be associated with the reduction in cross-sectoral content, as task teams did not have the extra time needed to introduce content into projects that fell outside their area of expertise. The creation of SDN, on the other hand, was expected to increase cross-sector work, as it removed an organizational barrier. But this was not the case in practice. Development policy loans are designed as more cross-sectoral and cross-thematic than investment loans, and increasingly so. The average pro- portion of development policy operations devoted to unaffiliated sectors and themes rose steadily over the period, exceeding 70 percent for both for the three year average over FY08-10. But even for DPLs, the policy focus inher- ent in the operation makes it easier to indicate linkages to multiple sectors without actual collaboration across sector units, as demonstrated in IEG’s evaluation of Poverty Reduction Support Credits (IEG 2010). Differences by Sector Board Excluding Sector Boards without associated sector codes (whose portfolio there- fore is allocated entirely to unaffiliated sectors), unaffiliated sectors are most prevalent in operations mapped to the urban development (91 percent) and social protection (50 percent) Sector Boards, and least prevalent in operations mapped to the energy and mining (15 percent) and transport (17 percent) Sec- tor Boards, as shown in Table J.3. Excluding Sector Boards without associated thematic codes, unaffiliated themes are most prevalent in operations mapped to the poverty reduction (91 percent) and water (84 percent) Sector Boards, and least prevalent in operations mapped to the health, nutrition, and population (15 percent) and environment (17 percent) Sector Boards. Mainstreaming of Corporate Priorities Table J.4 summarizes the extent to which environment, public sector gov- ernance, and gender were mainstreamed into the Bank’s portfolio of opera- tions. Among the three corporate priorities that were analyzed, PSG is the most prevalent in the portfolio, identified as a sector in about two-thirds of all operations not mapped to the PSG Sector Board over the full period and a theme in about one-third. Comparable figures for environment and social development and gender themes were 21 percent and 24 percent, respectively. The average allocation of operations to PSG as a non-affiliated theme in op- erations mapped to other Sector Boards was 10 percent over the period FY07- Appendix J 229 Table J.3 Time Trend of Cross-sectoral and Cross-thematic Work by Sector Board Percentage of operations allocated to non-affiliated sectors Approval FY ARD ED EMT ENV EP FPD GIC HE PR PSG SDV SP TR UD WA FY95-98 50 18 15 100 100 52 74 16 32 100 63 17 95 17 FY99-02 54 26 21 100 100 45 52 18 24 100 54 22 95 22 FY03-06 40 22 17 100 100 43 20 34 100 20 100 50 18 95 20 FY07-10 38 11 10 100 100 31 35 23 100 21 100 39 12 84 16 FY95-10 44 18 15 100 100 44 42 24 100 24 100 50 17 91 18 Percentage of operations allocated to non-affiliated themes FY95-98 63 58 84 33 67 41 100 31 42 51 60 84 53 81 FY99-02 65 60 89 44 80 43 100 36 41 50 59 83 58 88 FY03-06 61 40 80 30 80 47 100 26 92 34 63 61 78 47 85 FY07-10 44 23 73 21 86 40 100 17 98 25 47 59 75 35 83 FY95-10 57 44 80 32 79 43 100 27 95 36 54 60 79 45 84 Operations.(#) 772 440 446 358 390 465 36 403 60 382 90 336 482 308 301 Source: World Bank databases. Note: Sector Boards for which 100% of projects are assigned to non-affiliated themes do not have affiliated themes. ARD=agriculture and rural development; ED=education; EMT=energy and mining; ENV=environment; EP=economic policy; FPD=financial and private sector development; GIC=global information and communications technology; HE=health, nutri- tion and population; PR=poverty reduction; PS=public sector governance; SDV=social development; SP=social protection; TR=transport; UD=urban development; WA=water. 10, and to PSG as a non-affiliated sector was 19 percent; it was 7 percent for the environment theme and 4 percent for the social development and gender themes. Environment Based on the sector/thematic allocation, environmental themes do not ap- pear to have been mainstreamed across the Bank’s portfolio, and there is some evidence of a declining trend over time. ENV has been flagged as a theme in 19 percent of the operations approved during FY07-10, which is less than the 26 percent in FY95-98 period. Operations mapped to other Sector Boards allocated an average of 7 percent to the ENV theme in FY07-10, compared to 9 percent in FY95-98. Moreover, projects mapped to the ENV Sector Board are becoming more specialized over time as stand-alone environmental projects, rather than closely linked to activities in other sectors. Allocations to non-affiliated themes among operations mapped to the ENV Sector Board declined from an average of 33 percent during FY95-98 to 21 percent during FY07-10. 230 The Matrix System at Work Table J.4 Three Measures of Mainstreaming of Corporate Priorities Considers other Considered in operations of other Mainstreamed in operations of other themes in own Sector Boards Sector Boards operations Average percentage al- located to non-affiliat- Percentage of non-affiliated operations with ed themes of operations any portion assigned to following sector or Average percentage allocated of non-affiliat- mapped to following theme ed operations to following sector or theme Sector Boards Approval PSG as PSG as SDV & PSG as PSG as SDV & FY ENV sector theme Gender ENV theme sector Gender ENV PSG SDV FY95-98 26 79 31 25 9 9 20 7 33 42 * FY99-02 19 81 34 32 6 10 22 9 44 41 * FY03-06 22 70 37 27 8 12 20 7 30 34 63 FY07-10 19 48 29 16 7 10 16 4 21 25 47 FY95-10 21 68 32 24 8 10 19 6 32 36 54 Source: World Bank databases, IEG. Note: ENV=environment; PSG=public sector governance; SDV=social development. *There were too few operations mapped to the gender and social development Sector Boards in the first two periods to provide meaningful data. Public Sector Governance Public sector governance is the most prevalent of the corporate priorities among operations of other Sector Boards; this has not changed significant- ly over time. PSG has been flagged as a non-affiliated theme in 29 percent and a non-affiliated sector in 48 percent of the operations approved during FY07-10, compared to 31 percent and 79 percent in FY95-98. Operations of other Sector Boards allocated an average of 10 percent to the PSG theme during FY07-10, a slight increase from 9 percent during FY95-98. But when considering PSG as a sector, the trend is reversed, with an average of 16 percent of other operations allocated to PSG sectors during FY07-10, compared to 20 percent during FY95- 98. Like in environment, operations mapped to the public sector governance Sector Board are becoming more specialized over time. During FY07-10, 25 percent of allocations for operations mapped to the PSG Sector Board were assigned to non-affiliated themes, a share that has declined from 42 percent during FY95-98. Social Development and Gender The prevalence of Social Development and Gender as themes in op- erations mapped to other Sector Boards appears to have declined over Appendix J 231 time. SDV and gender have been flagged in 16 percent of the operations approved during FY07-10, a substantial decline compared to 25 percent during FY95-98. Operations of other Sector Boards allocated an average of 4 percent to the social development and gender themes during FY07-10, com- pared to 7 percent in FY95-98. Note 1 The sector/thematic distribution of the portfolio can also be assessed weighting the shares of projects associated with sectors/themes by the size of operations. Since the results using this method typically only differed from those obtained without factoring in the size by less than two to five percentage points, this annex only reports on data in terms of cumulative shares of projects. Reference Independent Evaluation Group (IEG). 2010. Poverty Reduction Support Credits: An Evaluation of World Bank Support. Washington, DC: World Bank. 232 The Matrix System at Work Appendix K World Bank Sector Strategy Papers Sectors typically produce new sector strategy papers every 8–10 years. Recent strategies are listed in Table K.1. Sector Strategy Papers (SSPs) are not strategies in the usual sense because they are designed to cover all of the Bank’s clients. The SSPs examined for the matrix evaluation were those for which a prior Independent Evaluation Group (IEG) sector or thematic evaluation had been conducted in the last five years. Those SSPs were approved between FY01-07. They typically did not include an implementation plan, a results framework with indicators and targets, a monitoring and evaluation plan, or a budget for implementation aside from occasional trust fund resources. Because of their global coverage, SSPs have little direct influence on country programs. The primary influence of SSPs on future Bank activities comes from the participatory nature in which they are prepared and the integration of sector priorities by the sector staff on the country teams. In most cases, the Bank is not able to fully implement its sector and corporate strategies. A review of 10 recent IEG evaluations of the Bank’s performance in individual sectors or subsectors found a substantial or high degree of implementation in only three, as shown in Table K.2. This reflects less on the performance of sector staff and managers in the Regions, and more on the efficacy of the strategy itself. Given the primacy of client demand in the design of country programs, it is unreasonable to expect an unfunded global strategy to accurately predict the needs of countries as far as a decade in advance. Table K.1 Recent Sector and Corporate Strategy Papers Year Network Sector Board Document 1998 HDN Health, Nutrition and Health, Nutrition and Population Sector Strategy Paper Population 1999 HDN Education Education Sector Strategy Paper 2000 INF Energy Fuel for Thought: Environmental Strategy for the Energy Sector INF Urban Development A Strategic View of Urban and Local Government Issues: Implications for the Bank 2001 FPD Financial Sector Strategy for the Financial Sector HDN Social Protection Social Protection Sector Strategy: From Safety Net to Springboard PRM Gender Integrating Gender into the World Bank’s Work: A Strategy for Action ESSD Environment Environment Strategy Paper: Making Sustainable Commitments 2002 FPD Private Sector Development Private Sector Development Strategy—Directions for the World Bank Group ESSD Rural Development Reaching the Rural Poor: An Updated Strategy for Rural Development INF Information and Information and Communication Technologies Communication Technologies 234 The Matrix System at Work Table K.1 Recent Sector and Corporate Strategy Papers (Cont’d) Year Network Sector Board Document 2004 INF Water Supply and Sanitation Water Supply and Sanitation Sector Business Strategy 2005 ESSD Social Development Empowering People by Transforming Institutions: Social Development in World Bank Operations 2007 FPD Financial Sector Financial Sector Strategy for the World Bank Group HDN Health, Nutrition and Healthy Development: The World Bank Strategy for Health, Nutrition, Population and Population Results PRM Public Sector Governance* Strengthening World Bank Group Engagement on Governance and Anticorruption* 2008 SDN Transport Sector Safe, Clean, and Affordable Transport for Development. The World Bank Group’s Transport Business Strategy, 2008–12 2009 SDN Energy* Development and Climate Change: A Strategic Framework for the World Bank Group* 2010 SDN Urban Development Systems of Cities: Harnessing urbanization for growth and poverty alleviation 2011 HDN Education Education Strategy 2020: Learning for All Investing in People’s Knowledge and Skills to Promote Development   PRM Trade The World Bank Group Trade Strategy, 2011–2021: Leveraging Trade for Development and Inclusive Growth Source: World Bank database and OPCS. Note: In 2006, ESSD and INF were merged into SDN and *Denotes Corporate Strategies. Appendix K 235 Table K.2 Record of Implementation of Recent Sector and Corporate Strategy Papers Sector Strategy Evaluation Key priorities of Bank sector (year) (year) strategy Implementation of sector strategy Assessment Transport A Decade of The 1994 WDR proposed Private sector involvement has improved Substantial (2002) Action greater use of the private sector accountability and efficiency in general, on Transport through promoting effec- but the extent of investment has been dis- (2007) tive competition and giving appointing; the Bank erred in reducing its stakeholders more voice and commitments in the sector despite warn- responsibility. ings from the 1996 paper that filling the The 1996 transport strategy gap through private sector funding was proposed reducing the func- unlikely to succeed; “many projects� have tion of the public sector as helped establish the regulatory framework a supplier in transport and necessary to safeguard the public; man- increasing its function as a agement and accountability improving; regulator. It also highlighted Bank’s support to institutional develop- the importance of economic, ment has had mixed success; transport financial, environmental, and projects with environmental objectives social sustainability. have performed reasonably well The 2002 urban transport strategy concentrated on social exclusion aspects of urban inaccessibility. Environ- Environ- The strategy sought to: There was growing attention to envi- Negligible ment mental mprove the quality of life (i) i ronmental issues in country strategies. (2001) Sustaina- (enhancing livelihoods, Results along the three dimensions of the bility (2008) preventing and reduc- strategy follow: ing environmental health analytic work on poverty and pro-poor (i)  risks, and reducing vulner- growth has lagged, and has not been ability to natural hazards sufficiently strategic in addressing en- and climate change); vironmental linkages with health and mprove the quality of (ii) i vulnerability or in reducing vulner- growth (supporting policy, ability to hazards and climate change; regulatory, and institu-  he Bank’s record on helping countries (ii) t tional frameworks for improve environmental governance sustainable environmental has been mixed; management and PSD); global public goods tend to receive (iii)  protect the quality of (iii)  insufficient priority the regional and global commons HNP Improving   The share of multi-sectoral projects Modest (2007) Effectiveness doubled during the period to half of all and Outcomes projects. Projects supporting SWAps rose for the Poor to 13% of the portfolio. The share of lend- in Health, ing with objectives to reform the health Nutrition, system dropped by nearly half. and Population (Continued) (2009) (continued) 236 The Matrix System at Work Table K.2 Record of Implementation of Recent Sector and Corporate Strategy Papers (Cont’d) Sector Strategy Evaluation Key priorities of (year) (year) Bank sector strategy Implementation of sector strategy Assessment Urban Improving The strategy put municipal management Management of municipal service Modest (2000) Municipal and service provision at the center of provision was a priority in all focus Management the urban strategy.For service provi- MDPs, although there were shortcom- for Cities sion: Municipalities were seen as key ings in results for many.Planning was to Succeed providers of local services for improving not a consistent priority in Municipal (2009) livability for the poor in cities; Development Projects despite its For planning: Capacity building at the emphasis in the urban strategy.Most municipal level was to help achieve MDPs addressed the financial dimen- good governance; and competitiveness sion of municipal management (and highlights City Development Strategies there were better results here than in to help urban markets work better. other areas). For municipal finance: bankability was to help municipalities achieve sound finances and creditworthiness through better municipal financial management. Gender (2001) Gender and The strategy planned to: The Bank failed to prepare the Negligible Development  repare periodic Country Gender (i) p number of country gender assess- (2010) Assessments; ments planned. Implementation of  evelop priority policy and opera- (ii) d the M&E system was inadequate. In tional interventions; practice gender issues were restricted (iii) monitor the implementation and to those sectors identified in the CAS results as relevant. Water (2003) Water and The 1993 strategy shifted the Bank’s The integration of water practices Modest Development approach from one based on discrete across sectors is well underway: a ma- (2010) investments to a multi-sectoral one. jority of water projects were overseen The 2003 strategy supported busi- by sector boards other than Water and ness as usual in areas with success Sanitation, but integration has made (including water quality, conserva- little progress in client countries. Dams tion, groundwater management), but and hydropower have become more focused on areas of poor performance important parts of the portfolio. or where issues are contentious. This Some activities (of the business as usu- included ‘principled pragmatism’ rather al variety) have become less important, than faith in private sector provision, including coastal zone management, re-engagement with big hydraulic pollution control, and to a lesser extent infrastructure projects, and assuring groundwater conservation.The Bank’s consistency with CASs through country complementary strategies for the water water resources assistance strategies. sector have been broadly appropriate, but their application has underempha- sized some of the most difficult chal- lenges set by the 2003 strategy. Social Social Safety The strategy focused on a “social risk Institution-building was seen as the Substantial Protection Nets (2011) management� framework, which would strongest aspect of Bank support for (2001) identify the key sources of risk faced SSNs, but it was stronger in MICs than by households and seek to develop LICs.In 80% of country cases the Bank a range of instruments, rather than was involved in analysis and dialogue free-standing projects, to address those on targeting; and in 75% of cases staff risks. were knowledgeable about the politi- cal factors around targeting.AAA and research is more nuanced in apprecia- tion of different types of poverty and their implications for program design. (continued) Appendix K 237 Table K.2 Record of Implementation of Recent Sector and Corporate Strategy Papers (Cont’d) Sector Strategy Evaluation Key priorities of (year) (year) Bank sector strategy Implementation of sector strategy Assessment Rural Growth and The strategy supported a broad rural The Bank supported projects focused Substantial Development Productivity agenda, highlighting the importance beyond agricultural production in (2003) in of: the rural sector. Many rural projects Agriculture (i) a focus on the rural poor; adopted CDD in which agricultural de- and Agri- (ii) fostering broad-based economic velopment was one of many priorities. business growth; This suggests that the ARD strategy (2011) (iii) addressing rural areas comprehen- was more or less pursued, even though sively; the later 2008 WDR and 2009 Ag forging alliances with all stake- (iv)  Action Plan suggested a focus on agri- holders; and cultural productivity was the highest addressing the impact of global (v)  priority, a conclusion with which IEG developments on client countries. evaluation implicitly concurs There was less focus on enhancing agricultural productivity and competitiveness. Information Capturing The strategy highlighted four objec- Notable contributions in support Modest and Technology tives: to sector reforms and in private Communi- for Develop-  roadening and deepening sector (i) b investments for mobile telephony in cations ment (2011) and institutional reform (integrat- difficult environments and poorest Technology ed policy framework for converg- countries; (2002) ing technologies, legislative and in other priority areas, contributions regulatory changes to facilitate were e-government, postal sector policy limited; reform); targeted efforts to go beyond what is ncreasing access to information (ii) i commercially viable has been largely infrastructure (by leveraging unsuccessful; private sector and through efforts ICT skills development is important going beyond what commercial constraint but has received little at- providers would be willing to do); tention in WBG operations; (iii) supporting ICT human capacity; significant role in ICT applications, and but WBG’s record has been modest (iv) supporting ICT applications Governance World Bank Objectives: To support poverty re- Bank CASs planned to support insti- Modest and anti- Country Level duction by developing capable and tutional strengthening in 3x’s the corruption Engage- accountable states undertaking sound countries post-GAC. (2007) ment on policies, improving service delivery, es- Number of projects relying on gover- Gover- tablishing rules for markets, combating nance and political economy analysis nance and corruption to ensure that its funds are upstream increased. Anti- used for their intended purpose. Greater use of country systems in corruption Principles: It set out to do this with the weaker-capacity countries. (2011) following principles: (i) focusing on But CAS commitments not matched by a capable and accountable state, (ii) increase in institutional strengthen- country ownership and leadership are ing components in projects. key; Operational responses varied when (iii) remaining engaged so the poor do countries experienced governance not pay twice; (iv) consistent approach downturns. across countries; (v) engage a broad Greater efforts to include political set of economy stakeholders; (vi) strengthen rather analysis but quality did not improve. than bypass country systems; (vii) Partial improvement of Bank’s reputa- harmonization tion on GAC issues. Source: IEG Evaluations. Appendix L International Finance Corporation 2013 Change Initiative The International Finance Corporation’s (IFC’s) most recent restructuring— IFC2013—was reviewed to identify lessons IFC has learned from the experience. It is not within the scope of this review to evaluate this reform nor would it be possible given how recently it has been introduced. This review answers the following questions: (a)  what was the impetus for this change? (b) what is the new organizational structure and levers used to enact this change? (c) what systems were enacted for communication and monitoring the change initiative? (d) what are the commonalities and differences between the IFC and Bank matrixes? (e)  what lessons can be learned? IFC2013 is part of ongoing reforms at IFC intended to align its organizational structure, processes, and incentives with its strategic priorities1. The previous reorganization, IFC’s Vision 2010, aimed to bring IFC closer to its clients largely through decentralization. IFC2013 seeks to address improvement opportunities identified through a diagnostic review of organizational effectiveness conducted by the IFC Management Team with the assistance of McKinsey & Company in 2009. The organizational changes implemented in FY11 as part of IFC2013 are the focus of this review. This section is largely informed by key documents describing IFC2013 and its implementation as well as interviews with IFC staff.2 Impetus for Change The 2009 diagnostic review—which included a staff survey, interviews, and small-group meetings—identified a growing amount of complexity in the organization and its processes at the same time investment processes were becoming more decentralized. The five main improvement opportunities identified in the staff survey were as follows:3 • “Improve efficiency/productivity particularly in the investment process� • “Clarify accountabilities and simplify the matrix� • “Further define the global in global/local—develop an IFC knowledge ‘marketplace’� • “Strengthen talent management and leadership development� • “Promote stronger line ownership� The challenges identified in the 2009 diagnostic review were the drivers behind the IFC2013 change initiative which aims “to strengthen central functions while achieving sufficient scale in the field so that staff can have better networks and mentors; and sufficient specialization to acquire global knowledge and know-how more quickly and effectively than in the past.�4 The specific objectives identified as part of IFC2013 were to “expand IFC’s Development Impact, strengthen IFC’s ability to be financially sustainable, improve Client Satisfaction, strengthen talent management and people development, and better leverage Global Knowledge in a decentralized 240 The Matrix System at Work environment.�5 The resulting key elements of IFC2013 were threefold: (a) to adopt specific development goals in order to drive growth in a financially sustainable manner while maximizing development impact in sectors which IFC has a comparative advantage; (b) launch a new performance management approach which is aligned with the strategy and budget cycle and accounts for organizational and staff role changes; and (c) revise the investment departments’ organizational structure to clarify roles and accountabilities through the following changes: • create global industry groups mandated to develop global and industry leadership in areas of IFC comparative advantage and deliver measurable development results • create regional client service groups supported by operations centers fostering client responsiveness and efficiency • increase the functional specialization of front-line staff for efficiency gains and added accountability • align advisory, risk, and non-client-facing operations staff in the new structure Incentive Revised Organizational Structure and Associated Business Processes and ­ Mechanisms IFC’s organizational structure has been changed significantly as part of the IFC2013 change initiative. Figure L.1 depicts investment services’ matrix related organizational structure as of October 2011 with the most significant changes being the implementation of global industry departments headed by one vice president and regional client service groups headed by three vice Figure L.1 International Finance Corporation Investment Services’ Matrix Related Organizational Structure IFC Executive Vice Presidency Global Industries Regional Vice Vice Presidency Presidencies (3) Global Industries Regional Industry Regional Departments (3) Departments (9) Departments (8) Source: Adapted from IFC Organizational Structure (IFC 2011). Appendix L 241 presidents—each overseeing both regional industry departments and regional departments—and a revision of the budgeting process. Global Industries Groups staff are based in headquarters and responsible for managing global clients, developing sector strategies, ensuring closer advisory and investment services collaboration, managing partnerships, and developing global knowledge networks. As part of its global knowledge mandate, IFC is reviewing the Bank’s experience as a global knowledge provider to inform its own organizational model on global knowledge and a knowledge strategy committee was “set up to oversee the formulation of a clear strategic direction for knowledge, creation of an enabling governance structure and information technology platforms, and introduction of processes and incentives to promote a sustainable knowledge culture.�6 Subsequently, a global head of knowledge was appointed to develop and drive the execution of the knowledge management strategy and heads up a global knowledge office. Global practice groups are also being piloted for areas of IFC’s strategic focus. As transaction-level support is decentralized, global industry groups are responsible for sharing knowledge with global industry directors involved in the investment project review process and responsible for overseeing global industry performance. Regional client service groups are composed of regional departments and regional industry departments. Staff are based in regional operations centers or country offices with regional departments responsible for managing country strategies, business development, and client relationship management and regional industry departments responsible for portfolio and transaction management, providing functional specialization, improved responsiveness, efficiency, enhanced decision making, and risk management. The regional operations center is central to the model and contains a full complement of operational and support staff (including regional industry, credit, legal, risk, environment, and social staff) that are expected to facilitate the execution of an investment program in an efficient and timely manner. In order to achieve efficiency and scale, investment staff roles were to be further specialized instead of one staff member having responsibility for multiple types of activities such as client-relationship management, deal structuring, and portfolio management. A revised budgeting process was rolled out to accompany the structural changes. Beginning with the FY11 process, vice presidents were involved more in the planning phase in setting parameters which determined their budget envelopes and then later in formulating/allocating budgets at the department level within their vice presidencies. Technical staff were also mapped either to global or regional industry units to assist in the allocation of vice presidency and department level budgets via a matrix budgeting system. The matrix budgeting system for investment services has, and continues to be, based on a matrix between industries and Regions. With the reorganization and move to regional client service teams, instead of a matrix with two components (global industry departments and regional departments) the matrix now has three components (global industry departments, regional departments, and now regional industry departments). In order to deliver an investment program, staff from each 242 The Matrix System at Work component of the matrix must be identified and the associated budget negotiated and allocated to each department. Alignment of IFC’s performance management system, or staff incentives, with strategic goals. As part of IFC2013, managers’ objectives are being adjusted to focus more on development and financial results, client orientation, people management and development, and contribution to corporate initiatives with the performance management cycle being integrated into the corporate strategy process. IFC intends to continue and ramp up various rewards based on individual and team-based performance for staff performance, project delivery and longer-term project and portfolio outcomes, and knowledge sharing. IFC will also continue existing practices of more rapid promotions enabled by a higher number of grade levels for professionals and staff. In addition, IFC will expand training and development opportunities, adjust its talent management process, and align human resources with the World Bank. An area in which IFC is likely to be particularly challenged in developing the right incentives is in encouraging the involvement of global industry department staff in fulfilling their roles in addressing the global knowledge mandate as well as providing technical expertise to specific projects given their competing incentives of overseeing global investment lending operations. Communication and Monitoring Systems To support IFC2013, a communications strategy and a milestone and efficiency monitoring systems were put in place. Given the scale of change, efficiency metrics were identified as part of the cost-benefit analysis and their monitoring is ongoing. A communication strategy was developed to keep staff informed and to create feedback loops between staff and management which included announcements, newsflashes to address issues raised by staff, town hall meetings, an IFC2013 web portal housing key information, discussions between supervisors and direct reports to communicate implications to staff, VP conversations series, and Info Sessions housed by VPs and directors. Through these staff feedback mechanisms, staff concerns were identified and the timetable for implementing IFC2013 organizational changes was delayed in order for the concerns to be addressed. Following the eventual roll-out, staff concerns were still being raised prompting the management team to form a Corporate Task Force to identify staff concerns and help guide the IFC2013 change initiative’s implementation. The Task Force findings indicate significant concerns still remain to be addressed (see Table L.1). Key Similarities and Differences Between IFC and Bank Matrixes Both the IFC and World Bank have ongoing organizational changes currently being implemented. While IFC2013 is in its infancy as a vast restructuring with long- term implications, the World Bank’s current changes are not as extensive. Table L.2 describes the key similarities and differences between the IFC and Bank matrix structures and how they are managing the change process. Appendix L 243 Table L.1 Task Force Recommendationsa Operational Recommendations Strategic Recommendations • clarify structure define and align strategy around a •  • foster global-local knowledge compelling vision • enhance HR processes • build one IFC • strengthen leadership • rebuild trust and respect • enhance career propositions • align the management team • build collaborative culture • actively manage change/transition a IFC Change Process–update at informal meeting with the Board of Directors on October 20, 2011. Lessons Learned from the IFC2013 Change Initiative IFC2013 continues to experience and address significant challenges as it progresses through the early stages of the change process highlighting the importance of both a sound strategy from the outset as well as ability to Table L.2 Key Similarities and Differences Between IFC and Bank Matrixes IFC2013 Change Initiative World Bank Ongoing Changes Common Technical staff assigned either to global industries or a)  a)  Technical staff assigned either to network Character- regional industries anchors istics b) complex organizational structure noted by staff or regions b) complex organizational structure noted by staff Differences Single matrix structure with three entities (global a)  Double matrix of anchor sectors versus regions a)  industries, regional departments, regional industry and within regions–country versus sector units departments). Advisory Services staff not included as  ll country client product delivery budgets b) A part of the matrix structure negotiated with regional CMUs holding  ll country client product delivery budgets managed b) A budget authority and SMUs negotiating for through VPs but investment services work programs funds. Network anchors play no role except are negotiated between the three matrix entities responding to cross support requests with budgets allocated directly to each entity F PD restructuring accompanied by strategy c)  Strategy for change initiative exists with c)  document and communication plan communication strategy Monitoring of organizational effectiveness d)  Monitoring of metrics for success/stated commitment d)  introduced in FY11 scorecard to make changes as necessary No noted change in staff incentives, only e)  C hange in staff incentives and mandate from above e)  mandate from above (compensation based on (compensation program aligned with strategic qualitative assessments) objectives and achievement of quantifiable 39 percent of staff in country (of 11,534 total f)  development goals) staff forecasted as of end of FY11) 55 percent of staff in country offices (of 3,429 total f)  g) 37 staff per sector manager (as of end of FY11) staff forecasted as of end of FY11)a g) 19 operational staff per managerb a IFC’s FY12 Business Plan & Budget (IFC 2011). b Source: IFC human resources department. Note: Includes only staff in operational VPUs (Global Industry and Regional Vice Presidencies) not including front office functions. Managers include those in IFC’s official managerial cadre as end of FY11. 244 The Matrix System at Work address ongoing challenges as they arise. The IFC2013 team noted the following as important lessons when completing a restructuring:7 • “Engagement with staff, managers, and directors is critical at every stage� • “Important to take time to get it right and ensure buy-in from those affected� • “Critical to be willing to adjust and re-calibrate as necessary� • “Testing and piloting new concepts is an effective way to ‘work out the kinks’� • “Unity and clarity in communications from the top down is essential� It is too early to evaluate if IFC2013 has been successful in addressing the problems originally identified and much will depend on how effectively management continually monitors progress and makes the necessary adjustments. In the interim, one can learn much from the significant challenges the IFC2013 change initiative has already faced in its short lifespan and how it has chosen to address them. Notes 1 A full description of IFC 2013 can be found in IFC Road Map FY11-13- Maximizing Impact, Unlocking Our Potential (IFC 2010). 2 Key documents include IFC Road Map FY11-13- Maximizing Impact, Unlocking Our Potential (IFC 2010); FY11 Business Plan and Budget (IFC 2010); and IFC’s FY12 Business Plan & Budget (IFC 2011). 3  See Box 8 in (IFC 2010: 41). 4  (IFC 2010: 41). 5  (IFC 2010: 42–44). 6  IFC Road Map FY11-13-Maximizing Impact, Unlocking Our Potential(IFC 2010). 7  IFC 2013 Change Initiative-update at informal meeting with the Board of Directors on September 16, 2010). Appendix L 245 Appendix M Staffing Data Table M.1 Regional Vice Presidencies’ Staff Head Counts, FY11 VPU ETT ETC GA-GD GE+ Total AFR 55 140 633 1,068 1,896 EAP 42 90 317 649 1,098 ECA 38 51 273 608 970 LCR 51 94 210 545 900 MNA 17 33 142 313 505 SAR 30 45 277 514 866 Regions 233 453 1,852 3,697 6,235 Source: World Bank Database (snapshot of 06/2011), HR Query. Note: Head counts of staff mapped to the units exclude SPAS i.e. staff on special assignments Tables M.1 and M.2 show the distribution of Bank staff across regional Vice Presidential Units (VPUs), network VPUs and selected central VPUs (i.e. Development Economics (DEC) and the World Bank Institute, WBI) as of June 2011. Table M.3 shows the head counts and shares of sector staff mapped to network VPUs as opposed to regional sector units. Table M.4 shows the distribution of technical staff—at grade GE and above—between network anchor and regional units by network. Table M.2 Staff Head Counts of Network Anchors (NVPUs) and Selected Central VPUs, FY11 VPU ETT ETC GA-GD GE+ Total FPD 3 16 45 248 312 HDN 4 36 35 143 218 PREM 3 30 27 123 183 SDN 12 74 118 553 757 OPCS 6 10 28 116 160 Total NVPUs 28 166 253 1,183 1,630 DEC 2 39 37 187 265 WBI 5 21 38 152 216 Other VPUs 72 137 729 2118 3056 Bank-wide Total 340 816 2909 7337 11402 Source: World Bank Database (snapshot of 06/2011), HR Query. Note: Head counts of staff mapped to the units exclude SPAS i.e. staff on special assignments. IBRD FPD staff head counts only. 248 The Matrix System at Work Table M.3 OPCS GE+ Staff in the Regions–Head Counts FY02-FY10 FY02 FY03 FY04 FY05 FY06 FY07 FY08 FY09 FY10 % FY02-10 TOTAL Regions 192 310 323 374 367 417 450 487 523 172% AFR 77 90 93 102 102 119 128 143 166 116% Financial Management 27 28 28 28 31 41 58 62 Procurement 37 36 38 38 45 45 50 63 Operations Services 77 9 9 10 11 7 7 6 4 –95% Other 17 20 26 25 36 35 29 37 EAP 14 47 53 54 58 64 70 77 83 493% Financial Management 27 Procurement 43 Operations Services 14 47 53 54 58 64 70 77 13 –7% ECA 20 67 63 67 69 70 64 64 64 220% Financial Management 21 Procurement 29 Operations Services 20 67 63 67 69 70 59 57 1 –95% Other 5 7 13 LCR 27 41 42 48 47 63 74 79 76 181% Financial Management 13 15 16 23 23 28 26 28 26 100% Procurement 11 26 26 25 23 29 30 30 29 164% Operations Services 3 1 3 3 3 2 –33% Other 3 15 18 19 MNA 23 26 31 56 42 36 40 41 49 113% Financial Management 12 11 12 14 16 Procurement 18 14 15 14 18 Operations Services 23 26 31 56 3 3 3 3 4 –83% Other 9 8 10 10 11 SAR 31 39 41 47 49 65 74 83 85 174% Financial Management 17 23 22 23 22 27 30 36 37 118% Procurement 12 13 17 21 24 28 34 37 36 200% Operations Services & Quality 1 1 2 3 1 1 Other 2 3 2 2 2 8 7 9 11 450% TOTAL Financial management 30 65 66 74 85 97 109 136 189 530% TOTAL Procurement 23 76 79 84 103 116 124 131 218 848% Source: World Bank databases. Note: Snapshots of June of each fiscal year. Headcounts of both ETCs and regular staff; exclude SPAs i.e. staff on special as- signments. Other units include Development Effectiveness, and Quality regional units. Appendix M 249 Table M.4 Anchor and Regional Staff Graded GE and Above by Network, FY11 Anchor staff Staff in the Regions % Staff in the Anchor Network GE+ ETC GE+ ETC GE+ GE+ and ETC FPD 189 11 177 42 52% 48% HDN 143 36 508 82 22% 23% PREM 123 30 465 98 21% 21% SDN 496 71 1196 119 29% 30% Source: World Bank Database. Note: Within the table GE+ refers to regular staff graded GE and above. Data for FPD excludes Doing Business (GIADB) and the Consultative Group to Assist the Poor (CGP) staff; data for SDN excludes the Carbon Finance unit (ENVCF) staff. Staff decentralization data is presented in Tables M.5 and M.6: Table M.5 shows counts of locally- and internationally- recruited staff based in headquarters and the field; table M.6 showcases similar data for staff graded GE and above only. Tables M.7 and M.8 show the counts and shares of headquarter-appointed net staff (i.e. full-time staff on open-ended or term appointments) graded GF or above in operational units, by duration of service in their current unit. Table M.5 IBRD Staff Decentralization Data   AFR EAP ECA LCR MNA SAR TOTAL HQ-based IRS 322 152 250 283 158 153 1,318 HQ-based LRS 348 122 167 261 117 104 1,119 Field-based IRS 267 143 51 47 41 81 630 Field-based LRS 959 681 498 309 189 528 3,164 Unassigned 0 0 4 0 0 0 4 Total 1,896 1,098 970 900 505 866 6,235 Source: World Bank databases, HR Query. Note: Headcounts exclude SPAs i.e. staff on special assignments, but include ETTs and ETCs. 250 The Matrix System at Work Table M.6 IBRD Staff Decentralization Data (Staff Graded GE and Above Only) AFR EAP ECA LCR MNA SAR TOTAL HQ-based IRS 322 152 250 283 158 153 1,318 HQ-based LRS 189 64 84 148 63 48 596 Field-based IRS 267 143 51 47 41 81 630 Field-based LRS 430 380 271 161 84 277 1,603 Unassigned 0 0 3 0 0 0 3 Total 1,208 739 659 639 346 559 4,150 Source: World Bank databases, HR Query. Note: Headcounts exclude SPAs i.e. staff on special assignments, but include ETCs. Table M.7 FY09 Time in PMU for HQ-Appointed GF-GI Net Staff Less than 3 Years 3–5 Years 5–7 Years 7 or more Years 5 Years or more # % # % # % # % Total # % AFR 223 41.4% 98 18.2% 81 15.0% 137 25.4% 539 218 40.4% EAP 124 48.4% 43 16.8% 45 17.6% 44 17.2% 256 89 34.8% ECA 147 46.4% 46 14.5% 31 9.8% 93 29.3% 317 124 39.1% LCR 163 47.1% 60 17.3% 48 13.9% 75 21.7% 346 123 35.5% MNA 91 47.4% 40 20.8% 32 16.7% 29 15.1% 192 61 31.8% SAR 121 51.7% 49 20.9% 34 14.5% 30 12.8% 234 64 27.4% Regional Units 869 46.1% 336 17.8% 271 14.4% 408 21.7% 1,884 679 36.0% FPD 56 53.8% 19 18.3% 6 5.8% 23 22.1% 104 29 27.9% HDN 46 54.1% 13 15.3% 8 9.4% 18 21.2% 85 26 30.6% PRM 47 60.3% 12 15.4% 9 11.5% 10 12.8% 78 19 24.4% SDN 126 48.3% 48 18.4% 24 9.2% 63 24.1% 261 87 33.3% Network Anchor 275 52.1% 92 17.4% 47 8.9% 114 21.6% 528 161 30.5% Units DEC 20 14.4% 15 10.8% 20 14.4% 84 60.4% 139 104 74.8% OPC 50 64.9% 5 6.5% 7 9.1% 15 19.5% 77 22 28.6% QAG 1 33.3% 0.0% 1 33.3% 1 33.3% 3 2 66.7% WBI 36 30.8% 15 12.8% 24 20.5% 42 35.9% 117 66 56.4% Other Operational 107 31.8% 35 10.4% 52 15.5% 142 42.3% 336 194 57.7% Units Networks and 382 44.2% 127 14.7% 99 11.5% 256 29.6% 864 355 41.1% Other Ops Total Operational 1,251 45.5% 463 16.8% 370 13.5% 664 24.2% 2,748 1,034 37.6% Units Source: HR Analytics. Appendix M 251 Table M.8 FY11 Time in PMU for HQ-Appointed GF-GI Net Staff Less than 3 Years 3–5 Years 5–7 Years 7 or more Years 5 Years or more # % # % # % # % Total # % AFR 269 46.7% 119 20.7% 68 11.8% 120 20.8% 576 188 32.6% EAP 150 54.5% 53 19.3% 32 11.6% 40 14.5% 275 72 26.2% ECA 159 47.9% 79 23.8% 29 8.7% 65 19.6% 332 94 28.3% LCR 165 46.7% 81 22.9% 46 13.0% 61 17.3% 353 107 30.3% MNA 114 54.5% 44 21.1% 24 11.5% 27 12.9% 209 51 24.4% SAR 145 59.2% 46 18.8% 26 10.6% 28 11.4% 245 54 22.0% Regional Units 1,002 50.4% 422 21.2% 225 11.3% 341 17.1% 1,990 566 28.4% FPD 66 48.9% 34 25.2% 15 11.1% 20 14.8% 135 35 25.9% HDN 57 57.6% 19 19.2% 10 10.1% 13 13.1% 99 23 23.2% PRM 58 62.4% 15 16.1% 11 11.8% 9 9.7% 93 20 21.5% SDN 148 52.7% 54 19.2% 34 12.1% 45 16.0% 281 79 28.1% Network Anchor 329 54.1% 122 20.1% 70 11.5% 87 14.3% 608 157 25.8% Units DEC 42 27.1% 10 6.5% 14 9.0% 89 57.4% 155 103 66.5% OPC 57 62.0% 22 23.9% 5 5.4% 8 8.7% 92 13 14.1% WBI 64 55.7% 9 7.8% 10 8.7% 32 27.8% 115 42 36.5% Other Operational 163 45.0% 41 11.3% 29 8.0% 129 35.6% 362 158 43.6% Units Networks and 492 50.7% 163 16.8% 99 10.2% 216 22.3% 970 315 32.5% Other Ops Total Operational 1,494 50.5% 585 19.8% 324 10.9% 557 18.8% 2,960 881 29.8% Units Source: HR Analytics. 252 The Matrix System at Work Appendix N Managers Interviewed Names Title Andrew Steer Vice President Isabel Guerrero Vice President Sanjay Pradhan Vice President Tamar Manueyan Atinc Vice President Annette Dixon Country Director Asad Alam Country Director Carlos Felipe Jaramillo Country Director David Craig Country Director John McIntire Country Director Martin Raiser Country Director Onno Ruhl Country Director Stefan G. Koeberle Country Director Susan Goldmark Country Director Adolfo Brizzi Country Manager Geoffrey H. Bergen Country Manager Ousmane Diagana Country Manager Xavier Devictor Country Manager Alfred Nickesen Director Edward Okere & Preeti Director Cristian Baeza Sector Director Emmanuel Jimenez Sector Director Jamal Saghir Sector Director James Warren Evans Sector Director John Henry Stein Sector Director John Roome Sector Director Juergen Voegele Sector Director Laszlo Lovei Sector Director Laura Tuck Sector Director Linda Van Gelder Sector Director Marcelo Giugale Sector Director Mayra Buvinic Sector Director Michal J. Rutkowski Sector Director Steen Lau Jorgensen Sector Director Akiko Maeda Sector Manager Anand Rajaram Sector Manager Bassam Ramadan Sector Manager C. Sanjivi Rajasingham Sector Manager Dina Umali Deininger Sector Manager Ede Jorge Ijjasz-Vasquez Sector Manager Ethel Sennhauser Sector Manager (Continued) 254 The Matrix System at Work Names Title Eva Jarawan Sector Manager Francis Ato Brown Sector Manager Guang Zhe Chen Sector Manager Guenter Heidenhof Sector Manager Helena G.Ribe Sector Manager Herbert Acquay Sector Manager Idah Z. Pswarayi Riddihough Sector Manager Joel Hellman Sector Manager John Kellenberg Sector Manager Jose Luis Irigoyen Sector Manager Juan Pablo Uribe Sector Manager Julie McLaughlin Sector Manager Junaid Ahmad Sector Manager Karin Erika Kemper Sector Manager Linda Van Gelder Sector Manager Luis Constantino Sector Manager Lynne D. Sherburne-Benz Sector Manager Magda Lovei Sector Manager Mansoora Rashid Sector Manager Mark Woodward Sector Manager Michel Audige Sector Manager Ming Zhang Sector Manager Simeon K. Ehui Sector Manager Veronica Zavala Lombardi Sector Manager Vijay Jagannathan Sector Manager Wael Zakout Sector Manager William Leslie Dorotinsky Sector Manager Xiaoqing Yu Sector Manager Vipul Tandon Manager Samantha Forusz Program Manager Mariam Roshini Jacob Sr. Resource Management Officer Appendix N      255 Bibliography Barnett, Michael N. and Martha Finnemore. 1999. “The Politics, Power, and Pathologies of International Organizations.� International Organization 53 (4): 699-732. ———. 2004. Rules for the World: International Organizations in Global Politics. Ithaca: Cornell University Press. Birdsall, Nancy and Devesh Kapur, eds. 2006. Rescuing the World Bank. Washington, DC: Brookings Press. de Bono, Edward. 1969. New Think: The Use of Lateral Thinking in the Generation of New Ideas. New York: Basic Books. Easterly, William. 2002. “The Cartel of Good Intentions: The Problem of Bureaucracy in Foreign Aid.� Journal of Policy Reform 5 (4):223-50. Einhorn, Jessica. 2006. “Reforming the World Bank,� Foreign Affairs 84 (1):18 Hobbs, Nathaniel. 2005. “Corruption in the World Bank Financed Projects: Why Bribery is a Tolerated Anathema.� DESTIN Working Paper No. 05-65. London School of Economics. http://www.lse.ac.uk/collections/DESTIN/workingPapers .htm Hofstede, G. 1991. Cultures and Organizations: Software of the Mind. London: McGraw- Hill. IEG (Independent Evaluation Group). 2008. Environmental Sustainability: An Evaluation of World Bank Group Support. Washington, DC: World Bank ———. 2009. Annual Review of Development Effectiveness: Achieving Sustainable Development. Washington, DC: World Bank. ———. 2012. World Bank Group Impact Evaluations: Relevance and Effectiveness. Washington, DC: World Bank, forthcoming Kapur, Devesh. 2002. “The Changing Anatomy of Governance of the World Bank.� In Reinventing the World Bank, ed. Jonathan R. Pincus and Jeffrey A. Winters. Ithaca: Cornell University Press, pp. 63–67. Lusthaus, Charles, Marie-Helene Adrien, Gary Anderson, and Fred Carden. 1999. Enhancing Organizational Performance: A Toolbox for Self-Assessment. Ottawa: International Development Research Centre. Mallaby, Sebastian. 2005. “Saving the World Bank.� Foreign Affairs 84 (3): 75-85. Stone, Diane and Christopher Wright, eds. 2007. The World Bank and Governance: A Decade of Reform and Reaction. New York: Routledge. Bibliography      257 World Bank. 1997a. “ECA Renewal: Serving Clients Better and Working More Closely with Partners.� A note for clients and partners. Europe and Central Asia Region. November 1997. Washington, DC. ———. 2000. “Strategic Compact Assessment: Informal Meeting with Executive Directors.� 13 December 2000. Washington, DC: World Bank. ———. 2001a. “Knowledge Management.� Washington, DC: World Bank. ———. 2001b. World Bank Group Strategic Framework. 24 January 2001. Washington, DC: World Bank. ———. 2003. “Annual Report on Portfolio Performance.� Washington, DC. World Bank, Quality Assurance Group. ———. 2004. “Sector Board Assessment. Summary Report.� Quality Assurance Group. March 16, 2004. Washington, DC. ———. 2005. “Organizational Effectiveness Task Force: Final Report.� February 15, 2005. Washington, DC. ———. 2009a. “The World Bank Matrix Management Structure.� Technical Briefing to the Board. Operations Policy and Country Services. July 28, 2009. Washington, DC. ———. 2009b. “The World Bank Matrix Management Structure. Current Challenges?� Informal briefing to the Board. December 15, 2009. Washington, DC. ———. 2010a. “Africa Region Staffing Report, FY05-10.� Organizational Effectiveness Task Force. Washington, DC. ———. 2010b. “Knowledge Strategy Recommendations. Transforming the Knowledge Agenda� Knowledge Strategy Group. January 25, 2010. Washington, DC. ———. 2010c. “Transforming the Bank’s Knowledge Agenda: A Framework for Action.� Knowledge Strategy Group. February 3, 2010. Washington, DC. 258 The Matrix System at Work RECENT IEG PUBLICATIONS The World Bank Group’s Response to the Global Economic Crisis, Phase II World Bank Country-Level Engagement on Governance and Anticorruption: An Evaluation of the 2007 Strategy and Implementation Plan Assessing IFC’s Poverty Focus and Results Capturing Technology for Development: An Evaluation of World Bank Group Activities in Information and Communication Technologies IEG Annual Report 2011: Results and Performance of the World Bank Group Earnings Growth and Employment Creation: An Assessment of World Bank Support in Three Middle Income Countries Growth and Productivity in Agriculture and Agribusiness: Evaluative Lessons from the World Bank Group Experience MIGA’s Financial Sector Guarantees in a Strategic Context Social Safety Nets: An Evaluation of World Bank Support, 2000–2010 Timor-Leste Country Program Evaluation, 2000–2010 Trust Fund Support for Development: An Evaluation of the World Bank’s Trust Fund Portfolio Peru: Country Program Evaluation for the World Bank Group, 2005–09 Water and Development: An Evaluation of World Bank Support, 1997–2007 Climate Change and the World Bank Group, Phase II: The Challenge of Low-Carbon Development Cost-Benefit Analysis in World Bank Projects To see all of IEG’s work, please visit our web site: http://ieg.worldbankgroup.org/ publications