Social Protection & Jobs DISCUSSION PAPER NO. 1803 Rapid Social Registry Assessment: Malawi’s Unified Beneficiary Registry (UBR) Kathy Lindert, Colin Andrews, Chipo Msowoya, Boban Varghese Paul, Elijah Chirwa, and Anita Mittal November 2018 Rapid Social Registry Assessment: Malawi’s Unified Beneficiary Registry (UBR) Kathy Lindert - Colin Andrews - Chipo Msowoya Boban Varghese Paul - Elijah Chirwa - Anita Mittal November 2018 Abstract This paper reports on a rapid assessment of Malawi’s integrated social registry, known as the Unified Beneficiary Registry (UBR). The timing of the assessment was ripe given the upcoming round of continued expansion of the UBR and a planned shift in registration targets (from 50 percent to 100 percent of households). As such, the objectives of this assessment are to: (a) review the UBR experience to date; (b) identify strengths and areas for improvement; (c) provide short-term recommendations to support the upcoming expansion, including implementation adaptations that would be needed to accommodate the revised registration targets; and (d) support the longer-term strengthening of the UBR. While primary audience for this paper includes the core stakeholders in Malawi, the report is also of potential interest to other countries interested in developing social registries and/or carrying out social registry assessments. Malawi’s UBR has many strong fundamentals. The Government has taken the lead in designing, managing, and implementing the UBR with strong ownership across the core agencies involved. Implementation is carried out by existing decentralized institutional structures, which is a major strength. Implementation processes and information systems are effective, and most importantly, data quality is robust and registration coverage is rapidly expanding. Nonetheless, the report identifies key short-term and longer-term actions that could address challenges and strengthen the effectiveness of the UBR, including in the areas of institutional arrangements, implementation processes, information systems, data quality, links to user programs, communications, and a possible rebranding of the UBR to support better understanding of this powerful tool for inclusion and coordination in social protection and beyond. JEL Codes: D60, D70, D80, H41, I38, O55 Keywords: Malawi, social registries (aka unified registries, single registries, unique registries, registration and eligibility systems), integrated social protection information systems, integrated social information systems, delivery systems, social assistance, transfers, social protection, Azerbaijan, Brazil. 2 Contents Abstract..................................................................................................................................................... 2 Acknowledgements .................................................................................................................................. 5 Executive Summary .................................................................................................................................. 8 Acronyms ................................................................................................................................................ 12 1. Introduction and Objectives............................................................................................................ 14 2. Conceptual Framework and Rapid Assessment Approach ............................................................. 15 Conceptual Framework ...................................................................................................................... 15 Rapid Social Registry Assessment: Approach and Prioritization of Topics ......................................... 17 3. The Origins, Evolution, and Expansion in Coverage of the UBR ..................................................... 18 Origins and Evolution of the UBR ....................................................................................................... 18 Expansion in Coverage of Registered Households in the UBR ........................................................... 21 4. Institutional Arrangements for the UBR ......................................................................................... 24 Central Roles for Management and Operations ................................................................................ 25 Local Roles for Implementation of Registration Processes ................................................................ 28 Institutional Relationship between the UBR and User Programs ...................................................... 31 Financing and Coordination with Development Partners .................................................................. 34 5. UBR Implementation Processes ...................................................................................................... 35 Implementation Processes for Phase 1 (“As-Is” Guidelines) .............................................................. 35 Sensitization, Training, and Outreach Processes .......................................................................................... 38 Intake and Registration (Data Collection) ..................................................................................................... 39 Assessment of Needs and Conditions to Determine Potential Eligibility for Social Programs ..................... 41 Strengthening Implementation Processes Going Forward ................................................................ 42 Implications of Shift to Full Registration on Implementation Processes ........................................... 44 6. The Quality of UBR Data.................................................................................................................. 46 Data Quality Checks ............................................................................................................................ 46 Oversight, Monitoring and Reporting in the UBR .............................................................................. 52 Systems Integrity Checks .................................................................................................................... 54 7. Information Technology Aspects of the UBR .................................................................................. 54 Systems Architecture and Management ............................................................................................ 55 ICT Infrastructure ................................................................................................................................ 57 Data Security and Privacy ................................................................................................................... 57 The UBR’s Evolution within an Integrated Information Management Framework ........................... 58 8. User Programs and the UBR’s Potential as a Powerful Tool of Social Policy .................................. 60 3 Multi-Program Use of Social Registries .............................................................................................. 60 Key Ingredients for Multi-Program Use of Social Registries............................................................... 62 Updating and Using the UBR for Shock-Responsive Safety Nets ....................................................... 64 9. Strategic and Operational Communications for the UBR ............................................................... 65 Strengths and Challenges of the UBR’s Communications .................................................................. 65 Toward a Communication Strategy for the UBR ................................................................................ 69 Branding the UBR and Managing Expectations .................................................................................. 71 10. Projected Cost of implementing UBR (Forward-Looking for Phase 2)............................................ 73 11. Conclusions and Recommendations ............................................................................................... 76 Institutionalization of the UBR ........................................................................................................... 77 Implementation Processes and Oversight, Monitoring and Reporting (OMR) .................................. 78 Information Systems ........................................................................................................................... 78 Strategic and Operational Communications ...................................................................................... 79 References .............................................................................................................................................. 83 Appendix 1: Social Registries’ Conceptual Framework and Assessment Tool ....................................... 88 4 Acknowledgements This paper was prepared by a World Bank team led by Colin Andrews, Senior Social Protection Specialist, under the technical guidance of Kathy Lindert, Global Lead for Social Protection and Jobs Delivery Systems, with contributions from Chipo Msowoya (Social Protection Specialist), Boban Varghese Paul (Young Professional), Anita Mittal (Consultant) and Elijah Chirwa (Consultant). Overall guidance was provided by Dena Ringold (Practice Manager), Robert Chase (Practice Manager) and Greg Toulmin (Country Manager). The paper benefitted immensely from extensive technical meetings, interactions and consultations, with different stakeholders in Malawi, including principally management and staff of the Department of Economic Planning and Development in the Ministry of Finance, Economic Planning and Development, especially the Social Protection and Poverty Reduction Division; the Local Development Fund Technical Support Team; the Ministry of Gender, Children, Disability and Social Welfare; and other Government ministries, departments and agencies. We also thank the district teams and the communities in Dowa district for hosting a field visit during their community meeting for the Social Cash Transfer Program. This paper was prepared in close partnership with Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH (GIZ) in Malawi. The team would also like to thank representatives of the European Union (EU); Germany’s GIZ and Kreditanstalt für Wiederaufbau (KfW); the World Food Programme (WFP); the Embassy of Ireland; the United Kingdom’s Department for International Development (DFID); the Food and Agriculture Organization (FAO) and the United Nations Children’s Fund (UNICEF) in Malawi who took out time to discuss their thoughts on the UBR. Valuable peer review feedback was provided by Ralf Radermacher (GIZ), Benedicte Leroy De La Briere (Lead Economist, Social Protection & Jobs) and Lucia Solbes (Social Protection Specialist). The team is especially grateful for ongoing technical inputs and discussions from Ivan Drabek (Senior Social Protection Specialist), Emanuela Galasso (Senior Economist), and Niall Kelleher (Consultant). Elizabeth Schwinn provided careful editorial assistance. The publication of this report would not have been possible without the generous support from the Rapid Social Response Trust Fund Program, which is supported by Australia, Norway, the Russian Federation, Sweden, and the United Kingdom. The findings and interpretations expressed here are those of the authors and do not necessarily reflect the views of the World Bank Group, its Executive Directors or the countries they represent. 5 Glossary of Technical Terms Used in this Paper Assessment of Needs Systematic processes for determining the needs and conditions of registered individuals or and Conditions households for the purposes of (a) determining potential eligibility for specific programs and/or (b) informing the determination of benefits and services that may be rendered by the programs. The primary inputs include: (a) complete information on the applicant(s), verified & validated (from the intake and registration phase); as well as (b) direct qualitative assessment by the social worker / employment officer. The primary output from this phase is the classification or profiling of the applicant(s) / registrant(s) according to various assessment tools (such as socio-economic welfare measures, risk profiles, labor profiles, and so forth). Beneficiary Monitoring Beneficiary monitoring is a phase along the Delivery Chain that involves actions and information flows related to the operational monitoring and management of the program. Depending on the type of program, beneficiary monitoring may include: basic monitoring, oversight of errors-fraud-corruption, grievance redress, monitoring of conditionalities, monitoring of accompanying measures, and so forth. Beneficiary Registries Beneficiary Registries track information on beneficiaries of specific programs based on program enrollment decisions. They underpin the information system for beneficiary management and benefits administration. Beneficiaries contain information on beneficiaries (not all registered households). Census-Sweep Mass registration of households into the social registry. With the census sweep approach, all Registration Method or most households in specific areas (or the entire country) are registered en masse. In contrast with the on-demand approach (see below), with the census-sweep approach, enumerator teams go to the communities and conduct intake and registration using door-to-door methods. Citizen Interface The access point to service delivery for individuals, families, and households (potential or current beneficiaries,) as well as to queries, grievances, and user feedback. Structurally, the citizen interface or “front-office” can include ‘infomediaries’ such as facilitators, social workers, community agents and mobile teams, as well as digital service windows, citizen service centers or one-stop-shops at local offices and so forth. For labor and social insurance, this interface would also cover interactions with employers (e.g., for contributions, records- keeping job matching services, labor inspectorate, etc.). [Note that in our definition, the term “citizen” does not refer to nationality, but rather to people in a given country—individuals, families, households]. Decisions on benefits Setting benefit levels (for cash or in-kind benefits) and/or defining the service package (for and service packages services) that will be provided to eligible beneficiaries of social program(s) and establishing and notifying beneficiaries of such decisions (and any associated conditions on their participation). Determining Eligibility Process for comparing the profile of an individual, family, or household that emerges from the assessment of needs and conditions with the eligibility criteria for a specific program. Enrollment Decisions Decisions taken by social program administrators to admit individuals, families, and/or households into that specific program. Those decisions may be made on the basis of the assessment of needs and conditions, eligibility criteria, as well as other program-specific factors (such as fiscal space). Intake Intake refers to the process of initiating contact, receiving the person/family, explaining the purpose, interviewing, and gathering information. Intake and registration usually happen simultaneously. Integrated Information Integrates all of an organization's systems and processes in to one complete framework, Management enabling an organization to work as a single unit with unified objective. It links information Framework across different services/systems and integrates information across agencies for a given user. 6 Notification involves informing applicants of their enrollment decisions (in, waitlisted, out) and onboarding involves finalizing the enrollment process for those who have been selected Notification and (conducting orientation, collecting additional information, providing option to opt-out, etc.). Onboarding On-Demand On-demand methods allow anyone to register their information in the social registry at any Registration Method time. With on-demand methods, households go to local offices (or apply online) to register their information. Most demand-driven social registries also use active outreach methods to ensure that marginalized groups are informed and reached. Outreach Outreach typically involves interactions to inform people about social programs, build awareness, and encourage the intended population to apply, provide their information, engage, and eventually participate in the program if deemed eligible and enrolled. Outreach can also involve two-way communication to better inform program design by gathering inputs, views, feedback from citizens and other stakeholders (see also communications). “Active Outreach” is often used to proactively reach vulnerable groups that may otherwise be uninformed about social programs or their rights. Payments Transactions for the delivery of cash or near-cash benefits from social programs (whether they are contributory or non-contributory benefits). [Note that social registries have no direct relation to payments as the households in social registries are not beneficiaries of social programs. Payments administration builds on the information base from beneficiary registries.] Registration Registration is the process of recording and verifying the information collected from the intake process. It can also involve pulling additional information from other administrative systems. Intake and registration usually happen simultaneously. Registration quotas Explicit caps (limits) on the number of households that can be registered in a specific district. Registration targets Planned number of households that would be registered in a specific district, but without operating as a fixed or rigid quota (cap or limit). Social Registry Information systems that support the processes of outreach, intake and registration, and assessment of needs and conditions to determine potential eligibility for social programs. They contain and maintain information on all registered households regardless of whether they eventually benefit from a social program. As such, we do not refer to households in social registries as “beneficiaries” but as “registered households.” Source: Leite et. al. (2018) and World Bank Global Solutions Group for SPJ Delivery Systems 7 Executive Summary This paper reports on a rapid assessment of Malawi’s integrated social registry known as the Unified Beneficiary Registry (UBR). The paper assesses the Phase 1 rollout of the UBR in 10 districts supported by World Bank financing: Lilongwe, Ntchisi, Kasungu, Rumphi, Chiradzulu, Nkhota-Kota, Blantyre, Karonga, Ntcheu, and Dowa. The UBR is a ground-breaking initiative in Malawi, allowing households to register and be considered for potential inclusion in social programs based on a transparent assessment of those households’ needs and conditions. To date the UBR has registered and collected data for over 800,000 households (or 4 million people) in the 13 districts where it has been rolled out. The timing of the assessment was ripe given the upcoming expansion of the UBR and a planned shift in district-level registration targets (from 50 to 100 percent of households). The objectives of the rapid assessment are to: (a) review Malawi’s UBR experience to date; (b) identify strengths and areas for improvement; (c) provide short-term recommendations to support the upcoming expansion, including implementation adaptations that would be needed to accommodate the revised registration targets; and (d) support the strengthening of the UBR over the long term. Topics covered include institutional arrangements, implementation processes, information systems, data quality, user programs, and communications. While the primary audience for this paper includes the program’s core stakeholders in Malawi, the report is also of potential interest to other countries interested in developing social registries or carrying out rapid social registry assessments. Social registries can serve as powerful social policy tools. They provide a gateway for households to register and be considered for potential inclusion in social programs based on an assessment of their needs and conditions. That assessment usually considers measures of socio-economic status, categorical factors or a combination of both, which are common criteria used by programs to prioritize eligibility for benefits and services. Many countries use social registries as integrated platforms to support registration and determination of potential eligibility for multiple programs, such as cash transfers, public works, subsidized health insurance, social energy tariffs, education and training vouchers, fee waivers for child care, emergency assistance, financial inclusion services, pro bono legal services, and so forth. This can have three advantages: reducing burdens on households who don’t have to apply for numerous benefits and services separately, reducing administrative costs and boosting efficiency for user programs, and improving the coordination of social policy and programming. In Malawi, the need for an integrated social registry was conceived as a response to growing concerns about the fragmentation of the social protection system. The Malawi UBR was created to serve as a consolidated source of harmonized information on the socio-economic status of households to determine their potential eligibility for social programs. In doing so, it serves as a social registry by supporting the processes of registering households and determining their potential eligibility for multiple user programs in a coordinated way. A key strength is the way the Government has taken the lead in designing, managing, and implementing the UBR with strong ownership across the core agencies involved. Centrally, the UBR is owned and coordinated by the Ministry of Finance, Economic Planning and Development (MoFEPD) and managed by the UBR Taskforce, which is composed of representatives of key central stakeholders including the Local Development Fund and the Ministry of Gender, Children, Disability, and Social 8 Welfare, among others. Implementation is carried out using district- and community-level structures according to a detailed Implementation Guideline and training materials. The process for designing and implementing the UBR has been highly consultative, and the reliance on existing government structures—particularly for local implementation—is a major strength. Nonetheless, a priority going forward would be to institutionalize the central-level unit, since the UBR Taskforce was originally created as a temporary arrangement. The UBR was designed to serve two flagship social programs, with a view to serving additional programs over time. The initial programs are the Social Cash Transfer Programme (SCTP) and the Public Works Programme. The UBR is also considered a potential source of information on the socio-economic status of households for use by other programmes, such as the Farm Input Subsidy Programme, the Village Savings and Loan Programme, microfinance, nutrition, scholarships, and humanitarian assistance. The use of the UBR by multiple social programs hinges on a set of harmonized eligibility concepts and a common intake questionnaire (called the “Harmonized Targeting Tool,” or HTT). That questionnaire was developed on the basis of extensive consultations with various stakeholders. Another ingredient for the success of the UBR is the use of information technology, which has been effectively leveraged by the UBR to register households, manage their data, and exchange relevant information with user programs. A key strength of the UBR information system is that it was largely developed and maintained by in-house resources with open-source software. The registration of household data is carried out using mobile tablets and Open Data Toolkit software. The UBR is a web- based information system with a back-end database to maintain household information. The UBR uses cloud infrastructure, which brings lower cost, scalability, and security. Data exchange with user programs is facilitated through web services and an API on the basis of data sharing protocols (which have been finalized but need to be put into practice). Implementation has been rapid, particularly in 2017. The UBR is being rolled out across the country in phases, using the “census sweep” door-to-door registration approach. Registration has already been carried out in 13 districts, and the UBR now covers over 800,000 households (or over 4 million people). Continued expansion is planned for 15 additional districts (14 in 2018 and 19, and 1 in 2020; updating of 2 existing districts is planned for 2020), with registration projected to reach a total of about 2.8 million households (14 million people), or 70 percent of the national population, by 2020. The UBR has performed well on measures of data quality. This assessment reviewed three aspects of data quality. First, the UBR’s data quality was assessed in terms of its completeness, internal consistency, and external consistency (UBR’s data was cross-checked with the nationally representative household survey, the Integrated Household Survey (IHS). Second, the assessment reviewed structures and processes for oversight, monitoring and reporting. It identified some areas for improvement, particularly in the standardizing of checklists for oversight and monitoring, as well as in reporting formats and analytics. Third, the assessment reviewed systems for integrity checks, and found that data validation and verification processes are incorporated into all stages of the UBR, although there are some areas for improvement (e.g., in automation and coding). As part of the rollout, the UBR is planning to shift its registration targets from 50 percent of households in each district to 100 percent. Both targets involve trade-offs. The initial target was based on Malawi’s poverty rate. The advantages of setting lower registration targets include: (a) managing 9 expectations and limiting waiting lists for user programs, since the budgets of most social programs are limited; and (b) limiting the administrative costs to implement the UBR. There are disadvantages to using targets, however, particularly when they are implemented as registration quotas (caps), including: (a) perceived lack of fairness and transparency in who is included or excluded from interviews and registration (a potentially divisive issue in communities); (b) the potential to replicate existing local inequalities, whereby certain disadvantaged segments are excluded because they are less informed or less connected); and (c) the likely exclusion of some poor households from registration, which effectively bars them from being considered for social programs. The planned shift in targeting would allow all households to register their information, thus avoiding the exclusionary disadvantages associated with a 50 percent target. Another advantage is that the larger registry would allow the UBR to serve a wider range of individuals, connecting even those who are not eligible for social assistance to programs that may have more extensive potential coverage, such as humanitarian assistance, and child-based programs such as nutrition and other early childhood interventions. The disadvantages of full registration include: (a) the risks of raising expectations for a large segment of the population, the majority of whom would not receive any form of social benefits given limited program budgets; and (b) higher implementation costs in terms of time, staffing, material inputs, and financing (though economies of scale of interviewing all households may reduce the per-household interview cost). The planned increase in registration has ramifications for implementation processes. One implication is the role of communities and “Community-Based Targeting” (CBT). CBT enters twice into UBR processes and complements the Proxy Means Test (PMT) scores that are estimated on the basis of information collected in the UBR. Communities first determine the poorest 50 percent of households to be registered, and then validate the ranking of households based on their PMT scores. Both CBT and the PMT add to the credibility of the UBR in assessing potential eligibility of households for social programs. However, with the shift to 100 percent registration targets, the focus of the first community meeting would shift to ensuring the inclusion of all households rather than prioritizing and selecting the poorest 50 percent. The goals of the second community meeting may also need to be adapted, because discussing the rankings of all households may prove too cumbersome. Alternatively, the community could validate the PMT scores of some subset of households or of groups of households (for example grouped by quintile), flagging households that they think are misclassified. Either way, the implementation processes and training manuals would need to be updated to reflect these changes. In sum, the UBR has many strong fundamentals. Some of these include government ownership of UBR (both the process and the IT systems), the use of existing decentralized institutional structures, good relationships with various stakeholders across ministries and donor partners, functional implementation processes, effective information systems, and most importantly, robust data quality and rapidly expanding coverage. These fundamentals need to be maintained and strengthened as the UBR considers future phases of expansion. Nonetheless, the UBR still faces areas for improvement to gear up for the next phases of rollout to new districts. An immediate priority would be to enhance and institutionalize the central agency, since the UBR Taskforce was set up as a temporary structure. A related challenge is the need to clarify the central team’s roles in coordinating and monitoring implementation, as well as the urgent need to invest in end-to-end implementation planning for continued rollout. Implementation processes would also need to be adapted to the new 100 percent registration target. Other priority actions to strengthen implementation would include adjusting the staffing model for enumerators and spot-checkers (for 10 random re-reviews); standardizing oversight, management and reporting checklists and guidelines, and formalizing steps for handling grievances and appeals. From an information systems perspective, there is a need to automate some processes, staff up for some key roles and responsibilities, enhance information and communications technology infrastructure to ensure that disaster recovery and business continuity platforms are available and can guarantee a robust and resilient UBR information technology environment, and ensure real-time data replication from the cloud to a local backup facility. Data-sharing protocols also need to be enforced, and communications with user programs could be facilitated with a dedicated focal point, service desk, or hotline. Other long-term actions are also needed. These include establishing a formal legal standing and framework for the UBR, as well as a coordinated and sustainable financing strategy. Standardizing reporting and analytics would support a more robust flow of information for strategy, policy, and quality operations. Embedding the UBR into a broader management information framework for social protection with a common data exchange platform would facilitate links to user programs, Malawi’s National Registration and Identification System, geo-referenced information systems, and so forth. The Government should also assess the frequency and scope of updates to take into account seasonality and changes in living conditions over time and to meet the ability of user programs to respond to these changing needs. An information strategy could prioritize shock-prone areas for more frequent updating, which would facilitate use of the UBR by humanitarian response programs. Finally, a communication strategy, with a possible rebranding of the UBR’s name, logo and tagline, is needed to address communications challenges that could hamper the UBR’s effectiveness. Among the challenges are a lack of awareness of the UBR among communities and other stakeholders, confusion regarding the role of the UBR vis-à-vis the user programs—particularly the SCTP—and misperceptions about the UBR’s data quality, which detract from the UBR’s credibility. The key objectives of strategic (external) and operational (internal) communication would include: (a) improving awareness and understanding of what the UBR is and does (objectives, purpose, functions); (b) improving understanding of how the UBR works (processes, functioning); (c) reducing confusion about the relative role of the UBR vis-à-vis the user programs, particularly the SCTP; (d) boosting credibility of the UBR with regards to information quality and validity as an “honest broker” for information on the demographic and socio-economic status of households; (e) ensuring two-way flows of information between the UBR and various stakeholders for improved transparency and responsiveness to queries; (f) boosting support for the UBR for financing, use, and policy coordination; and (g) standardizing messaging around the UBR for credibility, transparency and consistency. Furthermore, the name “UBR” is a misnomer, since it is a social registry (not a beneficiary registry) and since households included in the UBR are not all beneficiaries. Given the lack of understanding of the UBR—and the liability of the term “beneficiary” in its name, it would be worth considering a rebranding exercise to develop and test alternatives for the name, logo, and tagline, to see which would be better understood by various stakeholders. 11 Acronyms AEC Area Executive Committee API Application Program Interface COMSIP Community Savings and Investment Programme CONGOMA Council for Non-Governmental Organizations in Malawi CSO Civil Society Organization CSSC Community Social Support Committee DoDMA Department of Disaster Management Affairs DRM Disaster Risk Management DSSC District Social Support Committee DTT District Training Team E2E-IP End-to-end Implementation Plans (E2E-IP) EPD Department of Economic Planning and Development EU European Union FAO Food and Agriculture Organization FISP Farm Input Subsidy Programme GIS Geographic Information System GIZ Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH GoM Government of Malawi HTT Harmonized Targeting Tool for PWP and SCTP IA Irish Aid IHS Integrated Household Survey ICT or IT Information and Communications Technology KfW Kreditanstalt für Wiederaufbau LDF Local Development Fund Mechanism LDF-TST Local Development Fund–Technical Support Team MF Microfinance programme MFI Micro-Finance Institutions 12 MoFEPD Ministry of Finance, Economic Planning and Development MoGCDSW Ministry of Gender, Children, Disability and Social Welfare MGDS Malawi Growth and Development Strategy MIS Management Information System MNSSP Malawi National Social Support Programme MVAC Malawi Vulnerability Assessment Committee NGO Non-Governmental Organization NSO National Statistics Office NRIS /National ID National Registration and Identification System NTT National Training Team ODK Open Data Kit OPC Office of the President and Cabinet PMT Proxy Means Targeting PR&SP Poverty Reduction and Social Protection Department PWP Public Works Programme SCTP Social Cash Transfer Programme SDG Sustainable Development Goals SMP School Meal Programme VSL Village Savings and Loans Programme UNICEF United Nations Children’s Fund UNDP United Nations Development Program 13 1. Introduction and Objectives Social registries provide a gateway for households to register and be considered for potential inclusion in one or more social programs based on an assessment of their needs and conditions. That assessment usually considers measures of socio-economic status, categorical factors or a combination of both, which are often criteria used by programs in prioritizing eligibility for benefits and services. Many countries use social registries as integrated platforms to support registration and determination of potential eligibility for multiple programs. Such integrated platforms can reduce burdens on citizens (who don’t have to apply for numerous benefits and services separately ), reduce programs’ administrative costs and boost their efficiency, and improve coordination of social policy and programming. Indeed, integrated social registries can serve as powerful platforms that extend well beyond social assistance. Many countries use integrated social registries to support a range of other interventions, some targeted and some universal in nature. Some examples include subsidized health insurance, social energy tariffs, education and training vouchers, fee waivers for child care, financial inclusion services, and pro bono legal services. Operationally, social registries are information systems, and their basic architecture includes data intake and exchange, software applications to support both front-office and back-office functions, database management and interoperability (in some cases), and information and communications technology (ICT) infrastructure. Malawi’s Unified Beneficiary Registry (UBR)1 provides a consolidated source of information on the socio-economic status of households to determine their potential eligibility for social programs. In doing so, it serves as a social registry by supporting the processes of registering households and determining their potential eligibility for user programs in a coordinated way. It was designed to support these processes for the Social Cash Transfer Programme (SCTP) and the Public Works Programme (PWP), but it is also a potential source of information on the socio-economic status of households for use for other programs, such as the Farm Input Subsidy Programme (FISP), humanitarian and shock-response interventions, the Village Savings and Loan Programme (VSL), microfinance, nutrition programs, bursary scholarships, and so forth. Currently, safety net programs in Malawi are estimated to cover 25 percent of the population, compared to the average of 10 percent covered by safety nets across the Africa region. Trends in chronic and transitory poverty highlight the relevance for a UBR. Over time, Malawi’s poverty rate has remained persistently high, declining only marginally from 52.4 percent in 2004 to 50.7 percent in 2010. Transitory shocks have the potential to exacerbate rural poverty, pushing an additional two out of every five households below the poverty line (Dang and Dabalen, 2017). Demographic trends further show the potential demand of delivery systems such as the UBR in the future. Malawi’s population is expected to double in approximately two decades, increasing from 17.2 million in 2015 to 34.4 million in 2038 (UNDSA, 2015). At present, 56.2 percent of the population is under 19 years of 1 The literature (see Leite et.al. (2017) and Barca (2017)) distinguishes between “social registries,” which support the proces ses of intake, registration, and determination of potential eligibility and maintain data on all registered households (regardless of whether they are eventually deemed eligible for enrollment in social programs) and “beneficiary registries,” which support beneficiary and ben efits administration and include information only on beneficiary households who are enrolled in specific programs. As such, a more apt name for Malawi’s UBR would have been the Unified Social Registry (USR) rather than Unified Beneficiary Registry since not all h ouseholds included in the UBR are beneficiaries and the primary functions of the UBR are to register and consolidate information on hou seholds’ socio-economic status to determine potential eligibility for social programs. See Glossary of Key Terms and Appendix 1 for terminology and conceptual framework for social registries. 14 age, with the working age population constituting a mere 38.8 percent of the total (World Bank, 2016). The overall policy, design and management oversight of UBR falls under the Ministry of Finance, Economic Planning and Development (MoFEPD), which is also responsible for coordinating social protection in Malawi. The Directorate for Poverty Reduction and Social Protection in MoFEPD oversees UBR’s coordination and management, while its technical aspects are the responsibility of the UBR Taskforce, composed of representatives of MoFEPD, the Local Development Fund (LDF, coordinators of the Malawi Social Action Fund (MASAF) IV/PWP, supported by the World Bank), the Ministry of Gender, Children, Disability and Social Welfare (coordinators of SCTP), and Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH (GIZ, a key technical supporter of UBR), among others. Implementation is carried out using existing District and community level structures, as discussed in more detail below. The UBR is being rolled out across the country in phases. To date, half of Malawi’s households have been registered in 12 of 28 rural areas, as well as all households in a 13 th district (see Table 1 below). Further expansion is expected in 14 additional rural districts from 2018 through 2019, and the government and donor partners for those districts (Germany and the European Union) are planning to register all households in those additional districts. Subsequently, additional districts will be expanded or updated with the World Bank’s support in 2020 (one additional district and updates in two districts; see Table 1 for details). Given the upcoming expansion of the UBR, the doubling of the registration target, and potentially the eventual need to update the data for existing districts, the World Bank was asked to conduct a rapid assessment of the UBR, with the following objectives: 1. Reviewing users’ experiences with the UBR to date; 2. Identifying strengths and areas for improvement; 3. Informing the expansion of the UBR to new districts and its eventual updating and expansion within existing districts—and considering adjustments that would be needed to support the shift to 100 percent registration; and 4. Supporting the long-term evolution and strengthening of the UBR. While the primary audience for this paper includes the UBR Taskforce, user programs, and other stakeholders in Malawi, the paper may also interest an international audience as an example of a rapid social registry assessment.2 As such, it could serve as an example for other countries interested in developing social registries or carrying out social registry assessments and as an input to the ongoing collaboration by an international working group on an ISPA Tool for Social Information Systems.3 2. Conceptual Framework and Rapid Assessment Approach Conceptual Framework The conceptual framework for the assessment is adapted from a recent paper on social registries by 2See Leite et. al. (2017). 3ISPA is the Inter-Agency Social Protection Assessments initiative. The Working Group for the ISPA Social Information Systems Tool is coordinated and led by the World Bank, with participation of DFAT, DFID, GIZ, UNICEF, WFP, EC, FAO, OECD, Finland, and the EC. 15 the World Bank team (Leite et. al. 2017). It takes a functional approach, anchoring key concepts in their functions along the delivery chain (Figure 1). Such a delivery chain is common to most benefits and services aimed at providing support to individuals, families, and/or households. Indeed, most social programs require some form of outreach, intake and registration, assessment of needs and conditions to determine potential eligibility. Once those steps have been accomplished, the programs determine whom is eligible (by comparing the profile of needs and conditions to their program-specific eligibility criteria, take enrollment decisions, and set their benefits levels (or service package). The program then notifies and onboards beneficiaries. The program then goes through a recurring implementation cycle, which involves the processes of transactions for payments (or service provision) and beneficiary monitoring —that is, updating information, monitoring of program conditionalities (if relevant), linkages and referrals (if relevant), grievance redress and appeals, and so forth. Within that framework, social registries are the information systems that support outreach, intake and registration, as well as the determination of assessment of needs and conditions (the blue sections of Figure 1). Registries include information on all registered households (regardless of whether they are later enrolled as beneficiaries of social programs). In contrast, beneficiary registries include information only on those households that are enrolled in specific programs. Their function is to support beneficiary management and benefits administration (the red sections of Figure 1). Given those definitions, the name of Malawi’s social registry— “Unified Beneficiary Registry”—is technically a misnomer, since: (a) In terms of objectives, the UBR’s primary functions are to capture, store, access, retrieve, and share data on households’ needs and conditions (socio-economic profiles, consistent with the blue segments of Figure 1). These functions can inform the determination of eligibility for specific programs, but the programs themselves have the mandate for taking eligibility and enrollment decisions (red sections of Figure 1) and then managing program operations such as payments and monitoring (purple segments of Figure 1). (b) In terms of the population covered, the UBR collects and maintains information on all registered households—not just those who are selected as beneficiaries of specific programs. A more apt name for Malawi’s UBR might have been Unified Social Registry (USR) rather than Unified Beneficiary Registry. Households whose data are included in the UBR should not be referred to as UBR beneficiaries, since the UBR does not grant benefits and not all households will end up as beneficiaries of social programs. This paper sticks with the acronym “Malawi’s UBR” but recognizes and assesses the system as a social registry, given its functions and the population it covers. 16 Overview & Approach to the UBR Assessment Conceptual Framework Figure 1—Information Systems Support Implementation Phases and Processes along the Delivery Chain for Social Programs Provide Assess Enroll Benefits / Potential Demand Services 7 1 2 3 4 5 6 8 Monitor “Intended” Potentially Applicants Eligible Beneficiaries Population Applicants 4 Source: Leite et. al., (2017), with further updates by the World Bank’s Global Solutions Group for SPJ Delivery Systems Rapid Social Registry Assessment: Approach and Prioritization of Topics The conceptual framework and assessment tool from the recent paper on social registries by the World Bank team (Leite et al. 2017) served as the grounds for the initial terms of reference (TORs) and approach for this review. Based on feedback from government and development partners, as well as extensive desk review of the literature on UBR, that framework was then adapted to contextualize these TORs and prioritize the core relevant topics for the rapid assessment of Malawi’s UBR. The resulting topics form the outline for this report: • Institutional Arrangements • Processes and Implementation • Data Quality Checks • Information Systems Aspects • Monitoring and Evaluation of the UBR • User Programs and the Potential of the UBR as a Powerful Social Policy Tool • Strategic and Operational Communications • Projected Cost of implementing the UBR The rapid social registry assessment builds on an extensive review of documents related to the UBR (see bibliography); interviews with numerous stakeholders during a technical mission in March 2018 (including the UBR Taskforce, associated ministries and agencies, district and community officials, user programs, and development partners); a field visit to Dowa to observe the community meeting for the selection and enrollment of beneficiaries in the Social Cash Transfer Program (SCTP) following recent data collection and community meetings for the UBR; and some data checks and analysis. The technical 17 mission is also part of the World Bank’s ongoing dialogue and engagement on this topic, including numerous prior missions. The review was conducted by assessing the UBR in its current “as-is” situation. It also considers options and recommendations for the UBR going forward, taking into account lessons learned and planned changes (including the shift to registration of all households), prioritizing between measures that would be needed in the near term versus those for the long term. The assessment is “rapid” in that it was timed to be executed in a relatively short time, building on the lessons learned from the recent phases of implementation to inform implementation in the subsequent phases. This rapid assessment is focused on identifying strengths, challenges, and opportunities for improvement to provide recommendations for the short- and longer-term. It is not a comprehensive evaluation and does not evaluate the overall design, costs, or final outcomes of the UBR. Rather, it is focused on practical aspects of the UBR’s implementation and its associated institutional structures, processes, systems, data quality and uses. 3. The Origins, Evolution, and Expansion in Coverage of the UBR Origins and Evolution of the UBR The unified social registry was conceived as a response to growing concerns about fragmentation of Malawi’s social protection programs. Under the first Malawi National Social Support Programme (MNSSP, 2012–2016), several programs were prioritized as the main social support programs: the SCTP, the Public Works Program (PWP), Village Savings and Loans programme, the Microfinance programme, the School Meal Programme, and the Farm Input Subsidy Programme. An assessment4 carried out under MNSSP specifically pointed to a lack of a harmonized social registry to facilitate coordinated registration and the determination of eligibility for various social protection interventions. In response, the Malawi Government embarked on an effort to strengthen and harmonize social protection delivery systems in the country. Between 2013 and 2014, under the MNSSP, the Directorate of Poverty Reduction and Social Protection in the Ministry of Finance, Economic Planning, and Development (MoFEPD) consulted with stakeholders on the need for a social registry that would facilitate and harmonize processes for registration and determination of eligibility (“beneficiary targeting”) for various social support programs. Because the SCTP and PWP were the flagship social protection programs, the registration and eligibility processes for these two programs were harmonized first. The social registry was first developed as a pilot “Harmonized Targeting Tool” (HTT) as a precursor to the UBR. HTT—and subsequently the UBR—was developed, managed, and implemented by a locally mobilized intragovernmental team with donor and consultant support. From 2014 to 2015, the Government mobilized technical personnel from various ministries, departments and agencies (MoFEPD, the Local Development Fund, the National Library Service, the Community Savings and Investment Program and the Ministry of Gender) to form the current UBR Taskforce as the core team to drive these efforts. Between 2015 and 2016, the Taskforce, with support from one international 4 Malawi Government, Ministry of Finance, Economic Planning and Development (2016). 18 consultant and Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH as the technical assistance partner, rallied around the development of HTT as the precursor to the UBR. In 2015, a pilot of HTT was carried out in two districts to support further expansion of the SCTP. HTT was designed in line with the Government’s directive to collect data for 50 percent of households in the districts of Dedza and Nkhatabay (see Table 1). That target was based on Malawi’s average poverty rate. Building on the lessons of the HTT pilot in Dedza and Nkhatabay, the UBR Taskforce, with support from one international consultant who developed the software and other inputs for a subsequent pilot in Phalombe. With financing from the Food & Agriculture Organization (FAO), the registration target for Phalombe was set at 100%, percent although the actual share of households registered was 75 percent of the official National Statistics Office (NSO) estimate of total households in the district (see Table 1).5 Building on the lessons of the two pilot registries, the UBR was further developed. Those efforts were led by the UBR Taskforce, with support from an international consultant and other development partners. They included consultations and the preparation of a document on the requirements of user programs, as well as the development of a formal concept note on the UBR and the Implementation Guideline to guide the processes of collecting data (registration and determination of potential eligibility (see bibliography). In 2017, Phase 1 of the UBR was rolled out in 10 districts, with rollout for additional districts scheduled in 2018 and 2019. During Phase 1, 50 percent of households in 10 districts were registered, in line with the original Government directive and the national estimates of poverty. As discussed in Section 5, the selection of the households to be interviewed and registered was carried out during an initial community meeting. In 2018–19, Phases 2 and 3 of the UBR are expected to be rolled out to cover an additional 14 districts (Table 1), and the Government and donor partners for those districts plan to register 100 percent of households in those expansion districts to avoid the exclusionary disadvantages associated with registration quotas. Another important step in the evolution of the UBR include developing its capacity to transfer data to and from the main user programs (SCTP and PWP) and then expanding UBR to other user programs. Data transfer between the UBR and SCTP’s Management Information System (MIS) was piloted, certified, and achieved in March 2018 with the Application Program Interface (API). At the same time, PWP’s MIS was upgraded to make it more compatible with the UBR. Although some delays and glitches were experienced in the initial data transfer (such as the need to calibrate the mapping of geographic zones and clusters in the API between the SCTP and UBR classifications), such common “teething problems” occur with most new systems development and are being ironed out. The UBR Taskforce has hired five additional employees to further enhance human resource capacity. Most 5 Differences between registration targets and actual households registered could arise for various reasons. The first involves discrepancies between central and local data. The National Statistics Office (NSO) estimates of total households only go as far as the Traditional Authorities (T/As) level, which is one level higher than Group Village. As such, discrepancies can be found upon verification with actual village registers. The second, is that the situation on the ground can differ from central and local data. The location, formation, dissolution, and composition of households change over time. Some households may simply be missed, or others may decline to be registered. Finally, in the case of Phalombe, there are suggestions of some coordination and implementation challenges in the Phalombe pilot, although the World Bank team has not yet received the official report on that experience from FAO. 19 important, the SCTP is now using data from the UBR in the 10 expansion districts to enroll eligible households in the program so that they may soon receive cash assistance. Table 1—Phased Rollout of the UBR (and its pilot precursor HTT) Year of UBR Source of Name of District Registered Households (HHs) Roll Out Financing HTT Pilot (2 districts) HTT as Registration target = 50 percent Dedza, Nkhatabay precursor to World Bank 50.5% of total HHs were registered UBR, carried GIZ* (112,187 registered out of 222,121 total) out in 2015 100% UBR Registration Registration target = 100% FAO Pilot Early 2017 75% of total HHs were registered GIZ* Phalombe (72,641 registered out of 95,948 total) Phase 1 of UBR (10 districts) Lilongwe, Ntchisi, Kasungu, Registration target = 50% World Bank Rumphi, Chiradzulu, Nkhota- 2017 50.2% of total HHs were registered GIZ* Kota, Blantyre, Karonga, (652,867 registered out of 1,301,366 total) Ntcheu, Dowa Phase 2 of UBR (7 districts - 6 new, 1 updating) Registration target = 100% Germany (Through Likoma, Mangochi, Planned for Estimated 784,708 to be registered, which is KfW) (primary) Machinga, Mchinji, Salima, 2018 equal to the total # of HHs in those districts World Bank and Chitipa, with partial data (estimates assume full target met) collection in Phalombe (again) Phase 3 of UBR (7 districts) Registration target = 100% Mulanje, Mwanza, Neno, Estimated 808,964 to be registered, which is EU (Through KfW Planned for Chikwawa, Nsanje, Mzimba equal to the total # of HHs in those districts (primary)) 2019 and Zomba (estimates assume full target met) World Bank Phase 4 of UBR (3 districts - Registration target = 100% Planned for World Bank / 2 updating, 1 new) Dezda, Estimated 407,059 to be registered, which is End 2019- Government of Nkhatabay, Thyolo equal to the total # of HHs in those districts Early 2020 Malawi (estimates assume full target met) Estimated total of 2,839,619 registered / to be registered Total 28 Districts By 2020 All of the above Representing 70.3% of total households in Malawi Sources: UBR Taskforce, Ministry of Gender, Children, Disability and Social Welfare, National Statistics Office * GIZ has provided ongoing technical and capacity support to UBR development and roll out during the first phase. With future expansion plans, the UBR has become more visible, especially among local decision makers. Going forward, the vision for UBR’s future is to continue its expansion to serve other social protection programs and beyond. The UBR is increasingly well-positioned to support additional programs such as the Village Savings and Loans program, the Farm Input Subsidy Program, legal aid services for the poor, education bursaries, health services, and humanitarian assistance programs. Furthermore, the UBR’s evolution places it at the center of implementation of the second Malawi National Social Support Programme (MNSSP II, 2017–2022), given that program’s focus on the life cycle approach, emerging areas of shock responsiveness, resilient livelihoods, and linkages with other socio-economic programs. To achieve this goal, the UBR will need to build on existing foundations and strengthen areas addressed in this review: institutional structures, implementation processes, data quality, systems architecture, user program linkages and communications. 20 Expansion in Coverage of Registered Households in the UBR With the rollout of the UBR to additional districts, the social registry’s coverage of households has expanded rapidly. To date, the UBR has registered and collected data for over 800,000 households (over 4 million people) in the 13 districts where it has already been rolled out. This represents 21 percent of the total population in the country (household population projections for 2018), as depicted in Figure 2 below. Given registration targets, the projected share of total households to be registered would reach around 70.3 percent after the UBR is rolled out in all 28 districts (Figure 2). Figure 2—Actual and Planned Expansion of Registered Households with Phased Rollout of UBR Actual and Planned Expansion of Registered Households with Phased District Roll-Out of the UBR 3,000,000 80.0% 70.3% 70.0% 2,838,426 2,500,000 60.2% 60.0% 2,431,367 2,000,000 50.0% 40.2% 1,500,000 40.0% 1,622,403 30.0% 1,000,000 20.8% 837,695 20.0% 500,000 2.8% 10.0% 4.6% 112,187 184,828 112,187 72,641 652,867 784,708 808,964 407,059 - 0.0% 2015 HDCT Pilot (2 districts) 2017 Phalombe Pilot 2017 Phase 1 UBR (10 2018 Phase 2 UBR (7 2019 Phase 3 UBR (7 2020 Phase 4 UBR (3 districts) Districts) Districts) Districts) # Households Registered Cumulative Total # Cumulative # of Registered HHs as % of Total HHs in Country Source: Authors calculations using data from UBR Taskforce The planned increase in district-level registration targets from 50 to 100 percent is a significant step, which will require new design trade-offs. It also has important implications for implementation planning and execution, in terms of staffing, resources and distribution of field teams, as well as adaptations in key implementation processes (as discussed below). The initial 50 percent registration target was based on the poverty rate in Malawi. About half of the population was classified as poor using a needs basket in 2010 (around 57 percent in rural areas). The advantages of using that registration target included: (a) managing expectations and limiting waiting lists for user programs, since the budgets of most social programs are limited (for example with the SCTP, which currently provides cash transfers to 4.4 percent of the population);6 and (b) limiting administrative costs to implement the UBR. There are disadvantage to using such targets, however, particularly when they are implemented as registration quotas (caps), including: (a) There is a perceived lack of fairness and transparency in decisions as to who gets included or excluded from 6The SCTP currently pays cash transfers to 177,595 households, equivalent to 4.4% of the population. However, it hasn’t expand ed to cover all districts yet. The per-district targets are to cover the 10% poorest households in each district nationwide. 21 interviews and registration (which can divide communities); (b) Existing local inequalities may be replicated—certain disadvantaged segments may be excluded because they are less informed or less connected; and (c) Some poor households may not be able to register at all, which would bar them from being considered for social programs. The planned reforms would allow all households in each district to register their information, thus avoiding the exclusionary disadvantages associated with registration quotas. A larger registry would allow the UBR to serve a broader range of programs. Currently, the focus is on the SCTP, which targets the poorest 10 percent of households, that are also labor-constrained, and the PWP, which targets the next poorest 15 percent of households, whose members are not labor-constrained. Together, the two programs would be expected to provide benefits to the poorest 25 percent of households. However, other social protection programs may have broader target groups or potential coverage—and thus may need socio-economic information on a larger share of households. For example, a more universal social registry would facilitate shock-response efforts because all households in any given area would be visible to the system and thus could all potentially receive emergency assistance or disaster response aid. Another example would be the use of the social registry for gathering demographic and socio- economic information that could support child-based programs, such as nutrition and other early childhood interventions, which cover a broad swath of the population. The disadvantages of full registration include the risk of raising unfulfilled expectations for a large segment of the population, the majority of whom would not receive any social benefits; and higher implementation costs in terms of time, staffing, material inputs, and financing (though economies of scale related to interviewing all households may reduce the per-household interview cost). These current and projected registration rates for Malawi’s UBR are within the range of experience in other countries, particularly those that adopt census sweep approaches to registration (Figure 3). It is important to note that some of the social registries in Figure 3, like Malawi’s UBR, use census sweep methods with mobile interview and registration teams (such as Colombia, Djibouti, the Dominican Republic, Indonesia, Mali, Pakistan, the Philippines, Senegal, Sierra Leone, and Yemen). With the census sweep approach, all or most households in specific areas are registered en masse using door-to-door methods. Once these census-sweep social registries reach full scale, they tend to cover large shares of the population, as is seen for Pakistan (87 percent of households registered nationwide), the Philippines (75 percent), and Colombia (73 percent). Malawi’s projected 70.3 percent average for nationwide coverage of the UBR would be in a similar range to the social registries in those countries. The census sweep method can make sense when rolling out a social registry for the first time and when there are administrative and fiscal constraints (both for the social registry and for the user programs). The disadvantages of the census sweep approach include the infrequency of such mass registration waves, the associated static nature of the data, which can become out of date, the rising potential for exclusion of newly formed or changing households that cannot register or update their information in the interim years, and the high, unevenly distributed costs of infrequent mass registration episodes. In other countries, social registries primarily use on-demand methods of registration, which tend to be more frequently updated (such as those in Azerbaijan, Brazil, Chile, China, Georgia, Macedonia, Mauritius, Mexico, Montenegro, and Turkey). The on-demand methods allow anyone to register or update their information at any time. With on-demand methods, households go to local offices (or apply online) to register their information. Most demand-driven social registries also use active outreach methods to ensure that marginalized groups are informed and reached. Most nationwide on- 22 demand social registries cover a relatively small share of the population,7 such those in Turkey (50 percent), Mexico (47 percent), Brazil (40 percent), Montenegro (35 percent), and Mauritius (11 percent). The reason is that on-demand registries usually involve some degree of self-selection, whereby richer households tend not to bother taking the time to register because they are unlikely to qualify for the programs that use the social registry. Figure 3—Actual and Projected Coverage of Malawi’s UBR with International Comparisons Coverage of Social Registries: International Comparisons Registered households as % of total, circa 2015-18 Pakistan NSER 87% Dominican Rep SIUBEN 85% Philippines Listahanan 75% Chile RSH 75% Colombia SISBEN 73% Malawi UBR (projected after full roll-out by 2020) 70% Lesotho 52% Turkey ISAS 50% Mexico SIFODE 47% Brazil Cadastro Unico 40% Indonesia UDB 40% Georgia TSA Registry 36% Montenegro SWIS 34% Senegal RNU 30% Yemen SWF Registry 30% Gabon 27% Djibouti RSU 25% Macedonia CBMIS 25% Malawi UBR (current national avg % of total households) 21% China RPHR 15% Cabo Verde 14% Benin 12% Mauritius SRM 11% Azerbaijan VEMTAS 10% Tanzania 10% Mauritania 8% Sierra Leone SPRINT 5% Congo, Rep. 3% Comoros 3% Mali RSU 3% Cote D'Ivoire 3% Ghana 2% Chad 1% Burkina Faso 1% Nigeria 1% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Sources: Authors/UBR Taskforce for data on Malawi, Leite et. al. (2017) and Coudouel et. al. (2018) for international comparisons 7The exception is Chile’s RSH, which covers 75% of the population. Self-selection out of the registry is lower in Chile (and thus a higher share of households registered) because the RSH serves over 80 programs, some of which are near-universal in their coverage). 23 4. Institutional Arrangements for the UBR Institutional arrangements were assessed in four areas. The first is the role of management and operations. In other countries, these functions are typically centralized, but specific institutional arrangements vary between countries (Box 1). The second is the implementation of outreach, intake and registration processes. These processes are carried out locally to interface with citizens (households or individuals). But the specific delineation of institutional responsibilities varies significantly between countries depending on whether they are carried out on demand or conducted via census sweeps (Box 2). The third is the relationship of a social registry to user programs. In Malawi, a fourth area assessed was financing and coordination with development partners, since a significant share of financing for the UBR comes from donors. To contextualize this discussion, Figure 4 provides an overarching summary of the main roles and responsibilities for managing, operating and implementing the UBR. Figure 4—Main Roles and Responsibilities for Managing, Operating, and Implementing Malawi’s UBR Agency / Body Primary Roles Ministry of Finance, • Host Agency (official “owner” of the UBR) Central Management of Economic Planning and • Policy and Design Development (MoFEPD): • Management and Oversight Directorate for Poverty • Oversee Implementation UBR Reduction and Social Protection UBR Taskforce • Implement day-to-day technical operations • Manage UBR Information System National Training Team (NTT) • Supervise and guide field work for data collection processes oversight of Field (Group of technical officers from key • Central Support Geographic Pre-Mapping of Districts Work line ministries, departments, agencies for implementation of SS • Community sensitization, training of district teams, etc. programs) District Social Support • Overall management of data collection processes Committee (DSSC) • Quality checks for data collection District-Level Field Work (15 members including Head of Sectors + 2 members from CSOs) District Social Support • Coordinate implementation of data collection Secretariat (DSSS) • Consolidate UBR geographic pre-mapping data (Implementation unit; headed by • Approve data entered electronically using tablets District Social Welfare Officer plus various staff) • Collect completed data forms and ensure they are stamped by data entry team • Ensure proper and systematic filing system for data collection forms District Training Team (DTT) • Conduct training of local AEC teams (Team of officers from various • Supervise and support community sensitization meetings district ministries & departments) • Support selection of CSSC members 24 Area Executive Committee • Conduct household interviews for UBR, with support of CSSCs (AEC)8 • Ensure that all forms have been completed and stamped by GVHs (Main field implementation team; • Ensure that all completed forms are submitted to DTTs on daily basis consists of government & NGO extension workers at T/A level) • Mobilize community, support community sensitization meetings, orient CSSCs Local (village) Level Community Social Support • Mobilize communities Committee (CSSC) • Act as link between households and District (Team of community members • Validate pre-screened household listing chosen by the community to work on all SS programs) • Guide & support AECs during community meetings, household interviews • Ensure that all UBR household receipts are signed and stamped by GVHs Group Village Heads / village • Validate UBR geographic pre-mapping data heads (GVHs) • Call for community meetings • Sign UBR list of interviewed households to confirm that the interviewed households are members of their villages Sources: Compiled from Malawi’s UBR: Concept and Proposal (October 2015); Guidelines for UBR Process (November 2016); and discussions with UBR Taskforce and others during March 2018 mission. Central Roles for Management and Operations At the national level, the institutional arrangements for managing the UBR are complex, involving cooperation across numerous agencies (Figure 5). The overall policy, design and management oversight of UBR falls under the Ministry of Finance, Economic Planning and Development (MoFEPD), which also has the responsibility for coordinating social protection in Malawi. MoFEPD is thus the official “owner” of the UBR and its Directorate for Poverty Reduction and Social Protection oversees UBR coordination and management, which is carried out by the UBR Taskforce (see below). Institutional Arrangements: UBR: Institutional Arrangements National Level Figure 5—Organizational Chart for the UBR UBR : Institutional Arrangements at National Level - Strengthen coordination of National Social Support programmes Parliamentary committee on Policy issues to Cabinet Social and Community and Affairs Policy & Oversight Policy and resource mobilization Chair: Chief Secretary (OPC) National SS Secretariat: MoFEPD (EPD, as host Steering agency) Committee Technical Coordination Management and (implementation Guidance and Operations UBR policy recommendation EPD, LDF, National SS Chair: PS(EPD) Taskforce Technical Secretariat: EPD Ministry of Gender, GIZ Directors, DP Rep, CONGOMA, Committee Private Sector User Programs PWP Thematic SMP TWG Microfinance SCTP TWG VSL TWG Working Group TWG 9 Source: UBR Taskforce 8 AECs are mostly Government frontline staff also known as extension workers at the community /Traditional Authority level e.g. Community Development Assistants; Health Surveillance Assistants; Agriculture Extension Workers. Sometimes NGOs also dispatch their own frontline staff. 25 The central institutional arrangements for the UBR reflects other countries’ experience, although with some differences. As detailed in Box 1, central arrangements typically comprise four different models: (a) managed and operated by a Central Social Agency; (b) managed by a Central Social Agency and operated by an Operating Agency (c) managed and operational by a different type of Central Agency; or (d) managed and operated by a specific Program Agency. The institutional arrangements for the UBR are most similar to the second model, although in Malawi’s case the central agency is within the Ministry of Finance. Each program adopting the central institutional model applies a census sweep approach to registration. Box 1–Central Institutional Arrangements for Managing Social Registries: Examples The central institutional arrangements for managing and operating social registries vary between countries, though there are some common patterns based on a sample of 20 countries. • Managed and Operated by Central Social Agency. In many cases the central social agency—usually the ministry of social protection or social development —hosts, manages and operates the social registry. Examples include: Azerbaijan VEMTAS, Chile RSH, Djibouti RSU, Georgia TSA Registry, Macedonia CBMIS, Mauritius SRM, Mexico SIFODE, Philippines Listahanan, Senegal RNU, Sierra Leone SPRINT, Turkey ISAS, and Yemen SWF registry. • Managed by a Central Social Agency, Run by a Separate Operating Agency. In other countries, the central social agency hosts and manages (oversees) the social registr y as the “owner” of the system, but outsources the operations to a specific “operating agent.” In Brazil, for example, the Ministry of Social and Agrarian Development manages and owns the Cadastro Unico, with full data access, while the Caixa Economica Federal serves as the operating agent for software and systems management via a performance contract (the Caixa is also the payments agent). In Mali, the host managing agency of the Unified Social Register (RSU) is the Ministry of Solidarity and Humanitarian Action, while the operating agent is the Technical Unit of the RSU. The RSU is also guided by a Steering Committee (a political body) and a Technical Committee. In Montenegro, the host agency for the Social Welfare Information System is the Directorate for Information and Analysis within the Ministry of Labor and Social Welfare, while operations are contracted out to an IT company. • Managed and Operated by some Other Central Agency. In some countries, the social registries are managed by a central agency (not social). One example is the National Planning Department, which manages Colombia’s social registry. Similarly, in the Dominican Republic, the Social Cabinet within the Vice Presidency, manages the social registry. In Indonesia, the National Team for Accelerating Poverty Reduction “TNP2K” (under the Vice Presidency) has managed the UDB, but those arrangements were intended to be temporary, until the Ministry of Social Affairs builds capacity to undertake these functions as per its legal mandate. • Managed and Operated by a Specific Program. In Pakistan, the National Socio-Economic Registry is hosted, managed, and operated by the Benazir Income Support Program, even though it serves many social programs in different agencies. Source: Leite et. al. (2018) The UBR was developed under the authority of several high-level committees for policy and resource mobilization (Figure 4). At the highest level, the Parliamentary Committee on Social and Community Affairs provides broad policy guidance and oversight to the Cabinet. The National Social Support Steering Committee (under MNSSP) sets policy, provides oversight, and mobilizes resources for the UBR. It is chaired by the Office of the Presidency and Cabinet (OPC), with the secretariat in the Department of Economic Planning and Development (EPD), and membership from Permanent Secretaries of the Line Ministries, Civil Society Organizations, the Council for Non-Governmental Organizations in Malawi (CONGOMA), and the private sector. Under the guidance of the Steering Committee, the National Social Support Technical Committee provides more specific policy guidance, 26 recommendations, and implementation guidance. It is chaired by EPD, with the Secretariat also in EPD, with participation from Directors from various line ministries (including transport, gender, and agriculture), development partners, CONGOMA (a coalition of NGOs), and the private sector. The UBR Taskforce is the main national body for the UBR, comprised of representation of MoFEPD, LDF (coordinators of MASAF IV/PWP, supported by the World Bank), Ministry of Gender, Children, Disability and Social Welfare, (MoGCDSW, coordinators of SCTP), and GIZ (key technical supporter of UBR), among others. The representatives from these agencies fulfil specific roles within the UBR Taskforce, including two members in leadership roles (from EPD and LDF), two for quality assurance (MoGCDSW), and three for technical leadership (from GIZ, COMSIP, and the National Library Service). These representatives each maintain their functions in their respective home ministries, but most devote considerable time, talent, and expertise to the management and operations of the UBR. The UBR Taskforce was originally conceptualized as a technical working group, but has taken on the roles of hosting, managing, and operating the UBR in the absence of a more permanent structure. In the original Concept Paper for the UBR, the Taskforce was to provide guidance on the technical development of the UBR.9 However, in the absence of a permanent structure and until the recent recruitment of the UBR management unit, the Taskforce is managing and operating the UBR, with responsibility not only for its design and development, but also for hosting it, operating the information system, overseeing day-to-day operations, applying the harmonized targeting criteria to the data, and interfacing with social protection programs. The UBR Taskforce deserves huge credit for designing, developing, and managing the UBR so effectively in such a short period of time. There are many advantages to institutional arrangements in terms of government leadership, ownership, and coordination between key agencies. First, rather than outsource functions to external agencies or consultants, these arrangements are Government-led and Government-owned. Second, the model builds on extensive multi-sectoral coordination across central agencies, such as MoFEPD, MoGCDSW, and local development funds. This means that the ownership and knowledge of the UBR spans a broad spectrum of interested agencies, including those responsible for social protection programs. Third, the design and development of the software for the UBR information system was built in house, using open-source software components, rather than being contracted out to external firms with their own proprietary software.10 This has built on—and built up—internal capacity for designing, managing, and maintaining the system and allows for much greater flexibility as the system evolves over time. The temporary arrangements for UBR’s hosting and staffing risk positioning the registry as merely a “project” rather than a well-grounded anchor of Malawi’s social protection system. Social registries fulfill lasting functions that need ongoing and enduring institutional structures. They are not single- round surveys or static databases. Rather, they are permanent, “living” information systems that will continue to expand, evolve, and be regularly updated. 9 Malawi Government, Ministry of Finance, Economic Planning and Development (October 2015). The Malawi Unified Beneficiary Registry (UBR): Concept and Proposal. 10 The Taskforce was also supported by an international “support consultant,” who provided three months of development support a nd 7 months of maintenance support. 27 As a first step in institutionalizing the UBR, the Taskforce is recruiting and establishing a UBR Management Unit under Ministry of Finance, Economic Planning and Development, but temporarily placed and supported under MASAF IV coordinating agency-LDF. The roles and reporting arrangements of the UBR Management Unit and the UBR Taskforce (and associated agencies) are still being clarified, but the vision is that the Management Unit would eventually take on the day-to-day management and operations of the UBR, and the Taskforce would shift to providing guidance, knowledge transfer, and oversight. The unit has recently appointed four staff members: a UBR manager and three data management assistants. Further recruitments are in progress for a technical IT specialist and a capacity-building role. Additional positions and skills are needed in the areas of operations management, planning and implementation; business analytics; database management; and IT systems development. Another essential step is to establish the foundational legal framework for the UBR. Currently, the UBR has an operational manual that provides guidance on implementation processes (see Section 5), as well as data sharing protocols guided by the Access to Information Act (see Section 6). However, the UBR has no formal legal standing, such as a foundational law or decree. Such a legal framework is important to provide explicit guidance on: (i) the objectives, purpose and use of the UBR; formal institutional arrangements and responsibility for the UBR; (ii) the distinct relationship of the UBR to social protection programs and other agencies; (iii) the rules governing the use of the information provided; (iv) the rights and responsibilities of the population providing the information; (v) and data privacy; (vi) and similar ground rules. Foundational legislation should also be complemented by permanent institutional arrangements, including: the UBR’s legal home (host agency); roles and responsibilities for management, operations and implementation; administrative reporting arrangements; and supervision and oversight. The planned and ongoing work on a social protection law offers a huge opportunity because it is a logical vehicle on which to base the UBR’s legal framework. However, considering that social protection as an area is increasingly evolving and dynamic, it is also crucial that such a framework allow for some flexibility in terms of the legal instruments used (laws, decrees, regulations, operations manuals, etc.). Finally, the UBR lacks a financial strategy for managing its many diverse sources of funding (government, user programs, and development partners) for the long-term. One element of such a strategy would be to establish a formal budget line in the Government’s own administrative budget that could anchor and provide formal jurisdiction to the UBR. Another element would include donor coordination, as discussed further below. Local Roles for Implementation of Registration Processes Too often, countries treat social registries as mere databases for which data are collected only once through a “one-off” survey. Some countries outsource data collection to consultants or develop parallel structures to carry out these processes (Box 2). Such approaches may have the short-term benefit of moving quickly, but they fail to build on local capacities for citizen interface. They neglect the core functions of social registries, which are to serve as living information systems to support the processes of outreach, intake and registration, and determination of potential eligibility for social programs. Not so with Malawi’s UBR, which has invested significantly in the formalization of those 28 processes by developing solid implementation guidelines (see Section 5) and in the institutional arrangements used to carry them out. Box 2—Local Institutional Arrangements for Implementing Social Registries: Examples The institutional arrangements for these local roles vary significantly for on-demand applications and census-sweep registration methods. Some examples of arrangements for on-demand registration methods include: • Local Offices of the Central Host Agency. These local offices are managed, staffed and funded by the central agency. Examples of countries with this arrangement include: Georgia, Macedonia, Mauritius, Montenegro, and Turkey. Local Municipal Government Offices. These arrangements typically require a memorandum of understanding between the central agency and each autonomous local government (such as a municipal government). Examples include the social registries in Brazil, Chile, Colombia, and China. • Common Application Form for All Social Programs. Mexico’s social registry uses a common application form, and households can apply to one program but be considered for many programs with that common form. • Online Digital Service Windows. Households can apply for benefits and services using a common application form and an online digital service window for the social registries in Azerbaijan, Chile, and Turkey. Some examples of arrangements for census-sweep registration waves include: • Contracted Field Teams. These include the social registries in the Philippines, Colombia (which uses both census-sweep and on-demand methods), the Dominican Republic, and Yemen. • Communities + Contracted Field Teams. These include social registries in Djibouti, Mali, Sierra Leone • Outsourced to Firms. Pakistan National Socio-Economic Registry (contracts bid out to firms on a regional basis, another firm hired to supervise quality of implementation). • Contracted NGOs. Social registry in the Dominican Republic • National Statistics Office. Indonesia’s Unified Database (2015 registration wave) Source: Leite et. al. (2018) A major strength of the UBR is its use of existing local institutional structures for implementation. These include the District Social Support Committees (DSSCs) and the District Training Teams (DTTs), as well as the Area Executive Committee Members (AECs), Community Social Support Committees (CSSCs), and community leaders. The advantages of this institutionalization of implementation with existing local structures cannot be overstated. First, these local arrangements put a known “face” on the social registry for communities—and can allow for a point of contact for citizens (e.g., for queries, grievances, and appeals). Second, they build understanding, ownership, and credibility of the UBR as a social registry at the district level (rather than engendering suspicion by having an outside unit conduct registration, as we have seen in other countries). Third, these arrangements avoid the costly, confusing, and inefficient duplication associated with the use of parallel structures. Fourth, they strengthen ties and interaction between the central and local actors. Finally, this institutionalization recognizes the permanent and core function of social registries in social protection and positions the registry for sustained operations, including future updates and the potential for eventually shifting toward a more dynamic on-demand model. District-level actors cover the roles of coordination, training, and supervision, while field implementation is carried out by Area Executive Committee Members (AECs). Figure 6 displays the official organizational chart for implementation of data collection. At the district level, the main actors include (a) the DSSC and its associated DSSS, which are responsible for overall coordination of district- level implementation, consolidating geographic pre-mapping data (key preparatory step), and 29 overseeing data collection. The UBR also makes use of the National Training Team (NTT) and the DTTs to carry out training and sensitization activities for implementation at the district and community levels. The primary field implementation teams are the AECS, which consist of government and NGO extension workers for each T/A (Traditional Authority).11 The AECs mobilize the community, support the community meetings, conduct the household interviews, ensure that all forms are completed and submitted, and so forth. Community-specific actors include the CSSCs, which are teams of community members chosen by each community to work on social support programs, and the Group Village Heads or other village heads in each community. The CSSCs’ main roles are to serve as a link between households and the district, to mobilize communities, and to guide and support the AECs during community meetings and household interviews for the UBR. Again, these structures all exist and operate organically in the districts and communities in Malawi independently of the UBR, which built its implementation processes around them. Figure 6—Organizational Chart for UBR Implementation of Data Collection National Level District Level Source: UBR Taskforce presentation, and also presented in the Implementation Guidelines. One aspect of the UBR that could warrant further attention is the role of its central authorities, such as the UBR Taskforce, in planning, coordinating, and overseeing implementation. Indeed, the lack of mention of the UBR Taskforce on the implementation organizational chart and the few mentions of it in the Implementation Guidelines are somewhat surprising. It seems that the central management and operations of the UBR information system were conceived separately from implementation of the primary social registry functions of outreach, intake, and registration. In practice, the UBR team did get involved in ensuring implementation. However, the UBR could play crucial roles to support 11 Traditional Authorities (TA) are among the highest levels of traditional leaders or chiefs. They act as custodians of cultural and traditional values of groups of communities covering a particular geographic zone in a district. They have control of customary land; perform a semi-judicial function settling customary disputes over land; lead and mobilize their communities to participate in development initiatives and act as chairpersons of Area Development Committees (ADCs). Geographic areas under a TA are also sometimes called TA. 30 implementation of these core processes, including undertaking comprehensive implementation planning, coordination, and monitoring, as well as supervisory and oversight functions to ensure and check for quality. Institutional Relationship between the UBR and User Programs Many countries, including Malawi, use social registries as a common gateway for multiple social programs. Indeed, the UBR was conceived under the MNSSP 2012–16 as a response to concerns about fragmentation in social protection programs. As such, it was designed to serve multiple social programs in order to coordinate social policy, harmonize “targeting” (eligibility) concepts and criteria, and reduce costly duplications in registration processes (both for administrators and for citizens, who would have to provide the same information to numerous programs in the absence of such a common gateway). The UBR was initially designed to serve two flagship interventions, with a view to expanding the set of potential user programs over time. The flagship user programs are the SCTP and the PWP. Potential user programs include FISP, VSLP, microfinance, nutrition programs, scholarships, and humanitarian assistance. To support these efforts, the UBR Taskforce consulted with the various stakeholders, including user programs and their host agencies, in order to gauge their needs in terms of variables and information. The needs of the various user programs were then incorporated into a core Programme Requirements Document.12 On that basis, the Harmonized Targeting Tool (HTT) questionnaire was developed as the main instrument for collecting data in the UBR with the intention of providing core demographic and socio-economic information needs to the various user programs (as discussed in more details in Section 8 and Box 5). Clarity in the institutional relationship between social registries and user programs is crucial. Several factors come into play in this relationship. The first is the formal “home” for the social registry vis-à-vis the agencies managing the social programs. A key question in this regard is the degree to which a social registry can serve as an “honest broker” or custodian of data for multiple users and institutions. The second factor is clarity in jurisdiction and roles governing the specific functions along the delivery chain. Whereas the social registry supports the specific functions of outreach, intake and registration, and determination of potential eligibility (blue segments of Figure 7), the user programs determine actual eligibility and enrollment (red parts of Figure 7), as well as implementation of payments, service provision, and case management (purple parts of Figure 7). As such, it is important to note the distinction between the determination of potential eligibility (which is supported by the data collected in the social registry) and enrollment decisions (which are the jurisdiction of user programs). The social registry is not responsible for enrollment decisions (or payments or other implementation) because those are under the purview of the social programs. The third factor pertains to the policies governing information sharing between the social registry and the user programs. In Malawi’s case, the institutional relationships between the UBR and the user programs are indeed distinct. The UBR Taskforce and associated steering and oversight committees have the advantage of “joint ownership”—comprising representatives of both the MoFEPD (which doesn’t have 12UBR Taskforce/Ministry of Finance, Economic Development and Planning (2015). “Programme Requirements Document: The UBR—A Common Database for Social Support Programmes.” 31 implementation ties to any single social program) and various line ministries and agencies, including those hosting the user programs (such as local development funds, and others). However, there has been some blurring of the jurisdictional lines, roles and branding between the UBR and user programs. This institutional and branding confusion seems to stem from several factors. First, from the outset, the conceptual distinction in jurisdictional roles vis-à-vis user programs could have been more explicit. As shown in Figure 7 and Figure 8, which was included in the original Concept Paper and subsequent official documents, the diagram suggests that the UBR itself makes decisions on eligibility and enrollment and then forwards program-specific beneficiary lists to the user programs. This perception is a liability for any social registry because the responsibility for determining eligibility and making enrollment decisions (and considering appeals and grievances) is the mandate of the user programs (which reside in other host agencies). Moreover, the social programs may use their own criteria, combinations of variables, and may supplement with additional criteria (such as educational records for a needs-and-merit-based scholarship)—to determine actual eligibility and take enrollment decisions for their own beneficiary rosters. Second, as discussed in Section 9, the name Unified Beneficiary Registry implies that it is a registry of beneficiaries (when in fact many or most households in the UBR are not and will not become beneficiaries of the SCTP or other specific programs). This naming issue is not just a semantic issue: it perpetuates a conceptual confusion about the role of the UBR and creates a liability for the UBR in terms of its primary roles vis-à-vis the mandate of the user programs, as discussed in Sections 2 and 8. Third, the reliance of the UBR on financing that is channeled through user programs and the fact that the UBR does not have its own line item in the budget (nor a formal permanent institutional structure) further contributes to the institutional blurring of the lines between user programs and the UBR. That confusion has been even stronger for the UBR vis-à-vis the SCTP. The blurring of the lines between the UBR and with the SCTP derives from at least three factors. First, the SCTP is the dominant primary initial user of the UBR, which has given more prominence to the SCTP over other (future) users. Second, the SCTP used to operate its own data collection and in fact, the software in its MIS still has a targeting module) for “targeting.” As such, many of the recent discussions have focused on the degree to which the UBR would capture data on existing SCTP beneficiaries, and whether those would still be considered poor or otherwise eligible for the SCTP. Third, the SCTP has disparate implementation arrangements amongst development partners, which introduces different requirements and considerations in planning. AS UBR roll out extends to districts with ongoing SCT program; there are considerations by Ministry of Gender in collaboration with some SCTP donor partners to combine the UBR ‘census sweep” registration with recertification of existing SCT beneficiaries. In the absence of a strategic document that guides SCTP recertification, this is being done to pilot an exit or re-entry procedure for SCTP in those districts. The downside is that with such a back to back implementation of processes that are supposed to be distinctly undertaken by the social registry and a user program-SCTP, there is potential that this might be another source of confusion. Certainly, the UBR could have an agreement with user programs to make explicit efforts to locate beneficiaries of the STCP and other programs when it undertakes registration efforts in any given district. However, there is no guarantee that any updated registration effort would locate precisely the same households—or that the communities or eligibility assessment tools (such as the PMT) would choose the same households that previously qualified for the SCTP, since the demographic and socio-economic situations of both the SCTP’s beneficiaries and its non-beneficiary households might have changed in the interim years. 32 Overview & Approach to the UBR Assessment Conceptual Framework Figure 7—Integrated Social Registries Serving Multiple User Programs Provide Integrated Enroll Social Registry Assess Monitor Potential Provide Demand Enroll Monitor Enroll Provide Monitor Integrated social registries support registration & assessment User Programs then use the socio-economic data from the social of needs and conditions to help determine potential registry to make decisions on eligibility, enrollment, & benefit levels eligibility for multiple programs (sometimes combined with other criteria) 6 Source: Leite et. al. (2017). Figure 8—Institutional and Technical Framework of the Malawian UBR Source: The Malawi UBR Concept and Proposal—October 2015. 33 Financing and Coordination with Development Partners During the assessment, the role of development partners emerged as an important consideration in the implementation and legal setup of the UBR. Donors have continuously financed a significant share of UBR operations, from software development and IT support to rollout. Government pays the salaries of staff working on the UBR. Much of the donor funding for the UBR has been channeled through donor financing of user programs. In the next phase of the UBR’s rollout, the European Union (EU) and Germany Government (both through Kreditanstalt für Wiederaufbau (KfW)) as well as the World Bank, through their commitments to the SCTP, have committed to finance the UBR’s operations in 2018 and 2019. Support from development partners has provided several important advantages to the UBR. The partners’ financing has allowed for the rapid expansion of the UBR process. Development partners have also been able to access technical expertise and advisory services, particularly GIZ. Going forward, there are inherent challenges as the UBR works with a broader set of financing partners. This is a broader challenge in social protection more generally. On the program side, the SCTP’s experience illustrates this concern. The SCTP is financed by five separate agencies, each of which has different financing and implementation modalities. For example, Germany and the EU (both through KfW) have a mixed system of partially using the Government system and reliance on third- party vendors, relying on them for fiduciary co-management and increasingly for program technical support. While this facilitates program implementation, it also introduces transaction requirements for the Government and has the drawback of promoting structures outside of the Government system. By contrast, one of the UBR’s core strengths has been its ability to build on national structures, and this core strength should be maintained. In the future, clearly enshrined roles, responsibilities and processes will be important as the UBR continues to expand and to engage different financing partners. Strong coordination arrangements will also be required within the UBR, to minimize transaction costs and leverage expertise. Going forward, it will also be important to establish a line in the national Government budget to help formalize domestic financing. Donor dominance in financing the UBR broadly reflects donor dominance in financing social protection in Malawi. The unsustainability of such a financing situation cannot be overemphasized. It is therefore recommended that to formalize domestic financing of the UBR, Government establish a budget line, at least for the central management of the UBR. It would be important that such a budget line be included from fiscal year 2018 and onward. Finally, the UBR should allow for continued leveraging of funds from various donors and agencies, with multi-sectoral support. As highlighted above, the existing arrangements, especially for beneficiary assessment sub-processes (outreach, data collection, intake and registration) allow donors and agencies interested in using the UBR for other user programs to finance some operational costs. This can be formalized by exploring how UBR’s costs could be shared with programs that are consistently using it. 34 5. UBR Implementation Processes The assessment of the UBR’s implementation processes covers three areas: (a) understanding and reviewing the processes in the “as-is” model of implementation (Phase 1 implementation in the 10 districts), as it will likely serve to inform Phases 2, 3, and 4; (b) identifying and building on the lessons learned from Phase 1and considering the ways the processes could be strengthened; and (c) considering the implications of the planned shift from to registering half of all households in Phases 2, 3, and 4, how that shift will affect implementation processes, and which processes may need to be modified as a result of that shift. Implementation Processes for Phase 1 (“As-Is” Guidelines) A key step in designing the UBR was the development of an implementation manual for core processes. Implementation processes are detailed in the “Guideline for Unified Beneficiary Registration Process” or referenced as the Implementation Guideline elsewhere (November 2016). The Guideline is very thorough and well presented, with a solid description of institutional roles and core processes, and then with Annexes to further elaborate on details. The process maps in the Guideline could be improved by indicating the institutional responsibilities and actors for each step. Figure 9 shows the process map included in the Guideline, which is useful in introducing the sequential steps involved in implementation. However, it is not clear from Figure 9 who will perform each step. As an alternative visualization, Figure 10 assigns a “swim lane” (row) to each level of institutions (central, district, and village). It then maps the sequential steps from Figure 9 to specific actors within each box. This mapping helps assess the robustness of the processes by identifying who does what and when they do it for each of the core steps. Unique role assignments are crucial for clarity and accountability, and the term “swim lane” is a management tool to symbolize the concept that each actor “stays in his own lane,” without crossing lanes and thus confusing roles. This also allows one to understand the complexities that derive from multiple layers of implementation and the potential time lags that may arise from the back-and-forth across different layers. When the Guideline and associated training materials are updated, we recommend that the process steps be mapped according to these “swim lanes” for each actor or level (central, di strict, village). Unfortunately, due to rapid nature of this assessment the team was not able to delve deeper into the time requirements at each step. As reflected in the recommendations section, this is a crucial input to understand the efficiency of implementation. 35 Figure 9—Overview of Implementation Processes for Malawi’s UBR Source: Guideline for UBR Process (November 2016) The core steps in implementing the UBR can be grouped into the three main phases for social registries: sensitization, training and outreach (shown in light green in Figure 10); intake and registration (or data collection, shown in light yellow); and assessment of needs and conditions to determine potential eligibility for social programs (shown in orange). 36 Figure 10—Mapping of Main UBR Implementation Processes by Actor and Level UBR: Main Implementation Processes mapped by “actor” (level) Basic Steps - “As-Is” Preparatory Steps: 9. UBR applies PMT to Central UBR Train NTT & Carry out district 7. UBR applies PMT updated set of 10. UBR Capacity Assessments to all households households and assigns Data ready with “complete” classifications for use by Social status and assigns (after various data Programs 1. UBR & NTT classifications cleaning, consistency Conduct checks) District Sensitization Preparatory Meeting Step: District Level DSSS carries Supervision & Quality Oversight: out Geographic • 2 DTTs supervise teams of 6 AECs Pre-Mapping of • DSSC broader supervision 2. NTT carries out District with • NSO oversight, provide feedback to DTTs & monthly reports training of DTT UBR team to list • M&E team from EPD/MoG/LDF quarterly reports to UBR TAs, GVHs, Villages 3. DTT carries out training of AECs 5. AECs, with DTT, 6. AECs conduct 8. AECs, with DTT, conduct 2nd Village Level conduct 1st community 4. AECs & DTT household community meeting with support meeting for community carry out interviews with of CSSC & Community validates to elect CSSCs and select orientation of support of CSSCs; + PMT ranking; AEC conducts and rank “pre-listing” of communities & data collection, entry remaining HH interviews, data entry 50% of households based Chiefs & ODK upload for appeals cases (outside 50%) 24 on “ultra-poor” filters Sources: adapted from The Guideline for UBR Implementation Process + discussions with UBR Taskforce and others during mission 37 Sensitization, Training, and Outreach Processes The Guideline outlines key preparation activities for the UBR, including the sensitization, training, and outreach depicted in Figure 9, as follows: • Preparatory Steps. The UBR Taskforce undertakes initial preparatory steps, including training of trainers with the National Training Team (NTT) on the UBR’s objectives, processes, and institutional arrangements. During Phase 1, the Taskforce also carried out a capacity assessment of the districts to take stock of the various institutional actors in each district. This district capacity assessment has now been formalized as part of the UBR procedures when scaling up the UBR to new districts. However, for districts with ongoing social protection programs especially SCTP; the districts will have benefited from existing SCTP procured equipment and community coordination structures. As a result, the capacity assessments are not as rigorous as the assessments done during the first phase of UBR roll out in the totally new districts. • District-Level Stakeholders Sensitization Meeting (Step 1 in Figure 10). After the preparatory steps, the District Commissioner calls a one-day stakeholders’ meeting, which is facilitated by the NTT using stakeholder sensitization guidelines (presented in Annex 1 of the Guidelines). The key stakeholders include members of the District Executive Committee (DEC), Council Members, Traditional Authorities, sub-Traditional Authorities, and Members of Parliament for that district. The objectives of the meeting are to: (a) provide an orientation for these high-level participants about the UBR (“the harmonized data collection for MNSSP”); (b) spread awareness about upcoming activities for the UBR’s process; and (c) select members of the District Social Support Committee (this is done by the DEC). This sensitization meeting is viewed as a very helpful step to guide the implementation of the UBR in each district. • Training Activities (Steps 2 and 3 in Figure 10). Two sets of training activities are then carried out. First, the NTT conducts a five-day training of the DTT’s trainers. The training focuses on the HTT questionnaire (see Section 8), data collection process, guidelines, and household identification. The objectives of the training are to: (a) explain the objectives of the UBR within the Malawi National Social Support Programme (MNSSP) framework; (b) define concepts to be used during the harmonized data collection process; (c) present the selection characteristics of households in line with the MNSSP; (d) describe the roles and responsibilities of various stakeholders; and (e) demonstrate how to fill out the HTT questionnaire. Additional topics are contained in an annex to the Guideline. Second, the DTT conducts a five-day training of members of the AECs at the Traditional Authority (T/A) level. The objectives and topics of this training are similar to the DTT training, but focus on the role of the AECs. • Orientation of the CSSC Members and Chiefs (Step 4 in Figure 10). One AEC member, supported by a DTT member, then conducts a one-day orientation for six selected Community Social Support Committees (CSSCs), five Group Village Heads (GVHs), and their associated Council Chiefs, at a village cluster. The training introduces the UBR, its place in the MNSSP framework, the roles and responsibilities of CSSCs and Chiefs, and household characteristics based on the MNSSP. The orientation also shares geographic pre-mapping information (see below) with the GVHs and Council for validation. The steps outlined above are all crucial, but more attention could be paid to communications and outreach, particularly for communities. Observations during the mission and findings from other field 38 studies13 indicate that the acronym UBR is not well known at the community level and the purpose of the UBR is poorly understood. Further effort is needed to ensure that communities and households understand the UBR and its relation to social programs. (For more details, See Section 9.) Intake and Registration (Data Collection) Intake and registration involve the process of collecting information to register the intended population for consideration of potential eligibility for social programs. This phase represents the formal entry point for households into the social registry system. Because these data collection processes are costly, there are significant efficiency gains from carrying them out in a harmonized manner for multiple programs (rather than separately). Yet such harmonization also raises the stakes for quality in implementation processes because the information collected will influence households’ eligibility for many programs, not just one. The UBR processes for data collection in Phase 1 includes three key steps, primarily involving district and community actors. Broadly speaking, the process involves a preparatory step to map all communities and households in the district. Each community selects which households will be registered; and the AECs then conduct interviews and home visits to collect data, using the HTT questionnaire from those households that were selected by the community (see Section 8). The data collected during those home visits and interviews are then entered into the UBR ’s information system (see Section 4). More specifically, data collection for the UBR includes the following steps (yellow- shaded steps in Figure 10): • Geographic Pre-Mapping (preparatory data collection step). This step involves the process of demarcating each district into geographical reference areas, including collecting the names and codes for the district, T/As, Gazetted Village Heads (GVHs), and villages. The challenge is that Malawi does not have a physical address system for households and not all villages or settlements are formally designated or “gazetted.” The clustering of settlements evolves over time. For example, some settlements split off from GVHs, setting up their own village heads, without gazetted status. Further complicating the situation is the fact that the various social support programs may use different codes and even mapping methodology for the same geographic locations and villages. Part of the harmonization that is needed is to align geographic coding among the UBR and the various user programs. In terms of institutional responsibilities, geographic pre- mapping is carried out by the DSSS, in consultation with other actors, including extension workers and GVHs, to validate the information that has been collected. Efforts are made to ensure completeness and consistency of names and spelling. To establish a roster of households in each district, T/A, GVH and village, the DSSSs also use secondary data, such as village registers, census data from other sectors, information from the NSO and the National Registration Bureau (NRB, which manages the national ID). Key outputs of this step include the geographic coding of all T/As, GVHs, and Villages and the establishment of a roster of households in each of those areas in the district. The geographic codes are then programmed in the UBR information system and the software that is fed into the mobile tablets used for data collection. • First Community Meeting and preliminary listing of households to be registered (Step 5 in Figure 10). A crucial—and influential—step in implementation is the First Community Meeting because it 13 King and Tranchini (May 2017). 39 will result in determining which households will be registered in the UBR. The Guideline outlines the numerous activities that take place during the First Community Meeting, which, as the community’s initial activity, combines both outreach and data collection. 14 Participants include the AEC members who facilitate the meeting, GVH and village heads, and the members of their respective villages. The participation of community members is usually 95 percent or more, according to discussions with officials during the assessment mission. The Guideline specifies that the AEC, with the DTT’s support, will lead the meeting as the community at the level of each village head identifies poor households in the village, whether or not those households are present at the meeting. The Guideline also states that community members shall identify the households that meet at least one of the following criteria for being poor:” the household (a) has on average only one meal per day; (b) survives from begging; (c) is undernourished; (d) does not possess any valuable assets; (e) does not receive any monetary help, food, or gifts from others; (f) survives on piecework; (g) has no access to credit loans; and (h) has no shelter or a house in poor condition. Finally, the CSSC “validates household listing of 50 percent of the total households that are poor to be interviewed at village level.” The core output of this First Community Meeting is the list of poor households that will be registered in the UBR. • Registration: Household interviews, data collection and entry (Step 6 in Figure 10). The core implementation phase—and the ultimate objective of the UBR—is to collect demographic and socio-economic information on households that could potentially be eligible for multiple social programs. As discussed above, the households to be registered are the 50 percent prioritized by the community, although the Guideline also states that, during the visit, the AEC members may also visit households which in their view meet the eligibility criteria for identifying the poor (see above), but who are not on the provisional Community Listing. The respondent for the interviews is the household head. The home visits and interviews are carried out by the AEC members, with the CSSC member serving as the liaison to the community and assisting the AEC team in locating the households on the list. As the main field implementation team, the AECs consist of government and NGO extension workers. Typically, there are six AECs for each cluster of villages, and they are assisted by six CSSCs. The interviews are guided by the HTT questionnaire (preloaded in the ODK software on the tablets). The Guideline provides detailed instructions to the AECs on data collection, with many good-practice principles: the AEC member must explain to the household that this visit does not guarantee enrollment into any specific program, thank the interviewee for his or her time, and leave a receipt as proof of a household visit, among others. The Guideline also includes instructions for verifying completeness of all mandatory fields, stamping the questionnaires, entering the data, filing the forms, and so forth. With the goal of making quality data a fundamental output of the UBR, supervision is carried out by two DTTs for each team of six AECs. Each cycle of interviews in the village clusters typically takes the AECs eight days, while the DTTs put in five days of supervision. A full-time official from the National Statistics Office (NSO) is assigned to each district to observe each AEC in the field, and to give feedback to the DTTs and to the UBR Taskforce via monthly reports. While the DTTs verify the completeness of the data, the DSSSs are responsible for authorizing its transfer into the UBR’s information system. The core 14 According to the Guideline, the many objectives of the First Community Meeting are to: “inform the m embers of the respective GVH in detail about the objectives and data collection processes; assist the GVH to elect the CSSC; agree with the CSSC on the schedule of harmonized data collection activities and on venue and logistics for training; undertake preliminary listing of 50% of poor households to form the basis for harmonized data collection process (criteria for identifying the poor households); and confirmation of zones, clusters, catchments, and EPAs and reflected in the Pre-Geographical Pre-Mapping Data (sic).” 40 output of this step is the household demographic and socio-economic data that is collected and registered in the UBR system. Assessment of Needs and Conditions to Determine Potential Eligibility for Social Programs Potential eligibility for social programs is assessed using a combination of Community-Based Targeting (CBT) methods and Proxy Means Test Scores (PMT). PMTs are calculated using data collected from UBR using the HTT questionnaire (see Box 5 in Section 5). The role of the communities and CBT in the UBR processes is twofold: communities first determine the households that will be registered (Step 5), and then validate the rankings of the PMT (Step 8). As discussed above, this joint CBT-PMT process for the UBR only determines potential eligibility for social programs, which is distinct from actual eligibility and enrollment decisions. Those decisions are the jurisdiction of the social programs, not the social registry. The specific steps for assessing potential eligibility for social programs are shaded in orange in Figure 9. They include the following: • First-round PMT applied to UBR household data (Step 7 in Figure 10). The UBR applies PMT scores to all households with “complete” data status and assigns classifications. The calculation of PMT scores and the ranking classification is automated in the UBR information system and overseen by the UBR team. • Second Community Meeting to Validate the PMT (Step 8 in Figure 10). The AECs then conduct the Second Community Meeting, during which the community validates the PMT ranking and identifies any households they believe were either misclassified, excluded, or have appeals cases. The AECs manage this meeting with the support of a DTT member. The CSSCs, other respected members of community, council members, GVHs, VHs, and community members (including those interviewed) are all invited. The meeting’s objectives are to: (a) remind the community of the data collection process and its objective; (b) present to the community the village list with each household’s classification; and (c) identify and add any other households that may have been erroneously excluded. Each household is discussed, and the list is updated with a “community validation” column that notes whether the community is in agreement (and its reason). Households that want to appeal their status or believe they were erroneously excluded from the list can interviewed and registered by the AEC member. • Second-round PMT applied to the updated UBR data (Step 9 in Figure 10). After the second community meeting and any follow-up interviews, the UBR applies PMT to the updated set of households and assigns classifications (Step 9). The calculation of this second round of PMT scores and ranking classification is automated in the UBR information system and overseen by the UBR team. • The UBR data are then readied for use by social programs (Step 10 in Figure 10). The PMT scores and UBR data are all tested for data quality (see Section 7) and then readied for use by social programs. For example, data on the poorest 10 percent of the UBR’s households is sent to the SCTP because that is the group that program covers). 41 Box 3—Political Interference in SCTP enrollment process A recent report commissioned by the Local Development Fund and the MoGCDSW found some instances of political interference in enrollment decisions for the SCTP. Although such interference does not reflect directly on the UBR itself, in some instances gaps in implementation of the UBR could give room to subjectivity in the identification and registration of households (for example, there are reports that in some TAs, the AEC didn’t reach out to all households for UBR registration and/or SCTP enrollment due to difficult terrain. Political interference in enrollment for the SCTP was particularly observed in Lilongwe district. In Lilongwe, traditional leaders and Councilors created their own list of beneficiary households for the SCTP rather than selecting them from the UBR. In some TAs and village clusters in the same district, Councilors and Chiefs exerted pressure CSSC members to conduct the community meetings in a way that biased the results of the community selection process. In the Kasungu District, the Group Village Head in the Suza Cluster under the Kaomba Traditional Authority (TA) protested the decision to not elevate him to the sub-TA level, and prevented households within his jurisdiction from participating in the SCPT and any similar program until his status was elevated. In the Rumphi District, households in one constituency (Rumphi North, covering several clusters) were coached by political structures to falsify information so that they qualify to be within the 50 percent threshold for UBR registration, and subsequently SCTP enrolment. Source: IDEAL Consulting & Business Services (2018); Limbe (2018); Chitekwe (2018) Strengthening Implementation Processes Going Forward In general, the overall approach to implementation processes for the UBR is sound, with ongoing improvements and updates. The Implementation Guideline lays out a solid set of processes, and discussions during the assessment mission confirmed the utility of these processes. The UBR Taskforce team is currently updating the Guideline with a detailed list of updates to the specifics of each step in the process, building on the lessons learned from the pilots and Phase 1. The rapid assessment does not get to that level of granularity, but rather focuses on several key areas for improvement and/or updating (in addition to other suggestions discussed above): • Strengthening outreach and sensitization • Adjusting the model for the AECs, with the option of appointing and training more specialized teams of rotating AECs that would cover a larger number of villages within each T/A • Standardizing and enhancing supervision and oversight for data quality and appointing a team of spot-checkers • Formalizing steps for handling appeals and grievances • Clarifying the role of the UBR Taskforce (or UBR Management Team) in coordinating, planning, and overseeing implementation • Establishing and tracking productivity targets in practice and for coordination, planning, and monitoring • Developing a comprehensive end-to-end implementation plan with key targets and indicators for monitoring and reporting for Phases 2 and beyond. One area for potential improvement is to extend the role of the AECs. Currently, the model is to select multiple AECs (e.g., six) for each cluster of villages, train them and pair them with same number of CSSCs (representing each village). This means hiring, selecting and training numerous AECs across the village clusters in each T/A, in all T/As in each district. The AECs are paid lunch allowance as an incentive for each day of work on the UBR. Yet there is a learning curve for each AEC in terms of training, experience with the interviews and the tablets, and quality. As such, one measure under consideration for Phases 2, 3, and 4 is to reap economies of scale and improve quality with rotating teams of AECs 42 (fewer in total) that would gain more experience and specialization in the UBR processes (questionnaire, interviews, tablets). These rotating teams of AECs would work on the UBR for a longer time and would cover more households and villages, allowing them to apply their knowledge, rather than stretching training and supervision across so many AECs, each of whom work with just a few villages for a short time. Another improvement would be to standardize and enhance supervision and oversight for data quality. A key finding of the assessment points to the strength of data quality, which is discussed in more detail in Section 6. As the UBR’s coverage expands—and registration targets potentially increase—it is of utmost importance to maintain high data quality. No social registry is 100 percent free of errors, but the key is to have systems that can detect and remedy them. Various aspects of data quality are discussed in Section 6, but one that bears emphasizing in terms of implementation processes is the need to standardize and enhance supervision of data collection. Currently, the DTTs and NSOs are responsible for overseeing data collection processes. However, the Implementation Guideline would benefit from the inclusion of guidance and standardized checklists on what quality checks to perform, what errors DTTs and NSOs should look for, and how to remedy observed errors. It would also be useful to have standardized field reports that the DTTs and NSOs could send regularly to the UBR Taskforce to facilitate broader monitoring and coordination, as discussed in Section 6. Another supervisory recommendation would be to formally appoint a team of spot checkers to carry out random re-reviews for a sample of households. The proposal would involve appointing a rotating team of AECs (or other enumerators) who could serve as spot checkers that would re-review a random sample of households on a continuous basis. The spot-check team would then submit the data from the HTT questionnaires to the DTTs (or other supervisory body) for comparison (using the household head’s ID number or UBR functional identification number). If the comparison yielded significant differences, the DTT would accompany a separate AEC to visit the household to verify which set of data was accurate and the source of the errors. Simply having the rotating team of spot checkers would raise the bar for quality implementation by the AECs and for supervision by the DTTs. Moreover, the act of spot-checking would help identify and resolve errors, thus contributing to the overall quality and credibility of the UBR. In this process, there may also be an opportunity to use qualitative methods to garner feedback on the UBR process from beneficiaries and communities alike. It would also be helpful to formalize and strengthen the process for handling appeals. There are many forms such grievances can take. Households can appeal community decisions that excluded them from the set of households registered as poor. They can also appeal their PMT score and their associated eligibility for social programs.15 Or they can request a re-review of their information (which can also involve updating it, if their situation has changed radically). The notion of appeals is mentioned in the Guideline for the Second Community Meeting (step 8), noting that households can appeal to be included or correct errors. However, there is limited guidance on the process for indicating, recording, resolving, monitoring or reporting on those appeals. Moreover, the appeals seem to be initiated via 15 However, since the user programs apply “relativity” in their thresholds for eligibility (such as the bottom 10% of households registered), there no guarantee that an appeal of PMT score or re-review of household information would allow for that household to enter the program. The reason is that with relative eligibility thresholds, everyone’s position for eligibility depends on everyone else’s. If all spots in the program are already filled (to meet the 10% program quota), another household would have to exit to permit the appealing household to enter. These relative eligibility thresholds are inherently “static” as they require a full re -ranking of the PMT to determine eligibility, rather than allowing for everyone below an absolute cut-off score to be eligible and entitled to the benefits. This is a concern in many countries that use relative eligibility criteria. 43 the community discussions (which could replicate exclusionary decisions from the first community meeting in the second), but there is no formal channel for an individual household to appeal directly to the AEC/DTT team. Thus, another recommendation would be to formalize and strengthen the process for managing appeals. A broader recommendation would be to clarify the roles of the central UBR team for planning, coordinating, and overseeing implementation. As discussed in Section 4, the UBR central team does not appear on the organizational chart for data collection—and is relatively absent in terms of its roles in the Guideline. It makes sense that the UBR wouldn’t directly be involved in data collection per se because those roles are decentralized (as in many countries). Nonetheless, it should have a role in coordinating, planning, and monitoring implementation across districts. A key recommendation that merits attention from the UBR Taskforce is the establishment and analysis of productivity indicators and targets. These productivity indicators could include variables such as the number of interviews per AEC per day; the number of interviews per week; the amount of time needed to complete key steps (such as steps 5, 6, 8); and the amount of time needed from start to finish for each village, T/A or district. Clearly, the results of these indicators will depend on context (the remoteness of the villages or households, terrain, etc.). Yet it is helpful to measure, analyze, and report on these indicators as a key input for coordinating, planning and monitoring of implementation. The UBR’s implementation involves many “moving parts” and requires significant investment in end - to-end implementation planning. As the team prepares for Phase 2 and beyond, it should map out the full end-to-end implementation plan (E2E-IP) for each district (and across districts) from start to finish, building on the productivity indicators and identifying the time and resources needed for each step and each locality. E2E-IP is a common management tool that takes the UBR process to the next level, mapping even more specific details for the specific steps for all processes in sequence, including the actors and resources used for each step. E2E-IP brings together all of the process steps, actors, schedules, and resources needed to implement one or more programs, in order to enable management to monitor and control all of these elements under a single plan. Specifically, it involves the sequential scheduling of processes and responsibilities, addressing: (a) the main implementation phases and detailed process for each phase? (b) who does what in each process; (c) when each process will be carried out and how long it will they take; (d) the resources needed for each step (e.g., full-time equivalents in terms of staff time, materials, vehicles and other inputs); and (e) how the “system” calibrates timing, scheduling, target dates and deadlines, handoffs between actors, transitions, and other factors? Basic project planning software packages can be quite useful in developing E2E implementation plans. When a basic E2E-IP is developed as a “baseline,” it can be used for analytical and monitoring purposes. For example, actual implementation can be monitored and compared to the E2E Implementation Plan to check for delays and efficiency. The results of this monitoring can also be communicated to the various stakeholders (for example to development partners and other agencies to keep them updated on progress). Implications of Shift to Full Registration on Implementation Processes The planned increase to 100 percent registration targets has a number of ramifications for implementation. One of those implications is the role of communities. Community participation is critical as a means of supporting the effectiveness of the UBR, both in terms of its overall credibility 44 and in terms of the effectiveness of the UBR itself. Currently, the community plays a particularly important role in “Community-Based Targeting” (CBT) approaches. CBT enters twice into the UBR’s processes and complements the PMT scores that are estimated on the basis of information collected in the UBR. From a political-economy standpoint, these two methods serve complementary objectives. CBT methods seek to enhance the community’s ownership of the UBR (which in turn becomes the basis for determining potential eligibility for social programs with limited budgets). PMT methods inject an objective and empirical measure of poverty and well-being into the process. Both methods add to the credibility of the UBR and the quality of the data.16. In this manner, CBT and PMT cross-validate each other. PMT was introduced as a way to reduce subjectivity in the selection process, with community serving as a validation mechanism for PMT rankings. As the UBR targets all households for Phase 2 and beyond, the implementation processes involving community participation need to be revisited. An overarching challenge is to maintain transparency around the UBR, which current community engagement allows. Specifically, two key aspects need to be considered: • What is the inclusion function of the first community meeting (step 5 in Figure 10)? In the existing processes, the communities select which households would be included in the UBR’s 50 percent registration target. With the shift to 100 percent registration, there is no need for the communities to carry out this step. Rather, the core function of the communities at this stage would be to review the register of households identified by geographic pre-mapping and ensure that all households are included. This actually has simplified implementation processes both for the first community meeting in step 5 (politically, it is easier to make sure everyone is included than to discuss which households would be excluded) and for implementation of data collection in step 6, as the AECs would just go door-to-door for all households, rather than interviewing some and skipping others. • What is the ranking function of the second community meeting (step 8 in Figure 10)? In the existing processes, the communities discuss and validate the initial first-round PMT ranking from the UBR in the second meeting. This is already time-consuming because communities have to discuss the ranking of each household that falls within the 50 percent registration target. If they discuss the ranking of every single household (as would happen with 100 percent registration), it would presumably take even longer to carry out the second community meeting (step 8). This could be perceived as cumbersome and redundant by the community. Moreover, wealthier households may not want to be registered or have their status discussed in a public meeting, since they probably wouldn’t be eligible for the benefits and services offered by the user programs. Several options could be considered for adapting the community’s role in implementation processes for the UBR’s new registration target. First, the role of the community could shift to focusing on ensuring that all households are included in an adaptation of step 5, but would no longer validate the PMT ranking (dropping step 8). That option might reduce UBR’s credibility, however, especially given the historical importance of CBT in Malawi. Moreover, it could reduce the quality of data in the UBR. This became apparent in the Phalombe pilot, which was implemented with a 100 percent registration 16 Many donor partners also require some form of community participation and objective poverty measures for the programs that they finance. 45 target but without community validation of the rankings. Reportedly, the quality of the data collected was substandard (see Box 3, Section 5). Second, the community could maintain both its inclusion role in making sure no households are missed (an adaptation of step 5) and its ranking validation role by: (a) ranking all households (which would be cumbersome, as discussed above); or by (b) ranking some subset of all households (for example, the bottom 50 percent); or by (c) ranking groups of households according to their PMT ranges (for example, households could be grouped by quintile), with the community flagging households it thinks are misclassified. Third, options (b) and (c) could be combined, with the community ranking groups of households, and devoting more time and emphasis to discussing those in the poorest two (or three) quintiles. No matter the option, the Guideline and associated training materials would have to be updated to ensure understanding and proper implementation of these revised processes. Finally, it should be stressed that the shift in registration targets implies changes in terms of staffing, resources and distribution of field teams. The overall cost of data collection will need to be carefully reviewed, which goes beyond the current scope of this assessment. 6. The Quality of UBR Data The assessment took three approaches to examine data quality: (a) checks on the actual quality of the data in the UBR, verified through both internal consistency checks and cross-checks with the most recent national household survey (IHS4); (b) structures and processes for oversight, monitoring, and reporting (OMR); and (c) systems integrity checks. Internal and external data quality checks were applied to UBR data from the 10 districts in Phase 1. The analysis also includes the Phalombe pilot, for which data was also available. Data Quality Checks Since information is the primary input and output for social registries, data quality is fundamental. The main inputs to the system include demographic, geographic, and socio-economic information collected from households. The primary outputs of social registries are data that have been transformed into standardized formats or aggregations that permit assessment of needs and conditions against program eligibility criteria (such as means-tested incomes, PMT scores, etc.). The quality of these inputs and outputs depends on many factors, including implementation processes, monitoring and oversight, and systems integrity. Quality of data was assessed by three measures. These are: (a) its completeness (the number of missing variables); (b) its internal consistency (whether poverty classifications are consistent with key household characteristics); and (c) its external consistency (cross-checks between the UBR and the nationally-representative household survey, IHS). The vast majority of registrants have complete information. Apart from a few variables that were missing for many registrants, there were no missing values. Interestingly, for the missing values, the finding was in line with expectations and does not trigger concerns. For example, values were missing for the use of an irrigation system and type of organic fertilizer used. These are not common practices among the poorest 50 percent of Malawians, and hence not observed in their responses. Thus, it comes as no surprise to find missing values for those variables. 46 In terms of household composition, as expected, the poorest households in UBR have fewer working members or fit-for-work adults (see Figure 11). The ratio of household members who are fit for work per household, as well as the ratio of working to non-working members, is lower for poorer households. This fits with the narrative that poorer households have fewer members who are able to generate income for the household. Figure 11—Internal Consistency Checks of UBR data on Household Composition Source: authors’ analysis of UBR data for Phase 1 Districts + Phalombe (partial data available for Phalombe) UBR data are internally consistent in terms of the characteristics of households by income. As shown in Figure 12, poorer households rely more on ganyu (low paying casual labor) and begging as livelihood sources. About 67 percent of the poorest households depend on ganyu and begging, whereas that Figure 64 percent, 58 percent, 50 percent and 45 percent, respectively for the other four quintiles. Similarly, a greater share of the poorest Malawians live in traditional houses, use mud, compacted earth or grass for walls and use predominantly grass for the roof. 47 Figure 12: Internal Consistency Checks of Poverty Classification, Livelihoods, and Assets in UBR Data Source: authors’ analysis of UBR data for Phase 1 Districts + Phalombe (partial data available for Phalombe) 48 Figure 12 (contd.): Internal Consistency Checks of Poverty Classification, Livelihoods, and Assets in UBR Data Source: authors’ analysis of UBR data for Phase 1 Districts + Phalombe (partial data available for Phalombe) UBR data are externally consistent. For this analysis, the characteristics of households belonging to the poorest 50 percent of the nationally-representative sample in the IHS3 (2010–11) and IHS4 (2016– 17), were compared with the characteristics of the households in the UBR. The findings suggest that the UBR’s poverty data is high quality. First, comparing the two surveys, the poorest 50 percent of 49 households in the 2016–17 IHS4 have better characteristics (assets, livelihoods), on average, than the poorest 50 percent of households in the 2010–11 ISH3. Second, households in the UBR have either comparable or worse characteristics, on average, than the poorest 50 percent of households in the 2010–11 IHS3, and fare worse than those in the 2016–17 IHS4 (which is the similar time period as when the UBR data were collected). This implies that, on average, households in the UBR belonged to the poorest of the bottom 50 percent of households in 2016 and 2017. For example, 84 percent of UBR households use grass for the roof, compared with 67 percent of those surveyed in the IHS4. Figure 13 presents additional comparisons consistent with this finding. Figure 13: External Consistency Checks: Comparison of Households in the UBR with the Poorest 50 Percent of Households in the IHS3 (2010–11) and IHS4 (2016–17) Source: authors’ analysis of UBR data for Phase 1 Districts + Phalombe (partial data available for Phalombe) and of the IHS3 and IHS4 datasets. 50 Figure 13 (contd.): External Consistency Checks: Comparison of Households in the UBR with the Poorest 50 Percent of Households in the IHS3 (2010–11) and IHS4 (2016–17) Source: authors’ analysis of UBR data for Phase 1 Districts + Phalombe (partial data available for Phalombe) and of the IHS3 and IHS4 datasets. Over all, the findings on data quality give grounds for confidence in the UBR as an information source. The data quality is generally good, with a minimum level of missing entries. The data is comparable to the poorest 50 percent of household in IHS3 and IHS4. One notable finding from the data analysis relates to lower data quality in Phalombe, a finding which seems to reflect the different implementation arrangements adopted there (see Box 3). The remaining subsections consider the structures, processes and systematic mechanisms that were put in place for data quality control. Box 4—Examining Data Quality in Phalombe This rapid assessment assessed the quality of data for the 10 World Bank financed districts plus Phalombe. The findings suggest that data quality has been consistently strong across all districts – with the exception of Phalombe. Data cleaning has taken longer for Phalombe, and hence data uploaded to the UBR dashboard has been delayed vis-à-vis other districts. While information on most households in the 10 districts (85 percent of households registered across 10 districts) was available for download using the API data sharing platform, only 12 percent of Phalombe’s data were available. While the causes of the delay are not certain, some evidence from the implementation plan for this district points to two possible causes: • Lack of formal oversight, monitoring and reporting (OMR) in Phalombe: OMR was implemented across all 10 districts - but not in Phalombe, primarily due to funding issues. Such formal mechanisms ensure that data enumerators are aware that their work will be quality-checked in their community or district and any mistakes will be discovered and potentially penalized. Without such mechanisms, this accountability relationship between the enumerators and the UBR team is broken. • Lack of community-level checks in Phalombe: Since Phalombe operated with 100 percent registration targets, the community selection method used in the other districts was deemed unnecessary and hence eliminated from its implementation cycle. Beyond validating the selection of the poorest 50 percent of households after PMT scores have been generated, community-level checks also provide an informal check on the quality of data and related procedures. For example, if data quality were bad in the other districts, the ranking generated by the PMT would not correctly reflect the relative ranking of households within the community, something that would come up in the community discussion. The finding highlights the importance of strong checks and balances through OMR systems, as well as by the community. Moreover, it provides important lessons for the UBR in its next phases of operation, when 100 percent of households are slated to be registered. 51 Some data transmission issues have been reported, primarily, due to the lack of synchronization of UBR and user program geographic mapping. A technical audit17 of the SCTP enrollment process in May/ June 2018 highlighted challenges in the smooth transfer of data from the UBR to SCT MIS. Besides the usual errors in geographic pre-mapping (e.g., zones of a specific cluster being classified as belonging to another cluster, clusters with no zones, etc.), a significant reason for this was the use of different zoning criteria on SCT and UBR. While SCT and UBR’s zoning criteria could be made the same, it is not possible to do that across all programs as they all vary by themselves from each other. What is necessary, therefore, is to have user programs synchronize their zoning criteria using that in UBR. This is not a major data error as such but important in ensuring accurate use of UBR data by user programs, hence highlighted here. Oversight, Monitoring and Reporting in the UBR OMR systems, an important determinant of data quality, appear to be well-embedded within the UBR. UBR’s OMR system builds on four levels of monitoring (Figure 14), in addition to community validation during the community meetings as part of the implementation processes.18 These monitoring levels are ad-hoc/ informal and not hierarchical with the UBR taskforce being responsible for reviewing findings from the monitoring visits and taking appropriate action to rectify issues. • NSO: The NSO plays an important role in oversight of the quality of data collection across the T/As in each District. An NSO member is assigned to each district for the first three months of the UBR’s rollout. This staff member is responsible for (a) providing oversight of the implementation process from community meetings to enrollment; (b) observing how each enumerator collects data during the first three days of data collection and providing feedback to DTT on the same; and (c) submitting monthly monitoring reports to the UBR office. • DTTs: The DTTs have the primary responsibility for field supervision. Two DTTs are allocated per six AECs and six CSSCs. DTTs undertake monitoring activities during the first 5 out of 8 days allocated for data collection in each community. This includes observing style and language used for asking questions, accuracy of data recorded and interview duration. • Central UBR Team: A combination of UBR taskforce members and broader social protection operations team from the central-level. As a monitoring team they also assist during the implementation process by going out in teams of three members for every two districts. Their mandate is similar to that of the NSO and they provide quarterly reports with their observations to UBR office. • UBR Taskforce: The UBR Taskforce itself also undertakes regular monitoring visits to different districts. They make observations of every step in the implementation including observations from the monthly and daily logs available in each district. Although these OMR structures are in place, there are no standardized checklist guidelines or reporting formats for each supervisory body. This limits the extent to which findings can be compared across various levels, teams within levels and time. Although the teams did try to incorporate findings 17 (IDEAL Consulting & Business Services, July 2018) 18 There is little documentation of the framework for OMR, as a result of which the analysis of the framework was based primarily on interviews with the UBR operations team and analysis of monitoring reports submitted during one monitoring visit. While this is a limitation of the assessment, the team has supplemented this with quality checks on the data collected to deductively verify the quality of OMR. 52 from different monitoring activities in Phase 1, limited assimilation is a lost opportunity to make midstream adjustments to the implementation process. This is even more important because ongoing data analytics are limited during implementation, which limits the ability of the UBR central and local teams to remedy errors along the way. Figure 14—Four Levels of Oversight, Monitoring, and Reporting in the UBR (current) Sources: Discussions with UBR Taskforce; various documents on the UBR (see bibliography) Standardized OMR protocols are needed to support the UBR expansion as it is rolled out, particularly in light of the increased registration targets in 14-plus districts. Standardized checklists and reporting formats will enable more effective OMR, which can provide crucial feedback to the UBR —and opportunities to remedy errors and improve processes—during the course of implementation. These reporting formats would cover various aspects of the processes, including: the quality of implementation processes (e.g., outreach, intake and registration, etc.); the quality of AECs, in terms of their interactions with the communities and ability to carry out the interviews; quality of enumeration, in terms of data accuracy and administration of questions; resources, in terms of time and money used for enumeration (frequency of enumeration and duration, as well as the cost per enumeration); and quality of OMR activities at the district, T/A and community levels. Specific roles for OMR need to be allocated to each actor (NSO, DTT, Central Team, UBR Taskforce) according to its purview, with an associated reporting system and protocols (e.g., frequency of reports, end-user of reports, and assimilation methodology). Additionally, a dashboard with basic data analysis and data quality checks would allow for midstream assessment of data. A dashboard with even basic analysis of data collected and stored in the UBR database would be useful for the team to keep track of progress and quality of data in each district. In the long term, the UBR Taskforce should consider exploiting the UBR’s potential to carry out more extensive analytics. The richness of data collected in the UBR, combined with the potential to merge it with other datasets, yields immense potential for data analytics on the UBR and on social policy more 53 broadly. To this end, a framework for analytics should be developed, covering the types of analytics that could be carried out, institutional responsibilities for analytics, frequency and cost of analytics, data sharing protocols and the need for anonymization of personal data. Systems Integrity Checks Systems integrity checks include quality-control processes that are built into the system to ensure that data errors are minimized. This assessment of systems integrity covered both validation and verification processes, and was undertaken through a combination of interviews with the IT lead on UBR and hands-on testing. The validation process ensures that data entered on the UBR are logical (no pregnant men) and do not have extreme values (such as unviable land holdings) or entries generated through negligence (for example, if data-entry personnel enter data into the actual UBR system while learning the system). The verification process ensures that the data have been verified with respect to other existing databases or have been independently verified through various systems internal and external audits and independent reviews. Data validation and verification processes are incorporated into all stages of the UBR. For data entry, this assessment tested the ODK toolkit, which has many error checks already programmed into the software. The name of the enumerator, date and time of registration, are also recorded, which is important for tracking and accountability. Adequate controls are in place for data capture on mobile tablets, which accounts for 97 percent of registrations. These controls also check for duplicates before data are uploaded into the UBR database. Post data entry, the UBR team has also incorporated a combination of manual and automated checks into the system to further detect and remedy any data duplications or inconsistencies. It has also undertaken various system-level audits and independent reviews, such as the assessment of the API interface between the UBR and the SCTP information system. The availability of audit trails and transaction logs is also available for monitoring and oversight. Verification with external data from other administrative systems is limited, however, due to the lack of other similar information systems. Nonetheless, some aspects of the UBR validation and verification system could be strengthened. First, it would be important to automate most error checks to reduce the chances for human errors and to ensure all checks are carried out in a systematic manner. While most error checks are automated for data entry, many are still manual post-data entry. Examples where further automation could help include additional checks for extreme values and detection of duplications, among other efforts. Second, the lack of a National ID for individuals in the UBR has hampered checks for duplications as well as external verification cross-checks with other administrative systems. Third, the lack of common national standards for coding of geographic units used by the UBR and the user programs hampers interoperability between those systems. Finally, the HT questionnaire needs to be translated into local languages, with both the English and local-language versions of each question uploaded into the ODK software. 7. Information Technology Aspects of the UBR The assessment considers three aspects of UBR’s information technology. First, systems architecture, management and oversight; second, ICT infrastructure; and third, data security and privacy. Strengths of the system include its development and maintenance by in-house resources, use of cloud 54 infrastructure which lowers cost and allows for scalability, effective data protection systems and quality checks related to data entry. These will be discussed in more detail below, along with areas that require further strengthening. Systems Architecture and Management Information technology has been leveraged effectively by the UBR to register households, manage their data, and exchange select information with user programs. The registration of households is mostly done using mobile tablets, while a web-based application interface is used to enter information collected on paper forms from a small minority of households. The UBR is a web-based information system with a back-end database to maintain household information. Exchange of information with other information systems is facilitated through web services. A key strength of the UBR information system is that it was largely developed and maintained by in- house resources with open-source software. The IT system was developed and maintained by the UBR Taskforce19 with some inputs from external IT support consultants. The team used Open Data Toolkit (ODK), an open source software, on mobile tablets to capture registration data in electronic form, minimizing data errors. This approach has been a strength for the UBR because it has leveraged and strengthened in-house capacity. Using ODK also avoids vendor lock-in and unnecessary licensing costs. The UBR system is based on a service-oriented architecture comprising four main sub-components. The system is web-based and hosted on a cloud platform that has sufficient processing power and storage capacity, as well as adequate versatility. The four main sub-components include: registration, data cleaning, UBR web application, and the data sharing interface with user programs (Figure 15). • Registration (Data Collection). The registration of households in the UBR system is done primarily through the mobile application on the tablets using ODK (97 percent of the registrants). The data captured in the tablets in the offline mode is then uploaded to the aggregate server (a component of ODK), which is hosted in the cloud. The 3 percent of the applications captured on paper are entered through registration forms available on the UBR web application to enter the details of the applicants. The geo-coordinates are being captured for the households, which would help in leveraging this information by the Geographic Information System (GIS) component. • Data Cleaning. Household data are cleaned using semi-automated processes to check for duplicates. Data from the aggregate server are downloaded as Excel files and compared with data already uploaded in the UBR to check whether (a) the record has already been added to the UBR database; (b) the household is a duplicate—registered under another ID but matching household head details; or (c) household members have been duplicated in the two systems under different households. This is a semi-automated task with workflow and activities managed by a UBR IT team member with help from a software utility program developed to aid in checking for duplicates. At times, this process has been a bottleneck when the number of data uploads is higher than the capacity to clean and upload them in database, as discussed below. • UBR web application. The UBR web application provides a menu of options to query the data and view reports using a system of user access control to various functionalities. The system has audit 19 As discussed in Section 4, the UBR Taskforce includes four IT professionals from different parts of government—National Library Service, Ministry of Gender, Local Development Fund and Community Savings and Investment Program (COMSIP)—and one IT specialist from GIZ. A new project management team is being put in place for the next phase as part of the institutionalization process. 55 logging, archival, reports and some basic analytics for effective monitoring and management of the system. • Data Sharing Interface for User Programs. Web Services or APIs are used to extract data from the UBR database by the user programs according to the Data Sharing Protocols and the specific program eligibility criteria. APIs are also envisaged for providing data updates from user programs to the UBR (e.g., demographic data updates on household composition). Existing System Information Systems Architecture Figure 15—UBR Systems Architecture UBR System Architecture 32 Source: UBR Taskforce Despite these strengths, human resource capacity to manage, maintain, and operate the UBR is inadequate. The UBR IT team is severely constrained in terms of capacity. It lacks a project manager to plan and monitor IT activities and risks, which may lead to an informal operational style, leaving potential for exposure to unwarranted risks. There is also a lack of human resource capacity for other critical roles and functions, such as service desk management, feedback and defects reports, standard operating procedures, risk monitoring, and so forth. Moreover, the segregation of duties is not specified in some operations, such as systems and database administration versus quality assurance and web security administration for oversight and controls. Finally, with human resources stretched, roles and responsibilities are not clearly assigned, which leads to “everybody doing everything as needed” which has worked out in the short term but is not a sustainable situation for the long run. It can also lead to errors of omission and potential security issues, for example, if everyone is given the formal authority to create user profiles and share data. Key recommendations would thus include augmenting the team with additional skills (including software development, database administration, project management, operations management, and business process analyst) and establishing dedicated staff for the UBR with defined processes, roles and responsibilities. Additionally, the architecture of the UBR does not yet allow for seamless data flow and quality control. Currently, the data collected through the mobile tablets are stored in an ODK server. Data are then downloaded and cleaned by the UBR IT team with the help of a software utility program, and then uploaded to the UBR database. This data cleaning process itself has multiple manual interventions, 56 which could be standardized and automated. For example, if the data upload is interrupted for any reasons (such as electricity failure or a computer crash), when the data are reloaded, the system doesn’t have automated checks for duplicates. Such checks are currently manual but should be automated. Also, the data back-up process from the cloud to local server is undertaken manually on a weekly basis. This lack of seamless data synchronization on the local server could lead to inadvertent errors, especially when the system scales up both horizontally, to additional districts, and vertically, to cover all households in the additional districts. Additionally, although the system was designed to use APIs for two-way data sharing with user programs, data feedback from the SCTP to the UBR has been carried out to date using “cron jobs” (batch matches) rather than the API. It is recommended that API be used for data sharing in both directions. ICT Infrastructure The UBR uses the cloud infrastructure, which brings lower cost, better scalability, and security. The UBR is hosted on a cloud platform in Germany for database and web servers, with a local server replicating data from the cloud to serve as back-up and for strategic control over data. This local server has sufficient capacity even for next UBR rollout phase. However, some of supporting IT infrastructure capacity is limited —notably the ODK server and internet bandwidth. The ODK server is experiencing issues, including a recent crash, due to the increased number of registered households, which brought the number of records to over three million. This three-million-record limit is a problem inherent in ODK, not specific to UBR. A solution would require additional ODK server space. Moreover, internet bandwidth is limited, particularly outside Lilongwe. While none of these server capacity issues have led to data losses, they may impede processing speeds and UBR rollout to additional districts. Finally, the local server is not currently protected with a firewall or any intrusion-detection mechanisms. Although these risks are contained, given that exposure of the local server is limited, it is recommended that the UBR introduce such intrusion-detection mechanisms to further strengthen security. Data Security and Privacy Information security is the preservation of confidentiality, integrity and availability of information. Information Security is achieved by implementing a suitable set of controls, including policies, processes, procedures, organization structures as well as software and hardware functions. The assessment of the UBR focused primarily on system and application security; data access controls; and data sharing, privacy, and protection. In terms of the system itself, the hosting of the UBR using a cloud-based platform brings security, but the local server needs protection. As discussed above, the local server is not currently protected with a firewall or mechanisms for detecting intrusions, which should be introduced to further enhance security. The UBR’s operationalized access controls are also in place, though structured assignation of roles and responsibilities—and segregation of duties—would further enhance data protection. The UBR has implemented appropriate identification, authentication, authorization and accountability. Every user is uniquely identified and authenticated by the system with password control, distinct levels of data access by user type, password locking on three unsuccessful attempts, and authorization 57 requirements for API-based data sharing to other departments. The UBR has a logging functionality that records all activities of the subjects and for anomalous behavior or activities. Moreover, the system maintains audit trails, documentation of technical specifications, and processes for regular back-up and archiving. Nonetheless, data protection and access controls would be enhanced with the formal structuring of staff, roles and responsibilities—and segregation of duties—as discussed above. These steps would ideally be taken to support the rollout of Phase 2 of the UBR. Data-sharing protocols have been finalized, but need to be put into practice. Currently, both data sharing and privacy issues are legally protected through the Access to Information Act 2016 and the Electronic Transactions and Cybersecurity Act 2016. Protocols for data sharing have been developed, but enforcement measures are yet to be designed. These include tailoring of the data sharing protocols to specific users (e.g., in terms of which types of users would have access to personal data rather than anonymized data); installation of the SSL certificate to ensure data security and privacy; measures to ensure that the data-sharing protocols are signed, followed, and monitored; staff sensitization on privacy and security; assigning clear responsibility to a single person or team to undertake all interactions with user programs (including user creation and approval, grievance redressal, etc.). The UBR’s Evolution within an Integrated Information Management Framework The UBR Taskforce has been discussing further evolution of the system within the context of an integrated information management framework (Figure 16). Figure 16—Future Vision for an Integrated Information Management Framework Source: UBR Taskforce and GIZ 58 While many discussions are at a preliminary stage, and thus beyond the purview of this assessment, current dialogue points to a number of potential enhancements that could be quite valuable, including: • UBR’s role as social registry serving multiple user programs. The vision for UBR’s future emphasizes its role as a social registry, whose function would be to support the processes of registration and determination of potential eligibility for multiple user programs. It also envisages an expansion in the number and type of user programs that would use information from the UBR. This is discussed in more detail in Section 9. • Integration with National ID and Civil Registry of the National Registration Bureau. The next phase of the UBR rollout in 14 districts with 100 percent coverage would capture the National ID of the household members who have been issued the ID (see Box 4). Integration with the Civil Registry would help keep the information in the UBR updated with respect to births and deaths, and hence ensure that deceased people do not receive benefits and that UBR’s household data is current. • GIS platform. Integration of the GIS platform with the UBR system would have many advantages, including enhancing the use of the UBR as part of a shock-responsive social information system for humanitarian aid. • E-Payment System for User Programs. Integration with the National Payment switch will enable user programs to make e-payments to beneficiaries, which would reduce transactions and administrative costs in the payments system. Eventually, that capture of households’ financial addresses or account numbers would enable benefit disbursement during disasters like drought, flood or other emergency response situations. Since the UBR is a social registry, it doesn’t have direct links to the payments system, so this is primarily an enhancement vis-à-vis the user programs, not the UBR itself. • Other modules. Other enhancements could include a monitoring and evaluation framework, analytics and reporting, and an enterprise service business, which would facilitate a service- oriented architecture, unified grievance and complaint management system, and modules to support integrated case management and the interface for outreach and extension agents. Box 5—Malawi’s National ID and the UBR A national ID establishes legal proof of identity to confirm that a person is who he or she claims to be. It is the foundational system for proving a person’s identity based on a unique set of attributes, such as biometrics, and provides a common identifier for each individual across various data platforms in accordance with national and international standards. A national ID can support many functions in a social protection system. First, it can validate the identity of individuals included in a social registry. Second, the national ID can de-duplicate beneficiaries who are recorded more than once (or eliminate ghost beneficiaries) in a social program beneficiary registry. Third, a national ID can support authentication of identity for payments and improve social service delivery through eHealth and ePayment systems. Fourth, it can link information for data exchange across various information systems, which can reduce the amount of information that people must provide at the point of service delivery (e.g., when registering in a social registry), support coordination across programs, facilitate integrated case management, and better monitor ‘last mile’ service delivery. Beyond social protection, proof of identity is increasingly required to open a bank account, secure a loan, and apply for a job. Malawi has recently (and rapidly) rolled out a new National ID, which is managed by the National Registration Bureau (NRB) under the Ministry of Home Affairs and Internal Security, in cooperation with United Nations Development Program (UNDP). The National Registration Act of 2010 mandates NRB to implement, coordinate, manage and maintain 59 the National Registration and Identification System (NRIS) in Malawi. NRB has the following deliverables: birth certificates; national identity cards for bona fide Malawians; identity cards for foreigners; and marriage and death certificates. Malawi has issued biometric national identity smart cards (NIDs) to 9.1 million people based on a mass registration exercise from May to November 2017. Malawians ages 16 and up have been issued NIDs, while younger citizens qualify for the National Birth Certificates, which will be linked to the NID during the rollout of biometric registration kits at health facilities in all 28 districts of the country. Continuous registration and distribution of NIDS is currently underway at each of the allocated National Registration Bureau offices across Malawi. With respect to the UBR, the National ID information will be captured in Phase 2 of the UBR during registration by scanning the 2D barcode on the ID card, which contains not only the unique ID number of the individual, but also user details like name, date of birth and district. This will greatly facilitate the UBR, saving time and avoiding data errors while capturing data. It will also help the UBR connect to other information systems, including user programs. Within the UBR, a functional household ID is available that links individuals to the household unit (which is the base unit for the UBR and many user programs).20 During Phase 1 of the UBR, the National ID was not used because it had yet to be rolled out on a nationwide basis. The UBR system has been assigning a functional UBR ID to individuals, and a household ID that links people to households. Those functional individual and household IDs are generated by the UBR system. Each user program also generates a functional beneficiary ID, which is mapped to the UBR ID. There are several options for linking existing registered household members to the new National ID: (a) the UBR could perform an automated lookup in the ID data base to find matches based on form number, name, date of birth and address/location details and fill in the ID number in UBR system; (b) the UBR could capture the ID number during interface with individuals by UBR team or user programs (as is already occurring with the PWP and is planned for the SCTP); or (c) the UBR could provide a mechanism for individuals to share their ID number with the UBR, via portal, mail, mobile SMS, mobile USSD or visit to NRB offices responsible for continuous registration. Sources: UBR Taskforce, National Registration Bureau, UNDP, d’Albore (December 2017) 8. User Programs and the UBR’s Potential as a Powerful Tool of Social Policy The assessment considers three aspects of the UBR’s linkages with user programs and its potential to serve as a broad policy tool: multi-program use of social registries; key ingredients of multi-program use of social registries; and updating and the potential use of the UBR for shock-responsive safety nets. The discussion also considers potential linkages with the National ID initiative. Multi-Program Use of Social Registries Social registries, such as Malawi’s UBR, can serve as a common gateway for coordinating registration and eligibility processes for multiple social programs. Integrated social registries combine the processes of outreach, intake and registration, and the assessment of needs and conditions to determine potential eligibility for those multiple programs. The agencies responsible for the social programs then make program enrollment decisions, taking into account the information on eligibility 20Within the UBR, each household is assigned a unique number during registration called a form number, which represents the household in the database and is used mainly for querying the database for specific households. Within the database itself, a unique auto incremental ID is assigned to each household, and each household member is assigned to this auto incremental ID within the database. Within the household members, each member is also assigned a unique ID number. The internal ID(s) allow for machine-to-machine data exchange, e.g. API transfers for enrollment. 60 plus other factors (such as budgetary space and other prioritization criteria), as illustrated in Figure 6 in Section 5 above. The use of social registries for multiple programs can offer many advantages for user programs, policy makers, and people.21 For user programs, integrated social registries can generate economies of scale, efficiencies, and savings on administration costs—which can be significant because the processes of registering and determining potential eligibility of individuals or households can be quite costly. Integrated social registries can also be used to support planning and estimating the costs of interventions, assessing potential demand, monitoring and evaluation, reporting, and other analytics. For households, common intake and registration procedures reduce the burden of having to navigate complex bureaucracies and provide similar information and documentation to multiple user programs. Integrated social registries serve as powerful platforms that support access to benefits and services that can extend well beyond the sphere of social assistance.22 In many countries, these user programs go well beyond social assistance, providing services such as social tariffs for electricity, subsidized health insurance, education and training vouchers or materials, emergency assistance, subsidized child care, financial inclusion services, and pro bono legal services (see Figure 17). As such, social registries can evolve to support many types of programs, going way beyond targeted programs that serve a subset of the population, and supporting universal services, such as health insurance subsidies or social energy tariffs. Countries are also increasingly looking to social registries, in combination with other geo- spatial information systems, to support shock response interventions. Figure 17—Social Registries as a Powerful Social Policy Tool for Social Assistance and Beyond All countries In sample Chile, Georgia Chile, Colombia, Georgia, Cash Transfers Mauritius, Mexico, Social Pension Philippines, Turkey Legal Aid & other Services Social Brazil, Chile, Colombia, allowances Georgia, Mauritius, Chile, China, Colombia, Registries Mexico, Pakistan, Health DR, Djibouti, Georgia, Philippines, Productive Mali, Pakistan, like the UBR Sierra Leone Inclusion Benefits Philippines, Turkey, Senegal can serve as powerful Chile, China, Social Chile, China, Colombia, Colombia, Labor & Integrated Services DR, Djibouti, Georgia, policy tools Philippines, Employment Social Mali, Pakistan, Philippines, Sierra Leone … well Registry Turkey, Senegal Emergency beyond Education & Assistance / Chile, China, DR, Djibouti, Georgia, Mexico, Training Shock-Aid social Pakistan, Philippines, Turkey, Sierra Leone Brazil, Chile, China, protection DR, Georgia, Energy & In-Kind Indonesia, Brazil, Chile, Georgia, Indonesia, Other Programs Philippines, Turkey Mexico, Pakistan, Philippines, Turkey Subsidies Housing Leite et. al. (2017) Brazil, Chile, DR, Benefits Brazil, Chile, China, Colombia, Djibouti, Mauritius Georgia, Indonesia, 54 Pakistan, Philippines, Turkey Macedonia, Turkey Source: Leite et. al. (2017) 21 Leite et. al. (2017). 22 Ibid. 61 Indeed, Malawi’s UBR was conceived as an integrated social registry that would serve as a common gateway for harmonizing the registration and eligibility processes for multiple social programs. As highlighted earlier, it was designed to serve multiple social programs so as to coordinate social policy, harmonize “targeting” (eligibility) concepts and criteria, and reduce costly duplications in registration processes (both for administrators and for citizens This expansion in users would be in line with multi- program use in other countries (Figure 18). Figure 18—Malawi’s UBR and other Multi-Use Integrated Social Registries Single or Multi Program Use of Social Registries: Approximate Number of Social Programs Using Social Registries, Circa 2015-18 90 80 80 70 70 60 52 50 40 30 31 28 30 24 22 19 20 16 17 8 9 9 10 6 4 4 5 1 1 1 1 1 1 1 2 2 2 2 2 2 2 3 3 3 - Burkina Faso Georgia TSA Registry Ghana Yemen SWF Registry Chad Turkey ISAS Gabon Congo, Rep Cabo Verde Benin Nigeria Pakistan NSER Cote d'Ivoire Malawi UBR (Current = SCTP + PWP) Tanzania Chile RSH Montenegro SWIS Mali RSU Indonesia UDB Mauritania Brazil Cadastro Unico Comoros Azerbaijan VEMTAS Lesotho Senegal RNU Philippines Listahanan China RPHR Djibouti RSU Mexico SIFODE Sierra Leone SPRINT Dominican Rep. SIUBEN Macedonia CBMIS Mauritius SRM Colombia SISBEN Malawi UBR (Current + Potential Users) Sources: Authors/UBR Taskforce for data on Malawi, Leite et. al. (2017) and Coudouel et. al. (2018) Key Ingredients for Multi-Program Use of Social Registries Several key ingredients require special attention for the effective design and operation of multi-use social registries. These include harmonized eligibility concepts and a shared intake questionnaire, institutional arrangements, capabilities for data exchange, and regular updating of information is needed by particular user programs. 62 The UBR’s design hinges on a set of harmonized eligibility concepts and a common intake questionnaire to collect information on Malawian households. To develop this questionnaire, a series of consultations were held with the various stakeholders, including user programs and their host agencies, in order to gauge their needs in terms of variables and information. The needs of the various user programs were then incorporated into a core Programme Requirements Document.23 On that basis, the HTT questionnaire was developed as the main instrument for collecting data in the UBR with the intention of providing core demographic and socio-economic information needs to the various user programs (see Box 5). Box 6—The HTT Questionnaire and the Proxy Means Test (PMT) Measure of Well-Being The UBR collects using the Harmonized Targeting Tool (HTT) questionnaire. In print form, the questionnaire is three pages long and has 48 variables. It was developed through a process of assessing the data needs for the main user programs (see Programme Requirements Document). The questionnaire comprises five blocks of information on households: (a) Geographic location (district, TA, GVH, village and geo-codes); (b) Household information (such as name, age, gender, disability status, chronic illnesses, education); (c) Household characteristics and assets (such as home ownership status, housing condition, water source, land ownership status and asset ownership); (d) Food reserves and food security (such as number of meals eaten in past week and source, assistance received, type and acreage of crops grown); and (e) Economic characteristics (such as livelihood source, savings and credit information). The HTT questionnaire was designed for multiple programs supported by the MNSSP (SCTP, PWP, school meals, VSLP, and microfinance), MVAC, and FISP. More recently, the Joint Emergency Food Aid Programme (JEFAP) targeting manual was used to identify additional variables that could be added to the HTT questionnaire to help with selection of MVAC beneficiaries. Similarly, as new programs start using the HTT, it may need to evolve to incorporate additional variables to respond to the needs of these programs. The UBR team would need to decide if these new variable requirements are met on a regular or case-by-case basis, depending on its usefulness to the range of user programs that use the UBR. Additionally, the choice of variables also needs to be made based on a thorough analysis of the determinants of poverty. Data generated by the UBR via the HTT questionnaire are very rich, with a wide range of variables that could be used for determining eligibility for social programs, planning and budgeting, monitoring, analytics and other purposes. One use of the data from the UBR is the calculation of a Proxy-Means Test (PMT) index of well-being. The PMT was developed by user programs to objectivize the selection of poor and vulnerable households, and has been included in the UBR to cater to such user programs. The PMT is used to create a proxy score of weighted variables that are highly correlated with household consumption (measure of well-being). By design, PMT is a measure of chronic poverty as it uses variables that do not change rapidly. The variables and weights for the PMT that is currently being applied to UBR data for households in Phase 1 districts were derived from statistical analysis of data in the (Integrated Household Survey 3 (IHS3) from 2010– 11. The PMT is currently being updated to reflect the characteristics of chronic poverty as observed in the latest household survey, the (IHS4), 2015–16. The updated PMT is expected to be used in the next few phases of UBR rollout. User programs can set their own eligibility criteria and thresholds on the basis of the HTT. The SCTP and PWP currently use the PMT to identify the bottom 10 percent and penultimate 15 percent of households in each district, respectively. This was done to ensure an objective selection criterion that could complement the community based targeting method. Other programs could use other thresholds and/or other combinations of variables from the UBR. Institutionally, the jurisdictional boundaries of UBR end where those of user programs begin, as discussed in Section 4. While UBR is endowed with the task of registering households and determining potential eligibility for user programs, selecting and enrolling beneficiary households is the task of user programs. This provides user programs with the independence to add other criteria that would qualify a subset of households deemed eligible by the UBR based on user program-specified criteria. In this way, the UBR is not overburdened to collect and store information that may be resource-heavy to collect and of use to only one program. Instead, the UBR can focus on a harmonized set of criteria that 23UBR Taskforce/Ministry of Finance, Economic Development and Planning (2015). “Programme Requirements Document: The UBR—A Common Database for Social Support Programmes.” 63 are useful to many different programs. For example, the school bursary program may wish to provide benefits to students on the basis of both merit and need. While the list of students in need would be identified through the UBR, the merit criterion could be added at the school level. For the UBR to collect information on merit would be costly, especially as it would likely only be used by the bursary program. Capabilities for data sharing are also central to the UBR’s relationship with user programs. As discussed in Sections 4 and 6, key elements for data exchange include mutual agreement on data sharing protocols, as well as web services and APIs. Information-sharing could potentially be bi- directional between the UBR and user programs. The most common direction is to share information from the UBR to the user programs, with the UBR providing data on demographics, the socio-economic situation, and poverty status of households. Yet user programs could also potentially feed information back to the UBR. Such data could include the list of beneficiaries of user programs but could also extend to other demographic details such as the birth or death of members. However, data quality and consistency need to be evaluated when considering feedback of information from user programs to the UBR. A challenge with the backward flow of information from user programs to UBR will be verification of such information. Moreover, there may be inconsistencies in methodology between user programs for collecting the same set of UBR variables. In addition, this backward information flow would be limited to beneficiaries of user programs. Presently, the UBR data transfer platform, API, allows for transfer of beneficiary status information from user programs to UBR. In defining any further backward information flows, the UBR team will need to carefully consider the challenges of data quality and completeness that will come with this process. Updating and Using the UBR for Shock-Responsive Safety Nets Regular updating of information in the UBR is important—both for households and for user programs. The demographic and socio-economic characteristics of households change over time. Some variables change more frequently than others. For example, changes in household composition and demographics occur frequently (with births, deaths, marriages, aging, and other life events within the household—or the formation of new households). Sources of 1ivelihood and food security status can change on a seasonal basis for a large share of households. Illness and disability status can also change frequently (or at least idiosyncratically). Other variables may change less frequently, such as asset ownership, educational status, and chronic poverty status. The requirements of user programs may also differ in terms of the frequency of updates. For example, the SCTP updates its beneficiary enrollment every four years, whereas MVAC updates its information every year (albeit in specific districts and geographies, as also explained below). Hence, to meet MVAC’s needs, the UBR may need to be updated more frequently in specific districts. The current UBR updating policy is for Government to update the UBR every four years, making the UBR a rich data source upon with which the Population and Housing census can establish strong linkages, especially if the current UBR 100% coverage policy extends to both urban and rural areas. However, in future when the system becomes more dynamic, households should be able to update their information anytime (including when they move). However, Malawi has yet to incorporate such updating or re-registration mechanisms in the UBR. A possible improvement to the UBR would be the prioritization of updates in shock-prone areas. Data collection is costly and time-consuming, and frequent updating via the census-sweep approach may not be possible across the entire country (after the initial rollout across districts). Some countries have 64 moved toward dynamic updating using on-demand registration and interoperability (e.g., Brazil’s Cadastro Unico and Chile’s social household registry, among others).24 However, dynamic on-demand social registry systems require several key ingredients that may be elusive, such as fiscal and administrative capacity. Thus, many countries resort to full census sweeps every three to five years. Yet that frequency may not be regular enough for programs such as MVAC, which deliver humanitarian assistance in shock prone areas. As such, the UBR could adopt an innovative strategy to use geospatial maps to prioritize which micro-areas (districts or T/As) for more frequent (perhaps annual) updating based on their sensitivity to shocks. Two key ingredients would facilitate the updating and use of the UBR for MVAC and other types of shock-response assistance: (a) the geotagging of households in the UBR (which is already underway); and (b) the 100 percent registration target, which would provide visibility to a greater share of households in each area, so that they could be considered for humanitarian assistance in the aftermath of shocks. Additionally, information on the coverage of different social support programs by households give a better sense of whether a household affected by a shock has any form of external coping mechanism. 9. Strategic and Operational Communications for the UBR The assessment considers three aspects of UBR’s communications: First, its strengths and challenges; second, the role of communication strategy; and third, whether UBR’s branding helps it meet expectations. The increased visibility and importance of the UBR necessitates a solid strategy for strategic and operational communication. This is particularly important considering continued expansion and the upcoming Phase 2 rollout of the UBR, as well as various misperceptions about it that were identified during the assessment. Strengths and Challenges of the UBR’s Communications Within the current setup, there are strong elements upon which to build improved communication for the UBR, including the ongoing work on a broader MNSSP communication strategy. These include: • The Government’s push to develop a broader MNSSP communication strategy, which aims to facilitate a harmonized approach to communication in the social support sector and comes at an opportune moment to provide impetus for UBR’s own commun ication strategy. • The direct working relationship between key government stakeholders and UBR structures. The intense involvement of stakeholders in the UBR Taskforce is a key strength as it builds their ownership and understanding of the UBR, and allows for direct communication. These include: the Poverty Reduction and Social Protection Division of MoFEPD, the Ministry of Gender, Children, Disability and Social Welfare (MoGCDSW), the Local Development Fund (LDF) and other key Government agencies. • The UBR’s use of existing decentralized structures at the district and local levels also enhances trust, understanding, and operational communication channels. These include the DTTs, AECs, and CSSCs. • With financing (either of the UBR or user programs) from many development partners, it is also evident that there is goodwill and interest in understanding and supporting the objectives, 24 See Leite et. al. (2017) for a more in-depth discussion of dynamic social registries. 65 functioning and current and potential uses of UBR, which is crucial to having successful strategic communication for the UBR. • A distinct logo and tagline that distinguishes the UBR from user programs (albeit with some concerns about branding, which are discussed below). Nonetheless, communications challenges could hamper the effectiveness of the UBR, both strategically and operationally. These include: (a) lack of understanding of the UBR’s objectives, purpose, and functioning—both at the community level and among other stakeholders; (b) confusion about the role of the UBR vis-à-vis the user programs, particularly the SCTP; and (c) misperceptions about the quality of data in the UBR, which detract from its credibility. In the absence of proper communications, understanding of the UBR, its objectives, functions, and uses will be limited. This has led to fragmented and conflicting information, which has at times created unwarranted pessimism about the UBR. Communities are not entirely clear about what the UBR is and does. This was evident during the field visit to a community in the District of Dowa for this assessment (carried out in March 2018) and in the findings of a qualitative study by King and Tranchini (May 2017). The name “UBR” is not well known at the community level: “No one had heard of the term UBR and . . . [after translation, the study team] referred to it as ‘the government’s list’ of the poorest people.” Most participants remembered the UBR data collection process but there was some confusion with other data collection efforts and participation by the community and chiefs.25 The UBR is also perceived as a “Government initiative” produced by a “higher authority” or a “computer.” This view was reinforced by AECs, who referred to the UBR as “the machine” during discussions with the field -visit community.26 AECs and CSSCs explained to members of the community that it was a “machine” that selected beneficiaries for the SCTP.27 During a community meeting to announce and enroll beneficiaries in the SCTP, villagers also expressed concerns about families who were excluded from the SCTP given the 10 percent enrollment limits for that program. Some showed UBR receipts that indicated that they had been registered, with the misunderstanding that registration for the UBR would guarantee them for benefits in the SCTP. Further discussions revealed that some of those who were complaining about exclusion from the SCTP were actually beneficiaries of the PWP (which made sense since they were clearly not labor- constrained). Others expressed dismay that they had not been registered in the UBR at all, even though decisions regarding which households met the UBR’s registration target rest with the communities in the first community meeting. The AECs and CSSCs need to be trained in standardized responses to these types of queries and complaints, and given a deep understanding of the operating procedures for both the UBR and the user programs. Misunderstandings also extend to other stakeholders, particularly a blurring of the lines between the UBR and the user programs. Continuous misuse of terminology by various stakeholders (including development partners) further perpetuates these conceptual and functional misunderstandings. Various development partners regularly use vocabulary such as “beneficiary households in the UBR” or “enrolling UBR beneficiaries.” Not all households in the UBR are beneficiaries, and the UBR doesn’t “enroll” or “pay” them. Rather, the UBR registers households (which may or may not become 25 King and Tranchini do note that their study took place in early pilot districts and that a communication strategy has since been developed for the rollout of the UBR in subsequent districts. 26 Field visit, March 2018 27 Field visit, District of Dowa, March 2018 66 beneficiaries of social programs) by collecting their identifying, demographic, and socio-economic information to determine their potential eligibility for social programs. Those user programs then make their own decisions to enroll beneficiaries based on information from the UBR (and other factors). Furthermore, various stakeholders discuss (digital) payments from the UBR, even though payments are the responsibility of the user programs, not the UBR. Finally, various stakeholders refer to the UBR as a “program” or a “project.” It is neither. It is not a program, in the sense of a user program like the SCTP or PWP. And it is not a project, in the sense that it is not a temporary one-off activity. Rather, it is a social policy tool with specific functions that can support social protection interventions and other uses. That confusion has been even stronger for the UBR vis-à-vis the SCTP, as the primary initial user of the UBR. In the absence of a UBR communication strategy, communication hinges strongly on SCTP structures. As a result, current UBR communication has been driven more by personnel implementing the SCTP. This raises the potential risk that UBR communication gets mixed up with those for SCTP, thereby, losing the much-needed clarity and creating confusion on communication between the UBR and user programs. The risk of mixed messages spans both high level strategic communication and operational communication. Lack of knowledge and understanding has perpetuated misconceptions regarding the quality and validity of the UBR. Rumors and misperceptions regarding the UBR abound (see Box 6 for examples we heard during the assessment). Such misunderstandings are common with social registries, particularly when they are relatively new or undergoing changes in design or implementation. Sometimes, the rumors have a basis in reality, perhaps reflecting the experience of one of the districts (such as the Phalombe pilot), errors that occurred at one time but have since been resolved, or errors in some subset of the data, to name a few. However, as those rumors spread, they may be generalized to the UBR as a whole. In reality, all social registries (and all information systems) will always have some errors. The goal is to have systems in place to minimize, detect, and remedy the errors —and then the communication strategy to report on them and respond to queries in an accurate, consistent, and transparent manner. Some countries, such as the Philippines, have also developed “talking points” to respond to common misperceptions about their social registries, including data quality concerns. Box 7–The Importance of Managing Perceptions: Example Myths & Rumors About the UBR Rumor Response “We heard that the • Not so. Quality tests (including those conducted for this assessment) suggest high quality of the data in levels of data quality (see Section 6). the UBR is bad.” • The UBR is also working to further automate processes for cross-checking and detection of errors (see Sections 6 and 7). • When pressed to explain this concern, the Government official stated, “Well they’re bad quality because a lot of households were left out.” That allegation then doesn’t actually refer to the quality of data collected, but rather to the process that communities used to determine which households would be included in the 50% registration target. “We heard that power As with many rumors, there are some elements of truth in these concerns, but the issues outages and server have been resolved (see Sections 6 and 7). crashes have led to duplications in entries.” • Power outages are a common phenomenon across many sectors in Malawi, and beyond the control of the UBR. • The UBR also experienced difficulties with the server’s ODK software when the number of records exceeded 3 million (this is a challenge inherent to ODK) 67 • In some instances, when the data are reloaded after an interruption (such as for a power outage or server glitch), the system doesn’t yet have automated checks for duplicates. To date, these checks have been done manually. • In general, the findings of our assessment suggest that the quality of data in the UBR is good (see Section 6), including in tests for consistency and duplications in data. “We heard that there is • This occurred at an early stage of the SCTP enrollment, and harmonized mapping has a mismatch between the since been done within the Application Program Interface (API). geographic mapping in • A mismatch doesn’t imply that the UBR’s data quality is poor, but rather that the UBR the UBR and that of the and various user programs need to harmonize the geographic classification of districts, user programs, such as T/As, etc. the SCPT.” • A hotline or front office would help the UBR respond such issues. “We heard that the • In most districts to date, the UBR has registered around 50 percent of households—a UBR failed to find all larger percentage than any single user program has registered in the past. This makes of the households in the sense since the UBR will serve multiple user programs with different target SCTP.” populations. • Certainly, district-specific poverty profiles vary in relation to the uniform SCTP targeting threshold of 10 percent. As a result, strictly running the SCTP eligibility criteria on UBR registrants is bound to yield varying results, with the possibility of finding more or less than 10 percent of households in different districts. However, a policy decision has been made to continue with the 10 percent uniform threshold. Future policy decisions might revolve around whether SCTP targeting thresholds need to be redefined. • The UBR could reach agreement with user programs to make explicit efforts to locate precisely the same set of households already included in the programs (such as collecting data on all SCTP households, in addition to the 50 percent prioritized for registration by the communities). • However, there is no guarantee that any updated registration effort would locate precisely the same set of households that had been registered four years prior for the SCTP (or other user programs), since the demographics and socio-economic situations of all households (both beneficiary and non-beneficiary) would have changed in the interim years, and new households would have formed that may face worse conditions than those of households currently in the SCTP. “It takes way too long • These concerns are common among development partners and user programs. Social to implement the registries take time to develop and implement, particularly when countries are rolling UBR.” them out for the first time. • Based on international experience, the speed at which the UBR has been designed and developed from scratch is admirable. The UBR concept was developed in October 2015 based on the initial pilot, and registration of 11 districts was carried out in 2017 (Phalombe pilot + Phase 1 districts), reaching over 50 percent of households in those districts. That represents 21 percent of the national population. The UBR Taskforce is now refining and preparing for the Phase 2 and 3 rollouts, and is projected to reach about 40 percent of Malawi’s total population by the end of 2018. This is in line with the time it has taken to pilot and roll out social registries in other countries. • This concern may be partly related to implementation planning and communication. A key recommendation of this assessment is for the UBR team to undertake solid implementation planning with time and resource targets for each district, T/A and community—and then to monitor and report on progress regularly to the various stakeholders (see Sections 5 and 6). Another recommendation is to strengthen strategic and operational communication with stakeholders (see Section 9). Inviting the development partners to see the process on the ground could also be a way to improve their understanding of what it takes to implement the social registry. “Data will get outdated • This concern is valid, and countries around the world have adopted different very soon and will mechanisms for updating data and opening registration. Some countries use on- cease to be useful for demand methods, with registration open permanently such that anyone can register or targeting purposes.” update their information any time (and with maximum periods for such updates). 68 Other countries use the census-sweep approach, with infrequent updates (usually every 3 to 5 years). The concern with the census sweep is that households that are not registered, or whose situation changes, would be left out —particularly if the social registry is used to assess eligibility for many social programs. • The issue of updating is crucial—and the UBR team will have to address this soon, particularly for the districts from the pilots and Phase 1 (and because of the increased registration target in later phases of the UBR). • However, the counterfactual scenario of each program collecting its own socio- economic information in a systematic manner does not produce more regular updates (the SCPT waits four years for updates), and the updating needs of programs vary (for example, the MVAC does require annual updates). • One option could be to prioritize shock-prone areas for more frequent updates because the socio-economic status of households in those areas may change more frequently because doing so would serve the MVAC and other humanitarian assistance programs. “We heard that once • This view may be held by households being registered on the UBR. The use of the our information is in word ‘beneficiary’ in the name could be one potential reason for this confusion but the UBR, we will get another is the selection of the poorest 50 percent of households, particularly true about benefits.” in phase I. • The UBR is meant to only collect and store information on households and is not responsible for selecting beneficiaries. Selection of beneficiaries and provision of benefits is the responsibility of user programs, and they work independently of UBR. Toward a Communication Strategy for the UBR The UBR’s communication challenges reflect the broader communications landscape for social protection in Malawi. The current situation reflects the gaps in communication about social protection programming in Malawi. The UBR does not have clear guidelines for communicating strategic or operational aspects of the tool. However, even without formal communication guidelines, UBR implementation structures at both the national and district level have made an effort to disseminate strategic and operational information about the UBR. Dedicated resources should be allocated to develop and implement a communication strategy for the UBR. The Government has already recognized the need for communications and is starting to work on a communication strategy for the MNSSP and social protection more broadly. Within that context, efforts should be made to consolidate resources to develop a communication strategy specifically for the UBR. The strategy should capture both strategic communication (functions, objectives and uses of the UBR), as well as operational communication (such as implementation processes, messaging at the local level, and communication between actors). Some attention should also be given to updating the UBR’s branding. The key objectives of strategic and operational communication would include: (a) improving awareness and understanding of what the UBR is and does (objectives, purpose, functions); (b) improving understanding of how the UBR works (processes, functioning); (c) reducing confusion about the relative role of the UBR vis-à-vis the user programs, particularly the SCTP; (d) boosting credibility of the UBR with regards to information quality and validity as an “honest broker” for information on the demographic and socio-economic status of households; (e) ensuring two-way flows of information between the UBR and various stakeholders for improved transparency and responsiveness to queries; (f) boosting support for the UBR for financing, use, and policy coordination; and (g) standardizing messaging around the UBR for credibility, transparency and consistency. 69 That strategy should be developed for the full range of stakeholders. In developing a communication plan for the UBR, key strategic stakeholders to be considered include Government ministries, agencies and officials, policy makers, and user program coordinators, as well as implementers at both central and local government levels. Development partners are another key strategic partner providing financial and technical support to SP and UBR. The public, civil society and media form the backbone of the information conduit for facilitating public dialogue and social awareness, and provides an accountability mechanism for enhancing implementation. User programs play a unique role, cutting across aspects of both operational and strategic communication. As end users of the UBR’s information, they are crucial in shaping operational aspects of the UBR and therefore the related operational messaging and communication. Additionally, user programs remain a crucial stakeholder in continuously shaping the functioning and use of the UBR. On the operational side, beneficiary communities and implementers—at both central and local government levels—are key operational stakeholders that need to be at the core of a UBR communication strategy. Examples of the types of communications issues that may arise for these stakeholders are included in Box 7. A communication strategy needs to address the specific perspectives of each stakeholder. Building on the key stakeholders identified above, it is important that for each stakeholder, the communication strategy emphasize the following: (a) target audience; (b) purpose and objectives; (c) key tailored messages; and (d) communication channels and messengers. It is imperative that each stakeholder is viewed as a messenger. Indeed, there are many “spokespeople” for the UBR, including politicians and government officials, the UBR Taskforce, district officials, user programs, AECs, CSSCs, and the communities themselves. The strategy needs to recognize that consistent and accurate information must be communicated by each of these actors. And the actors need to realize their role in speaking on behalf of the UBR. For that, efforts are needed to boost their understanding, including outreach, sensitization, talking points and FAQs (including “talking points” on key issues), regular bulletins, training, focal points, and hotlines to respond to queries or complaints by stakeholders. Messaging needs to evolve as the UBR does and within a changing context. One issue that needs to be clearly communicated with stakeholders is the shift to 100 percent registration targets for Phases 2, 3, and 4. This shift has implications for messaging about policy, such as the broader potential use of the UBR for social protection and other uses. It also has operational implications that clearly need to be communicated to all actors for training, sensitization, and implementation processes (see Section 5). Another implication is the potential to use the UBR as a potential tool to facilitate humanitarian assistance and disaster management. Box 8—Common Communications Issues for Social Registries Government What is the strategic role of the Social Registry in broader social policy? What are its objectives, agencies and uses, and functions? How does it relate to the user programs? What resources are needed to officials (policy implement it? guidance, resource mobilization) Implementing How does the Social Registry work? What’s my role in implementing it? How can I get answers to Agencies (central, operational questions? How to respond to questions about the Social Registry (e.g., from citizens)? local) What’s the relation of the social registry to user programs? 70 User Programs How do we access data from the Social Registry? What are the confidentiality and data sharing protocols? Who do I talk to in the Social Registry Agency if I have an operational or technical question? Citizens, What is the Social Registry and what does it mean for me? What benefits and services could I Communities potentially receive via the Social Registry? Am I guaranteed to get benefits if I register? Where do I go to register? Who do I talk to if I have a question about the Social Registry? When will I be informed if I get approved for benefits and/or services? What if I disagree with the decision? (Queries, grievances) Development What is the strategic role of the Social Registry in broader social policy? What are its objectives, Partners (e.g., uses, and functions? How does it relate to the user programs? What resources are needed to donors funding the implement it? SR or user programs) General public, What is the Social Registry? What’s it for? How is it different from the flagship social program? media How can the media be informed to create interest and awareness of the UBR? Branding the UBR and Managing Expectations As the UBR expands, the Government may want to weigh the options for enhancing the UBR brand. As discussed, elsewhere, the very name of the UBR—the Unified Beneficiary Registry—is a misnomer since (a) the households included in the UBR are not all beneficiaries and have no guarantee that they will become beneficiaries of the user programs; and (b) the UBR’s primary function is that of a social registry, that is, to support the processes of registration and the determination of potential eligibility for social programs. There are pros and cons to changing the UBR’s name. The main advantage of changing the name relates to the challenge of managing expectations. The concern with the name is not just a technical point—it is also a political one. The name itself can inappropriately raise households’ expectations by implying that just by registering their information, they would be guaranteed some form of benefits and services. The need to manage such expectations will become even more important as the UBR shifts to the 100 percent registration target. Moreover, the name itself could be perpetuating much of the confusion among stakeholders that was discussed previously. Nonetheless, there are also drawbacks to changing the name, notably that such a change would conflict with the desire for continuity and speed for the Phase 2 rollout, as well as the potential to create temporary confusion. These advantages and disadvantages should be weighed against each other. The UBR already has a distinct logo and tagline (See Figure 19). The logo includes an icon for people (households) as well as a map of Malawi to signal the UBR’s (eventual) nationwide coverage. The tagline is “Capture. Share. Coordinate. Harmonize.” One concern about the tagline is with the word “capture,” which has alternative interpretations (such as capturing people rather than capturing data). Another concern is the word “share,” which could raise concerns about data privacy and security. As Malawi moves toward a rebranding, it may want to test these potential concerns and/or an alternate tagline on the various stakeholders. 71 Various factors should be considered in rebranding the UBR. The new name—and any associated logo and taglines—should be carefully considered and tested with various stakeholders. Ideally, the name, logo and tagline would convey the UBR’s core philosophy and values (such as inclusion, accuracy, and fairness). They should also convey the UBR’s function as a social registry (not as a beneficiary registry), and not over-promise. Ideally, the name and tagline would use locally understood languages. The name, logo, and tagline would also be clear, simple, recognizable—and easy to remember. Finally, they should allow the social registry to be easily distinguished from the user programs. Box 7 below presents some examples of how other countries have developed and market-tested names, logos, and taglines to convey the core values and functions of social registries. Figure 19—The Logo and Tagline of the UBR Box 9—On the Branding of Social Registries Brazil’s Cadastro Unico and the Philippines’ Listahanan Integrated social registries can be powerful social policy tools, as they provide a gateway for households to apply and register for numerous benefits and services in social assistance and other uses. For social registries to be effective social policy tools, however, they need to be recognized and understood. Countries are increasingly making explicit efforts to develop distinct branding efforts to foster that understanding. That branding typically involves: • Adopting a recognizable name that will be remembered and understood. • Developing a logo and tagline to communicate the core functions and values of the social registry. • Promoting “product recognition” by using the name, logo, and tagline on all materials related to the social registry, such as questionnaires, stickers, brochures, banners, uniforms for enumerators, and carrying bags. In Brazil, for example, it was important to distinguish between the logo of the Cadastro Unico (integrated social registry, top image at right) and the logo of the original flagship program, Bolsa Familia (bottom image at right). Moreover, the Cadastro Unico adopted a clear tagline “Conhecer para incluir” which means “Know them to include them.” This tagline is a core philosophy of the Cadastro Unico, which has represented the primary way for many poor families to register their information in a formal system that would allow them to access social benefits and services. The basic concept is one of visibility for the poor and their situations: “If we don’t know them, how can we include them?” In the Philippines, a Communication Assessment showed significant confusion about the name and role of its social registry, originally called the “National Household Targeting System for Poverty Reduction” (NHTS-PR). Even staff couldn’t remember the name or acronym (let alone the general public or other stakeholders), and many people thought it was only related to the Pantawid Conditional Cash Transfer program, even though the registry serves over 50 programs. Therefore, in advance of the next round of registration, the Department of Social Work and Development (DSWD) developed a rebranding, marketing and communication strategy for the social registry in order to inform people, improve understanding, and promote unity and consistency in messaging nationwide. One aspect of that strategy was to hold discussions in a branding workshop around the values that should characterize the branding of the social registry. Those values were identified as: nationwide coverage; champion of 72 the poor; balance of “authority” and “compassion;” accurate and reliable: scientific, internationally accepted methodology (referring to the PMT and data collection); non-political (must not be partisan); and that the taglines should not over-promise—meaning that the social registry is a tool for social policy, not a program or set of benefits. The next step was to develop options for the name, tagline and logo for the social registry. Twelve options were developed, and four were then market tested (see options 1–4 below) to see if they would be recognizable and understood. Based on feedback from the market testing, DSWD opted for a version of Option 1. That logo and tagline now appear on all materials, including T-shirts for staff and banners for the Listahanan social registry. Option 2 Option 3 Option 1 • Visuals: checkmark = • Visuals: House Option 4 • Visuals: checkmark = symbol of symbol of accuracy; sun shaped by caring • Visuals: arrow = symbol of accuracy; house instead of a = hope, uplifting; colors hands; red = a feeling targeting tick-box of national flag color and from palette • Tagline: conveys “what it • Tagline: Correct number, • Tagline: Correct of DSWD is” (NHTS) with simpler correct support; emphasis on number; correct support; • Tagline: Caring acronym than NHTS-PR; accuracy and response, not over- Tamapinas = correct, message on purpose conveys ultimate goal: “aim promising; simple fair Philippines; registry “Giving a name and to reduce poverty” of the poor face to poverty” Sources: MDS, DSWD (May 2013), Lindert et. al., (2007) 10. Projected Cost of implementing UBR (Forward-Looking for Phase 2) A critical concern of the UBR has centered on the understanding overall costs. Gathering an accurate cost breakdown was a challenge. This challenge is not unique to Malawi’s UBR. Estimating and comparing the costs of social registries is not straightforward for several reasons.28 First, there are many different types of administrative costs, both at the local “front end” (intake and registration, data collection) and in the “back office” (information systems management, coordination), and covering, both physical and human capital cost relating to staff salaries, method and frequency of data collection and diversity in institutional arrangements. Second, costs of social registries are spread out over time, varying significantly, because countries rarely design, build and operate these systems from scratch with a single investment. In the case of Malawi’s UBR, these challenges were compounded by additional factors. First, many parties were involved in funding the costs of the UBR, including numerous donors and government agencies – and some costs were provided in kind (such as tablets, staff time, etc.). As such, there isn’t a central accounting system keeping track of all such costs. Second, the modality for carrying out the UBR varied across districts registered to date, with some districts registering 100% (such as Phalombe) and others registering 50% of households (Phase 1, with communities prioritizing which would be registered). As a result of these challenges, we focused instead on gathering and presenting an illustrative set of cost projections for rolling out the UBR across the seven districts in Phase 2 (since trying to piece 28 Leite et.al. (2017). 73 together costs of past registration waves was not possible). It is estimated that rolling out the UBR in one district during the second phase will cost approximately $1.74 per households (Figure 20). The total cost of rolling out the UBR across 785,000 households in seven districts in Phase 2 is projected to total approximately $1.27 million, or $182k per district. Although international comparisons are fraught with caveats, for the reasons stated above, this projection is comparable to estimates in other countries (Box 10). It is slightly above the cost of developing Turkey’s social registry at $1.3 per household, and lower than the registry costs per household in Brazil (costing between $2.03-2.06 after the initial merging of registries) and Colombia ($2.25), as discussed in Box 10. While this projection for Malawi is the estimated cost of registering 100 percent households in the districts in the next phase, the cost of registering 50 percent of households in the previous phase would have been lower but not half. This is because some time saved in terms of having to only register half the households would be offset, at least partly, by the additional time required to search and locate the poorest 50 percent of households. Figure 20 – Projected Costs of Implementing the UBR in Phase 2 (per household) Source: Authors estimates with data provided by the UBR management unit and other counterparts. Note: Cost of the UBR is budgeted costs associated with roll-out in 7 districts in Phase 2, i.e., 100 percent household coverage; Technical support & capacity development, fixed assets and cost of UBR office rent, etc. is based on roll-out in 10 districts in Phase 1. Supervision costs represent the highest projected expenses associated with the UBR (Figure 20). Supervision costs in Phase 2 are associated with allowances for personnel and fuel charges to undertake oversight and monitoring function with the UBR. It is not surprising that supervision would be relatively expensive, as this is a crucial function in ensuring good quality of data. About three-quarters of these costs are associated with the supervision to be undertaken by the National Training Team, and the remaining with supervision activities to be carried out by the NSO. Training is the second highest cost associated with UBR. This includes training of the National Taskforce members and National Training Team, Training of IT team & MISOs, Training of District Training Team and finally Training of Areas Executive Committees (AECs) and Enumerators (includes recruitment of the 74 latter). This again suggests an appropriate allocation of resources as training is important to ensure proper implementation and data quality. Three quarters of these costs are associated with training of AECs and enumerators. Community meetings, a crucial component of the data collection process, is the third highest cost associated with UBR. This includes projected costs incurred during both the first and second community meetings. Given that Phase 2 of the UBR would register 100% of all households in each village, implementation of the first community meeting is expected to be less costly than in Phase 1 (since in the past, communities were expected to discuss and prioritize 50% of households to be registered). It is also expected to be less costly than the second community meeting, which involves ensuring that all households in the community are registered, as well as discussing and validating the ranking of the poorest households (similar to the first community meeting in Phase 1). About 86 percent of these projected community meetings costs would be incurred in the second community meeting. Equipment costs and cost of capacity development and technical support are the last big-ticket items in the budget. Equipment costs represent about 16 percent of overall projected outlays for Phase 2 of the UBR. This includes the cost of tablets, printers, cloud servers, internet, paper, etc. Some of these are fixed costs in that they can be reused over multiple phases. Technical support (from external consultants) and capacity development, on the other hand, cost about 14 percent of overall cost of UBR. This is an important aspect of ensuring that staff have the right skills and technical knowledge to implement the UBR. Other costs include rent; resources for communication and advocacy; data entry support; and costs involved in pre-mapping and validation of the harmonized tool. These costs are about 10 percent of total projected expenses. These though are important procedures as it ensures political buy-in (through advocacy) and future usability to the UBR data across various user programs (pre-mapping and harmonized tool). Items not included in this projected budget include the salaries of government functionaries involved in this work, as well as cost of hiring short term consultancies to advise UBR on specific technical areas. These salaries are already paid for by the government and is therefore not included in the overall cost of the UBR. Moreover, besides the UBR management unit, government functionaries involved in the UBR are also involved in other activities within their respective departments in government. The short-term consultancies include technical advice on IT systems, API based data transfer systems, etc. Box 10 – Examples of Cost Estimates for Implementing Social Registries in Other Countries Turkey:29 Even though Turkey started from a base of extensive administrative information systems and program information systems, the development of the integrated information system represented a substantial investment. The total estimated cost of ISAS development was US$13.1 million, an amount that is considered reasonable compared to other countries that have developed similar systems. Since the system currently covers about 40 million people, or 10 million households, this represents a system development cost of US$1.3 per household in the period. Turkey was able to reduce development costs by contracting TUBITAK, a public agency, t o develop the system “in house” and provide 29 See Ministry of Family and Social Policy (April 2017, and 2017). It is important to note that ISAS is a full integrated social protection information system, including modules for the Social Registry functions to support intake, registration, and determination of eligibility, as well as beneficiary management and payments administration. 75 ongoing maintenance. A breakdown of the costs by type includes: US$5.3 million for hardware (computers, servers, security systems, and system rooms), and US$7.8 million for analysis, technical design, and software development. The contract with TUBITAK included ongoing maintenance through 2015, and the IT department of the Ministry of Family and Social Policy provides continued daily maintenance. The cost efficiencies that were generated by ISAS outweigh the costs of developing and operating the system. ISAS has the capability to identify these cost efficiencies, for example: identifying and eliminating 10 percent of assistance benefits that were duplicated, reducing paper costs to the tune of processing 2.3 million fewer documents per month, and reducing processing time. With respect to the latter, for the time needed to process applications from registration to enrollment decisions was reduced by 20 percent. Moreover, it is estimated that the system generates a savings of one million full-time equivalent person days per year. Finally, ISAS estimates overall resource saving of $39 million per year – far higher than the $13.1 million invested to develop the system. Colombia: The set-up cost of SISBEN was estimated at around USD 2.2530 per family registered in 1995, of which 73% was related to household data collection. In 2001-2002 SISBEN administrators defined the new strategy for SISBEN, the SISBEN II, when a new questionnaire and procedures we defined to be applied nationwide from January 2003. SISBEN II continued using an en masse approach in selected poor areas identified by the cities and municipalities, to be followed by application on-demand for those not included in the initial survey of 1995. Over the SISBEN II cycle for registration, 2003-2006, the estimated cost per family registered was USD 2.3 for a total of 8 million families. Then, SISBEN III registration cycle cost about USD 2.52 per family, including data collection and front/back office investment for improving interoperability. As technology evolved some data collection costs are declining due to use of electronic instruments, the SISBEN IV cycle cost for updating and registering new families, more than 10 million families, dropped to USD 1.27 per family. Brazil: During the phase of the consolidation of four programs31, 2003-2005, into the Bolsa Familia program that formed the initial largest base for the Cadastro Único (Cadúnico) the estimated cost per family in the CadÚnico was USD 0.53 per household, because it mainly consisted of merging pre-reform registries (data already existing for many households) and registering additional households using a simple form and system. As the Cadúnico matured32, becoming the gateway for benefiting from “low income families” social policies during the period 2006-2009, it required more human and physical capital investment increasing the cost per families to USD 2.03. Between 2010-2013 Cadúnico version 7 introduced online synchronization with the federal center and other systems as pensions systems, increasing the cost per family to USD 2.06 due to the physical infrastructure needed. Source: Leite et. al. (2017) 11. Conclusions and Recommendations The UBR is a potentially ground-breaking initiative allowing households to register and be considered for inclusion in social programs based on a transparent assessment of their needs and conditions. To date the UBR has registered and collected data for over 800,000 households (or 4 million people) in 13 districts where it has been rolled out. As the UBR gears up for expansion there are considerable strengths to build on and areas for improvement (see Table 2 for a summary). The UBR has many strong fundamentals. Some of these include government ownership of UBR (both process and IT systems), the use of existing decentralized institutional structures, good existing relationships with various stakeholders across ministries and donor partners, functional 30 See Castaneada (2005) 31 The Bolsa Familia program resulted from the consolidation of Bolsa Escola, Bolsa Alimentação, Cartão Alimentação and Vale- Gás. 32 See Baddini Curralero (2016) 76 implementation processes, effective information systems, and most importantly, robust data quality and rapidly expanding coverage. These fundamentals need to be maintained and strengthened as the UBR considers future phases of expansion. Nonetheless, there are some areas that need further strengthening. Some actions are more urgent for the short run (to support the continued expansion in Phases 2 and beyond), while others are longer- term measures. Some are already being taken into consideration by the UBR team, but they are still mentioned here for comprehensiveness (Table 2). Institutionalization of the UBR In the short term, UBR needs to enhance and institutionalize its central management team. As discussed in Section 4, a new UBR Management Team is being put in place to institutionalize the day- to-day operations of the UBR. This needs to be carried out, with explicit efforts to clarify institutional reporting lines and staff roles and responsibilities, ensure knowledge transfer, and provide continued guidance by the UBR Taskforce (which would shift to a more advisory role as the UBR Management Team builds capacity). Additionally, there is a need to enhance this team’s capacity by adding specialists for field operations and planning, business analysts, database managers and IT systems developers. In the short term it makes sense to have the UBR Management overlap with the UBR taskforce in the short to medium term, with continued support from MASAF IV until the transition is completed and the necessary human resource capacities have been added. Thereafter Government needs to seriously consider moving the management unit into the Ministry of Finance, Economic Planning and Development under the Poverty Reduction and Social Protection Division. This would require that PRSP Division already starts the process of requesting an organizational functional review that should allow the incorporation of UBR unit within the Ministry’s organogram. In the longer term, it will be necessary to establish a formal legal standing for the UBR, including a legal home that gives it jurisdictional independence and authority over its processes and functioning. A formal legal standing would help formalize the UBR and clarify its objectives, purpose and uses. It would also provide the UBR with jurisdictional and operational independence from user programs. This is important to maintain the integrity and security of households’ information. The social protection law is a potential platform for a legal framework for the UBR, although, given the evolving nature of social protection in Malawi, this framework should allow for some flexibility in terms of legal instruments used (laws, decrees, etc.). The UBR also needs to establish a coordinated and sustainable financing strategy. This includes the short-term action of introducing a budget line in the 2018–2019 or fiscal year 2018 national budget, as part of the government’s formal commitment to provide ongoing support to the UBR, recognizing its potential not just for the social protection sector but more broadly for other sectors as well. Additionally, there is a need to establish a formal longer-term financing strategy that includes external financial support such as from donor partners and their coordination. Financing modalities need to consider the immediate costs of data collection, as well as the longer term costs of keeping data current e.g. via on demand registration. 77 Implementation Processes and Oversight, Monitoring and Reporting (OMR) Another top-priority action involves clarifying the role of the UBR central team in implementation— and implementation planning. The UBR central team (Taskforce or Management Team) does not appear on the organizational chart for data collection—and is relatively absent in terms of its roles in the Guideline. Yet its roles in coordinating, planning, and monitoring implementation across districts are crucial. Similarly, the UBR team needs to invest in developing end-to-end implementation plans (E2E-IPs) for Phase 2 and beyond. These E2E-IPs are needed for each district (and across districts) from start to finish, building on the productivity indicators, and identifying the time and resources needed for each step and each locality. E2E-IP is a common management tool that takes the UBR’s process to the next level, mapping more specific details for the specific steps for all processes in sequence, including resources used for each step and actor. E2E-IP brings together all of the process steps, actors, schedules, and resources needed to implement one or many programs in order to monitor and control all of these elements under a single plan. In the short term, there is a need to revise implementation guidelines as per the new 100 percent registration target. Based on the decision to register 100 percent of households in the new districts, it needs to update the Implementation Guideline and associated manuals, in particular revising and assessing the quality as well as the processes that guide the role of CBT and the community meetings. Other priority short-term actions to strengthen implementation for Phase 2 and later would improve the quality of data collection and oversight. These include: (a) strengthening outreach and sensitization of communities; (b) adjusting the model for registration (interviews and data collection) by the AECs, with the option of appointing and training more specialized teams of rotating AECs that would cover a larger number of villages within each T/A; (c) standardizing guidelines and checklists for OMR and enhancing supervision (including spot checkers for random re-reviews); and (d) formalizing steps for handling appeals and grievances (including in the Implementation Guideline). Standardized reporting and analytics would also help in the longer term. An important aspect of a robust system is the flow of information across horizontal and vertical levels of UBR implementation channels. To enable this flow, the UBR management team needs to develop clear OMR guidelines (process and reporting) across all existing levels of monitoring and develop a dashboard that analyzes household information stored in the UBR to provide real-time updates on data quality. Information Systems In the short term, there is a need to automate processes and define roles and responsibilities. This includes seamless automation and integration of all key processes with minimal manual interventions; ICT infrastructure enhancements to ensure that disaster recovery and business continuity platforms are available and can guarantee provision of a robust and resilient UBR IT environment; and real-time data replication by the local back-up facility. At the same time, roles and responsibilities need to be clearly defined—for example, the roles of database administrator and web administrator need to be segregated. This would ensure accountability within the team. 78 In the longer term, the UBR would fit into a broader information framework for social protection, with a common data exchange platform to facilitate links to user programs, the National ID, geo- referenced information systems, and so forth. a. Interface with User Programs In the short term, the UBR should enforce data sharing protocols and set up a user-program–facing service desk or hotline. As discussed earlier, the UBR team has developed a data sharing protocol to guide all future data related interactions with user programs. It is important as a next step to start enforcing it. As part of this, it is important to establish a separate user-program–facing service desk that deals with all queries from user programs and monitors use of information. This is particularly important during the initial stages of information transfer because issues related to integration of data may arise from differences in data structure (for example, a user program may use zones instead of villages) between the UBR and user programs’ platforms. In the longer term, it is important for the UBR management team to assess the frequency and scope of updates. Information in the UBR is a snapshot of households at the time the UBR data collection exercise was undertaken. However, there is substantial seasonality in consumption levels in Malawi, with significant changes in living conditions both within a given year and across years. Moreover, climate shocks may change estimates of needs substantially whenever they occur. If programs are to respond to these changing needs, the UBR itself needs to plan the frequency and scope of updates. Currently, humanitarian aid is the only program that responds to rapidly changing needs. Similarly, the UBR needs to determine when, what and how information may flow from user programs to the UBR. While updating may make it easier for the UBR to stay on top of household information (albeit only for beneficiaries), it is worth considering how quality control measures could be applied to this process. Strategic and Operational Communications In the short term, further diagnostics could identify priority actions that could strengthen operational and strategic communications for the next stage of UBR rollout. Clear communication is part of the process of gaining confidence and setting expectations among stakeholders, including not only user programs but also households that are being registered in the UBR. The UBR team should consider conducting a communication needs assessment to better understand the communications needs of various stakeholders, and to prioritize areas that are most important in the next stage of UBR rollout. For example, clear communication regarding the selection procedures at the village level may stem some of the political pressure to bias selection but also highlight efforts from the UBR to stem such incidents and consequently, safeguard the integrity of the process. In the longer term, develop a clear communication strategy and consider rebranding. In the longer term, the UBR needs to have a coherent communication strategy that ensures clear communication with all stakeholders as well as training on that strategy for the UBR team. The key objectives of strategic and operational communication would include: (a) improving awareness and understanding of what the UBR is and does (objectives, purpose, functions); (b) improving understanding of how the UBR works (processes, functioning); (c) reducing confusion about the UBR’s role vis-à-vis the user programs, particularly the SCTP; (d) boosting the UBR’s credibility with regard to information quality and its validity as an “honest broker” for information on the demographic and socio -economic status 79 of households; (e) ensuring two-way flows of information between the UBR and various stakeholders for improved transparency and responsiveness to queries; (f) boosting support for the UBR’s financing, use, and policy coordination; and (g) standardizing messaging about the UBR’s credibility, transparency and consistency. Furthermore, it would be worth considering a rebranding exercise that clearly communicates to all stakeholders the mission and objectives of the UBR. This rebranding could involve changing the UBR’s name and updating its logo and tagline. The term “beneficiary” in the UBR name creates unfulfilled expectations and causes conceptual confusion. Moreover, the registry is often misunderstood and not well-known by communities and other stakeholders. One alternative that emerged in initial discussions was the term granary, which was one way that communities expressed their interpretation of the UBR—that is, as a granary to store information on households. Ideally, market testing would be used to try out various name/logo/tagline combinations to see which would be best understood by various stakeholders. 80 Table 2: Key Recommendations: What to Keep and What to Change in the Short and Medium-to-Long Term What to Keep What to Change What to Change Key Strengths Short Term Improvements Medium-Long Term Institutional • Government-led, • Institutionalize central management team (assign roles • Provide legal foundation to UBR— Structures ownership and responsibilities by function) including finalization to provide it a clear mandate, • Operationalized of agency that would permanently house the UBR protecting its jurisdictional exclusively using • Enhance HR capacity with additional specialists for independence and delineating its existing decentralized field operations & planning, business analyst, DB role from that of user programs and institutional structures manager, IT systems developers user program administration • Allows for community • Establish jurisdictional and role independence from user engagement and program programs consultation, including • Develop a sustainable financing strategy, particularly, local leaders. through introduction of a new budget line in the 2018– 19 national budget Implementation • Well-defined • Finalize agreement and approach for coverage targets • Develop clear M&E guidelines Processes implementation steps in (vs quotas) (process & reporting) across all Phase 1 • Revised implementation guidelines and update training existing levels of monitoring • Four different levels of manuals • Develop a dashboard that analyses monitoring and • Develop detailed implementation plans and timelines for collected data for reporting to M&E oversight (NSO, each district and overall for Phases 2 and 3 and oversight teams District Trainers, • Introduce additional M&E procedures such as spot- • UBR data analysis to inform future Central M&E team and checks, enumerator quality control, etc. and increase program design UBR taskforce) frequency of checks • Strong level of • Establish clear roles, responsibilities and protocols for community engagement engaging with development financing partners. Data Quality • Good quality data on • Automation of data quality checks • Develop a dashboard that provides UBR data • Commissioning of solid analysis and future assessment team live updates on data quality. • Strong oversight, building on program data This is even more important for the monitoring and • Finalize data sharing protocols, also for academic next phase when 100 percent of reporting; and system communityIntroduce real time quality checks, and households will be registered in the integrity checks qualitative reviews and spot checks with community, UBR as the volume of data being including re-interviewing random sample households processed would be much higher and regular monitoring & quality check of the data would ensure timely availability of good quality data for user programs 81 Information • System independence • Ensure seamless automation and integration of all key • Improve communication and Systems (open source, in-house processes with minimal manual interventions interaction between UBR and development and • Enhance ICT infrastructure to ensure disaster recovery other stakeholder systems such as maintenance) and business continuity platforms are available and can SP programs, govt. shared systems • Modular systems guarantee provision of a robust and resilient UBR IT and payments platforms, architecture environment potentially through a common • Cloud based hosting of • Define roles and responsibilities clearly to ensure Data Exchange Platform database system integrity, appropriate levels of access, info • Beyond UBR: Introduce common • Data sharing protocols security country geographic identification in place • Ensure real-time data replication from cloud to local coding systems, with common • Good quality control at backup facility geographic codes (or GPS data entry stage • Develop a domain name instead of IP address coordinates) that should be • Support mechanisms for integrated information common across the UBR, management e.g. links to national ID programs, the NSO, mapping agencies, post office, etc. This would greatly help with the quality of data and data sharing across UBR and user programs. User Programs • API links between UBR • Enforce application of data sharing protocols • Define frequency and scope of and SCTP have been differentiated by type of user updates (national versus specific established • Enable seamless integration of UBR and user programs districts) based on program needs • Large number of • Set up a service desk to deal with all user program • Define the need for and scope of variables collected from requests in a consistent manner within the legal and backward information flows from each household, and operational framework the user programs beyond available upon request information on whether a household received a particular benefit Communications • Strong existing • Communicate jurisdictional and role independence • Develop clear communication relationships between from user programs strategy catered to different different stakeholders • Carry out a communication needs assessment stakeholders (community, donors, & UBR structures, and • Identify and take action on priority communications ministries, etc.) goodwill that are crucial to Phases 2 and beyond. • Consider rebranding the UBR • Govt. pushing to develop a broader MNSSP communication strategy 82 References D’Albore, Antonio (December 2017). “Malawi Issues 9 Million national IDs in six months.” Embedded Security News. Ayala Consulting Corporation (November 2017). “Proposal for Retargeting, Recertification, and Exit of SCTP Beneficiaries: German and EU-Funded Districts.” Prepared for Ministry of Gender, Children, Disability and Social Welfare. Barca, Valentina (July 2017). “Integrating Data and Information Management for Social Protection: Social Registries and Integrated Beneficiary Registries.” Australian Government/DFAT and Oxford Policy Management. Castaneda, Tarcisio and Kathy Lindert, with Benedicte de la Briere, Luisa Fernandez, Celia Hubert, Osvaldo Larranaga, Monica Orozco, and Roxana Viquez. “Designing and Implementing Household Targeting Systems: Lessons from Latin America and the United States.” World Bank Social Protection Discussion Paper Series No. 0526. Chimutu, Louis, Magombo Galunzanga, Jiri Nemec. (November 2016). “Quality Assurance of Implementation of UBR.” Chirwa, Elija (19 December 2017). “Quality Assurance of the Application Programme Interface (API) between the Unified Benefiicary Registry (UBR) and Social Cash Transfer (SCT) Management Information System (MIS).” Chitekwe, Benard (July 2018). “Implementation Review for Unified Beneficiary Registry and Social Cash Transfer Program in Rumphi District- July 2018 SCTP Planning and Budgeting Session.” PPT. Ministry of Gender, Children, Disability and Social Welfare, Social Welfare Department. DSWD, National Household Targeting Office of the Philippines (May 2013). NHTO Branding Workshop. (Powerpoint and notes file) Faruq T, K Ahmad and Richard Chirchir (2015). “Assessment of Malawi’s Proposed Unified Beneficiary Registry and Information Management Systems for Social Protection in Malawi.” UNICEF, Irish Aid, ILO, Development Pathways. Government of the Republic of Malawi, Ministry of Finance, Economic Planning and Development. Malawi National Social Support Programme (MNSSP). Unified Beneficiary Registry (UBR). Harmonized Data Collection Tool: Household Questionnaire. Henninger, Jakob (November 2017). “Malawi: Social Protection Programme (SPP).” German Cooperation GIZ. 83 Henninger, Jacob and McNeil Mhango (June 2017). “One for All—Creating a Unified Register for Social Protection Programmes in Malawi.” Rural 21, GIZ, UBR Taskforce. Holmes, Rebecca and Cecilia Costella et. al., (2017). “Towards a Shock-Sensitive Social Protection System for Malawi.” London: ODI and The Hague: Red Cross Climate Centre. Kardan, Andrew. (January 2016). “Streamlining Targeting Mechanisms and Processes Across National Social Protection Programs: Developing a Concept.” Oxford Policy Management (OPM). IDEAL Consulting & Business Services (July 2018). “Malawi Social Action Fund IV: Strengthening Safety Net Systems in Malawi”. Technical Audit report submitted to Local Development Fund King, Diana (World Food Programme-Malawi) and Lodovica Tranchini (United Purpose-Malawi) (Final Report May 2017). “UBR MVAC Trial: A study on the feasibility of using a Social Registry for targeting a humanitarian response in Malawi” Leite, Phillippe, Tina George, Chanqing Sun, Polly Jones, and Kathy Lindert. (2017). Social Registries for Social Assistance and Beyond: Guidance Note and Assessment Tool. The World Bank Group. Social Protection and Labor Discussion Paper 1704. Limbe, Vivian (July 2018). “Implementation Review for Unified Beneficiary Registry and Social Cash Transfer Program in Lilongwe District-July 2018 SCTP Planning and Budgeting Session.” PPT. Ministry of Gender, Children, Disability and Social Welfare, Social Welfare Department. Lindert, Kathy (September 2015). “Communication and Social Registries: With Examples from the Philippines and Brazil.” PPT adapted from social safety nets core course session, the World Bank. Lindert, Kathy (March 2017). “Communications and Outreach in Social Safety Nets.” PPT from social safety nets core course session, the World Bank. Lindert, Kathy, Anja Linder, Jason Hobbs, and Benedicte de la Briere (2007). “The Nuts and Bolts of Brazil’s Bolsa Familia Program: Implementing Conditional Cash Transfers in a Decentralized Context.” World Bank Social Protection and Labor Discussion Paper No. 0709. Lindert, Kathy and Vanina Vincensini (December 2010). “Social Policy, Perceptions and the Press: An Analysis of the Media’s Treatment of Conditional Cash Transfers in Brazil.” World Bank Social Protection and Labor Discussion Paper No. 1008. Luhanga, Suzgo Judith (February 2018). “Institutional Arrangements for Managing and Promoting the Use of Social Registries in Malawi.” PPT. Ministry of Finance, Economic Planning and Development, Poverty Reduction and Social Protection Division. Luhanga, Suzgo Judith (February 2018). “Linking Complementary Services to Safety Net Beneficiaries: The Malawi Experience.” PPT. Ministry of Finance, Economic Planning and Development, Poverty Reduction and Social Protection Division. 84 Luhanga, Suzgo Judith (February 2018). “Main Lessons on Complementarities for Kenya.” PPT. Ministry of Finance, Economic Planning and Development, Poverty Reduction and Social Protection Division. Luhanga, Suzgo Judith (February 2018). “Malawi’s Social Cash Transfer Programme (SCTP): Linkages and Referrals’ Component.” PPT. Ministry of Finance, Economic Planning and Development, Poverty Reduction and Social Protection Division. Malawi Government, Ministry of Finance, Economic Planning and Development. Malawi National Social Support Program. “Training Plan Concept” and Training Templates.” Various training docs for UBR. Malawi Government, Ministry of Finance, Economic Planning and Development —Poverty Reduction and Social Protection, UBR website. “About the UBR in Malawi.” http://www.ubr.mnssp.org Malawi Government, Ministry of Finance, Economic Planning and Development (October 2015). “The Malawi Unified Beneficiary Registry (UBR): Concept and Proposal.” Malawi Government, Ministry of Finance, Economic Planning and Development—Directorate of Poverty Reduction and Social Protection. UBR Taskforce (2015). “Programme Requirements Document: The Unified Beneficiary Registry (UBR)—A common database for social support programmes.” Malawi Government, Ministry of Finance, Economic Planning and Development—Directorate of Poverty Reduction and Social Protection (December 2016). “Review of Malawi National Social Support Programme: A Stakeholder-Driven Review of the Design and Implementation of the Malawi National Social Support Programme (2012–2016). GIZ, UNICEF, ILO. Malawi Government, Ministry of Finance, Economic Planning and Development—Malawi Natinoal Social Support Program (MNSSP). “Harmonized Data Collection Tool.” HTT UBR Questionnaire. Malawi National Social Support Program. Geographic Pre-Mapping Tool. Form to fill out. Malawi National Social Support Programme website. http://mnssp.org Mhango, Mc Neil, Whitney Kabango, Anthony Waliala. (November 2016). “Open Data Kit Quick Installation User Guide.” PPT National Social Support Services (NSSS) (January 2015). Targeting Manual for Social Support Programmes. Republic of Malawi. Němec, Jiří (March 2017). “Supporting the quality assurance team during the development of the Unified Beneficiary Registry (UBR) for Malawi.” Republic of Malawi (November 2016). “Guidelines for Unified Beneficiary Registration Process.” Republic of Malawi (July 2016). “Interview Guidelines for Harmonized Data Collection Tool.” 85 UBR Core Team (February 2018). “Unified Beneficiary Registry—Operations.” PPT on implementation processes. UBR Core Team (February 2018). “Insights into the Malawi Unified Beneficiary Registry.” PPT on UBR systems. UBR Core Team (November 2017). “The Unified Beneficiary Registry Data Sharing Protocols.” UBR Core Team (May 2017). “Compliance Monitoring: Blantryre and Chiradzulu Report.” Internal document. UBR Core Team (June 2017). “Minutes of the UBR Implementation Progress Review Meeting, Held on 2nd June 2017 at the UBR Office at City Centre.” Internal document. UBR Core Team (2017). “Compliance Monitoring of UBR Data Collection Process.” Consolidation of issues from compliance monitoring reports. Internal document. UBR Core Team (2016). “Technical and Functional Specifications of UBR Information System.” Internal document. UBR Core Team (20XX). “Infrastructure Requirements—Core UBR APPLICATION.” Technical document. UBR Core Team (20XX). “UBR API Specification.” Technical document. UBR Core Team (20XX). “Mobile DATA capture Simulation Test Report.” Technical document. UBR Core Team (20XX). “UBR Malawi User Acceptance Test Checklist (Web Application)—for online and offline.” Technical document. UBR Core Team (20XX). “Integration Test Case Report.” Technical document. UBR Core Team (20XX). “Web Capture Test Case Report.” Technical document. UBR Core Team (20XX). “UBR Geographic Pre-Mapping Tool.” Spreadsheet form. UNICEF—Innocenti Research Brief (2018). “Malawi’s Social Cash Transfer Programme: A Comprehensive Summary of Impacts.” Vikan, Selvi, Rebecca Holmes, Edward Archibald, Caoimhe de Barra, Harry Mwamlima, and Daniela Cuellar (February 2018). “Managing Disaster Differently: Shock-Sensitive Social Protection in Malawi.” Webinar with socialprotection.org, organized by GIZ, Malawi Government, World Bank Group, and WFP. The World Bank (2017). “Malawi: Overview of ID System.” Informal note. 86 _________ (20XX). Malawi Unified Beneficiary Registry: Supporting the Development and Implementation of the Malawi Unified Beneficiary Registry for Social Protection Programs in Malawi: GIS Inputs.” Document on geographic information systems. The World Bank (February 2018). “Social Registry Assessment: Malawi’s UBR.” Initial Terms of Reference for this task. 87 Appendix 1: Social Registries’ Conceptual Framework and Assessment Tool33 Conceptual Framework and Terminology Social Registries are information systems that support outreach, intake and registration, and the assessment of needs and conditions to help determine potential eligibility for one or more social programs (benefits and services). They are both inclusion systems and information systems. As inclusion systems, the primary function of social registries is to provide a “gateway” for people (individuals, families) to register and be considered for potential inclusion in one or more social programs based on an assessment of their needs and conditions. That assessment usually takes into account measures of socio-economic status, categorical factors or a combination of both, which are often factors used by programs in prioritizing eligibility for benefits and services. More specifically, from a functional perspective, social registries support the implementation phases of outreach, intake and registration, and assessment of needs and conditions to determine potential eligibility for inclusion in selected social program(s) (see the blue segments of Figure A1.1). Many countries use “integrated social registries” to serve as gateways for registration and eligibility for multiple programs—not just one. This can have the advantages of reducing burden on citizens who don’t have to apply for numerous benefits and services separately, reducing administrative costs and boosting efficiency for “user programs,” and improving coordination of social policy. Indeed, integrated social registries can serve as a powerful platform that extends well beyond social assistance programs. Many countries use integrated social registries to support determination of potential eligibility for a range of other interventions, some targeted, some universal in nature. Examples include subsidized health insurance, social energy tariffs, education and training vouchers, subsidized child care, financial inclusion services, pro bono legal services, and so forth. Some social registries support dynamic inclusion, meaning that access to registration is open and continuous, usually via a combination of on-demand applications plus active outreach to vulnerable populations. In countries where social programs are relatively new, coverage is small, fiscal space is constrained, and administrative capacity is limited, registration and updating are carried out less frequently, usually with significant time lapses between “census sweeps.” The risk of this more static approach is rising errors of exclusion and inclusion with the passage of time, as the information becomes out of date. Operationally, Social Registries are information systems that support the flows of information on individuals and households and their socio-economic conditions to determine potential eligibility for social programs. As information systems, their basic architecture includes data intake and exchange, software applications to support front-office and back-office functions, database management and interoperability (in some cases), and ICT infrastructure. Social Registries don’t operate in isolation and are usually part of broader information systems supporting social programs, including other complementary functions and components, such as: 33 Based on Leite et. al., (2017). 88 beneficiary registries and administration systems, payments administration, and case management systems. Importantly, Social Registries are distinct from beneficiary registries in their purposes, population coverage, and functions. Social Registries include information on all registered households (not just beneficiaries of social programs) and support the “gateway” functions of intake, registration and determination of eligibility. In contrast, Beneficiary Registries include information only on those enrolled in specific programs to support beneficiary and benefits administration (see the red segments of Figure A1.1.) As such, the name for Malawi’s social registry as “Unified Beneficiary Registry” (UBR) is a bit of a misnomer, since: (a) In terms of objectives, the UBR’s primary functions are to capture, store, access, retrieve, and share data on households’ needs and conditions (socio-economic profiles, consistent with the blue segments of Figure 1). These functions can inform the determination of eligibility for specific programs, but the programs themselves have the mandate for taking eligibility and enrollment decisions (red sections of Figure 1) and then managing program operations such as payments and monitoring (purple segments of Figure 1). (b) In terms of the population covered, the UBR collects and maintains information on all registered households—not just those who are selected as beneficiaries of specific programs. As such, a more apt name for Malawi’s UBR would have been Unified Social Registry (USR) rather than Unified Beneficiary Registry. Households whose data are included in the UBR should not be referred to as “UBR Beneficiaries” since the UBR does not grant benefits and not all h ouseholds will end up as beneficiaries of social programs. This paper sticks with the acronym “Malawi’s UBR” but recognizes and assesses the system as a Social Registry given its functions and population coverage. Overview & Approach to the UBR Assessment Conceptual Framework Figure A1.1—Information Systems Support Various Implementation Phases Along the Delivery Chain Information systems support these functions all along the delivery chain for Social Programs Assess Provide Potential Demand Enroll Benefits / Services 1 2 3 4 5 6 7 8 Monitor UBR = Social Registry (information on all applicants, registrants) Beneficiary Registries & Beneficiary Operations Management Systems (information on beneficiaries) “Intended” Potentially Applicants Eligible Beneficiaries Population Applicants 5 Source: Leite et. al., (2017) 89 Social Registry Assessment Overview and Methodology Within this conceptual framework, the UBR assessment was adapted from the assessment tool elaborated in a recent World Bank Discussion Paper “Social Registries for Social Assistance and Beyond: Guidance Note and Assessment Tool” by Leite et.al., (2017), see Figure A1.2 below. The framework lays out five basic parts for the assessment of social registries: • Part 1: Characterizing and situating the main features of the Social Registry • Part 2: Structural Features, including Institutional and Legal Arrangements, Citizen Interface, and User Programs • Part 3: Implementation Processes: Outreach, Intake and Registration, Determination of Eligibility, Enrollment Decisions (by user programs), Updating, Grievance Redress and so forth • Part 4: Information Systems Aspects, including: data and information, software applications, database management, interoperability, ICT infrastructure, system strategy, information security and privacy • Part 5: Measuring Performance: inclusion (coverage), efficiency (including costs), data quality and accuracy, and transparency and accountability. Parts 1 and 2 will be carried out largely on the basis of a “light desk review” so as to allow a more detailed focus on Parts 3 (processes), 4 (information systems), and 5 (measures of performance). Social Registries Assessment Tool Figure A1.2—Social Registries Assessment Tool (for adaptation to Malawi UBR context) Characterizing & Situa ng the Main Features of the Social Registry (status quo & future) Part 1 Social Policy Roles – as Inclusion System Informa on Systems Aspects: 1. Single or Mul -Program Use? 3. Sources of Informa on? 2. Dynamic Inclusion of Popula on? 4. Degree of Interoperability? 5. Database Management Structure? Mostly Light “Desk Review” Structural Features Part 2 Ins tu onal & Legal Ci zen Interface User Programs Arrangements Delivery Processes Part 3 Outreach Intake & Registra on Upda ng Processes Grievance Redress Informa on Systems Aspects Part 4 Info Security, Data & So ware Database Inter- ICT System Privacy, More in-depth review Informa on Applica ons Management operability Infrastructure Strategy Confiden ality Measuring the Performance of Social Registries Part 5 Accuracy (Data Systems for Inclusion (Coverage & Efficiency (for ci zens, Transparency & Quality & Systems Measuring Equity, Access) administrator, user programs) Accountability Integrity) Performance Source: Leite et. al., (2017) Building on this framework, the team then elaborated a detailed Terms of Reference that adapted the assessment tool to the initial desk review plus on-going dialogue via earlier missions and VCs with counterpart team in Malawi (see TORs, World Bank (February 2018). Subsequently, after receiving further comments from clients and development partners, and conducting additional desk review of available literature, the team continued to adapt and tailor the assessment to Malawi’s context (“Malawinisation” of the task). The resulting framework is summarized the priority topics to be 90 explored during the mission with the matrix in Figure A1.3 below. During the March mission, the topics were further refined in dialogue with the UBR teams and development partners, including the enhanced focus on the 100 percent registration target and strategic communications. The final set of topics used in the assessment is discussed in the main text and forms the outline for this report. Figure A1.3—Revised Set of Priority Topics for Malawi UBR Assessment Overview & Approach to UBR Assessment Priority Topics Topic Areas Assessing the Current Experience “As-Is” Looking Forward to the “To-Be” (ST & LT) Current Institutional Arrangements Longer-Term Sustainability • Institutional Arrangements: Central, Local (implementation roles, citizen • ST: what are the planned institutional arrangements for the Expansion? Structural / interface), relationship to user programs o For UBR management? • Administrative aspects o For data collection? Institutional • Legal foundations • Institutionalization for (a) more “permanent” management structure & staffing for UBR (b) more permanent UBR registration mechanism • Legal framework, financing Planned vs Actual Implementation Processes Implementation Adjustments Going Forward • Core processes & roles (swim lanes): preparation, training, sensitization & • Updating of Implementation Manual and planning cycles Implementation & outreach, I&R, assessment, supervision & oversight • Detailed implementation plans for expansion • Implementation guidelines, training , cycles • Strategic communications, branding and outreach Efficiency • Communications, branding and outreach Efficiency measures (performance) Efficiency & costing going forward • Productivity • With updated cycles, productivity • Time, resources • Costing of the expansion Information Systems Information Systems Priorities Going Forward Information • Data privacy & security; data sharing protocols with users • System allows for flexibility in defining eligibility criteria • Systems architecture and data flows • Bidirectional data updates via API Systems • Database management • Scalability of the system – assessment, cost, resources • ICT infrastructure • Integration with ID system • Information systems institutional aspects Data Quality & Accuracy (performance) Data quality & accuracy going forward • Data validation processes: consistency checks, extreme values, illogical • Real time reports Data Quality entries, missing data etc. & Accuracy • Data verification: cross-checking with other systems, independent reviews Overview & Approach to UBR Assessment / spot checks, systems audits, etc. Priority Topics • Data quality measures (% error rates, audit findings, etc.) Topic Areas Assessing the Current Experience “As-Is” Looking Forward to the “To-Be” (ST & LT) Current use/ users Potential uses / users Use of UBR for • Questionnaire to meet needs for social program’s eligibility • What are other potential user programs? assessment (SCTP, PWP, FISP) - Adequacy • What types of information would they need? Social Programs • Data sharing • Potential role of UBR for “shock-responsive” aid • Broader info use for linkages & referrals system • Updates to UBR - frequency Targets & Indicators of Inclusion / Coverage Targets & Inclusion Strategy Going Forward • What were coverage/registration rates before UBR? • Registration targets going forward? • Registration 50% target (over/under?) • Registration targets - > 50%? 80%? 100% • Coverage of UBR total and by district: #HH and % of • Relation to uses and users (incl. shock-resp) Coverage & population (for existing 11 districts) • Role of CBT? Inclusion • Coverage of UBR by PMT group: #HH and % of population ✓ CBT as inclusion mechanism (don’t miss (for existing 11 districts) these households) • Characteristics of UBR HHs: HH size & composition, highest ✓ What is role if target is 100%? EDU level, LFP status, livelihood sources, housing, land and assets, food security: • Average and by PMT group • Compared with IHS for average & by PMT group Current systems for monitoring, reporting, analytics Monitoring, reporting, analytics going forward • Process monitoring reports from implementation? • Including links to ESW/shock response system, other admin Monitoring, • Standard monitoring reports - Content & Frequency systems Reporting, Analytics • Use for decision making • Accessibility of UBR for analytics - Anonymization • Analytics being carried out World Bank team, March 17, 2018 91 Social Protection & Jobs Discussion Paper Series Titles 2015-2018 No. Title 1803 Rapid Social Registry Assessment: Malawi’s Unified Beneficiary Registry (UBR) by Kathy Lindert, Colin Andrews, Chipo Msowoya, Boban Varghese Paul, Elijah Chirwa, and Anita Mittal, November 2018 1802 Human(itarian) Capital? Lessons on Better Connecting Humanitarian Assistance and Social Protection by Ugo Gentilini, Sarah Laughton and Clare O’Brien, November 2018 1801 Delivering Social Protection in the Midst of Conflict and Crisis: The Case of Yemen by Afrah Alawi Al-Ahmadi and Samantha de Silva, July 2018 1705 Aging and Long-Term Care Systems: A Review of Finance and Governance Arrangements in Europe, North America and Asia-Pacific by Laurie Joshua, November 2017 1704 Social Registries for Social Assistance and Beyond: A Guidance Note & Assessment Tool by Phillippe Leite, Tina George, Changqing Sun, Theresa Jones and Kathy Lindert, July 1027 1703 Social Citizenship for Older Persons? Measuring the Social Quality of Social Pensions in the Global South and Explaining Their Spread by Tobias Böger and Lutz Leisering, July 2017 1702 The Impacts of Cash Transfers on Women’s Empowerment: Learning from Pakistan’s BISP Program by Kate Ambler and Alan de Brauw, February 2017 1701 Social Protection and Humanitarian Assistance Nexus for Disaster Response: Lessons Learnt from Fiji’s Tropical Cyclone Winston by Aisha Mansur, Jesse Doyle, and Oleksiy Ivaschenko, February 2017 1614 Urban Social Assistance: Emerging Insights from Three African Countries by Vanessa Moreira and Ugo Gentilini, December 2016 1613 Issues for Civil Service Pension Reform in Sub-Saharan Africa by Anita M. Schwarz and Miglena Abels, November 2016 1612 How to Target Households in Adaptive Social Protection Systems? Relative Efficiency of Proxy Means Test and Household Economy Analysis in Niger by Pascale Schnitzer, October 2016 1611 Pensions for Public-Sector Employees: Lessons from OECD Countries’ Experience by Edward Whitehouse, October 2016 1610 Pension Systems in Sub-Saharan Africa: Brief Review of Design Parameters and Key Performance Indicators by Miglena Abels and Melis U. Guven, October 2016 1609 Household Enterprises in Fragile and Conflict-Affected States: Results from a Qualitative Toolkit Piloted in Liberia, Volume 2 – Annexes by Emily Weedon and Gwendolyn Heaner, August 2016 1608 Household Enterprises in Fragile and Conflict-Affected States: Results from a Qualitative Toolkit Piloted in Liberia, Volume 1 – Report by Emily Weedon and Gwendolyn Heaner, August 2016 1607 Benefits and Costs of Social Pensions in Sub-Saharan Africa by Melis U. Guven and Phillippe G. Leite, June 2016 1606 Assessing Benefit Portability for International Migrant Workers: A Review of the Germany- Turkey Bilateral Social Security Agreement by Robert Holzmann, Michael Fuchs, Seçil Paçacı Elitok and Pamela Dale, May 2016 1605 Do Bilateral Social Security Agreements Deliver on the Portability of Pensions and Health Care Benefits? A Summary Policy Paper on Four Migration Corridors Between EU and Non-EU Member States by Robert Holzmann, May 2016 1604 Assessing Benefit Portability for International Migrant Workers: A Review of the France-Morocco Bilateral Social Security Agreement by Robert Holzmann, Florence Legro and Pamela Dale, May 2016 1603 Assessing Benefit Portability for International Migrant Workers: A Review of the Belgium-Morocco Bilateral Social Security Agreement by Robert Holzmann, Jacques Wels and Pamela Dale, May 2016 1602 Assessing Benefit Portability for International Migrant Workers: A Review of the Austria-Turkey Bilateral Social Security Agreement by Robert Holzmann, Michael Fuchs, Seçil Paçaci Elitok and Pamela Dale, May 2016 1601 The Greek Pension Reform Strategy 2010-2016 by Georgios Symeonidis, July 2016 1507 Integrating Disaster Response and Climate Resilience in Social Protection Programs in the Pacific Island Countries by Cecilia Costella and Oleksiy Ivaschenko, September 2015 1506 Effectiveness of Targeting Mechanisms Utilized in Social Protection Programs in Bolivia by Ignacio Apella and Gastón Blanco, September 2015 1505 Kyrgyz Republic: Social Sectors at a Glance by João Pedro Azevedo, Paula Calvo, Minh Nguyen and Josefina Posadas, August 2015 1504 Entering the City: Emerging Evidence and Practices with Safety Nets in Urban Areas by Ugo Gentilini, July 2015 1503 Pension Patterns in Sub-Saharan Africa by Mark Dorfman, July 2015 1502 Social Protection in Fragile and Conflict-Affected Countries: Trends and Challenges by Mirey Ovadiya, Adea Kryeziu, Syeda Masood and Eric Zapatero, April 2015 1501 Defining, Measuring, and Benchmarking Administrative Expenditures of Mandatory Social Security Programs by Oleksiy Sluchynsky, February 2015 To view Social Protection & Jobs Discussion Papers published prior to 2013, please visit www.worldbank.org/sp. Abstract This paper reports on a rapid assessment of Malawi’s integrated social registry, known as the Unified Beneficiary Registry (UBR). The timing of the assessment was ripe given the upcoming round of continued expansion of the UBR and a planned shift in registration targets (from 50 percent to 100 percent of households). As such, the objectives of this assessment are to: (a) review the UBR experience to date; (b) identify strengths and areas for improvement; (c) provide short-term recommendations to support the upcoming expansion, including implementation adaptations that would be needed to accommodate the revised registration targets; and (d) support the longer-term strengthening of the UBR. While primary audience for this paper includes the core stakeholders in Malawi, the report is also of potential interest to other countries interested in developing social registries and/or carrying out social registry assessments. Malawi’s UBR has many strong fundamentals. The Government has taken the lead in designing, managing, and implementing the UBR with strong ownership across the core agencies involved. Implementation is carried out by existing decentralized institutional structures, which is a major strength. Implementation processes and information systems are effective, and most importantly, data quality is robust and registration coverage is rapidly expanding. Nonetheless, the report identifies key short-term and longer-term actions that could address challenges and strengthen the effectiveness of the UBR, including in the areas of institutional arrangements, implementation processes, information systems, data quality, links to user programs, communications, and a possible rebranding of the UBR to support better understanding of this powerful tool for inclusion and coordination in social protection and beyond. About this series Social Protection & Jobs Discussion Papers are published to communicate the results of The World Bank’s work to the development community with the least possible delay. This paper therefore has not been prepared in accordance with the procedures appropriate for formally edited texts. The findings, interpretations, and conclusions expressed herein are those of the author(s), and do not necessarily reflect the views of the International Bank for Reconstruction and Development/The World Bank and its affiliated organizations, or those of the Executive Directors of The World Bank or the governments they represent. The World Bank does not guarantee the accuracy of the data included in this work. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgement on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. For more information, please contact the Social Protection Advisory Service at socialprotection@worldbank.org or visit us on-line at www.worldbank.org/sp. © 2018 International Bank for Reconstruction and Development / The World Bank