Document of The World Bank FOR OFFICIAL USE ONLY Report No: ICR00004594 IMPLEMENTATION COMPLETION AND RESULTS REPORT TF015956, TF014278 ON A SMALL GRANT FROM THE MENA TRANSITION FUND IN THE AMOUNT OF US$4.7 MILLION TO THE REPUBLIC OF TUNISIA FOR A SOCIAL PROTECTION REFORMS SUPPORT PROJECT November 30, 2019 Social Protection and Jobs Global Practice Middle East And North Africa Region ABBREVIATIONS AND ACRONYMS AMG Assistance médicale gratuite (Free and Subsidized Health Insurance Card) CNAM Caisse Nationale d’Assurance Maladie (National Health Insurance Fund) CNRPS Caisse Nationale de Retraite et de Prévoyance Sociale (Pension Fund - Public Employment) CNSS Caisse Nationale de Sécurité Sociale (Pension Fund - Private Employment) CRES Centre de Recherche et d’Études Sociales (Center for Research and Social Studies) FM Financial Management GDP Gross Domestic Product GRM Grievance Redress Mechanism HMT Hybrid Means Test ICRR Implementation Completion and Results Report ICT Information and Communication Technology IT Information Technology M&E Monitoring and Evaluation MAS Ministère des Affaires Sociales (Ministry of Social Affairs) MDICI Ministry of Development, Investment and International Cooperation MENA Middle East and North Africa MIS Management Information System MOF Ministry of Finance MOU Memorandum of Understanding MTR Midterm Review PDO Project Development Objective PMT Proxy Means Test PMU Project Management Unit PNAFN Programme National d’Aide aux Familles Nécessiteuses (National Program of Assistance to Families in Need) SIFNRL Système d’information relatif aux familles nécessiteuses et à revenue limité (Information System for the National Program of Assistance to Families in Need) TF Trust Fund TTL Task Team Leader UGTT Union Générale des Travailleurs Tunisiens (Tunisian General Labor Union) UTICA Union Tunisienne de l'Industrie, du Commerce et de l'Artisanat (Tunisian Union of Industry, Trade and Handicrafts) Regional Vice President: Ferid Belhaj Country Director: Jesko Hentschel Senior Global Practice Director: Michal Rutkowski Practice Manager: Hana Brixi Task Team Leader(s): Yuko Okamura ICR Main Contributor: Matuna Mostafa TABLE OF CONTENTS DATA SHEET .......................................................................................................................... 1 I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVE ........................................................ 4 III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME ................................ 17 IV. WORLD BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME ........................................................................................................................... 21 V. LESSONS LEARNED AND RECOMMENDATIONS .............................................................. 25 ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS ........................................................... 28 ANNEX 2. PROJECT COST BY COMPONENT ........................................................................... 35 ANNEX 3. RECIPIENT, CO-FINANCIER AND OTHER PARTNER/STAKEHOLDER COMMENTS ...... 36 ANNEX 4. PROJECT IMPLEMENTATION STRUCTURE .............................................................. 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) DATA SHEET BASIC INFORMATION Product Information Project ID Project Name P144674 DTF: Tunisia Social Protection Reforms Support Project Country Financing Instrument Tunisia Investment Project Financing Original EA Category Revised EA Category Not Required (C) Not Required (C) Organizations Borrower Implementing Agency Republic of Tunisia, Ministry of Development and Ministry of Finance International Cooperation Project Development Objective (PDO) Original PDO The project development objective is to strengthen institutional capacity to design social protection reforms and improve targeting of safety net programs. Page 1 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) FINANCING FINANCE_TBL Original Amount (US$) Revised Amount (US$) Actual Disbursed (US$) Donor Financing TF-15956 4,700,000 4,700,000 4,693,527 Total 4,700,000 4,700,000 4,693,527 Other Financing Borrower/Recipient 1,000,000 0 0 Total 1,000,000 0 0 Total Project Cost 5,700,000 4,700,000 4,693,527 KEY DATES Approval Effectiveness Original Closing Actual Closing 23-Oct-2013 05-Nov-2013 30-Jun-2017 31-May-2019 RESTRUCTURING AND/OR ADDITIONAL FINANCING Date(s) Amount Disbursed (US$M) Key Revisions 20-Jun-2017 2.00 Change in Results Framework Change in Components and Cost Change in Loan Closing Date(s) Change in Procurement Change in Implementation Schedule 28-Dec-2017 2.03 Change in Results Framework Change in Components and Cost Change in Loan Closing Date(s) Change in Implementation Schedule KEY RATINGS Outcome Bank Performance M&E Quality Moderately Satisfactory Moderately Satisfactory Substantial Page 2 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) RATINGS OF PROJECT PERFORMANCE IN ISRs Actual No. Date ISR Archived DO Rating IP Rating Disbursements (US$M) 01 24-Jun-2014 Satisfactory Satisfactory 0.50 02 03-Feb-2015 Moderately Satisfactory Moderately Satisfactory 0.50 03 24-Nov-2015 Moderately Satisfactory Moderately Unsatisfactory 0.63 Moderately 04 23-May-2016 Moderately Unsatisfactory 0.83 Unsatisfactory Moderately 05 05-Dec-2016 Moderately Unsatisfactory 1.04 Unsatisfactory Moderately 06 28-Aug-2017 Moderately Unsatisfactory 2.03 Unsatisfactory 07 27-Feb-2018 Moderately Satisfactory Moderately Satisfactory 2.23 08 07-Aug-2018 Moderately Satisfactory Moderately Satisfactory 3.08 09 10-Feb-2019 Moderately Satisfactory Moderately Satisfactory 4.06 ADM STAFF Role At Approval At ICR Regional Vice President: Inger Andersen Ferid Belhaj Country Director: Neil Simon M. Gray Jesko S. Hentschel Director: Steen Lau Jorgensen Keiko Miwa Yasser Aabdel-Aleem Awny El- Practice Manager: Hana Brixi Gammal Task Team Leader(s): Hebatalla Abdelhamid Elgazzar Yuko Okamura ICR Contributing Author: Mayalen Claire Iron Page 3 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVE A. Context 1. The January 2011 popular uprising in Tunisia was a call from the public protesting high unemployment and regional disparities, which exposed the social and economic divisions within the country and called for a renewed social contract and a new vision of economic growth. These events also highlighted the limits of public policies, especially social programs to fight against poverty and reduce vulnerability and exclusion. With an uncertain gross domestic product (GDP) growth rate and fiscal commitments projected to exceed revenues by 2017, the newly elected interim government needed to ensure economic and social stability, as national elections were planned for October 2014. The Government prioritized reducing the country’s current expenditure, particularly for wages and subsidies, to free up resources for regional development and social protection. The subsidies, in particular, were exerting fiscal pressures with expenditure on food, fuel, and transport subsidies projected to grow and become unsustainable in the medium-term, from 4 percent of GDP in 2012 to 8 percent of GDP by 2013. The rationale to reform subsidies was clear, as subsidies provided universal support that was regressive and inequitable, particularly for fuel,1 with approximately 70 percent of the subsidies benefitting the wealthiest 20 percent of the population, to create fiscal space to reallocate more resources to expand targeted and progressive social safety net interventions. 2. Tunisia’s social protection system had several weaknesses, including the absence of an information system and reliable data for monitoring beneficiaries of social assistance programs. According to the Project Appraisal Document, in 2012, government expenditure on the social protection sector accounted for approximately 18 percent of public expenditure, or about 6 percent of GDP2.Tunisia’s social protection system is composed of two pillars. The first pillar, the contributory system (social insurance), comprises three schemes. The national pension and social insurance fund for the public sector (Caisse Nationale de la Retraite et de la Prévoyance Sociale, CNRPS) and the national social security fund, mainly for the private sector (Caisse Nationale de la Sécurité Sociale, CNSS), cover old age, invalidity, death, and family benefits. The National Health Insurance Fund (Caisse Nationale d’Assurance Maladie, CNAM) covers sickness, accident, and occupational disease for both public and private sector contributors. The publicly managed pension schemes for workers, the CNRPS and the CNSS, were assessed to be unsustainable, with fiscal pressures expected to emerge by 2014. The public pension scheme, CNRPS, was facing a larger deficit than the private pension scheme, CNSS, while the medical insurance fund, CNAM, was facing deficit from rising expenditures. 3. The second pillar consists of a noncontributory scheme (social assistance) which supports the poor and vulnerable population with special needs. The National Program of Assistance to Families in Need (Programme National d’Aide aux Familles Nécessiteuses, PNAFN) and indigent health insurance (Assistance médicale gratuite, AMG 1 and 2) are the country’s largest national programs. Under social pressure, the Government expanded the PNAFN, doubling the number of beneficiaries as well as benefit 1 The food subsidies also suffered from capture by the intermediary producers of goods and were highly inequitable due to their universal support, although this was not always apparent to the society at large. Direct Social Assistance: Towards Better Targeting of Monetary Poverty and Deprivations in Tunisia, September 2, 2014, African Development Bank. 2 This comprised pension expenditure which represented approximately 5 percent of the GDP and social assistance expenditure which represented 1.3 percent of the GDP (of which direct cash transfers accounted for nearly 0.5 percent). Page 4 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) level of transfer from 2011 to 2016. Currently, nearly 10 percent of the population receives cash transfers and free health insurance (AMG 1), and an additional 20 percent of the population receives health insurance at a subsidized rate (AMG 2). These programs, however, suffered from poor targeting, particularly due to the absence of a management information system (MIS) including an updated beneficiary registry. The beneficiaries were selected, based on the household’s reported income, size, number of people with disabilities, and family support, through two steps: (a) assessment by social workers and (b) decision by local/regional committees (till 2011) and by the Ministry of Social Affairs (Ministère des Affaires Sociales, MAS) regional offices (from 2011). Although priority was given to people living below the poverty line, the complexity of beneficiary selection and management was very difficult, due to the lack of a supporting MIS and objective/transparent selection criteria and process, making the PNAFN prone to a high degree of leakage. For example, the updating of the beneficiary list3 was slow and difficult without regular and systematic mechanisms to verify the affiliation with social security programs and to process payments to beneficiaries. An analysis of the 2005 household survey data from the National Institute of Statistics estimated that 60 percent of PNAFN beneficiaries were not poor and that targeting needed to be more effective and efficient. 4. The Ministry of Finance (MOF) developed a medium-term framework that included a vision for a gradual reform of food and fuel subsidies. Although there was public support for structural reforms, they remained a highly sensitive political issue, especially in a period of transition with economic and social uncertainty. In this context, in January 2013, following extensive consultations, the interim government entered into a new Social Pact, ‘the social contract’, with the labor unions (represented by the Tunisian General Labor Union [Union Générale des Travailleurs Tunisiens, UGTT]) and the private sector (represented by the Tunisian Union of Industry, Trade and Handicrafts [Union Tunisienne de l'Industrie, du Commerce et de l'Artisanat, UTICA]). The Social Pact set in place the principles for dialogue on key areas of reform for social protection, regional development, employment, and skills, as well as for the governance of the social dialogue. Subsequently, MAS, the agency implementing social assistance programs, began preparing for social protection reforms, and its Center for Research and Social Studies (Centre de Recherche et d’Études Sociales, CRES) led an assessment of the social protection programs, which included both social assistance (noncontributory) and social security (contributive). The study, “Towards greater equity: energy subsidies, targeting and social protection in Tunisia� (Report number 82712-TN) diagnosed issues related to the assistance program and recommended the following: (a) establishing a national dialogue on subsidy reforms; (b) strengthening the program to support vulnerable households by developing a targeting mechanism, creating interoperability between electricity billing, car registration, and social security databases and between the PNAFN and the AMG programs; (c) conducting an analysis of the mitigation measures to support those who could be adversely affected by subsidy reforms, especially the poor and vulnerable; and (d) reforming the electricity pricing structure. 5. The Social Protection Reforms Support Project (the Project) responded to the Government’s request for building institutional capacity for subsidy and safety net reforms. The Government’s rationale was to reform the regressive and inequitable universal subsidy programs into a progressive, well-targeted, and efficient social assistance program to benefit the poor and vulnerable. The Project is aligned with the priorities identified under the ‘social contract’; the objectives of the World Bank’s Middle East and North Africa (MENA) Transition Fund of (a) Enhancing Economic Governance and (b) Inclusive Development; and 3 In fact, more than half of program outflows were due to deaths. Page 5 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) the World Bank’s 2013–2014 Interim Strategy Note,4 including the pillars (a) Laying the Foundation for Renewed Sustainable Growth and Job Creation and (b) Promoting Social and Economic Inclusion. The Project is also aligned with the MENA Regional Strategy and contributed to the pillars of social inclusion, growth, and governance and supported the World Bank’s Governance in Social Sectors Technical Assistance Program (2011–2014). Given the Project’s importance for enabling structural reforms and ensuring fiscal, economic, and social stability in Tunisia, this was the first project approved by the MENA Transition Fund. B. Project Development Objective and Original Project Design 6. The Project Development Objective (PDO) was to strengthen institutional capacity to design social protection reforms and improve targeting of safety net programs. 7. The PDO remained unchanged throughout the life of the Project. However, the Project was restructured twice, and the activities, outcome indicators, and funds allocation were formally revised during the first restructuring, as discussed below. Table 1. Original and Revised Components and Cost Allocations Original at Approval Revised at First Restructuring (November 5, 2013) (June 20, 2017) Component Name Allocation Component Name Allocation Component 1 Subsidy and Safety Net Reform US$3.78 Subsidy and Safety Net US$2.15 Support million Reform Support million • Subcomponent 1.1: • Subcomponent 1.1: Technical assistance to the Technical assistance to Inter-ministerial Working the Inter-ministerial Group (US$0.82 million) Working Group • Subcomponent 1.2: • Subcomponent 1.2: Development of a unified Technical assistance to population database and improve social safety targeting system (US$2.12 net system million) • Subcomponent 1.3: Consensus-building and communication activities (US$0.84 million) Component 2 Strengthening Social Security US$0.40 System Integration US$2.30 Analysis and Planning million Support for Social million Protection Component 3 Project Management and US$0.40 Communication, Project US$0.25 Monitoring million Management and million Monitoring Reserve US$.12 US$0.00 (Miscellaneous) million Total US$4.7 million US$4.7 US$4.7 million million 4 World Bank. 2012. Tunisia Interim Strategy Note. Report No. 67692-TN. Washington, DC: World Bank. Page 6 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) 8. Original components/activities and costs. The original project included the following key activities which focused on providing support to the subsidy reform: (a) Component 1: Subsidy and Safety Net Reform Support (US$3.78 million). The original design included three subcomponents. (i) Subcomponent 1.1: Technical assistance to the Inter-ministerial Working Group (US$0.82 million) to assess economic and social impact of potential subsidy reforms, evaluate reform options for automatic fuel price adjustment mechanisms, and develop a plan and manual for a subsidy-compensation program for households (ii) Subcomponent 1.2: Development of a unified population database and targeting system (US$2.12 million) by consolidating existing databases linked to revenue collection, national identification number used by the Ministry of Interior, database for gas and electricity, and various identification numbers used for MAS’ programs (iii) Subcomponent 1.3: Consensus-building and communication activities (US$0.84 million), including enhancing existing grievance redress mechanisms (GRMs) for social protection beneficiaries (b) Component 2: Strengthening Social Security Analysis and Planning (US$0.40 million). This component aimed to provide training to build the capacity for (i) policy analysis of pension and health insurance reform options; (ii) economic and social impact scenario analysis for social contributions and protection for workers who become unemployed; (iii) policy analysis of labor legislation in relation to social security; and (iv) preparation, coordination, and dissemination of information policy reform options among social dialogue partners. (c) Component 3: Project Management and Monitoring (US$0.40 million). This component was to finance consultant services and equipment to support the Project Management Unit (PMU) within the MOF and the implementing agencies. C. Significant Changes Introduced during Implementation (Two Restructurings) 9. The project Grant Agreement was signed on November 5, 2013, and became effective the same day. The Project was formally restructured twice. The first restructuring was finalized on June 20, 2017; introduced important changes to the project activities and outcome targets and extended the closing date by six months from June 30, 2017 to December 31, 20175. The second restructuring was approved on December 28, 2017, to extend the closing date by another 17 months to May 31, 2019.6 Through the two restructurings, the implementation period was extended by a total of 23 months.7 5 To reflect new project activities and new closing date, the changes were also introduced in component cost, implementation schedule, and procurement plan. 6 To reflect a new closing date, implementation schedule and cost/procurement tables were also adjusted. In addition, the second restructuring also made slight corrections to the Results Framework and component names in line with the restructuring of June 2017 due to unexpected error in the World Bank’s system, because of which the changes to the names of the components and an outcome indicator did not get registered; these were corrected as part of the second restructuring. 7 The World Bank did not agree to the Government’s request of November 22, 2018, for a third extension of the closing date to Page 7 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) 10. The Project became effective on November 5, 2013, and several activities, including analysis related to the adjustment of fuel prices and accompanying costs, were launched before the signing of the Grant Agreement. However, effective implementation of the Project did not start until mid-2014 following the first reconstitution meeting of the new Project Steering Committee in February 2014, which was delayed primarily due to Government reshuffling that continued till mid-2017 and resulted in changes in key Government counterparts in the MOF, the Steering Committee, MAS, and the CRES. 11. During the first two years of the Project, the subsidy reform remained one of the Government’s priorities because of the high fiscal burden. However, the subsidy reform discussion did not move forward toward actual implementation for a number of reasons, including the following: (a) the implementation of subsidy reform would have required the implementation of project activities that were too complex and not possible with existing systems and data at the time, and (b) the fall in world fuel prices in 2014 reduced the urgency to undertake the sensitive energy subsidy reforms. This resulted in a Government decision to address the inadequacies of the social protection sector. From late 2014, the project activities started to strengthen the social protection system, as this could also inform the design of a compensation mechanism to mitigate the impact of the gradual removal of subsidies. As the two programs, PNAFN and AMG 1 and 2, remained high priority, the Head of the Government led the ministerial group in charge of Component 1. 12. Apart from the exogenous factors related to the subsidy reform, the original activities were also deemed too ambitious from a technical perspective, particularly the development of a unified database (Subcomponent 1.2). This activity required several prerequisite conditions and systems to be in place such as strong consensus and coordination across ministries and institutions, including the Ministry of Interior, MAS, and the utility companies; institutional and legal framework for secure data exchange between government agencies; a unique identifier that is functional and commonly used by all participating ministries/agencies; and the existence of a quality beneficiary database. 13. By the second half of 2015, the project’s Inter-ministerial Steering Committee8 recognized the potential need to restructure the Project, and a comprehensive consultation process started in October 2015 to assess the original project activities and its strategic alignment with the Government’s priorities. The assessment highlighted several limitations with the original project design, in particular the readiness of Information Technology (IT) and MIS systems to support analysis of subsidy reforms, and recommended (a) establishing an electronic social registry of poor and vulnerable households and (b) developing cross- referencing functionality between the Government’s social protection and social insurance databases (that is, between the contributory and noncontributory systems) to improve beneficiary identification and targeting mechanisms. The review also determined that the Government did not see this Project as the right vehicle for discussing reform options, as more time was needed for strategic analysis and implementation of subsidy reforms than was available through this Project. Government priority was to establish an e-pension system to enhance the transparency and efficiency of the insurance funds. 14. Based on these findings, at the end of October 2015, the Steering Committee agreed to revise the project’s activities to building the prerequisite tools and systems, particularly electronic MISs, to December 2019, confirming that no additional time was needed to complete the project. 8 The Inter-ministerial Steering Committee comprised the Office of President; MOF; MAS; the Ministry of Development, Investment and International Cooperation (MDICI); CRES; CNSS; CNRPS; CNAM; Ministry of Technology, Information and Communication; and Ministry of Commerce. Page 8 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) modernize the social protection system, which would also enable implementation of future reforms. During the six-week Midterm Review (MTR) (January 26–March 10, 2016), the World Bank team held a series of eight thematic workshops with Government counterparts and reached agreement on the terms of the project restructuring. On July 21, 2016, the Government sent the World Bank an official request to restructure the Project which reflected a disbursement rate of 18 percent. However, the restructuring agreement took nearly a year to be finalized and approved, due to lengthy negotiations to resolve unexpected and critical issues, notably (a) uncertainty related to the implementation of the social survey, as social workers threatened to generalize the boycott to claim better compensation to carry out the survey, and (b) the urgent need to appoint dedicated procurement and financial management (FM) staff in the PMU to promote faster implementation and speed up disbursements. First Project Restructuring (Level 2) 15. Although the Project had operated under the revised conditions since late 2015, the first restructuring was officially approved on June 20, 2017, 10 days before the original closing date (disbursement rate reached 42.6 percent, or US$2 million of the total Grant). The changes were aligned with the World Bank’s Country Partnership Framework for Tunisia 2016–2020 (Report No. 104123-TN); the World Bank Group Strategy for MENA 2015; the Tunisian Government’s Five-Year Development Plan 2016–2020 that focused on modernizing the social safety net system; and the National Program on major reforms. In line with the revised Project activities, the outcome indicators were changed, and the closing date was extended. 16. Revised PDO indicators are summarized in table 2, compared to the original ones. Table 2. Original and Revised PDO Indicators and End Targets Original PDO Indicators New/Revised/Dropped PDO Indicators Target Actual Achieved 1. Number of direct project beneficiariesa Revised: 1. Number of participants in 2,645 2,240 (a) Number of institutional capacity trainings which enhance knowledge and building capacity (b) Number of registration in unified population database 2. Citizens registered in new unified Revised: 2. Number of households in the 300,000 304,698 databaseb system who are updated with complete information 3. Number of safety net beneficiaries Dropped: Government priorities shifted n.a. n.a. from subsidy reform and social security 4. Subsidy and safety net reform plan reform, which did not take off, to focus on n.a. n.a. prepared and disseminated modernizing the safety net system. 5. Integrated social security reform plan n.a. n.a. prepared and disseminated New: 3. Number of assessed persons who 750,000 1,000,717 received unique social identifier New: 4. Development of targeting tool, Yes Yes socioeconomic classification applied to households registered in the database Page 9 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Note: a. It also included a total number of direct beneficiaries (sum of a and b) and the percentage of which were female. b. Captured total number, of which were female, and of which were vulnerable. 17. Revised project components. The first restructuring revised project activities to shift away from subsidy reforms to focus on modernizing the safety net system and building blocks for both contributory and noncontributory social protection programs. (a) Component 1: Subsidy and Safety Net Reform Support (original US$3.78 million; revised US$2.15 million; actual at project closing US$2.66 million). The revised component included two subcomponents. (i) Subcomponent 1.1: Technical assistance to Inter-ministerial Working Group. The original subcomponent title was retained, as the activity to support the inter- ministerial working group to evaluate the impact of subsidy reforms was already completed before the restructuring (a draft report was completed in January 2016 and received and validated by the Working Group.) The original plan to develop a subsidy- compensation program was eliminated since there was no reform envisaged in that context. (ii) Subcomponent 1.2: Technical assistance to improve social safety net system. This subcomponent was reformulated to establish a social benefits registry through a social survey, develop and implement a targeting tool, and build the foundations for an integrated MIS. (b) Component 2: System Integration Support for Social Protection (original US$0.40 million; revised US$2.30 million; actual at project closing US$1.84 million). The revised activities focused on the interconnectivity between the contributory social insurance funds and noncontributory social protection programs. The component invested in IT to enable interoperability between the two schemes, including (i) implementing a social identifier9 for beneficiaries of social protection programs; (ii) developing interoperability of IT systems across social insurance schemes and MAS’ social assistance program; and (iii) upgrading relevant MIS systems of the social insurance schemes to support system integration, including development of e-pensions platforms. (c) Component 3: Communication, Project Management and Monitoring (original US$0.40 million; revised US$0.25 million; actual at project closing US$0.19 million). The component included two subcomponents. (i) Subcomponent 3.1: Enhancing stakeholder communication and citizen engagement. Communication and civic engagement activities (including GRMs), which were 9 As the creation of a national unique database was too ambitious and politically sensitive, the Project decided to focus on supporting a social identifier for the social protection sector. The social identifier was initially only implemented for social insurance beneficiaries (contributory scheme), and the Project also supported integrating the social identifier to social assistance program beneficiaries (noncontributory scheme). Page 10 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) originally under Subcomponent 1.3, were moved to this component to align all cross- cutting project functions. (ii) Subcomponent 3.2: Project Management and Monitoring. This subcomponent remained the same and financed consultant services and the purchase of IT and other equipment for the PMU and MAS. 18. First extension of the project closing date by six months (from June 30, 2017, to December 31, 2017). For the first restructuring, the Government requested a one-year extension to the project closing date to June 30, 2018, and to increase the end target of the second revised PDO indicator to complete social surveys for 400,000 households. The World Bank agreed to a phased project extension: (a) the first six-month extension to be given on the completion of assessing 2,000 households by May 2017 and (b) a second six-month extension subject to the completion of assessing 30,000 households by September 2017. The World Bank recommended keeping the target for the social survey at 300,000, due to the high uncertainty surrounding the survey implementation. Second Project Restructuring (Level 2) 19. Second extension of the project closing date by 17 months from December 31, 2017, to May 31, 2019. The second project restructuring was approved on December 28, 2017, three days before the closing date (at which point 43.2 percent, or US$2.03 million, of the total Grant had been disbursed).10 In November 2017, the Government requested a 17-month extension to complete planned activities, particularly the social survey. The World Bank approved this request upon receipt of communication from the Government confirming its commitment to complete the assessment of 32,500 households by December 22, 2017. The Project was thus extended for a total of 23 months from the original closing date of June 30, 2017 to May 31, 2019. In line with the further extension of the project closing date, adjustments were made to the component costs, the project implementation, and the disbursement schedules. II. ASSESSMENT OF OUTCOMES 20. The PDO was to strengthen institutional capacity to design social protection reforms and improve targeting of safety net programs. For this assessment, the PDO is interpreted as follows11: the Project focused on establishing the tools and mechanisms required for reforming the social protection sector and for improving targeting of social protection programs through building relevant institutional capacity required to implement these. 21. The Project is assessed before and after restructuring12 against both the original and revised indicators, considering the following factors: (a) the restructuring made major changes to the project’s 10 Although the condition for the second restructuring was not met on time (that is, finalization of 30,000 household surveys by September 2017), there was sufficient progress made as MAS had completed assessing 16,000 households by October 2017. 11This interpretation is based on the description of the PDO and project beneficiaries in the Project Appraisal Document, which defines poor and near-poor households as main beneficiaries. 12 The restructuring introduced significant changes in outcome targets while the PDO remained the same. In addition to the reasons explained in this paragraph, the separate assessment for outcome ratings before and after the restructuring is to help explain the complexity of this Project as well as to bring more accountability, while this Project is not necessarily subject to the Page 11 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) outcome targets; (b) the restructuring was finalized just eight days before the original closing date; hence, the pre-restructuring implementation period was long (3 years and 8 months, as compared to the 23 months post-restructuring); and (c) the Project demonstrated very different performance/results before and after the restructuring. A. Project Achievements before the First Restructuring 22. Before the first restructuring, the Project made some progress toward implementing the original activities but did not achieve any of the outcome targets. With regard to a subindicator of the first outcome indicator, the number of direct project beneficiaries of institutional capacity building was 2,067, against a targeted 500. These 2,067 people were trained and contributed to enhancing the PMU staff’s knowledge of procurement and FM procedures and MAS’ knowledge of social safety nets, MIS, and use of IT, including computers and tablets to better control the new processes for the social surveys conducted by the social workers. No-one was registered in the unified database (another subindicator of the first outcome indicator as well as outcome indicator 2 was zero), while information on 2,000 households was updated in the MAS’ beneficiary registry. With regard to outcome indicator 3, Number of Safety Net Beneficiaries, the PNAFN covered 250,000 households by 2015, after extending the target to 25,000 additional beneficiaries in 2014/15. An impact assessment of various scenarios to reduce energy subsidies was conducted, and the progress made on achieving outcome indicators 4 and 5 included the following: (a) a consultancy firm was contracted in August 2014 and prepared an analysis of the potential targeting criteria for reforming the energy subsidies, and (b) a second consultancy firm was recruited to assess the various scenarios to reduce energy subsidies; the report was completed in early 2016. 23. The World Bank team also provided technical assistance to the Tri-Partite Social Dialogue Technical Team (government, labor, and private sector) and delivered the draft TUNISIA-specific SimPLE model (simulation of labor impacts of social security reforms) to the Working Group (developed between March and July 2014 and delivered through training during September 2014). However, as the Government gave no indication on whether it would pursue subsidy reforms, the proposal and operational plan for the subsidy-compensation program were not developed. With regard to the targeting, a plan for improve targeting was prepared, and the questionnaire for the social survey was developed and approved by the National Statistics Council in March 2016. Finally, the company Oxia/WEVIOO was recruited; started developing an enhanced MIS for MAS (including the development of the first three modules: registration, social survey, and service life cycle management); and purchased and installed servers and IT equipment for the consolidation of existing but insufficient resources. split rating because it is a small Recipient-Executed Trust Fund. Page 12 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) B. Project Achievements after the First Restructuring Table 3. Restructured Project - Achievement of Outcome and Intermediate Indicators Indicator Name At First Target Actual (Revised at Project Restructuring, June 20, 2017) Restructuring (May 31, Achievement at (New Baseline as 2019) Completion of June 20, 2017) (May 31, 2019) Outcome Indicator 1. Number of participants in training, which enhances 2,067a 2,645 2,240 knowledge and capacity 2. Number of households who are updated with complete 2,000a 300,000 304,698 information 3. Development of targeting tool, socioeconomic Institutional Yes Yes classification applied to households registered in the framework database formalizedb 4. Number of assessed persons who received unique social 280,000 750,000 1,100,717 identifier Intermediate Indicators 1.1. Number of registered households in the new MIS 0 400,000 602,619 1.2. Number of modules developed for the new MIS 0 3 9 2.1. Implementation of interoperable platform No Yes Yes 2.2. Renovation of pension system in the CNSS No Yes Yes 2.3. Re-engineering of procedures of pension management No Yes Yes system in the CNRPS 3.1. Communication plan prepared and implemented No Yes Yes Note: a. The baseline was 0 in the Project Appraisal Document. At the time of the first restructuring, the Project had supported the training of 2,067 staff and information on 2,000 households had been updated in the beneficiary registry. b. Before the first restructuring, the PMU, MAS, and the CRES signed a Memorandum of Understanding (MOU) to formalize the institutional arrangement for targeting work. Key Indicator 1: Number of participants in training, which enhances knowledge and capacity 24. Partially achieved (85 percent). This indicator measured the training provided to build the capacity of the Government to design social protection reforms and improve targeting of social protection programs. Training was provided to 2,240 participants, the majority of whom (nearly 90 percent, or 1,969) were social workers and MAS staff at the regional level, against a targeted 2,645 (85 percent). Knowledge and capacity-building activities included workshops as well as personalized training and covered the entire spectrum of project activities. These capacity-building activities contributed to develop a variety of skills for a comprehensive approach to assess social protection programs and reform strategies, including issues related to targeting methodologies. MAS has developed a better understanding of safety net programs. 1,300 MAS social workers were trained to help them enhance their communication with current and potential beneficiaries and collect data for the social survey using tablets. Training on the new Proxy Means Test (PMT) was provided to MAS teams, including the CRES. The PMT was developed through the Project and to conduct descriptive analysis using the MAS database. MAS, the CNSS, and CNRPS also received a wide range of IT-related trainings, including on the social identifier, social registry, interoperable database, web services, and MIS. As the development of five MIS modules (intermediate indicator 1.2) was delayed, the training on these modules that planned to reach about 330 local and Page 13 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) regional officials did not materialize before the project closing date, which resulted in underachievement of this outcome indicator. Key indicator 2: Number of households who are updated with complete information 25. Achieved (102 percent). Information on 304,698 households has been fully updated in the new MAS database since April 2017, against a targeted 300,000. Furthermore, a total of 602,619 households have been registered against a targeted 400,000 (151 percent) (intermediate indicator 1.2). Beyond the project, the MAS plans to continue this activity to expedite the data collection from a total of 900,000 beneficiary households,13 which will consist of all PNAFN and subsidized health card beneficiaries as well as those who are on the waiting list. Key indicator 3: Development of targeting tool, socioeconomic classification applied to households registered in the database 26. Achieved. The PMT model was identified as the first targeting methodology to be developed by the Project. The PMT was developed using data from the 2010 Household Budget and Consumption Survey from the National Institute of Statistics. In May 2019, a series of trainings on statistics was provided to the technical staff of the CRES and MAS to apply this new model to a small subset of the households registered in the MAS database (246,457 households or about 81 percent of the 304,698 households under outcome indicator 2). In addition to the PMT model, another model that is based on an income and asset test (Hybrid Means Test [HMT])14 is currently being developed to compare the estimated performance of the different models. The development of the HMT model, however, is outside the scope of measurement for this PDO indicator. Key indicator 4: Number of assessed persons who received unique social identifier 27. Achieved (147 percent). The Project supported the expansion of the social identifier (which was only previously used for the contributive system) to the social protection system (noncontributory system). By project closing, out of the 1.2 million persons registered in the new MAS database, over 1.1 million were assigned a social identifier, against a target of 750,000. The social identifier is being assigned to each member of the household and not just the head of the household, which substantially enhances cross-checking of beneficiary information between different social protection databases. For example, the implementation of the social identifier has enabled MAS to cross-reference and automatically verify the status of social safety net beneficiaries, 30 percent of whom were identified in the national social security database. The Project also supported the evaluation of the social identifier system to improve its performance and data quality. 13 The 900,000 households comprise 250,000 PNAFN beneficiaries, 600,000 AMG 2 beneficiaries, and 50,000 households in waiting list. 14 In mid-2017, a decision was reached to also develop the HMT model, which is based on an income and asset test, where formal income is self-reported by the households and verified by the Government, and informal incomes are estimated based on certain verifiable characteristics. The decision to develop the HMT is appropriate for countries with moderate informal sector such as Tunisia, as it tends to give good targeting results that are generally above the PMT. The development of the HMT model has taken longer because administrative data must first be reviewed at the CNSS, CNRPS, the tax office, the land office, and MAS to ensure relevant data are available. Page 14 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Achievement of Other Key Activities Intermediate indicator: 1.2. Number of modules developed for the new MIS 28. Achieved.. MAS completed the development of a new MIS (Système d’information relatif aux familles nécessiteuses et à revenue limité, SIFNRL), including nine modules, against a targeted three. The first three modules were developed by 2016 and the remaining six modules in 2019: registration, social survey, service life cycle management, cross-checking, scoring and decision, appeal and fraud (GRM), payment, citizen access, and card. The new MIS enables management of information of current and potential social safety net beneficiaries. Moreover, MAS institutionalized15 the Information and Communication Technology (ICT) Unit within the ministry and set up a helpdesk. The project also financed the purchase of IT equipment (both at central and local levels), including 900 computers, 1,400 tablets, servers, software, and interoperability equipment for MAS, and consultancy services to develop the manuals for the enhanced IT system. This has enhanced the efficiency of MAS, as it did not have IT equipment such as computers before the project. Government agencies have recognized the important role of new technologies needed to modernize the social protection system. Intermediate indicator: 2.1. Implementation of interoperable platform 29. Achieved. The sectorial interoperable platform was established and is operational. The specifications for the platform have been developed and protocols established for data exchange among MAS (SIFNRL), CRES, CNSS, CNAM, and CNRPS. MAS also signed a convention with the Ministry of Local Affairs and Environment in March 2019 to authorize data exchange between the social sector and Civil Registry using the social identifier. This will enable intersectorial web services for citizens’ data on death, marriage, and divorce status, which are important for improving identification of beneficiaries and monitoring their status. To this end, MAS is also continuing to provide IT training to partners to integrate and manage their web-based services. Starting September 2019, the CNSS no longer requires the workers to provide hard copies of their birth certificates, which are now verified through the interoperability platform. Intermediate indicators: 2.2. Renovation of pension system in the CNSS and 2.3. Re-engineering of procedures of pension management system in the CNRPS 30. Achieved - Improvement of the MIS of social insurance funds. The CNSS renovated the private pension system (achievement of intermediate indicator 2.2) and received an award for good practices from the Africa International Social Security Association in August 2017. The review of the business processes of the existing system and design for enhancements have been completed and the framework created for the development of e-services, including (a) pension simulation; (b) generating of certificates (fiscal, pension status, following-up requests related to pensions); and (c) downloading of pension-related forms and documents. Additionally, the CNSS e-services have been deployed and the CNRPS’ public pension management system has also been reengineered (achievement of intermediate indicator 2.3), 15The establishment of the ICT Unit was specified in Decree #340, which was published in the Government’ s Official Journal on April 9, 2019 (http://www.legislation.tn/fr/detailtexte/D%C3%A9cret%20Gouvernemental-num-2019-340-du----jort-2019- 029__20190290034032?shorten=m49D). Page 15 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) with the installation of an updated framework and building of a death benefits module for the e-pension system. Intermediate indicator: 3.1. Communication plan prepared and implemented 31. Achieved. The communication plan was developed and implemented (achievement of intermediate indicator 3.1). A communication plan and campaign strategy were developed by a consulting firm following extensive discussions with MAS/CRES key staff, including social workers, and the Steering Committee. The plan identified the target population and developed key messages and guidelines for the communication campaign. Following the Tunisian Parliament’s ratification of the country’s first social protection law, the Amen Law, in January 2019, MAS launched a national communication campaign with the slogan “Ensure your right, come to a nearby local office of the Ministry of Social Affairs to be registered in the social registry.� The campaign was broadcast via radio, television, and Facebook, with 20 campaign days held at regional levels involving the Minister of MAS, Governors, and MAS officials. The communication was well received, and within three months of the launch of the campaign, 100,000 new applicants were registered in the SIFNRL database. C. Outcome Rating 32. The outcome rating before the restructuring is Moderately Unsatisfactory. As discussed earlier, before the first restructuring, the Project only achieved (and exceeded the target substantially on) one sub-indicator of the first outcome indicator target by training 2,067 people against a targeted 500. The Project also initiated activities to support the analysis of subsidy reforms but could not develop a plan and manual for a subsidy-compensation program for households, as the Government did not indicate its commitment to this work. Despite the lack of government commitment and high turnover of staff in the implementing agencies, the Project moved forward and contracted consultancy services to start work on developing the PMT model, establish interconnectivity between the insurance and pensions systems, and design and implement the MIS. In addition, IT servers and equipment were purchased for the MIS and the social survey questionnaire was finalized. These activities also supported the revised project activities that were implemented from late 2015. Considering some progress mentioned above despite all the political challenges, initial overambitious project design, and the refocusing of activities from late 2015 toward modernizing the social protection sector, the outcome rating before the restructuring is rated as Moderately Unsatisfactory. 33. The outcome rating after the restructuring is Satisfactory. Following the restructuring, the Project was successful in achieving and/or exceeding three of the four outcome indicators; made substantial progress on the training outcome indicator, by training 2,240 staff against a targeted 2,645; and achieved or surpassed all six intermediate indicators, as discussed earlier. 34. As a result, the overall outcome rating is Moderately Satisfactory (table 4). This rating reflects the substantial achievements made after the restructuring, as discussed above. The Project supported several high-value and critical activities, such as development of the beneficiary database, social identifier, targeting methodology, an interoperable platform for social assistance and safety net programs, and MIS with a relatively small amount investment totaling US$4.7 million over a period of five-and-a-half years. The efforts put into improving institutional capacity and modernizing the social protection system will enable better targeting of beneficiaries and potential future reforms of the social insurance funds and Page 16 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) social assistance programs. Building blocks of delivery mechanisms supported the important transition towards an effective, efficient and transparent social system, which is part of a critical transformational reform package for Tunisia to achieve more inclusive growth. Table 4. Overall Outcome Rating Before the After Restructuring Dimensions Restructuring (June 20, 2017) 1 Outcome rating Moderately Satisfactory Unsatisfactory 2 Numerical value of the outcome rating 3 5 3 Disbursement US$2 million US$2.7 million 4 Share of disbursement 43% 57% 5 Weighted value of the outcome rating 1.29 2.85 (Row 2 × Row 4) 6 Final Outcome Rating (Split Evaluation) (sum of Row 5) Moderately Satisfactory - 4 (rounding from 4.14) Note: The rating scale: Highly Unsatisfactory (1), Unsatisfactory (2), Moderately Unsatisfactory (3), Moderately Satisfactory (4), Satisfactory (5), and Highly Satisfactory (6). D. Other Outcomes and Impacts 35. The achievements of this Project contributed to two larger World Bank lending operations as well as technical assistance on subsidy reform in Tunisia. The first is the Tunisia Investment, Competitiveness and Inclusion Development Policy Financing (P161483) in the amount of US$500 million, approved on June 27, 2018) and which supported two prior actions related to Tunisia’s new social protection law and new child grant program. Thus, the Digital Transformation for User-Centric Public Services Project (P168425) (GovTech Project in the amount of US$100 million, approved on June 14, 2019) will continue to support the initiatives launched under this Project, including progressing activities to improve targeting, finalizing the social survey, carrying on enhancements for database interoperability, implementing universal health cards, and correcting errors in the CNSS social identifier database. Furthermore, the World Bank continues fostering its engagement in subsidy reform to support the Government’s efforts to address fiscal constraints, as illustrated by a new technical assistance, launched in 2017 and funded by the Energy Sector Management Assistance Program (P167211). III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME Shortcomings in Project Preparation 36. The shortcomings in preparation were the key reasons for the delayed implementation in the early years of the project. The lack of proper assessment of the readiness of the MAS’ IT systems, the social registry, and relevant agencies’ databases for interoperability of information resulted in identifying activities that could not be undertaken immediately, without first investing considerable time and resources to developing critical basic tools and systems as well as protocols between government agencies. As the assessment of the existing systems required procurement of consultancy services and detailed reviews to take place to identify the steps to move forward, the implementation of the Project was substantially delayed during the first two years. Page 17 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Transitional Governments Delayed Initial Project Implementation 37. Government transitions, institutional restructures, and staff turnover. From project conceptualization in 2012 to 2017, Tunisia had seven different administrations, four MAS ministers, five Director Generals at the CRES, and recurrent changes in the technical staff at the CRES and the PMU. The transitional nature of the country, political instability, and high turnover of counterparts disrupted project implementation till mid-2017. As noted earlier, this affected the commitment, ownership, and/or knowledge of the Project in all partnering agencies, particularly during the initial years. A Steering Committee for the Project was established in May 2013, but a subsequent government reshuffle before the Project became effective in November 2013 led to a change in the chairmanship of the Steering Committee and weakened the support for the Project. The changes in government parties and high officials also led to shifts in priorities. The interim government had a strong focus on reforming energy and food subsidies; however, as global energy prices fell in 2014, the subsequent elected administrations were less committed to reforming the universal subsidies in favor of maintaining political and social stability. The time required to form new administrations and set new agendas further delayed project implementation, while the highly complex and dynamic political economy landscape made it difficult for the World Bank to navigate political support to expedite the Project when key actors kept changing. Coordination and Governance 38. The Steering Committee’s close oversight and guidance was key to the Project’s many achievements, including facilitating mobilization and cohesion among the implementing partners; demonstrating strong leadership by closely monitoring the Project’s implementation; and ensuring that issues were timely resolved, which helped complete key activities by project closing. Two critical elements facilitated the establishment of data interoperability: the decision to expand the social identifier to noncontributory social protection programs and the creation and adoption of a social identifier by all relevant agencies. To facilitate this, in early 2017, the CRES and the insurance agencies CNSS, CNRPS, and CNAM created the Technical Committee to manage the development of the social identifier for resolving issues and correcting data errors, communicating key messages, and examining the quality of data certification undertaken by the CRES for issuing the social identifier number. This technical group brought cohesion and enhanced the stakeholders’ commitment to the work. In addition, MAS signed a convention with the Civil Registry for data exchange using the social identifier, with more inter-ministerial agreements planned to be signed for data interoperability. Complex Implementation Structures and Limited Capacity Confused Responsibilities 39. The implementation structure was highly complex and made coordination of activities and decisions difficult. In accordance with the country structure, the MDICI was signatory to the Grant Agreement and maintained portfolio oversight. An Inter-ministerial Steering Committee, headed by a high-level official from the Office of the President, had overall governance and monitoring responsibility for the Project. The MOF housed the PMU because the initial planned reform of subsidies came under its responsibilities and because the staff also had some experience in managing finances; different implementing agencies within MAS were responsible for the implementation of project activities. A detailed diagram of the implementation structure is provided in annex 4. This implementing structure created a disconnect between the governing bodies and implementing agencies, for example, the Steering Committee Chair did not have the authority or the capacity to designate staff when human resource issues Page 18 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) arose within the PMU and MAS, nor the ability to allocate other financial resources. This multi-stakeholder arrangement complicated coordination and clearance processes, while the Steering Committee’s detailed involvement at times also overburdened the decision-making process and further delayed the execution of project activities. 40. Inadequate staffing arrangements of the PMU. The PMU was composed of MOF staff whose time was partly dedicated to the Project, as they were also entrusted with other responsibilities within the ministry. The project coordinator and team members were assigned from different directorates within the ministry and were periodically reassigned, and the entire structure of the PMU was changed thrice. Although, the World Bank provided periodic training on procurement and FM, the high turnover of staff made it difficult and time consuming to build institutional capacity and ownership. 41. Despite a challenging implementation structure and difficulties faced with fiduciary management, the PMU’s enhanced capacity gradually enabled it to support the project’s procurement activities. The PMU was especially resourceful during work overloads, in developing the bidding document based on the terms of reference prepared by the implementing agencies and the procurement of numerous goods consultancy services, which contributed to the execution of a large number of different activities and eventually accelerated project implementation. Focus on Targeting and Need for an Integrated Approach to Improve Eligibility Testing 42. As reflected in the PDO, the Project had a focus on improving the targeting, and this focus was initially rather narrow but expanded over time. While there was a consensus on the need to improve the targeting based on an estimation of targeting performance that used the 2010 Household Budget and Consumption Survey from the beginning, there was no robust analysis to inform or build a consensus on the optimal targeting approach in Tunisia due to a lack of data. At that time, a main proposal by the World Bank was to start with the PMT based on international experience. Succeeding discussion supported by the World Bank’s technical assistance broadened the scope of the targeting work from the development of one targeting model (that is, PMT) to explore different approaches. In addition, the restructuring broadened the Project’s focus to building blocks essential for improving the eligibility assessment and program monitoring, such as the beneficiary database, MIS, interoperability, and grievance redressal. Rather than the targeting methodology alone, the integrated approach for improvement of the whole system revealed that delivery issues were essential for improving eligibility testing; thereby also enabling the effective management of a social protection system. Project Restructuring Helped Set the Project on an Achievable Path 43. In-depth and open discussions during the six-week MTR workshops helped enhance clients’ ownership of the project. The workshops, held between January 26 and March 10, 2016, included the full participation of the Head of the Steering Committee, its members, and other stakeholders and facilitated a shared understanding of disbursement issues and how to address them; contributed to identifying the capacity-building needs of counterparts for monitoring and evaluation (M&E), procurement, FM, and disbursements; and enhanced trust and fluidity in the exchange of information. This collaboration facilitated refocusing the project activities from 2016, based on a sound diagnosis of IT systems and capacity, and was a key reason for the Project making substantial progress toward achievement of the project objectives. Page 19 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) 44. A phased extension of the closing date had positive implementation impacts despite the administrative burden. In their formal communiqué for the restructuring, the Government requested that the Project be extended for a year. Subsequently, agreement was reached for a phased extension of the Project. Had the first extension been approved for one year, the end target would have been set at 400,000, which the Project would not have achieved. This phased approach also strengthened the incentive to roll out the social survey nationwide. Despite burdensome administrative processes needed to expedite the two restructurings within six months, these prudent decisions were critical to the Project achieving the majority of targets by project closing. Challenges Related to the Delayed Implementation of the Social Survey 45. The unique approach to fully rely on MAS social workers delayed implementation, as they went on strike and delayed the fielding of the social survey. The alternative plan was for MAS to contract a private firm to conduct the social survey in 2014; instead in mid-2015, the ministry confirmed that its social workers would conduct the surveys to ensure data confidentiality, retain institutional capacity and ownership of the survey, and ensure sustainability of data and the application and use of the new policy and program management models. However, fielding of the survey was put on hold in the beginning of 2016, as the social workers, through the labor union (UGTT), went on strike seeking compensation for conducting this new activity. There may also have been misunderstanding among the social workers that use of the new PMT score would limit their ability to respond to specific cases or lessen their relevance in the selection of beneficiaries for social protection programs. The concerns of the social workers were addressed by (a) increasing communication to the social workers on the advantages of using modern technology to develop a beneficiary database that helps them implement their programs, (b) strengthening the internal working arrangements and communication within MAS by holding regular follow-up meetings with key stakeholders and disseminating unified messages, and (c) setting up a committee within MAS for close monitoring and oversight of the survey. The social workers ended their strike in October 2016, when the Minister of Finance approved a salary increase for them. Subsequently, the field activities were launched in April 2017 in 6 of the 24 governorates and expanded to the remaining governorates in February 2018. 46. A lengthy questionnaire and data collection process also delayed field implementation of the social survey. Although steps have been taken to improve logistical support to conduct the surveys, the survey instrument itself is lengthy, as it consists of around 170 questions that take on average about two hours to complete for each household. Combined with the implementation challenges, as discussed above, the length and the design of the questionnaire meant that data collection was extremely slow and long-drawn compared to other countries. Tunisia completed the data collection for 300,000 households after more than three years, while other countries usually plan to complete this process within one year to ensure data integrity and quality.16 47. The provision of improved logistics made household data collection less burdensome. An assessment of the PNAFN, in 2013, showed that it needed to use state-of-the-art technologies and provide a solid platform for program management; however, there was a lack of means and adequate incentives for social workers to collect information from beneficiaries. The Project provided pragmatic solutions to 16For example, the Takaful and Karama program, supported by the World Bank’s Egypt Strengthening Social Safety Net Project (P145699) registered over 1 million beneficiaries within the first year of implementation, from May 2015 to May 2016. Page 20 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) support the data collectors through the provision of resources such as rental cars and drivers for field visits, the use of laptops and tablets for electronic data collection, SIM cards to facilitate communication in the field, and Wi-Fi connectivity for real-time data sharing. These logistical improvements facilitated field implementation of the survey from February 2018. Since then, the Project has made steady but very slow progress. By project closing, the assessments for 304,698 households were completed, with an additional 68,000 households near finalization, out of a total of 900,000 households. IT Business Project Management Support from a Private Firm Was a Key Factor in Developing the MAS MIS Modules 48. An IT consulting firm, contracted in 2013, conducted an audit of the existing information systems at MAS, developed a road map for updating the system, and supported the management of implementing the system enhancements. This resulted in the development of the first three MIS modules (registration, survey, and services life cycle) by 2016. Due to a lengthy procurement process, the remaining five modules remained undeveloped until May 2019. MAS made a prudent decision to hire private firms to review and identify the remaining business processes and system specifications, and support IT enhancements, including development of the MIS modules. This partnership was critical, as it brought in expertise that did not exist within the Government and facilitated the transfer of knowledge from the private sector to the public sector and enhanced government recognition that outsourcing specific tasks to the private sector could achieve a greater result. This also enabled critical MIS activities to be progressed on the additional five modules: (a) decision (scoring and targeting); (b) data crosschecking (providing for interoperability with social security database); (c) payment; (d) fraud prevention and control; and (e) public access. IV. WORLD BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME World Bank Performance 49. The World Bank team was responsive to prepare and appraise the Project within a short time frame to meet the MENA Transition Fund deadlines. The World Bank evaluated the procurement and FM capacity of the PMU, but a critical review of the databases, IT, and MIS of the relevant agencies could not be conducted during project appraisal due to the tight time frame. This led to an overestimation of Government’s capacity and systems readiness and resulted in an overambitious project design. The World Bank also assessed the political and stakeholder risks to be high, and the mitigation measures that were identified proved to be inadequate given the challenging circumstances associated with the political economy.17 50. The World Bank team demonstrated a strong commitment from the beginning, including conducting 4 missions in the first year, 2–3 missions per year in the following years, and more frequent missions in the last years. Despite these frequent missions, they were still not sufficient or adapted to the 17The mitigation measures included the following: (a) the Government closely monitoring the country’s environment and engaging with political and civil society stakeholders; and (b) the World Bank providing technical assistance and communication support focusing on improving equity in subsidy expenditure and improving sustainability of pensions and health insurance. With the changes to governments, the government’s agenda and priorities, and frequent turnover of key counterparts, these measures did not adequately address the risks. Page 21 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) needs and expectations for an effective launch of the project activities. While the World Bank team provided ongoing technical assistance, the PMU needed more focused and direct support, especially given the numerous issues related to fiduciary management. The World Bank’s close implementation support during the six-week MTR mission led to identifying feasible activities for the project. The incoming Task Team Leaders (TTLs) participated in 6–8-week handovers to get a clear understanding of the Project and established strong relationships that built trust and commitment of government counterparts. Following the MTR, the World Bank provided closer monitoring and implementation support, which helped quickly address issues and make important progress toward achieving the development objective. The close implementation support provided, notably by the last TTL through regular virtual meetings, was central to the Project’s successful completion. Overall, M&E quality was rated Substantial.18 The World Bank was proactive and identified issues and worked closely with the Steering Committee and the PMU to find solutions, as required, and used adequate resources, including dedicated technical support from FM, procurement, and MIS specialists. Therefore, the World Bank’s performance is rated Moderately Satisfactory. Safeguards and Fiduciary Compliance 51. Safeguards compliance. The project was rated Category ‘C’ and thus did not trigger any safeguards policy. 52. Fiduciary compliance. The PMU structure and capacity affected fiduciary compliance. Initially, only one FM officer and one procurement officer from the MOF were assigned part-time to the PMU. The World Bank team provided substantial training on FM and procurement; however, the high turnover of PMU staff and their inability to fully dedicate time to the Project repeatedly delayed procurement activities and reporting, which improved after additional staff were assigned to the Project in early 2017. 53. Procurement. Procurement activities were delayed for much of the life of the Project, stemming from delays in the preparation of required documents, including terms of reference, justifications for new requests, and evaluation reports. In the early years, the PMU did not have a dedicated procurement officer, and there was also high turnover/reassignment of staff. In addition, the PMU was confused on whether to apply the Government’s new national procurement rules (adopted in 2014) or the World Bank’s rules and questioned the World Bank’s procurement procedures even though they followed international standards. Throughout the project life, the PMU continued to enforce its own national standards, which requires additional clearance process with much lower thresholds than the World Bank’s and resulted in lengthy delays. Procurement capacity and compliance improved following the hiring of a procurement assistant in January 2017, leading to a large number of contracts being signed at an accelerated pace, which also resulted in increased disbursements. Further training, however, was still required to enhance the PMU’s procurement capacity to ensure it fully understands the World Bank’s rules. 54. Financial management. All interim fiduciary reviews were submitted on time; however, there were systematic and notable delays19 in the submission of audited annual financial reports, all of which were qualified with reservations expressed by the auditor. The delays were partially due to the Tunisian 18 For M&E quality, the ratings are Negligible, Modest, Substantial, and High. 19 For example, the audit report for 2017 was received with four-month delay, on October 30, 2018. Page 22 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) law requiring audits to cover all procurement activities rather than a sample and the lack of MIS for collecting information from the regions that resulted in poor quality data and also delayed the generation of reports. In addition, the nature of this small grant project (in other words, not being a part of the Government budget) made it a lower priority for the MOF and the auditors’ office, which further delayed clearing of audit reports for submission to the World Bank. Fiduciary activities improved in early 2017 following the assignment of an FM specialist and two part-time FM assistants to support the FM specialist. Although the assistants were assigned part-time to the Project, they spent adequate time (more than 50 percent of their work time) to support project activities. 55. Progress report. The Project’s semiannual progress reports, which are due 45 days from the reporting period, were always submitted with a notable delay (around 2 to 4 months delay during 2016– 2018), while the updated results framework was shared with the World Bank during the implementation support missions. The project completion report has not been received to date (November 2019). Risk to Development Outcome 56. This Project was designed to start building the basic foundations for modernizing Tunisia’s social protection system. It has largely met its development objective through the achievement or overachievement of all but the first outcome indicator—Number of participants in training, which enhances knowledge and capacity—achieving 85 percent of the target. 57. The modernization of the social protection sector remains high on the agenda and the Government has given its full commitment to complete the activities started under this Project. These activities will be further supported by the World Bank GovTech project (P168425), which aims to operationalize the systems developed to improve the implementation of social assistance programs. In addition to finalizing ongoing activities, the GovTech project will support the simplification of business processes, apply case management, implement universal health/social protection cards, and build greater interoperability by linking the national identifier and the national platforms. The activities already launched by this project and to be completed are as follows: (a) For the social survey, the slowness of data collection is an outlier, requiring a number of interventions to accelerate its speed and ensure data integrity. Despite the progress, the social survey could take another three years to complete, based on the current execution speed of the survey. In comparison, this type of data collection is usually done within one year in other countries to ensure data integrity and relevance. Specifically, for the current round of data collection, the social workers would benefit from continued access to transportation for household visits and additional administrative support for manual paper work. In addition, for the next round of data collection, different approaches need to be introduced for faster implementation, including the simplification of the questionnaire20 and the creation of a portal for the first step of registration. (b) The singular approach to fully rely on social workers for the social survey delayed implementation. The implementation of the social survey experienced significant delays, because of the lengthy negotiations needed to end the social workers’ boycott and also due 20 The current questionnaire consists of 170 questions and takes two hours to complete. Page 23 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) to their heavy workload. At the same time, there was a huge heterogeneity in implementation across the country. For example, as of May 2017, three top governorates completed the survey for around 50 percent of the beneficiaries in their governorates; other bottom governorates completed less than 5 percent, while the national completion rate was an average of 15 percent. The overall performance would have been improved had different approaches been tested to inform what works better at the local level. To accelerate the data collection process in the future, it may be worth considering piloting of different approaches (that is, strengthening coordination with administrative staff at the local level, hiring temporary staff, partnering with private firms/nongovernmental organizations), while the social workers would continue to play a critical role in validating the information regardless of the approaches adopted. (c) Government’s official endorsement is needed to improve targeting of the social protection programs. The new targeting model (based on the PMT) has been developed, but this technical work has not resulted in a policy decision. Building upon the new social protection law, the new targeting methodology has to be adopted by the Government for it to be operationalized; the PMT has to be applied to all households registered in the MAS’ new databases and for MAS to select beneficiaries for social protection programs based on the new approach. For this, round tables/discussions with key stakeholders, including civil society, academia, and labor unions, are critical to gain their support and buy-in on the new targeting approach. Furthermore, the Government’s official endorsement is also essential for data access, as the quality of technical work depends on whether relevant data are made available and are accessible by various institutions, such as household survey data from the National Institute of Statistics and administrative data related to income (for example, tax, automobile registration). (d) Commitment from all stakeholder agencies is required to address the data quality issues with the social identifier and strengthen interoperability. Although the social identifier is fully functional, there are issues with data quality. A recent assessment of the social identifier system identified several action plans and highlighted the importance of institutional arrangements and closer coordination among stakeholders. Furthermore, real- time web services need to be deployed (including with the Civil Registry, as per the MOU signed with the Ministry of Local Affairs and Environment) through the interoperable platform that will automatically update beneficiaries’ data from the linked databases. Moving forward, close collaboration across stakeholders outside of the social protection sector is critical for facilitating data exchange through the national interoperable platform, as this will require linking the social identifier and the national identifier. (e) The new business process and MIS modules need to be deployed to the end users for actual utilization. Following the development of the new MIS modules, training on the new business processes for the social protection programs (planned for November 2019) need to be rolled out, including to partner agencies to help them integrate and manage their web- based services. Given that the new MIS will be able to generate beneficiary information more systematically, MAS will have better information to assess and monitor the beneficiaries for program management. Specifically, social workers could use this database to provide customized services to individual persons depending on their characteristics and Page 24 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) needs (case management). While the database is currently limited to the beneficiaries of the two main social assistance programs, the coverage and use of the database can be expanded to other programs (assistance for the disabled and scholarships) and the interoperability can be further enhanced to better assess and identify the needs of the people. V. LESSONS LEARNED AND RECOMMENDATIONS 58. The PDO which clearly reflected the government’s priorities allowed for greater flexibility and project adaptation. The PDO was formulated to support broader Government reforms, although the subsidy reform did not take off as initially envisaged. This definition of the PDO, consequently, enabled significant shifting of the project activities toward building the prerequisite delivery systems needed for potential social protection reforms while remaining aligned with the overall project objective. 59. Clarification about the terms of the Grant Agreement and a simpler implementation arrangement would have benefited the Project. The Grant Agreement was signed with the MDICI; MAS was the implementing agency; and the PMU, with responsibility for fiduciary management, was hosted within the MOF. This complex implementation structure was adopted with critical responsibilities resting with different government institutions due to the initial focus on the subsidy reforms. Inadequate clarification about the Grant Agreement during project preparation created confusion within the PMU, which believed it had more autonomy and flexibility on the use of the funds beyond the scope of the Grant Agreement. This confusion greatly delayed disbursement of funds by the PMU for procurement activities undertaken by MAS for the Project. The implementation structure was not revised through the restructurings, as other issues took priority (that is, low disbursement, PMU staffing, uncertainty about the social survey). However, the Project may have benefited from a simpler implementation structure, in which the head of the Steering Committee, fiduciary functions, and the lead implementation agency resided within the same ministry. 60. Human and financial resources were grossly underestimated. Given that this Project was implemented amid a highly transitional and uncertain environment with high staff turnover in government, more human and financial resources were needed for project management and administration. The composition of the PMU was changed several times, staff members were reassigned, and there were frequent turnovers of the head and technical staff of the CRES. The Project would have benefitted from management officially designating qualified full-time or part-time staff to the Project to ensure they allocated adequate time and support to project implementation. Subsequently, the Government and the World Bank team spent substantially more staff time and budget than initially allocated to implement the Project, including for the two-year extension. Considering this massive investment, it is highly recommended to capitalize on the human resources and capacity built by this Project to support future similar projects, particularly for fiduciary functions. 61. The Project achieved a greater value for money because of the Government’s counterpart budget. This Project achieved many important activities for a relatively small amount, as it started building the foundations of a modern social protection delivery system. This was made possible because the Government covered almost all human resource/staff cost, which is estimated to be almost equivalent to Page 25 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) the amount of the grant proceeds.21 In addition, the Government’s counterpart budget was also allocated for additional IT equipment, establishing the network and transportation cost related to the social survey. As a result, the Project was able to allocate available funds from its budget (that is, grant proceeds) to launch the new reform initiatives, which the Government could not allocate from the national budget in 2013. This financial arrangement led to realizing cost efficiencies while also ensuring that the capacity built through the Project stayed within the government structure. 62. Detailed assessment of the risks and mitigation measures would have helped better define the scope of this high-risk project. The tight timelines of the MENA Transition Fund limited project preparation, and activities were selected based on what the interim government wanted (a unified database and targeting mechanism) without first assessing the readiness of existing tools and systems to determine appropriate and feasible activities. Throughout the implementation period, the political economy of the country significantly affected this Project primarily due to the many government reshuffles and changes in agenda and key counterparts. For example, lengthy negotiations with the labor union to resolve compensation issues and end the strike of the social workers also delayed implementation of the social survey. The Project could have avoided these significant delays through early and continued high-level dialogue to identify appropriate mitigation measures and concrete alternative plans/activities. Furthermore, a thorough assessment by experts during the first year of implementation would have facilitated an earlier redesign of IT-related activities and lessened the duration of implementation, as this Project relied on building complex IT systems. 63. Government ownership at the highest levels and social communication are critical for implementing sensitive reforms. The project activities were designed in response to intense social pressure, with successive governments not communicating any information to the public about the potential sensitive subsidy reforms. Commitment from the highest level of Government was needed combined with communication to gauge public sentiment to identify and take appropriate measures for gaining public support. For example, following the approval of new social protection law in January 2019, the rollout of the national communication campaign increased awareness about the social protection programs and the social registry, which led to 100,000 new applications within three months. 64. The Steering Committee’s close oversight and guidance was key to the Project’s many achievements. The Project benefited greatly from relying on the high-level Steering Committee, which was both responsive and empowering and brought cohesion among the implementing partners at a time when they were undergoing institutional restructures, with changes in mandates, leadership, and staff. The Steering Committee demonstrated strong leadership by closely monitoring project implementation and quickly resolving issues, which facilitated completion of key activities by project closing. The close working relationship with the Steering Committee and regular interactions with the government team also helped enhance the capacity of counterparts and strengthened project ownership. The World Bank’s regular follow-up was time consuming and costly, but its broad and pragmatic support helped build trust and get buy-in from counterparts, which helped overcome implementation challenges and put the Project back on schedule. 21These include social workers and administrative agents who worked on the social survey, PMU staff, and central staff from various implementing agencies. Page 26 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) 65. Partnership with the private sector was essential for modernizing Tunisia’s social protection system. The Government collaborated with a management consulting firm and an IT firm to identify and develop the MIS for the social protection programs. This outsourcing enabled the Government to benefit from external ICT expertise and transfer of knowledge and establish connections with the constantly evolving private IT sector. Without this collaboration, the ICT enhancements would likely not have taken place. In addition, the partnership with the private sector was not limited to the IT sector, as all communication activities were implemented in close collaboration with private firms. Moving forward, this partnership model seems to be promising for the Government to achieve greater results, complemented by the specific expertise available in the private sector. 66. Ongoing activities should be continued to achieve the modernization of the social protection system. While this Project started building the foundational tools and systems toward modernizing the social protection system, most activities are only half developed and need further work to fully operationalize a modernized social protection system. These modernized systems can support the rationalization of social protection programs as well as the implementation of subsidy reforms in the near future. Further efforts and support, including collaboration with the private sector and international expertise are recommended to help implementing agencies complete the remaining critical activities (see section IV, Risk to Development Outcome). . Page 27 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS A. RESULTS INDICATORS A.1 PDO Indicators Objective/Outcome: To strengthen institutional capacity to design social protection reforms and improve targeting of safety net programs Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Development of targeting tool, Yes/No N N Y Y socioeconomic classification applied to households 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 registered in the database Comments (achievements against targets): Target met. The first PMT model has been developed and applied to a sample of households (246,475) registered in the database of the MAS. In this process, a series of customized and hands on training were organized for the researchers, statisticians, operational staff, and engineers of the CRES and the MAS so that they will acquire the statistical skills to handle data and to replicate the application of the targeting models as they get revised (link with another PDO indicator on training). In addition, the targeting module was developed as a part of the management information system (intermediate indicator 1.1). Furthermore, the development of the second model (HMT) is underway, with progress in analytical work using the labor market survey and on the estimation of agricultural income. Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Page 28 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Number of assessed persons Number 0.00 0.00 750000.00 1000717.00 who received unique social identifier 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Target exceeded (134 percent achievement against the target value). Out of the 1.2 million persons who are registered in the MAS' new database, social identifier was assigned to over 1 million persons (82 percent), achieving the target value of 750,000. (The validation is ongoing for the remaining persons (about 18 percent).) The implementation of the social identifier enabled the MAS verify the social security contribution status of social safety net beneficiaries, finding 30 percent of household members of social safety net beneficiaries were found in the database of social security systems. This achievement is very important, as Tunisia aims to establish social protection floor through reforming both non-contributory and contributory systems. It is also important to note that the evaluation study of the social identifier was successfully completed in March 2019, and its results and recommendations are used to improve the quality and the process of assigning the social identifier. Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Number of participants in Number 0.00 0.00 2645.00 2240.00 trainings which enhance knowledge and capacity 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Target was partially met at 85 percent of the target value. This indicator did not meet the target due to the delay in the development of modules of the MAS' management information system (intermediator indicator). Consequently, the trainings of new additional modules did not take place before project closing. The duration of training varied from half day to 2 weeks, and the biggest training about the social survey in which over 1,300 social workers were trained. There were a series of training related to the IT, and a total of nearly 900 staff were trained. Page 29 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Number of households who are Number 0.00 0.00 300000.00 304698.00 updated with complete information 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Target met (102 percent of the target value). Updating complete information was finalized for 304,698 households, with additional 68,000 households to be finalized shortly. A.2 Intermediate Results Indicators Component: To strengthen institutional capacity to design social protection reforms and improve targeting of safety net programs Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Number of registered Number 0.00 0.00 400000.00 602619.00 households in the new MIS (Management Information 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 System) Comments (achievements against targets): Target exceeded (151% achievement against the target value). Page 30 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Number of modules developed Number 0.00 0.00 3.00 9.00 for the new MIS (Management Information System) 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Target exceeded (300 percent achievement against the target value). With the support from the external management consulting services and developers, the MAS completed the development of 9 modules: registration; social survey; service life cycle management; cross-checking; scoring and decision; appeal and fraud (GRM), payment; citizen access; and card. Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Implementation of Yes/No N N Y Y interoperable platform 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Achieved. The interoperable system is operational with the specifications developed and protocols established for data exchange between different government agencies, including between MAS’ MIS system (MAS, SIFNRL), CRES CNSS, CNAM, and CNRPS. The MAS also signed a convention with the Ministry of Local Affairs and Environment in March 2019 to authorize data exchange between the social sector and Civil Registry using the social identifier. Indicator Name Unit of Baseline Original Target Formally Revised Actual Achieved at Page 31 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Measure Target Completion Renovation of pension system Yes/No N N Y Y in the CNSS 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Achieved. The CNSS renovated the pension system. The review of the business processes of the existing system and design for enhancements were completed, and the framework created for the development of the following e-services: (i) pension simulation (ii) generating certificates (fiscal, pension status, following-up requests related to pensions) (iii) downloading pension related forms and documents. The e-services were also deployed. Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Re-engineering of procedures Yes/No N N Y Y of pension management system in the CNRPS 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Achieved. The CNRPS’ pension management system has also been re-engineered, with the completion of the following two activities: (i) installation of the design and development framework; and (ii) development of module for death benefit which is a part of target e-pension system. Training was provided for the transfer of skills. Unit of Formally Revised Actual Achieved at Indicator Name Baseline Original Target Measure Target Completion Page 32 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) Communication plan prepared Yes/No N N Y Y and implemented 05-Nov-2013 30-Jun-2017 31-May-2019 31-May-2019 Comments (achievements against targets): Achieved. The social communication plan was developed and launched in January 2019. The plan identified the target population and developed key messages and guidelines for the communication campaign. The national communication campaign to raise awareness about the social protection program and social registry was launched by the MAS after the Tunisian Parliament ratified the country’s first social protection law, the Amen Law in January 2019. The campaign was broadcast via radio, television and Facebook, with 20 campaign days held at regional levels involving Governors and MAS officials. The communication was well received and within three months of the launch of the campaign, 100,000 new applicants were registered in the MAS’ database. Page 33 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) .. Page 34 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) ANNEX 2. PROJECT COST BY COMPONENT Revised Borrower Actual at Amount at Amounts Financing Project Percentage Approval at First (US$, millions) Components Closing of Approval (US$, Restructuring (US$, (%) millions) (US$, millions) millions) Total Financing 4.70 4.70 4.70 — — 1. Subsidy and Safety Net — 3.78 2.15 2.66 80 Reform Support 2. System Integration — Support for Social 0.40 2.30 1.84 9 Protection 3. Communication, Project — Management and 0.40 0.25 0.20 9 Monitoring Reserve (miscellaneous) 0.12 2 — Total allocated/expended 4.70 4.70 4.70 100 5.00 (estimated) Page 35 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) ANNEX 3. RECIPIENT, CO-FINANCIER AND OTHER PARTNER/STAKEHOLDER COMMENTS 1. In preparing this Implementation Completion and Results Report (ICRR), the World Bank benefited from the previous progress reports, other project documents, and interviews with the key government counterparts and the World Bank members. It is also worth noting that the Government has been preparing its own ICRR for the Project; however, the report has not been received by the World Bank to date (November 2019). The World Bank shared its draft ICRR with the Government counterparts and discussed it with them during the mission in September 2019 to reflect their comments in finalizing this document. Based on the Government’s last progress report (second semester of 2018), this annex summarizes the assessment of progress and the remaining priorities that are also relevant to the context of project closing. 2. Based on the progress of the Results Framework, the majority of the indicators’ targets have been achieved. In fact, the progress on the number of registered and surveyed persons is satisfactory. In parallel, the social identifier has started to cover the spouses and dependents and not only the household heads. In addition, the new information system of the social security funds has been implemented, allowing better data exchange. The communication effort/campaign has also made a remarkable impact on the local offices of MAS. Owing to the quality of the communication plan, as well as the content of the materials, there is an increased demand for social protection programs. Continued work, however, is needed for the remaining activities, including work on targeting, which is also needed for developing the modules on decision and scoring. Furthermore, there is a difference in the number of people surveyed between the regions that needs to be addressed. 3. It is important to devote efforts on the priority activities. Main recommendations are summarized as follows: • The CRES to speed up the development of the targeting model. This needs to be coordinated with MAS so that the new model is programmed into new information system. • Improve the social identifier system of the CRES by addressing and reducing the rejected cases. • Solve the compatibility problem of the software for data management. • Provide more intensive technical assistance to the MAS regional offices where the implementation of the social survey is slow. • Take advantage of new knowledge that the MAS staff have acquired through the training, and maintain the technical teams (both IT and operation) for the evolution of MAS’ new information system. • Expand the scope of interoperability, anchoring both national and sectorial projects, to benefit from better coordination across the public services (for example, health, education). • Evaluate the impact of the communication campaign. Page 36 of 37 The World Bank DTF: Tunisia Social Protection Reforms Support Project (P144674) ANNEX 4. PROJECT IMPLEMENTATION STRUCTURE Inter-ministerial Steering Committee MDICI Grant Signatory/Recipient Head: High-level representative from the Office of the President Members: MOF, MAS, MDICI, CRES, CNSS, CNRPS, CNAM, Ministry of Responsibility: Country’s portfolio Technology, Information, and Communication, Ministry of Commerce oversight and monitoring Responsibility: Guiding project activities Ministry of Finance Ministry of Social Affairs Grant Management Project Implementation Agencies Implementing Unit PMU Responsibility: Social survey/registry, SIFNRL, MIS, interoperable platform, Responsibility: Grant disbursement Fund communication, and grievance and project management, request procurement, FM, and M&E (from the CRES Office of President) Responsibility: Social identifier and targeting Fund model disbursement CNSS, CNRPS, CNAM Responsibility: Modernizing the insurance schemes, including MIS and interoperable platform Page 37 of 37