ADAPTIVE SOCIAL The delivery chain PROTECTION and shock response Gabrielle Smith and Thomas Bowen Graphic Design: Miki Fernandez/ULTRAdesigns, Inc. Cover photo: sadikgulec. Facing page: arindambanerjee. Back cover: UN Women/Allison Joyce. ADAPTIVE SOCIAL PROTECTION The delivery chain and shock response Gabrielle Smith and Thomas Bowen ADAPTIVE SOCIAL PROTECTION :: iii :: Contents Acknowledgments |vi| Preface |vi| Abbreviations |viii| Introduction |1| SECTION 1: ASSESS |10| Outreach: How will shock-affected households be informed that they may be eligible for support? |11| Intake and Registration: How should information on shock-affected households be gathered, in order to assess their needs and eligibility? |15| Assess Needs and Conditions: Among those that are registered, how will their needs be assessed to determine their potential eligibility for post-shock support? |21| CASE STUDY: The ‘Assess’ phases in Turkey |27| SECTION 2: ENROLL |28| Eligibility and Enrolment Decisions: Based on assessment results, who most requires support after a shock, and how will they be enrolled into the program? |29| Determine Benefits Package: Once enrolled, what kind of benefits will they receive that best address their post-shock needs |33| Notification and onboarding: How best to let the selected beneficiaries know that they will receive support? |37| CASE STUDY: The ‘Enroll’ phases in Kenya |39| SECTION 3: PROVIDE |40| Benefits provision: How to physically deliver the assistance to the selected beneficiaries in the post-shock setting? |41| CASE STUDY: The ‘Provide’ phases in the Philippines |45| SECTION 4: MANAGE |46| Beneficiaries Compliance, Updating and Grievances: Is the program functioning effectively and responding to post-shock needs, are adjustments needed? |47| Exit decisions, notifications and closing cases: When should the shock response program be wound down? Who needs ongoing long-term support? |51| Summary: Key Considerations |54| References |57| :: iv :: The delivery chain and shock response List of boxes, figures, and tables Boxes Box 1 Contingency planning for shock responsive Box 1.13 Why coverage of a social registry matters: social protection: Ethiopia, Mexico, and Kenya and Lesotho |26| Pakistan |5| Box 2.1 Waiving home visits: Turkey |45| Box 2 Linking with EWS for shock response through Box 2.2 Modifying proof of ID requirements during social protection: Kenya, Uganda, and enrolment: Yemen |47| Thailand |6| Box 2.3 Adapting enrolment to ensure accessibility for Box 3 Combining steps in the delivery chain: ‘One vulnerable groups: Yemen and Turkey |47| Stop Shops’ for rapid enrolment, notification and payment in Pakistan |7| Box 2.4 Pre-enrolment of vulnerable households: Kenya |48| Box 4 Coordination of shock responsive social protection: Mauritania, Tonga, and Uganda |8| Box 2.5 Setting transfer values: Lesotho, Mozambique, Nepal, the Philippines, and Tonga |34| Box 1.1 Modifying outreach processes to ensure business as usual: the Philippines |12| Box 2.6 Political and legal constraints to setting transfer values: Lesotho, Nepal, and Turkey |36| Box 1.2 Ensuring outreach is accessible for vulnerable groups: Yemen |13| Box 2.7 Modifying key messages: Turkey |38| Box 1.3 Experiences with digital communication Box 3.1 Modifying the payment schedule: channels: Nepal |13| Mozambique |42| Box 1.4 The importance of internal communications Box 3.2 Changing payment schedule: Ecuador |42| to support outreach: Kenya, Nepal, Tonga and Box 3.3: Importance of ensuring capacity for payments: Uganda |14| Nepal and Turkey |43| Box 1.5 Collecting data for revalidation ahead of Box 3.4 Ensuring liquidity for shock response payments: vertical expansion: the Philippines |16| Kenya |44| Box 1.6 Leveraging social protection capacity and Box 3.5 Modifications to payment interface for new systems for “Post Disaster Household Needs beneficiaries: Yemen and Turkey |44| Assessments”, experiences from Latin America and the Caribbean |18| Box 4.1 Waiving conditionality following a shock: Mexico, the Philippines, and Turkey |49| Box 1.7 Relaxing or simplifying demand registration processes: Kyrgyzstan and Mexico |18| Box 4.2 Challenges with monitoring after shocks: Kyrgyzstan and the Philippines |50| Box 1.8 Overcoming on-demand registration barriers: Kyrgyzstan, Madagascar, and Mexico |19| Box 4.3 Adapting grievance mechanisms to ensure they are accessible: Nepal and the Philippines |50| Box 1.9 Overcoming barriers linked to missing IDs: Kyrgyzstan |20| Box 4.4 Standardising terms of exit/scale-down: Mozambique, Nepal, and the Philippines |52| Box 1.10 Coordination of cash transfer and wider registration activities: the Philippines |20| Box 4.5 Verifying changes in vulnerability status for exit/scale-down: Kenya and Turkey |53| Box 1.11 Predefining modified rules and processes for assessing needs and conditions during Box 4.6 Incorporating new beneficiaries: Ethiopia, horizontal expansion: Kenya |24| Kyrgyzstan, and Mozambique |53| Box 1.12 Using Social Registries for Targeting Emergency Assistance: Dominican Republic, Malawi, Mauritania, Pakistan, and the Philippines |25| ADAPTIVE SOCIAL PROTECTION :: v :: Figures Table 2.1 Strengths and constraints of eligibility and enrollment processes for shock response |30| Figure 1: The Adaptive Social Protection Building Blocks |2| Table 2.2 Strengths and constraints of processes for determining benefits packages for shock Figure 2: Visualizing shock responsiveness through response |33| vertical and horizontal expansions |3| Table 2.3 Strengths and constraints of notification and Figure 3: Delivery chain for cash transfers |4| onboarding processes for shock response |37| Figure 4: Examples of financial risk layering in support Table 3.1 Strengths and constraints of routine payment of shock response |9| systems for shock response |41| Figure 1.1: Social registry coverage and utility for shock Table 4.1 Strengths and constraints of routine response |26| beneficiary data management, compliance verification, and grievance redress systems and processes for shock response |47| Tables Table 4.2 Strengths and constraints of routine exit and Table 1 Shock response through existing cash transfer notification systems and processes for shock programs and systems |2| response |51| Table 1.1 Strengths and constraints of routine outreach Table S.1 Effective shock response: key considerations systems and processes for shock response across the delivery chain |54| |11| Table S.2 Checklist for using or adapting the cash Table 1.2 Strengths and constraints of routine transfer delivery chain for shock response registration systems and processes for shock |55| response |16| Table 1.3 Types of data management systems used on cash transfer programs |22| :: vi :: The delivery chain and shock response Acknowledgments G abrielle Smith, independent social protection at the World Bank, including Asha Williams, social protection consultant, was the lead author for this report specialist, Carlo del Ninno, lead economist, Kelly Johnson, with coauthor Thomas Bowen, social protection senior social protection specialist, Sarah Coll-Black, senior specialist, World Bank. It is intended to serve as economist, and Colin Andrews, program manager. Reviewers a companion piece for and a bridge between two recent Ines Rodriguez Caillava, social protection specialist, and World Bank publications, Adaptive Social Protection: Building Asha Williams provided comments and inputs ahead of Resilience to Shocks (Bowen et al. 2020) and Sourcebook on finalization, and Ugo Gentilini, global lead for safety nets and the Foundations of Social Protection Delivery Systems (Lindert John Blomquist, global lead for delivery systems provided et al. 2020). oversight for the report. Lastly, the team would like to thank Lisa Ferraro Parmelee for her rigorous editorial work and This paper incorporates findings from an unpublished Guillermo Siercke, disaster risk management specialist, World background paper produced by Gabrielle Smith for Oxford Bank, for his inputs and assistance in finalizing the report for Policy Management (OPM), on behalf of the World Bank, which publication. fed into the drafting of Adaptive Social Protection: Building Resilience to Shocks. The background paper benefited from The team would also like to acknowledge the generous sup- inputs and editorial work by Valentina Barca, independent port of the Japan–World Bank Program for Mainstreaming consultant. We are also particularly thankful to Kenichi Chavez, Disaster Risk Management in Developing Countries, which is senior social protection specialist, World Bank, for his inputs financed by the government of Japan and receives technical into and guidance on the background paper. Useful support from the World Bank Tokyo Disaster Risk Manage- comments and insights were also provided by other colleagues ment Hub. ADAPTIVE SOCIAL PROTECTION :: vii :: Preface T his paper outlines the ways in which governments of the social protection response, of which over 50 percent can make use of and adapt operational systems and were cash transfers. Interventions have included vertical and processes for delivering cash transfer programs to horizontal expansion of existing programs, adaptations to the support shock response. Governments, develop- design of existing programs to make them more accessible ment partners, and humanitarian actors have become in- and effective, and the introduction of new programs using terested in adaptive social protection (ASP) and shock-re- social protection administrative systems (Gentilini et al. sponsive social protection (SRSP) largely out of recognition 2020). that shocks, particularly natural disasters, are increasing in As these responses are in their early stages, the inclusion of frequency and severity, and climate change is predicted to lessons learned from them in this paper would be premature. exacerbate these trends. Much of the global experience re- Experiences to date, however, highlight that, while the lated to ASP before 2020 came in responding to such natural specific challenges posed by COVID-19 may differ from those disasters, and it forms the primary focus of this paper. presented by other shocks (not least the constraints imposed Of course, in 2020 governments across the world have had to by social distancing on the implementation of each phase of respond to an unprecedented global shock—the Coronavirus the delivery chain), the key principles and considerations set Disease 2019 (COVID-19) pandemic—which has brought a out for policymakers in this document—emphasizing flexible remarkable surge in their use of social protection to mitigate and robust systems, the need to balance speed and accuracy, the social and economic impacts of the crisis. As of June, and the adaptations that are needed along the chain—are 195 countries or territories had planned or introduced just as pertinent. Indeed, the COVID-19 pandemic has social protection measures in response to COVID-19. Cash cemented even more firmly the imperative for governments transfers have been a huge part of this, with social safety and partners to invest in ASP delivery systems as a key factor nets accounting for 60 percent (representing 621 measures) of resilience building. :: viii :: The delivery chain and shock response Abbreviations ASP adaptive social protection ATM automatic teller machine CBT community-based targeting CCTE Conditional Cash Transfer for Education (Turkey) CGP Child Grants Programme (Lesotho) CNIC computerized national identity cards (Pakistan) COVID-19 Coronavirus Disease 2019 CSA Commissariat à la Sécurité Alimentaire [Commissioner for Food Security] (Mauritania) DDC District Development Committee (Nepal) DGMM Directorate General of Migration Management (Turkey) DNH Do No Harm DRDPM Department of Relief, Disaster Preparedness, and Management (Uganda) DRM disaster risk management DSWD Department for Social Welfare and Development (Philippines) ESSN Emergency Social Safety Net (Turkey) EWS early warning system FATA TDP ERP Federally Administered Tribal Areas Temporarily Displaced Persons—Emergency Recovery Project (Pakistan) FIBE Ficha Básica de Emergencia [Basic Emergency Sheet] (Chile) FRiT Facility for the Refugees in Turkey (Turkey) HSNP Hunger Safety Net Programme (Kenya) IDP internally displaced person IRC International Rescue Committee ISAS Integrated Social Assistance System (Turkey) IT information technology IVACC Índice de Vulnerabilidad ante Choques Climáticos [Index of Vulnerability to Climate Shocks] (Dominican Republic) JHDINA Household Disaster Impact and Needs Assessment (Jamaica) LIC lower-income country MEB minimum expenditure basket MIC middle-income country MIES Ministerio de Inclusion Economic y Social [Ministry of Economic and Social Inclusion] (Ecuador) MIS management information system NADRA National Database and Registration Authority (Pakistan) NDMA National Drought Management Agency (Kenya) NDWC National Disaster Warning Centre (Thailand) NECOC National Emergency Coordination and Operations Centre (Uganda) NGO nongovernmental organization ADAPTIVE SOCIAL PROTECTION :: ix :: NISSA National Information System for Social Assistance (Lesotho) NUSAF Northern Uganda Social Action Fund (Uganda) OPM Office of the Prime Minister OSA Observatoire de la Sécurité Alimentaire [Food Security Observatory] (Mauritania) OSS one-stop shop PATH Program of Advancement through Health and Education PETi Immediate Temporary Employment Program (Mexico) PIN personal identification number PMT proxy means test PSNP Productive Safety Net Program (Pakistan) SASF social solidarity foundation (Turkey) SASW State Agency for Social Welfare (Kyrgyzstan) SIUBEN Sistema Único de Beneficiarios [Unique Beneficiary System] (Dominican Republic) SMS short message service SOP standard operating procedure SRSP shock-responsive social protection SWF Social Welfare Fund (Yemen) TRC Turkish Red Crescent (Turkey) UBR Unified Beneficiary Registry (Malawi) UNICEF United Nations Children’s Fund VCI Vegetation Condition Index VDC Village Development Committee (Nepal) WCFC Watan Card Facilitation Center (Pakistan) WFP World Food Programme INTRODUCTION Photo: Dreamstime ADAPTIVE SOCIAL PROTECTION :: 1 :: Adaptive Social Protection and the Role of Cash Transfers in Building Resilience T oday’s global landscape is characterized by mated 26 million people fall into poverty every year because multiple, interconnected, and increasingly dev- of natural disasters—especially frequent floods and drought astating shocks, the impacts of which directly (Hallegatte et  al. 2017)—and the World Bank estimates the undermine the well-being of those they affect. ongoing COVID-19 crisis will result in an additional 49 million Between 1980 and 2012, for example, the an- people being pushed into extreme poverty in 2020 (Mahler nual frequency of natural disasters increased by 250 per- et al. 2020).  cent and the number of people affected by 140 percent Within this context, the concept of “adaptive social protec- (EM-DAT 2020). Climate change is predicted to exacerbate tion” (ASP) has emerged as a tool for building the resilience of these trends and, without climate-informed development, poor and vulnerable households to shocks. Interest is wide- will push an additional 100 million people into extreme pov- spread among governments and other development and hu- erty by 2030 (Hallegatte et al. 2016). In recent years, forced manitarian policymakers and practitioners in understanding displacement has also risen to record highs. The COVID-19 how to make use of and enhance social protection programs crisis in 2020 highlights the risk pandemics pose to the lives and livelihoods of millions in an increasingly global- and systems to build the resilience of poor and vulnerable ized world, leaving no country unaffected by its widespread households in this way. A new framework from the World sweeping health, economic, and social impacts. The deep Bank highlights the ability of ASP to build the resilience of interconnectedness of such shocks, their trends, and asso- poor and vulnerable households by investing in their capac- ciated risks has created a global environment of heightened ity to prepare for, cope with, and adapt to shocks, thus pro- complexity for households, policymakers, and practitioners tecting their well-being and ensuring they do not fall into or alike to navigate (World Economic Forum 2017). become trapped in poverty as a result of the impacts (Bowen et al. 2020). This definition of ASP promotes government-led Poor households tend to be particularly vulnerable to the investment via social protection programs in the three resil- impacts of covariate shocks. For them, factors such as hav- ience capacities of poor and vulnerable households for pre- ing limited or no savings, access to finance, access to formal paredness, coping, and adaptation, along the pre- and post- insurance, or safety nets can combine and contribute to this shock continuum (before, during, and after). excessive vulnerability and a generally limited capacity to cope with the impacts when a shock hits (see, for example, The new framework the World Bank has elaborated to Dercon 2005; Hallegatte et al. 2016). To protect short-term help guide the design and implementation of ASP focuses well-being and consumption after a shock, poorer house- on strengthening social protection systems ahead of future holds may turn instead to such negative coping strategies as shocks and crises. It conceptualizes ASP around four key removing children from school to work for extra household building blocks: programs; data and information systems; fi- income, availing themselves of high-interest loans, selling nance; and institutional arrangements and partnerships (fig- productive assets, and forced migration (Del Ninno, Pierre, ure 1). The framework highlights both the elements of existing and Coll-Black 2016; Hallegatte et al. 2016; Skoufias 2005). social protection systems that are the cornerstones for build- Such short-term coping strategies often work to the house- ing household resilience to shocks and the additional priori- hold’s longer-term detriment, however. At the aggregate lev- ties and core investments that are considered instrumental in el, shocks work to undermine poverty reduction efforts and enhancing these outcomes and making the social protection can cause a country to hemorrhage human capital, also to its system more prepared in advance of the next crisis. long-term disadvantage. The “programs” building block emphasizes the role cash Moreover, shocks routinely impoverish nonpoor households transfer programs can play in building resilience, including when their capacity to cope with the impact is overwhelmed. when they respond to shocks by helping people cope with Like the poor households especially vulnerable to covari- the impacts. Evidence highlights the vital support cash trans- ate shocks, near- and nonpoor households also struggle to fer programs provide in lessening households’ food insecurity, cope with such impacts, in some cases becoming at risk of smoothing consumption, and reducing the need to resort to impoverishing losses. Indeed, international experience over- negative coping mechanisms (Ulrichs and Slater 2016; Asfaw whelmingly shows poverty is both a driver and consequence and Davis 2018; Hidrobo et al. 2018; Knippenberg and Hod- of disasters (UNISDR 2015). Globally, for example, an esti- dinott 2017). This suggests the importance of preparing social :: 2 :: The delivery chain and shock response Figure 1: The Adaptive social protection building blocks protection programs that can continue delivering assistance to existing beneficiaries after a shock, as well as scale up to reach more people with more assistance. The options for shock responsive social protection are well elaborated in a Data and typology by Oxford Policy Management (summarized in table information 1), and these are highlighted where relevant throughout the paper.1 Taken together, these options, as visualized in figure 2, represent methods existing programs can use to respond to shocks by either increasing the amount of assistance pro- vided and/or the number of people reached. Finance Programs Several limitations, however, constrain the ability of cash transfer programs to meet the needs of poor and vulnerable households in practice. These include the following: ● Limited access to any form of social protection for many Institutional poor and vulnerable households, often exacerbated by lack arrangements and of access to identification systems, mobile phones, and partnerships bank accounts for receipt of electronic transfers and by physical remoteness from points of service ● Often, limited coverage among the poor, near poor, and Source: Bowen et al. 2020. nonpoor of households spatially vulnerable to the impacts of disasters ● Difficulties in maintaining delivery of cash transfers in the face of disruption caused by shocks Table 1: Shock response through existing cash transfer programs and systems Requirements for helping households cope with shock Shock response: OPM typology Service continuity post shock Small adjustments to a routine cash transfer program to Inclusion of those vulnerable to maintain the regular service for existing beneficiaries in a shocks in long-term social protection, Design tweak shock supporting preparedness and Alternatively, the cash transfers can be designed to address adaptation vulnerability to crisis (targeting, coverage). Temporary increase of the value or duration of a cash Meeting new needs of existing Vertical transfer (top-up payments/extra payments) to meet the beneficiaries in shock response expansion additional needs of existing beneficiaries Temporary inclusion of new beneficiaries from disaster- affected communities into a cash transfer program by Horizontal extending geographical coverage, increasing enrollment expansion of eligible households in existing areas, or relaxing the enrollment criteria Meeting needs of new beneficiaries in shock response Use of parts of an established cash transfer program’s systems or processes to deliver emergency assistance, Piggybacking through a dedicated emergency response program, to and emergency disaster-affected households—for example, making use of programs existing beneficiary list, social registry, payment mechanism, social welfare staff Source: Adapted from O’Brien et al. 2018. See O’Brien et al. (2018) for further details. O’Brien et al. highlight a fifth option in their typology, “alignment,” primarily for contexts where social 1 protection is still emerging or nonexistent and where humanitarian assistance prevails. As this paper focuses exclusively on government systems, alignment is not explored here in detail. ADAPTIVE SOCIAL PROTECTION :: 3 :: Figure 2: Visualizing shock responsiveness through vertical and horizontal expansions Benefit amount Vertical expansion Temporarily increased benefit amount Regular social Regular protection Horizontal benefit(s) system expansion amount parameters Core Those not in Population beneficiaries receipt of of social regular benefits protection but affected Source: Bowen et al. 2020. systems by a shock ● Rigidity in program design that prevents adjustment of pa- protection can be bridged in practice (Grand Bargain 2016). rameters to reflect changed needs To achieve the objectives of a cash transfer, either in normal ● Differences between long-term social protection and times or during shock response, the program must reach short-term “emergency” assistance in terms of design and and provide to the right people, in a safe and timely manner, implementation requirements and the difficulty of realizing suitable forms of assistance to meet their needs and the ob- both objectives within the same operational framework jectives of the program. Success in this regard depends on In overcoming these challenges and delivering a shock-re- the design, efficiency, and effectiveness of the cash transfer sponsive cash transfer program, the design, implementation, program’s operational systems and administrative process- and adjustment of the operational processes and systems es. Where these are poorly designed, overly bureaucratic, that underpin the program are critical determining factors. or overstretched, they will undermine the achievement of They are the focus of this paper. program objectives and increase exposure to risks such as exploitation, harassment, and fraud (Barrett and Kidd 2015; Scope of this Paper Harvey and Bailey 2011). These risks are amplified when pro- grams are rapidly mobilized during shock response. This paper examines the ways in which the operational sys- tems and processes for delivering cash transfer programs can Just as the design and execution of administrative process- support shock response. It explores cash transfer programs— es and systems are key to the successful implementation of as opposed to other social safety net programs—because cash transfers in normal times, they are fundamental to en- the evidence base for responding to shocks with this instru- suring programs can continue to function, and scale up, in ment is extensive, while it remains relatively thin for other response to shocks. A shock can have an impact on the tech- instruments. Moreover, interest in the use of cash transfers nological systems, institutions, and capacities underpinning for shock response is pronounced within not only the social cash transfer programs and create hazards and challenges in protection sector but among humanitarian actors, the share the wider operating environment. During or following one, it of whose interventions delivered in cash has been growing, is important that cash transfer systems and processes have accounting for US$2.8 billion (10 percent) of humanitarian been prepared and designed, or can be adapted as needed, to assistance in 2016 (CaLP 2018). For this reason, cash transfer maintain the regular service for existing beneficiaries, taking programs are often highlighted as an intervention by which these issues into account. Also important is that cash transfer the divide between humanitarian development and social systems and processes can support the efficient and effective :: 4 :: The delivery chain and shock response Figure 3: Delivery chain for cash transfers ASSESS ENROLL PROVIDE MANAGE Eligibility Determination Provision of Beneficiaries Exit decissions, Assessment and of benefits Notification benefits compliance, notifications, Intake of of needs and enrollment and service and and/or updating, and and case Outreach registration conditions decissions package onboarding services grievances outcomes RECURRING CYCLE PERIODIC REASSESSMENT Source: Lindert et al. 2020. expansion of cash transfers to meet new needs caused by the 3. Possible adaptations, or design tweaks, to these systems shock—or that they can be amended, simplified, or otherwise and processes to ensure continuity of program opera- supported to enable them to be used this way—and without tions, during or following a shock, in a way that is most overburdening the capacities of the system and staff. This re- useful and accessible to beneficiaries sponse could be one that meets the new needs of existing 4. Considerations for ways in which these systems and pro- beneficiaries, or it could reach new individuals and house- cesses may need to adapt to support response following holds affected by the shock, or both. Reflections on overcom- a shock (whether a vertical or horizontal expansion of ing these challenges form the primary focus of this paper. an existing program or implementation of an emergency program piggybacking on these systems and processes), Structure of this Paper highlighting any risks or constraints to be aware of and implications in terms of modifications to data manage- This paper looks in turn at each phase in the delivery chain ment systems, regulations, procedures, and capacities for cash transfer programs, from outreach through to exit (see (human resources and institutions) figure 3). The delivery chain constitutes the operational pro- cesses for implementing cash transfer programs, conceived as four phases common to most cash transfer programs: as- Cross-cutting, key messages sess, enroll, provide, and manage. They include nine discrete Before delving into the individual phases of the delivery chain, subsidiary phases: outreach; intake and registration; assess- it is important to highlight some key messages that are rele- ment of needs and conditions; determination of eligibility and vant throughout the implementation of a cash-based shock enrolment; decisions on the benefits package; notification response program: and onboarding; payments of benefits; beneficiaries compli- ance; updating and grievances; and exit decisions, notifica- ➤ Contingency planning and the establishment of standard tions, and case outcomes. Key actors, including people (ap- operating procedures along the delivery chain will enable plicants and beneficiaries) and institutions (both central and faster, more effective responses. local), interact all along the delivery chain. The interactions Many of the cash transfer programs that have responded to are facilitated by communications and information systems shocks to date were developed ex post, in an ad hoc manner. and technology, among other factors. While ex post programming can be effective in meeting needs stemming from shocks, a growing body of global experience For each of the nine phases along the delivery chain, the pa- shows that lack of planning or agreed-on ways of working per outlines the following: contributes to various challenges, including communication 1. The function and processes of this part of the cash trans- difficulties, regulatory bottlenecks, overburdening of staff fer delivery chain during “day to day” operations and systems, and, ultimately, delays in providing assistance. 2. Strengths of these systems and processes for use in Global experiences highlight the importance of preparedness shock response and any constraints to be aware of planning. Contingency planning for shock responsive social ADAPTIVE SOCIAL PROTECTION :: 5 :: protection is an essential preparedness measure increasing- on these indicators. Governments and partners in countries ly being adopted by governments. It is needed to define in prone to natural disasters increasingly are linking cash trans- advance things such as roles and necessary adaptations to fer program responses to early warning information and trig- cash transfer processes and systems, develop and train stake- gers for rapid response. These triggers can be built using the holders in standard operating procedures (SOPs), and articu- data generated by existing early warning systems (EWS) and late linkages to wider government disaster risk management climate forecasts (Bastagli and Harman 2015; O’Brien et al. (DRM) plans. 2018). They are typically designed to release funds and initi- ate early actions when preestablished thresholds are met and A good contingency plan will address all phases of the de- can lead to automatic responses; this implies frontloading the livery chain, outlining necessary modifications to processes decision-making process and directly linking climate forecasts and changes needed to systems and institutions (such as ad- to their potential consequences. aptations to the program management information system or capacity building for staff or other examples outlined in Several countries, including in southern and East Africa, have box 1), addressing many of the considerations set out in the had experiences with linking social protection responses to forthcoming sections of this paper. Ultimately, such planning early warning information; box 2 highlights the progress be- can ensure faster, more effective, and more coordinated im- ing made in slow-onset crises like drought. While some coun- plementation. tries such as Thailand are developing more robust EWS for rapid-onset disasters, work remains to be done to link them ➤ Early warning information and triggers can provide a effectively with social protection systems for shock response. rules-based approach to initiating contingency plans and promoting earlier action. ➤ Ensuring “business continuity” should be the overriding priority. An important part of the contingency planning process is to establish clear indicators that will trigger activation of the Experience highlights that using social protection programs plan and initiate the shock response program, underpinned as a response mechanism is contingent on the ability of their by robust institutional processes for monitoring and acting underlying delivery systems and processes to continue to op- Box 1: Contingency planning for shock responsive social protection: Ethiopia, Mexico, and Pakistan In Ethiopia, the Productive Safety Net Program (PSNP) identified the woreda (district) for collecting household information and designed contingency plans to ensure that, in the event of shocks, transitory and regular beneficiaries would receive support in the same manner. Steps in the contingency planning process include (1) context analysis, based on early warning information, historical data, and community needs; (2) scenario assessment, identifying hazards and their potential impacts on food security and estimating the number of potential additional beneficiaries; (3) response planning, including targeting of transitory beneficiaries, public works activities, and budgeting; (4) operational support planning, including setting up “shelf projects” for public works and identifying the necessary decision makers, resources, systems, and structures; and (5) revising contingency plans, based on new information, early warning systems, and annual updates. In Mexico, most cash transfer programs have SOPs that are updated annually and set out changes in operational processes in the event of a disaster, declared emergency, or epidemic. In the case of the former conditional cash transfer program, Prospera, these changes include, among others, paying cash transfers without verification of compliance with conditions, delaying reassessment of eligibility by one year for households in affected areas, and deploying Prospera personnel to affected areas. In Pakistan, the government developed a national strategy for managing catastrophic events, the Federal Disaster Response Action Plan, which outlines contingency plans and the minimum resources and swiftest approval processes required to respond to shocks. The plan clearly defines the cash response model for emergencies and the roles and responsibilities of the respective partner agencies essential to future responses. These include the national and provincial disaster management authorities, the Benazir Income Support Program, the National Database and Registration Authority, the Ministry of Finance, and commercial banks. The processes outlined in the Federal Disaster Response Action Plan have since been implemented during the response to floods in Sindh province (2012–13), as well as to internally displaced persons of the conflict-affected Federally Administered Tribal Area regions (2015). Sources: Coll-Black et al. forthcoming; Government of Mexico 2018; Bowen et al. 2020. :: 6 :: The delivery chain and shock response Box 2: Linking with EWS for shock response through social protection: Kenya, Thailand, and Uganda In Kenya and, more recently, in Uganda, the governments have been developing contingency plans that link the cash transfer programs with EWS and testing the use of triggers for social protection shock responses based on remote sensing data. Both countries are using the Normalized Difference Vegetation Index (NDVI) for this purpose. While this indicator is useful for drought-related shocks, it is not relevant for warning about other natural disasters that can affect populations, such as flooding. Uganda plans to try to test a secondary data source suitable for other crises or disasters. In Thailand, the National Disaster Warning Centre (NDWC) was originally conceived to provide warnings about earthquakes and tsunamis and subsequently extended to monitoring other hazards. Early warning aims to notify and alert governments and the general public of a significant likelihood of hazardous events. While the NDWC monitors natural hazards and issues alerts, however, no protocols or triggers currently exist for automatic responses (social protection or otherwise) based on these EWS data. In addition to the NDWC, a national DRM plan describes the role of the Ministry of Interior, the social protection ministries, and nongovernmental organizations in providing assistance to affected people. Sources: Government of Uganda 2016; Riungu et al. 2017; Beazley 2018. erate after a shock. In the aftermath of a fast-onset, destruc- for them to continue to do so. Similarly, international human- tive shock, ensuring existing programs are able to continue itarian assistance will be more appropriate in situations where to operate and deliver benefits among affected beneficiaries there is no functioning, legitimate state, or where the state is who are in need will be vital. The more disruptive the shock is not acting with impartiality and neutrality in the interests of to institutions, infrastructure, and systems, the more critical the affected population. The “Do No Harm” principle should this will become. Having clear procedures in place for timely also be considered from the perspective of the cash trans- restoration and/or modification of systems and procedures fer program and those it aims to serve. Responding to shocks that may have been undermined for operation in a post-shock through social protection systems should not overburden or environment is a prerequisite for initiating a shock response. undermine delivery of the regular cash transfer benefits pack- This paper explores modifications to the phases along the ages to existing beneficiaries. delivery chain that can help ensure business continuity. Only ➤ Flexibility along the delivery chain is necessary for after the continuity of these underlying processes is ensured shock response, but change does not always need to be or restored can responses of increasing complexity and scale significant. be pursued. The case studies in this paper demonstrate that a key enabler ➤ Relatedly, the response should build on the principle of of responsive cash transfer programs is flexibility in processes “Do No Harm,” with a focus on safeguarding existing and delivery systems. In other words, whether the response is beneficiaries. a vertical or horizontal expansion of a cash transfer program The “Do No Harm” (DNH) concept has been applied in hu- or an emergency program whose delivery is piggybacked on manitarian action for over 25 years. Originally an approach social protection systems and capacities, the ability to mod- to working effectively in conflict-affected situations, DNH ify and adapt design features, business processes, systems, subsequently emerged as a main principle underpinning good and mechanisms and to have the capacities to implement ad practice design and implementation in all aspects of aid. Put hoc variations to the delivery chain based on the post-shock simply, it means humanitarian interventions should not cause context is crucial. Notably, however, it is not only large-scale negative impacts to, or worsen the situation for, those the in- changes that enable this flexibility. As highlighted throughout tervention aims to support, or to the wider community. DNH the paper, marginal changes along the delivery chain, which is relevant to delivering shock response in that social protec- are quick to implement and require little in training or resourc- tion will not always be the most appropriate mechanism for es, can improve the program’s capacity to respond to a shock. the purpose. This is especially true where a shock is severe and ➤ To a reasonable extent, timeliness should be prioritized over the national social protection system is nascent, with limited accuracy. access for affected households. International humanitarian actors often take the lead in providing assistance to affected, In routine cash transfer programs, considerations of efficien- vulnerable households in lower-capacity and fragile-country cy and effectiveness focus heavily on maximizing targeting contexts; in many such contexts, it may be most appropriate accuracy. Program implementers aim to minimize inclusion ADAPTIVE SOCIAL PROTECTION :: 7 :: and exclusion errors, and much effort is put into implement- ➤ Individual phases in the delivery chain that often require ing downstream delivery chain processes (in the assessment multiple separate interactions with beneficiaries can be and enrollment phases) to this end. These processes can, as combined to expedite post-shock delivery. a result, be time consuming to implement. In contrast, while The timeliness of shock response can also be increased by targeting accuracy remains important when delivering assis- streamlining the phases in the delivery chain. Aside from tance to meet needs after a shock, greater emphasis is placed relaxing the premium on accuracy in targeting, individu- on the timeliness of the response. Broadly speaking, for the al phases along the delivery chain—for example, intake and sake of a faster response, inclusion errors should be accepted registration, assessment of needs and conditions, enrollment, and exclusion errors rapidly addressed. Widespread inclusion and payment—that otherwise may require multiple separate errors of those without significant need of post-shock sup- interactions between implementers and beneficiaries or po- port can come at the expense of those in greater need, how- tential beneficiaries can be combined. Doing so can speed ever, particularly in situations of limited resources and budget up implementation and deliver assistance to affected house- constraints. This presents a tradeoff for implementers, who holds faster. An example of this kind of streamlining can be will need to reconcile these competing priorities of speed of drawn from the “one-stop shops” model in Pakistan, outlined delivery versus targeting accuracy. in box 3. Box 3: Combining phases in the delivery chain: “One-Stop Shops” for rapid enrollment, notification, and payment in Pakistan In Pakistan, “one-stop shops” (OSS) respond to flooding and forced displacement by rapidly registering, assessing, enrolling, and paying affected households. Following widespread flooding in 2010, the National Database and Registration Authority (NADRA) established as part of the Citizens Damage Compensation Program 101 local offices called Watan Card Facilitation Centers (WCFCs) to cover the flood-affected districts. The WCFCs serve as one-stop shops, where the beneficiaries are enrolled, can register complaints and grievances, and often receive their payments via a point of sale machine. Biometric screening is used to verify the beneficiaries against their computerized national identity cards (CNICs). They are then enrolled and issued with a Watan card (see illustration below). This can be used at the point of sale desk or any of the payment service provider’s automatic teller machines (ATMs). In certain districts, the placement of a cash desk at the WCFC (that is, onsite cash storage) was deemed a security risk, and payments are instead processed at a local bank branch. The OSS model has also been used to provide Livelihood Support Grants and Child Welfare Grants to internally displaced persons as part of the Federally Administered Tribal Areas Temporarily Displaced Persons—Emergency Recovery Project (FATA TDP ERP). Process flow for Livelihood Support Grant for displaced persons NADRA’S Information NADRA’S Biometric ➁ Counter ➂ Counter ➃ Payment Counter UTC processing UTC payment Biometric verification Payment to beneficiary Filtration of ineligible of beneficiaries persons NADRA’S GR cases ➄ Grievance Counter EXIT UCT beneficiary ➀ (TDP only) Sources: Ovadiya and Costella, 2013; Islamic Republic of Pakistan n.d. :: 8 :: The delivery chain and shock response ➤ Multistakeholder coordination throughout implementation ment actors involved in the cash transfer delivery chain. will be essential to increasing effectiveness, reducing ● Coordination with external actors is needed, particularly duplication, and addressing capacity gaps. among international humanitarian actors that fund and de- Effective shock response with a cash transfer program will liver emergency cash and voucher responses and have over- depend on coordination among a large number of different lapping aims in addressing needs and negative impacts. stakeholders: While there is no one-size-fits-all for what constitutes good coordination, and mechanisms will depend on context, sev- ● Horizontal coordination involves national government de- eral good practices are emerging. As illustrated in box 4, ex- partments managing cash transfer programs, social regis- amples include the integration of social protection into DRM tries, and disaster response. structures and plans (as in Tonga and Uganda) and strategic ● Vertical coordination is needed among central government partnerships between government and international devel- bodies and those decentralized bodies and local govern- opment and humanitarian actors (as in Mauritania). Box 4: Coordination of shock responsive social protection: Mauritania, Tonga, and Uganda In Mauritania, the World Bank has financed development of a shock-responsive social protection mechanism since 2017. This has required collaboration across government departments, since the pilot mechanism is managed by the Commissioner for Food Se- curity (Commissariat à la Sécurité Alimentaire, or CSA) rather than the department managing the Bank-supported safety net project (Tekavoul), and the EWS is managed by the Food Security Observatory (Observatoire de la Sécurité Alimentaire, or OSA). The Prime Minister’s Office dedicated an advisor to oversee the process and established a technical working group to help coordinate activities. OSA also established an “early warning technical committee” to improve dialogue among relevant national institutions and partners and improve preparation. Development and humanitarian partners (World Bank and the United Nations World Food Programme, or WFP) have been engaged from the outset in a strategic partnership with the government, which has utilized complementary skills and expertise for effective planning and implementation of adaptive social protection (ASP). WFP leveraged its Capacity Needs Mapping methodology to assess the EWS and the CSA’s ability to manage preparedness and response planning. WFP also facilitated a South-South exchange mission to Niger and Senegal and is currently implementing a four-year capacity support plan to improve the EWS and preparedness and response planning. In 2018 and 2019, the World Bank, WFP, and OSA jointly developed a satellite data–driven decision tool to support the na- tional early warning system. An independent evaluation confirmed that WFP engagement was instrumental to the program’s success. In Tonga, the social protection system has provided cash transfers to poor and vulnerable people since 2012. The government of Tonga is making significant efforts to respond to natural disasters in a coordinated way. In response to Cyclone Gita in 2018, it established a cross-governmental National Emergency Committee that coordinated discussions with donors and international organizations, improv- ing coordination of emergency response and ensuring the availability and timely release of funds to line ministries. The institutional framework for disaster risk management mainstreamed ASP mechanisms within it, enabling top-ups to cash transfer program beneficia- ries (in the Social Welfare Program for the Elderly and the Disability Welfare Scheme) after the cyclone. This required coordination with administrative institutions—the Tonga National Retirement Benefits Fund and the Ministry of Internal Affairs. An evaluation recom- mended further improving coordination between central and local actors and government and external actors by developing standard operating procedures to guide future implementation of top-up grants. In Uganda, the government is piloting a “scalability” function for the Northern Uganda Social Action Fund (NUSAF) 3, which will expand the public works scheme to meet new needs at times of shock. To support this, it is developing a governance framework that aims to ensure coordination of all program elements delivered through a range of institutions at the central, district, and subcounty levels, overseen by the Office of the Prime Minister (OPM) as the overall implementing agency for NUSAF 3. This structure brings together the OPM’s Project Implementation Unit, which manages NUSAF’s various program components, including the public works component, with the government’s institutional structures for disaster risk management. The scale-up mechanism of NUSAF will, therefore, sit with- in the Department of Relief, Disaster Preparedness, and Management (DRDPM), another department under the OPM. The National Emergency Coordination and Operations Centre (NECOC) within the DRDPM is responsible for collecting and distributing EWS data (based on the Normalized Difference Vegetation Index) and will act as the data collection and analysis unit for the shock response com- ponent. The government acknowledges this structure will take time to put into place and will require capacity building for stakeholders at the central and district levels, including support from UN agencies. Sources: World Bank and WFP 2019; Parsons 2018; World Bank 2018b; Government of Uganda 2016. ADAPTIVE SOCIAL PROTECTION :: 9 :: Figure 4: Examples of financial risk layering in support of shock response Reserve Contingent Risk Budget Donor Humanitarian Country fund credit transfer reallocation finance resources Ethiopia ● ● ● ● Kenya ● ● Mexico ● ● ● Philippines ● ● ● Uganda ● Source: Maher 2018. Through good coordination, international humanitarian ac- early warning information, effective coordination with other tors can play an important role in supporting government-led actors, and the required flexibility in systems and processes, social protection responses at each phase of the delivery ex post resource mobilization to finance the cost of the pro- chain, offering a wealth of expertise to national social pro- gram will create a serious bottleneck in implementation. The tection actors who may not be well versed in responding to prevailing model for financing shock responses, through both humanitarian crises. More concretely, they can provide addi- the international humanitarian system and government-led tional human resources or financial and logistical support to cash transfer programs, remains overwhelmingly ex post in help manage new, resource-intensive shock response activi- nature, posing challenges to timely response. ties. At a minimum, the participation of or co-leadership by Analysis should be undertaken to determine the potential government social protection departments in humanitarian cost of response, and financial instruments should be prepo- cash working groups can help with coordination of processes sitioned and linked to the cash transfer program accordingly. and actions to overcome gaps in coverage and reduce du- To shift toward ex ante resource mobilization, the analysis plication. Furthermore, identifying the precise roles and re- should examine a variety of data sources, including historical sponsibilities of government and international humanitarian hazard data, to shed light on the anticipated contingent liabil- actors along the delivery chain can help to establish action- ity of using a safety net to respond to shocks. Second, build- able, operational partnerships for the delivery of cash transfer ing from these costing models, appropriate funding should response programs. be prepositioned. No single financial instrument, however, should cover the entire contingent liability created through ➤ To ensure quick liquidity is available to implement the the development of a shock-responsive safety net program. program, financial instruments should be prepositioned and Rather, a risk financing strategy will be required that estab- the rules for their use agreed on ex ante. lishes the rules for use to finance shock-responsive programs Implementation will often be significantly delayed if adequate and layers multiple instruments according to the frequency funding is not prepositioned to finance the post-shock cash and severity of the shock. Figure 4 provides examples of risk transfers. Even with good contingency planning, the use of layering of financial instruments.2 Risk financing for adaptive social protection is explored in greater detail in the finance building block section of Bowen et al. (2020). 2 :: 10 :: The delivery chain and shock response SECTION 1: ASSESS ASSESS ENROLL PROVIDE MANAGE Eligibility Determination Provision of Beneficiaries Exit decissions, Assessment and of benefits Notification benefits compliance, notifications, Intake of of needs and enrollment and service and and/or updating, and and case Outreach registration conditions decissions package onboarding services grievances outcomes RECURRING CYCLE PERIODIC REASSESSMENT Photo: MediaNation ADAPTIVE SOCIAL PROTECTION :: 11 :: Outreach How will shock-affected households be informed they may be eligible for support? O utreach is the first phase in the delivery chain for outreach may be carried out by the central implementing cash transfer programs. It involves certain inter- agency, local governments, supporting nongovernmental or- actions to build awareness of the program and ganizations (NGOs), the program’s field staff, or even private encourage potential beneficiaries to apply. These firms contracted as payment service providers (TRANSFORM interactions are often carried out as part of a comprehensive 2017; World Bank 2017a). communications strategy, comprising a mixture of mass me- Outreach systems and processes for routine cash trans- dia (both electronic and print) campaigns, local information fer programs can support shock response in several ways. sessions, word of mouth strategies, visual campaigns, and As summarized in table 1.1, these include making available social media. Messages describe the program, the intended trusted and knowledgeable staff and established digital com- population, its requirements, and the application process. munication systems to disseminate messaging to otherwise Outreach also serves to keep existing beneficiaries apprised hard to reach communities and vulnerable groups affected about regular interactions, providing information regard- by a shock. Constraints to bear in mind include disruptions ing, for example, the payment schedule and conditionality from the shock to staff and systems and challenges to access, verification. Depending on the institutional arrangements, especially for new beneficiaries. Table 1.1: Strengths and constraints of routine outreach systems and processes for shock response Strengths Constraints to be aware of • Where mobile penetration is high, cash transfer programs • Although digital technology is exponentially increasing com- can employ digital communications as part of their outreach munication penetration in low- and middle-income countries, strategy. Such systems can enable a highly effective mass service coverage is not uniform. Urban areas are likely to be communication strategy, with potential to reach large and more effectively covered than isolated and rural communi- dispersed populations and those in difficult to access loca- ties. Access to mobile communication and the internet is tions immediately. heavily restricted in some countries. • In disaster-affected areas, the institutions for cash transfer • A disaster can temporarily disrupt the infrastructure under- program administration at the local level can provide ready pinning digital communication platforms while also affecting and trusted “go to” points for receiving and sharing informa- program administrative staff. tion. • Existing communication channels and media used for a cash • Some social protection programs include processes for active transfer program will have been selected because they are outreach by social welfare staff—sometimes in partnership accessible to and trusted by the program’s usual target group. with civil society organizations—providing “last-mile” com- They may not be as accessible to other population groups af- munication to ensure particular vulnerable groups are effec- fected by a shock. tively informed. • Communication channels based on person to person inter- action (social welfare staff, hotlines, active outreach) depend heavily on human resources, which may limit their ability to scale up activities. :: 12 :: The delivery chain and shock response Ensuring continuity of outreach processes Key outreach considerations in shock response during or after a shock ➤ Key messages need to be modified for communication Outreach processes must continue to operate and reach ex- during shock response outreach activities. isting program beneficiaries after a shock to ensure they re- In the case of vertical expansion of a program in response to ceive as normal the regular messaging provided to them—for a shock, beneficiaries are already familiar with the program, example, notifications of payment dates and any instructions its design, and its business processes, but they will need to be for how to receive their payments. This will be most import- told how much their benefits will increase. Communication ant for programs that do not have a fixed date for each pay- will, therefore, focus on the top-ups: informing beneficiaries ment cycle. of plans to provide this additional support, its value and dura- The outreach processes themselves can be modified, or tion, and whether it is to be provided as part of existing pay- “tweaked,” to ensure they continue to be effective (that is, ments or as a separate transfer. Knowledge of how much they accessible to beneficiaries) after a shock. For cases in which will receive and when, and when support will end, influences a shock has temporarily disrupted digital communication the expenditure decisions of beneficiaries. Where the shock channels, for example, outreach strategies can include other has badly affected nonbeneficiary households as well, it will channels, such as word of mouth or town criers. Where ben- be important to justify to the wider community why program eficiaries have been extensively displaced, messages through beneficiaries are receiving top-ups and provide any details short message service (SMS) may be useful, along with “ac- about how nonbeneficiaries can receive assistance through tive outreach” through trusted social networks, program im- other channels. In the Philippines, for example, the provision plementers, and community leaders (see box 4). A program of emergency top-ups to beneficiaries of the Pantawid cash that employs a range of communication channels in normal transfer program after super Typhoon Haiyan created some times will be strongly placed to manage the changes needed tensions at the community level. It was challenging for imple- after a shock. menting staff to explain to nonbeneficiaries why beneficiaries were getting additional support, since they were already re- It will also be important to consider the human resource ca- ceiving the regular program payments in the month after the pacities for the outreach process, especially where addition- disaster (Smith et al. 2017). al sensitization of beneficiaries is needed and where face to face communication channels are to be the primary outreach ➤ Ensure outreach mechanisms are accessible to (new) mechanism. Any gaps can be addressed by bringing in addi- beneficiaries. tional administrative support or relying on other trusted lo- To be effective, communication channels must be accessible cal actors to disseminate messages, such as happened in the to the target population, and they must be trusted. In the Philippines following Typhoon Haiyan (see box 1.1). case of horizontal expansion and piggybacking, the program is reaching a new cohort of beneficiaries who may make use of different media and face different communication barriers Box 1.1: Modifying outreach processes to ensure than existing beneficiaries. The greater the difference be- business as usual: the Philippines tween the new cohort and the existing one, the more likely In the Philippines, beneficiaries of the Pantawid cash communication channels will need some modification. Com- transfer program were extensively displaced following munication channels (and messaging) should take into ac- super Typhoon Haiyan in 2013. For the program to count the language, level of education, literacy, social margin- continue operations, the Department for Social Welfare and alization, sex, and age of the beneficiary group, as indicated Development (DSWD) needed first to locate beneficiaries by the following examples: and inform them of a validation exercise for confirming the identity and documentation of affected households. ● In the case of noncitizens or refugees, the language of ex- Beneficiaries in certain affected areas also needed to be isting outreach mechanisms (whether program administra- informed of the switch from electronic to manual payments. tive staff, print media, television, or SMS) may need to be Program staff managed this through active outreach through the Parent Leaders’ Network—a community-based changed. communication and support structure set up as part of the ● Older people may have difficulty accessing information Pantawid program. DSWD also brought in program staff through SMS or social media; conversely, these may be ex- from other regions to assist. cellent channels for reaching younger demographic groups Source: Smith et al. 2017. and isolated populations (such as refugees, or those in hard to reach areas). ADAPTIVE SOCIAL PROTECTION :: 13 :: ● Particularly vulnerable or marginalized groups may need reach strategies: to be reached proactively through their trusted social net- ● The target population must be sufficiently “technologically works, as was the case in Yemen (box 1.2). literate” to make use of such channels or be provided with appropriate training and incentives. ● SMS channels require access to current phone numbers. Box 1.2: Ensuring outreach is accessible for Such data will need to be captured as part of registration vulnerable groups: Yemen procedures (see section 2, Intake and Registration, below). In Yemen, almost half the adult population is illiterate, ● People living in areas with poor connectivity will not re- and access to mass media is limited. In response to the ceive this information. 2015 civil war, with support from UNICEF and the World Bank, the Social Welfare Fund (SWF) ensured the 2017 Emergency Cash Transfer program’s messages reached the marginalized Muhamasheen communities by word of mouth Box 1.3: Experiences with digital communication communication through trusted channels. This included the channels: Nepal SWF staff, local leaders (aqels), and a local community- based organization with links to this group. Face to face In phase 2 of the response to the 2015 earthquake in communication with SWF staff was most favored by the Nepal (horizontal expansion of the program to new target group; in the areas where access was restricted by beneficiaries), the government and UNICEF piloted the conflict, an information hotline set up and managed by use of SMS alongside the traditional outreach channels of social welfare offices proved very useful. the social assistance system. This was possible because phone numbers had been collected from applicants Source: World Bank 2017b. during registration. The objective was to provide general information about the program before disbursements, as well as to promote the use of cash for improving children’s A single means of communication is unlikely to be sufficient, nutrition. The SMS campaign had limited success, however, and various mechanisms will likely be required, particular- since very few respondents—just 9.4 percent in the post- ly given the complexity of the operating environment after distribution monitoring sample—reported receiving the a shock. These should be tested ex ante and deployed de- messages. pending on what that environment is. Implementers may also Source: Merttens et al. 2017. need to use different approaches in different communities, depending on their access to particular media. Efforts to broaden communication channels and provide out- ➤ Adapt internal communication, and training, for those reach activities to new target groups will benefit from the involved in outreach during shock response. prepositioning of resources. Capacities in this area should be To be effective, those directly involved in raising awareness of built so as not to have detrimental effects on program ad- a response program’s core parameters or disseminating infor- ministrative staff or operations. Staff on teams tasked with mation during any shock response must themselves be fully outreach activities can be bolstered by bringing in translators informed about those parameters early on. Outreach pro- where needed or by adding logistical support from across gov- cesses that rely on person to person communication (wheth- ernment agencies, as well as from international humanitari- er through hotlines or face to face) are more at risk of error an partners. In the Philippines, for example, the department than other media, since those involved must have the neces- responsible for the Pantawid program brought in additional sary knowledge and skills for relaying messages. Speed and staff from regions of the country unaffected by the typhoon coherence in rollout will be enhanced by investing in estab- to ensure sufficient personnel were available to administer lishing the post-shock communication strategy, including key shock response outreach activities (Smith et al. 2017). In re- messages and distribution channels, and in training staff in its sponse to interethnic conflict in southern Kyrgyzstan in 2010, implementation, while prepositioning to the extent possible UNICEF funded the recruitment of additional social workers the communication materials themselves. The specificity and to support outreach activities in the horizontal expansion of clarity of the messages will be ensured at the outset of any cash transfers (EUD 2019a). response by refresher training and tailoring of the strategy to ● When adopting digital communication channels to support whatever post-shock context materializes and to the result- the response program, several prerequisites increase the ing program design features (see box 1.4). Lastly, a simpler likelihood of success. As outlined in experiences from Ne- program design will be easier than a more complicated one pal (box 1.3) the following constraints should be kept in to understand and communicate accurately, especially with mind when adding SMS or social media channels to out- regard to explaining who is eligible for assistance and why. :: 14 :: The delivery chain and shock response Box 1.4: The importance of internal communications to support outreach: Kenya, Nepal, Tonga, and Uganda Upon the horizontal expansion of Kenya’s Hunger Safety Net Programme (HSNP) in response to a shock, local chiefs are expected to hold community meetings to inform households which of them are included in the payments list and how much they will receive. Lessons-learned reviews of the drought responses in 2016–17 showed that delays in internal communication processes between the financial service provider and the project management unit resulted in local chiefs’ receiving this information late. The information was shared by SMS, which was also a challenge for chiefs in areas with poor connectivity. Consequently, the chiefs were not able to hold their community meetings before the scheduled payment dates. In Nepal, communication with beneficiaries about the vertical expansion of the cash transfer program in response to the 2015 earthquake took place through the staff of Village Development Committee (VDC) offices, local radio, Nepal Scouts, leaflets, and informal local networks. In practice, most people heard about the vertical expansion from VDC officers or by word of mouth. VDC officials had not, however, participated in the design of the response, and this was not communicated effectively to them. The result was an incomplete understanding on the part of local officials of the top-up program and of their own roles, which limited the effectiveness of communication to beneficiaries. Beneficiaries were often unclear about where the top-up money came from, its purpose, or the number of transfers to expect. In Tonga, Tropical Cyclone Gita in 2018 was the strongest to hit the country in over 35 years. It caused significant damage to Tongatapu and ’Eua islands and affected approximately 80,000 people—75 percent of Tonga’s population. In response, the government used its existing cash transfer programs to disburse disaster assistance to some of the most vulnerable people by vertically expanding the old age pension and disability cash transfer schemes. Details of this planned expansion were communicated to affected populations through community radio. While evaluation showed the government was able to reach large numbers of people in a short time by this means, such channels are usually most effective when coupled with face to face communication approaches, especially for the most vulnerable. Because of the state of emergency, the government did not have the resources to do this. The evaluation also highlighted that the outreach process lacked crucial information for the wider community regarding the rationale for the program and eligibility criteria. It recommended additional community sensitization to the response to inform the affected population of why existing cash transfer program beneficiaries are eligible for assistance and emphasized that, to do this, training must be made available to those responsible for managing the program, as well as to others—such as town and district officers—involved in distributing other relief. In Uganda, contingency planning for the shock response component of the Northern Uganda Social Action fund recognized the importance of timely and accurate communication to subnational stakeholders, including those engaged in the program and others involved in emergency response. This outreach has been included as a key step in the SOPs to ensure the necessary communication is institutionalized. When a scale up is triggered and approved, this must be formally communicated by the NUSAF director to the chief administrative officers of all relevant districts. They will receive full information as to the number of beneficiaries and the amount of funding approved, which is needed by district teams to inform and then select target communities and beneficiaries. In the pilot phase it was proposed that this communication be through a workshop in the relevant districts. The decision to scale up will also be communicated to the Ministry of Finance, the World Bank, the Donor Steering Group, and the National DRR Platform. Source: Riungu et al. 2017; Merttens et al. 2017; World Bank 2018b; Government of Uganda 2016. ADAPTIVE SOCIAL PROTECTION :: 15 :: Intake and Registration How should information on shock-affected households be gathered to assess their needs and eligibility? I ntake and registration is the process of collecting informa- The process of using the intake and registration mechanisms tion on registrants who will be considered for inclusion in of existing cash transfer programs for shock response pres- the cash transfer program. Depending on the program, the ents both strengths and constraints, which depend to a large subject or unit of analysis may be the individual, household, extent on the underlying approaches adopted—either on de- or community. Information collected in this phase will be used mand or census based—as broadly summarized in table 1.2. in subsequent ones to assess applicants’ or registrants’ needs Strengths include the advantages of having access to existing and conditions and determine eligibility. Some programs may data and networks of local staff with the capacity to under- also, as part of the registration process, require verification take intake and registration processes, including for post-di- of identity or validation of attributes through the collection of saster beneficiaries. Constraints include disruptions created supporting documents, household visits, or online database by the shock to the administering of these processes and the cross-referencing (World Bank 2017a). variable accuracy of or gaps in the existing data. Intake and registration take one of two forms: it may be on demand, or it may be census based or administrator driven. In Ensuring continuity of on-demand intake and on-demand registration, the interested and potentially eli- registration processes during or after a shock gible population is invited to apply, usually at social welfare With programs using on-demand registration, processes can offices and service centers. In census-based or administra- be modified, or “tweaked,” to ensure program operations tor-driven registration, a selected population is visited and continue and new applicants can be registered during or fol- registered en masse by survey teams. En masse registra- lowing a shock. Measures may include varying the location tion is often repeated periodically—for regular cash trans- of registration points or setting up temporary offices when fer programs, around every four or five years, on average. the permanent ones are damaged or made inaccessible and The information gathered is sometimes complemented by bringing in staff from unaffected areas. Where data are col- information pulled from government databases—for exam- lected and/or input digitally, recovering connectivity and en- ple, identification, tax, and land ownership data. Information suring the restoration of digital information systems may be may be collected manually (through paper surveys) or using necessary. Such measures will be most needed when shocks electronic tools (such as tablets or computers). The optimal cause displacement or damage to infrastructure. Outlining output from this phase is a list of applicants with verified such modifications in contingency plans will ensure smoother identity and validated attributes. and more efficient adjustment of the program. Increasingly, some programs make use of preexisting data- Maintaining continuity of on-demand intake and registration bases, such as other programs’ beneficiary registries or an processes may not always be feasible or desirable after a underlying social registry (see the section 4, Assess Needs shock. Often, programs decide to suspend new on-demand and Conditions, below), to select households for new assis- registration during crisis periods so as not to overburden staff tance. Importantly, while the on-demand approach enables and to ensure continuity for those already enrolled. This is more dynamic inclusion, tension exists between this aspira- in line with the key message of safeguarding existing ben- tion and what is feasible in practice. The choice of one model eficiaries’ welfare as an overriding priority throughout the over the other is usually highly dependent on each country’s response. Programs may also take this step where financing local administrative capacity and available budget (Lindert et is not available to cover the cost of the resultant increase in al. 2020). caseload until resources become available, which highlights the key message regarding prepositioned financing. :: 16 :: The delivery chain and shock response Table 1.2: Strengths and constraints of routine registration systems and processes for shock response Strengths Constraints to be aware of • Using the existing beneficiary list of an existing cash transfer • Household lists generated from the intake and registration program removes the need for new intake and registration processes of existing programs (whether beneficiary lists or activities post shock, enabling vertical expansion. social registries) will inevitably not capture all households af- fected by the shock. • Programs using census-based registration will often have col- lected and stored household data for a portion of the popu- • Since census-based intake and registration processes are lation beyond those enrolled in existing programs, reducing costly to implement, census sweeps tend to be carried out the need for those households to engage in new intake and infrequently (on average, every five years), meaning data col- registration activities post disaster. lected before the shock may have become outdated and less accurate over time. • The capacity to conduct large-scale, census-based intake and registration exercises can be drawn upon to undertake a ded- • If not prepared in advance, including through adapted survey icated post-shock intake and registration exercise. tools and SOPs, a dedicated post-shock intake and registra- tion exercise may take an excessive amount of time to mobi- • The institutional systems and processes established for lize and complete. on-demand registration can provide the administrative and logistical structures needed for rapidly undertaking new in- • On-demand registration can be time consuming and requires take and registration following a shock. a large network of staff at the local level. Where staff are con- strained, this will limit the potential for scaling up registration • In contexts where there is a shift toward wider e-government following a shock. and integration of government databases, screening and val- idation can be undertaken via interoperability with the other • Processes for registering new households and individuals can databases. be complex and bureaucratic, requiring submission of docu- ments and taking time—neither of which is conducive to rap- idly scaling up post disaster. Key intake and registration considerations for shock response Box 1.5: Collecting data for revalidation ahead of The main issue for intake and registration in shock response is vertical expansion: the Philippines how to use and adapt systems and processes to manage the In the Philippines, the scale of the Typhoon Haiyan disaster process in support of the horizontal expansion of an existing in 2013 led to the loss of identification documents, com- program or of piggybacking and emergency programming—in pounded by the deaths of parents and caregivers. DSWD other words, how to collect the information needed from the needed to conduct a large revalidation exercise to replace wider disaster-affected population who are not already bene- documents to ensure households could receive their regular ficiaries of cash transfer programs or included within existing payments and to update enrollment data with the named guardians for newly orphaned children. To continue with social registries. regular cash transfer payments, DSWD planned this valida- ➤ Adapt registration processes to revalidate data on existing tion exercise immediately after the disaster in the affected beneficiaries for vertical expansions. areas. The quick decision and effective systems in place to carry on proved instrumental to the successful implementa- In the case of vertical expansion, the intake and registration tion of the emergency cash transfer. The exercise took place phase is already complete; however, additional data may be over three weeks and was highly intensive for DSWD’s mu- nicipal and provincial staff, who were assisted by personnel needed post shock to ensure its accuracy. Contact details, drafted in from outside the affected area. For those Pantaw- banking information, or location of households, for example, id beneficiaries who had lost their ATM cards, replacement may need to be updated, particularly in cases of widespread (for which Land Bank waived the fees) took up to several displacement, as was the case in the Philippines following Ty- months. In the interim, beneficiaries could be paid over the phoon Haiyan (box 1.5).3 This will require activating procedures counter at bank branches. to locate displaced households, supported by a strong commu- Source: Smith et al. 2017. nication campaign to inform households and communities of how and where to report changes to their circumstances. Note this is true for vertical expansions as well. 3 ADAPTIVE SOCIAL PROTECTION :: 17 :: ➤ Complete post-shock intake and registration activities to been seen in Chile, Ecuador, and Jamaica (box 1.6). Running gather information on affected households, filling gaps in a new, post-shock intake and registration exercise after the existing data and information shock will mean that new household data can be used to up- date the existing social registry, potentially making its data on Census-based registration that has previously taken place, households in high-risk areas more relevant for future shocks. usually to populate social registries, is unlikely to include all affected and vulnerable households post shock. In the Phil- ➤ Modify on-demand registration systems to enable ippines and Pakistan, for example, coverage of households registration of new beneficiaries affected by the shock in the national social registry that underpins the main cash On-demand registration systems can provide more flexible transfer programs ranges from 70 to 90 percent, and, in many support to horizontal expansion than census-based systems. countries, coverage is significantly lower (Barca 2017; Leite et They can also, however, be overwhelmed by the scale of a al. 2017; Barca and O’Brien 2017). Furthermore, where peo- shock, with administrators outnumbered by applicants and ple have been displaced, national social registries are unlikely struggling to cope with physical damage to offices. Cash to contain records on these refugees or noncitizens. transfer programs are often constrained in terms of human Collection of new data, therefore, may be needed post shock and material resources even in normal times, and a disaster to capture changing household conditions and determine el- can reduce this capacity further, leaving them overburdened igibility for the response program. Blanket, geographical tar- in taking on additional registration duties for post-shock geting of all existing beneficiaries of a cash transfer program scale up. These departments will also need office space, in the affected area can be a pragmatic choice in severe con- equipment, and budget to cover these added requirements. texts where the scale of the disaster means most are likely to Assessing existing capacity and taking measures to bolster have been affected, and if program coverage is high there; it and address systemic weaknesses will be important. Do- this was the case following the super typhoon in the Philip- ing so may include, for example, bringing in extra personnel pines in 2013 (Smith et al. 2017) and the 2015 earthquake from other regions, government agencies, or government in Nepal (Merttens et al. 2017). This approach offers speed levels (including local-level staff) to support the registration and reduced workload in the acute phase of the response. In process; recruiting and training new social welfare teams and many cases, however, where coverage is low and household administrators to undertake data collection; and streamlining information contains gaps, it may be more appropriate to take or otherwise improving processes and systems for data col- the existing beneficiary lists or social registry data as a start- lection and management. ing point and then collect more data to verify vulnerability Relaxing or simplifying registration processes can help re- and eligibility for the emergency programs. This will require duce the burden of labor on staff and speed the processes developing procedures for data collection that specify how, up. Collection of some data could be waived, for example, in where, and by whom data will be collected and entered into the interest of time. In the case of horizontal expansion, this the program management information system (MIS). It will may require relaxing the program regulations and perhaps also require devising a form that captures the relevant fields also amending the processes for reaching eligibility decisions, of data to be gathered, and adaptation of the program MIS since these decisions will need to be based only on those data to include these fields. The Latin American and Caribbean fields collected (see also section 4, Assess Needs and Condi- region has had growing experience in undertaking such ex- tions). Good practice examples can be drawn from experienc- ercises in post-shock, household-level data collection, as has es in Kyrgyzstan and Mexico (box 1.7). :: 18 :: The delivery chain and shock response Box 1.6: Leveraging social protection capacity and systems for post-disaster household needs assessments: experiences from Latin America and the Caribbean In Chile, the Ficha Básica de Emergencia (FIBE) is a post-disaster needs assessment tool used to identify households affected by disasters. Information collected with FIBE supports government decision making on who should receive emergency assistance. The Ministry of Social Development manages FIBE’s database and supports local administrations on data collection. The MSD rather than the ministry with the disaster risk management mandate plays this role because it also manages the country’s social registry and cash transfer program beneficiary registries. Having a single ministry oversee these processes and systems ensures data collected through FIBE can be coordinated with other household data to inform targeting. MSD already has database interoperability agreements, for example, with other line ministries, government organizations, and local authorities. In Ecuador, the ministry in charge of noncontributory social protection (Ministerio de Inclusion Economic y Social, or MIES) registered the households affected by the 2016 earthquake. At the beginning of the crisis, other government organizations (such as the statistics office) were in charge of the data collection process, but after some operational challenges, it was decided that MIES, with its capacity at the local level and experience with vulnerable households, should lead the process. In addition, the Ministerio Coordinador de Desarrollo Social was designated to manage the registry database, precisely because of its experience managing the integrated social registry. Jamaica’s Household Disaster Impact and Needs Assessment (JHDINA) instrument is the main tool used to assess post-shock needs of households affected by disasters, assist in quantifying post-shock social protection needs, and inform social protection shock response by a range of government and nongovernment actors. The JHDINA is an instrument of the Humanitarian Assistance Committee of the National Disaster Risk Management Council. The committee is chaired by the Ministry of Labour and Social Security, which is responsible for social protection; also serving on it are representatives of Jamaica’s national DRM agency, other government agencies, and nongovernmental organizations, such as Food for the Poor and the Adventist Development and Relief Agency. The JHDINA is applied by multisector teams from the committee, led by ministry social workers. Communities to be assessed are prioritized following an initial damage assessment at the national level. The JHDINA was redesigned in 2017 to help address gaps identified in the previous version. Challenges included a reliance on paper- based data collection, variables too limited to inform appropriate social protection response across a range of actors, and limited coordination of the post-disaster household assessment process. The new instrument is available in multiple formats to enable quick decision making and integration of the multiple agencies engaged in disaster response in the country. Variables captured by the questionnaire included disaster type, location, demographics and family composition, health of household members, receipt of social assistance, damage and losses, and immediate needs. Jamaica has not experienced a national-level disaster event since the JHDINA’s redesign. The instrument’s predecessor, however, informed vertical expansion of the Program of Advancement through Health and Education (PATH) conditional cash transfer program and national insurance pensions and horizontal expansion to non-PATH affected households following Hurricane Dean in 2007. Sources: World Bank 2018a; Beazley 2017; Beazley et al. 2016; Beazley et al. 2019; Williams et al. 2020. Box 1.7: Relaxing or simplifying demand registration processes: Kyrgyzstan and Mexico In Kyrgyzstan following the 2010 interethnic conflict, the government signed a temporary regulation that relaxed for six months the registration requirements for two cash transfer programs in the two affected provinces. Under it, ad hoc local social commissions were established to assess applications rapidly for households without having to visit them. In Mexico, the Immediate Temporary Employment Program (PETi in Spanish) provides cash transfers for disaster-affected households, among other groups. Following a shock, the guidelines for intake and registration relax the procedures for registering these households, using a simplified version of the registration form, in the interest of speed and workload. The eligibility assessment is also modified so as to not implement a more time- consuming, full proxy means test. Sources: EUD 2019a; Diario Oficial de la Federacion 2017. ADAPTIVE SOCIAL PROTECTION :: 19 :: Box 1.8: Overcoming on-demand registration barriers: Kyrgyzstan, Madagascar, and Mexico In Kyrgyzstan, the process for registration for both the country’s cash transfer programs meant families had to apply in the social welfare offices at the district level, often a good distance (over 100 km) from their residences. During the scale up in response to the 2010 conflict, UNICEF supported the government in setting up mobile outreach services to take registration to communities, making the process more accessible for the poorest. This speeded up registration of people made vulnerable and impoverished by the crisis and also identified those erroneously excluded from the programs—that is, those who were previously eligible but had not known about the programs or had been unable to apply. In Madagascar in 2018, the government expanded its flagship cash transfer program, the Human Development Cash Transfer (TMDH), to new communities, providing assistance to households with children that were affected by drought. The government’s administrative capacity to manage the additional targeting was limited, and UNICEF and WFP provided logistical support and human resources to enable it. In Mexico, the on-demand registration protocol for the PETi program is flexible and allows for a combination of door to door registration, temporary registration camps, and registration desks in temporary shelters. The method selected depends on the type of shock and the context. During the extensive flooding in Tabasco and Veracruz in 2009, for example, whole villages were underwater for days, so most of the registration took place at temporary shelters. In the 2010 earthquakes in Oaxaca, a combination of door to door registration and registration camps was implemented. Sources: CaLP 2018; UNICEF2019; SEDESOL 2017a and b; EUD 2019a. ➤ Make places for on-demand registration accessible to reduced through such adaptations as producing application vulnerable groups after a shock forms in relevant languages, recruiting additional staff for registration who speak those languages, providing translation On-demand systems often require people to travel to partic- services, and providing additional help for people in vulner- ular places—such as the local offices of social welfare teams able households who struggle to understand the process. In or program administrative offices—to apply for the program. Kyrgyzstan, for example, where the 2010 conflict most heav- This means new applications can be made at any time, which ily affected Uzbek communities, the majority of social welfare offers flexibility for scaling up registration following a shock. officers were Kyrgyz. To ensure the initiative was inclusive It relies, however, on having a permanent and widespread of both communities, the government, with support from network of well-staffed registration points that applicants UNICEF, recruited both Kyrgyz and Uzbek social workers for are able to reach (Barca 2017; Leite et al. 2017). The existing mobile registration activities in affected communities (EUD offices may not be convenient for or accessible to the popu- 2019a). Many such activities have implications for human re- lation targeted by an expanded program, sometimes because sources for a program, since they will either require more staff they are far away or because they are cut off as a result of the time or the recruitment of additional personnel. shock, or because of constraints on their mobility. ➤ Modify proof of identity requirements for registering new Various adaptations to registration systems can overcome beneficiaries. this problem. Temporary offices can be set up and staffed in locations that are safe and accessible for the target group; Limited access or barriers to acquiring the necessary formal registration camps or doorstep services can take registration documentation among newly targeted beneficiaries can im- activities to communities; and transportation costs can be pede registration if ID is a requirement in this phase.4 This covered for vulnerable applicants so they can travel to social may be common in countries where civil registration or na- welfare offices elsewhere. Such activities have implications tional ID systems (whether foundational or functional) are for human resources and logistics budgets. Box 1.8 highlights not well developed, or in cases of refugee influx. It can also ways in which they have been successfully carried out in Kyr- be an issue where documents are lost as a result of the shock. gyzstan, Madagascar, and Mexico. To avoid this issue at times of crisis, requirements can be New applicants may struggle to complete registration pro- simplified. The number of data fields can be reduced, for ex- cesses as a result of insufficient education, language barri- ample, in a registration system that involves the collection of ers (particularly in the case of refugees), or discrimination by many variables, and requirements to provide documents or the staff responsible for registration. These barriers can be conduct household visits can temporarily be waived. This was Depending on the context and program, ID can be made a requirement for the registration or enrollment phase (see section 5). 4 :: 20 :: The delivery chain and shock response the case in Mexico for the Immediate Temporary Employment ➤ Coordination with other actors to fill gaps in intake and Program (PETi in Spanish), which accepted local authorities’ registration information affirmation in lieu of official documents during intake and Investment is needed to improve coordination of registra- registration activities after a rapid-onset disaster (SEDESOL tion activities between cash transfer programs and the wider 2017a, 2017b). As seen in Kyrgyzstan (box 1.9), such changes emergency response. Parallel registration activities can be, may require amendment of the program regulations, as well and often are, designed and implemented by other actors as some backend system adaptations to maintain the capacity (including government disaster risk management actors, Red to deduplicate applicants in the absence of unique identifiers. Cross or Red Crescent societies, and international humanitar- Where paper forms are used, this could require changing the ian organizations). They can help fill gaps in social protection form or, instead of using it, directly contacting applicants to registration systems, especially if the capacity of the imple- gather the needed data. Where screening of data in existing menting agency to administer its own new intake and regis- civil registries is part of registration, the screening process will tration process is limited after the shock (see the Philippines need to be adapted to make use only of the relevant fields or experience in box 1.10). perhaps to incorporate data from new registries. Alternative- ly, programs could help applicants obtain the necessary doc- uments—perhaps through intergovernmental collaboration. Box 1.10: Coordination of cash transfer and wider registration activities: the Philippines Box 1.9: Overcoming barriers from missing In the Philippines, some 16 million people were affected by IDs—Kyrgyzstan Typhoon Haiyan in 2013, with 4.1 million people displaced and 1.1 million houses damaged. During the response, over In response to the 2010 conflict in Kyrgyzstan, both 1.4 million people received cash assistance from govern- existing social protection programs required the applicant ment and a range of nongovernmental humanitarian orga- to provide extensive documents, which could take time to nizations. The coverage of the affected population by the collate and were not all easily accessible. In addition, some Pantawid program—the flagship cash transfer program conflict-affected households had lost civil documentation. of the Department of Social Welfare and Development Those applying to the social protection programs during the (DSWD)—was significant. In the affected regions, 805,000 horizontal expansion did not have to submit the necessary households were registered as beneficiaries, with up to half verification documents for six months. During this time, a the enrolled population in the poorest municipalities. With government taskforce involving the department responsible WFP’s support, DSWD provided all Pantawid beneficiaries for social protection, the vice prime minister responsible for with a cash “top-up” in addition to their usual cash transfers social affairs, and the state registration services were set up for two months. To reach those affected households that to fast track claims for replacing the requisite national ID were not Pantawid program beneficiaries but also in need and civil documentation. of assistance, WFP implemented a parallel program and Source: EUD 2019a. registered them through nongovernmental organizations and staff from the local government units. The lack of insti- tutional coordination between these two interventions did present challenges. Sources: Smith et al. 2017. Harmonized questionnaires for use across actors and multi- ple programs will improve the efficiency of the response as a whole and reduce the burden of repeated surveys on affected households. Often, following a shock, various institutions and actors do their own assessments to inform their post-shock transfers across government agencies and nongovernment actors. This increases expectations and creates fatigue among affected households. Agreement among all institutions to use data from a single instrument to the extent possible could help reduce duplication in these processes. ADAPTIVE SOCIAL PROTECTION :: 21 :: Assess Needs and Conditions How will the needs of those who are registered be assessed to determine their eligibility for post-shock support? This phase of the delivery chain encompasses the systemat- ment of each individual’s needs and condition. In the case of ic processes and methodologies for determining registrants’ screening based on disability, for instance, a medical exam needs, according to various assessment tools, using data col- administered by an authorized body is necessary to ensure lected during intake and registration.5 Its purpose is to deter- rigor and accuracy. mine their eligibility for cash transfer programs, as well as to Social registries can provide comprehensive socioeconomic inform the benefits package (covered under section 6, Deter- data about households that different cash transfer programs mining Benefits) (World Bank 2017a). can use when assessing needs and conditions. Several pro- Screening eligibility can involve some form of socioeco- grams, for instance, that include socioeconomic status as a nomic or aggregate welfare assessment. Most commonly in criterion can make use of the same PMT scores. Social pro- lower-income countries (LICs) and middle-income countries tection information systems are already fully institutionalized (MICs), the proxy means test (PMT) is used. This estimates in over 30 LICs and MICs worldwide, and about the same a family’s socioeconomic welfare using a composite measure number of countries are in the process of developing such or weighted score based on observable household character- systems (Barca 2017). These are increasingly being set up as istics, such as demographic structures, education levels, lo- social registries serving multiple programs. In addition to pro- cation and quality of the dwelling, and ownership of durable viding an efficient entry point for assessing needs and condi- goods and other assets. The respective weight for each char- tions across multiple cash transfer programs, they can supply acteristic represents the strength of the indicator for identify- data for use in other sectors, such as housing, utilities, civil ing welfare (defined in terms of expenditure or consumption) registration, and education, supporting a “whole-of-govern- and has been calculated through statistical analysis of income ment” approach (Leite et al. 2017). and expenditure household surveys. A poverty score for each The assessment phase also includes the periodic reassess- household is generated through automated algorithms built ment of needs and conditions. Among the factors that influ- into information systems. These socioeconomic assessments ence the periodicity of reassessment across programs and may be conducted on demand (at the time people apply for countries are the program objectives, characteristics of the the program) or be administrator driven (in line with new target population, and administrative capacity. With on-de- rounds of census-sweep registration). They are also often mand approaches, beneficiaries are typically notified in ad- combined with geographical and community-based targeting vance that they need to be reassessed so they can gather the methods. The latter leads to a ranking of households in the required documentation before their benefits expire. With database from poorest to richest. Disadvantages arise with administrator-driven approaches, reassessment is conducted poor-quality or outdated data, which can result in high occur- through a new en masse registration wave. rence of errors. It is essential to stress that the data management systems Another common approach to assessing needs is rules-based underpinning the assessment stage range widely in their set- categorization, using such categories as age, duration of un- up, size, function, and levels of integration, as illustrated by employment, or disability. The advantage of this approach is the typology in table 1.3. Differences in scope of data and that categorical rules are relatively simple to administer and program coverage greatly affect how these different types of easy to communicate to communities. It can also be useful registry can be used for shock response. when caseworker capacity and available information are lim- ited. The disadvantage is that it does not allow for assess- It should be noted that the design choices made for this phase of the delivery chain when responding to a shock ultimately dictate the exact 5 requirements for both registration (see section 3) and enrollment (see section 5). A shock response that relies on a self-assessment of needs and conditions via self-targeting, for example, has very different implications for these phases than one that requires the assessment of damage to assets and property. :: 22 :: The delivery chain and shock response Table 1.3: Types of data management systems used by cash transfer programs Serving one program Serving multiple programs Systems Beneficiary registries and MIS track data on benefi- Integrated beneficiary registries operate as data retaining ciaries and benefits to support program management warehouses that collect information from different data only on and implementation (payments, data updates, condi- social programs and their benefits administration beneficiaries tions monitoring, and grievance redress). They main- systems, allowing for monitoring and coordination of tain information only on beneficiaries of a specific “who receives what benefits” and identifying intend- program or programs. ed or unintended duplications across programs. Systems Social registries that support one social program Social registries that support multiple social pro- retaining combine processes of intake, registration, and as- grams combine the processes of outreach, intake and data on all sessment of needs and conditions to determine po- registration, and assessment of needs and conditions registrants tential eligibility for the program. The assessment to determine potential eligibility for these programs. usually takes into account measures of socioeco- They serve as platforms that support access to ben- nomic status, categorical factors, or a combination of efits and services that can extend well beyond the both. The registries contain information on all regis- sphere of social assistance. trants, whether or not they are deemed eligible for or enrolled in a particular program. The potential of systems and processes for assessing needs and coverage of the data available. Since the assessment of and conditions in support of shock response is limited in needs and conditions to determine eligibility for cash transfer some respects; table 1.4 summarizes their strengths and con- programs tends to focus on chronic poverty rather than vul- straints. While this phase of the delivery chain is not resource nerability to disasters, their methods and indicators may or intensive or logistically challenging, which makes it inherently may not be useful for assessing emergency needs, depending scalable, its overall utility is wholly dependent on the quality on the context and the type of shock. table 1.4: strengths and Constraints of Routine Processes for Assessing Needs and Conditions for Shock Response Strengths Constraints to be aware of • As this phase of the delivery chain is not resource intensive • Assessment of needs and conditions in routine cash trans- or logistically challenging to implement, it is well suited for fer programs is based on methods and techniques that are expansion during shock response (though this is wholly con- defined based on the characteristics of the target group (for tingent on the existence of data collected during the intake example, income or proxies for income demographic or geo- and registration phase). graphical characteristics). These are designed to assess needs in terms of chronic poverty. Depending on the context, they • This phase makes use of preexisting administrative proce- may or may not be good indicators of needs in terms of vul- dures and systems (data management systems, staff) for rap- nerability to a shock. id decisions. • Additionally, depending on the type or scale of the shock, • Social registries and poverty-targeted programs collect sub- priority support may be provided based on additional, stantial data on the socioeconomic characteristics of house- shock-specific parameters, including, for example, the scale holds, which can enable assessment of the vulnerability of of the impact of a destructive disaster on a household. The shock-affected households, within the program or separate- underlying decision-making procedures and screening pro- ly. This can even be undertaken ex ante. cesses are also based on these criteria. • In contexts where there is a shift toward e-government and • Data protection regulations may limit access of other depart- integration of government databases, data can be drawn ments and institutions to data. from these databases to verify or complement self-reported information and allow for an initial prescreening, which can • Periodic reassessment of needs and conditions requires be automated (and therefore rapid). gathering and updating data (see section 10, Beneficiaries Compliance, Updating, and Grievances). This has implications for staff time as well as program budgets. ADAPTIVE SOCIAL PROTECTION :: 23 :: Ensuring continuity of processes for assessing ability to a shock. This has implications for the process of as- needs and conditions during and after a shock sessing needs and conditions in any shock response. Following a shock, very little is needed to ensure the continu- ➤ With vertical expansions, assessment processes should be ity of assessment processes within the delivery chain. Since kept the same, in the interest of a rapid response. this phase focuses on the back-office application of target- With vertical expansions to existing beneficiaries, the assess- ing methods, it does not involve human resource–intensive ment tools of the underlying program can be adopted as the or logistically complex activities. A highly disruptive disaster basis for the shock response. This means no new assessment may necessitate recovering connectivity and ensuring digital of needs and conditions is needed, since all existing benefi- information systems are restored, and programs may imple- ciaries will be targeted. One advantage is that this can enable ment procedures to suspend registration of potential new more rapid response, as was seen in the Philippines during beneficiaries for routine social protection during crisis peri- the vertical expansion of the Pantawid program after super ods so as not to overburden staff and to ensure continuity Typhoon Haiyan (Smith et al. 2017). If the routine assessment for those already enrolled. This means the processes for as- of households accurately reflects disaster vulnerability, this sessing the needs and conditions of these potential routine will help ensure existing beneficiary lists accurately reflect beneficiaries will be temporarily suspended, as well. the needs and conditions of the population following a crisis. Key considerations in assessing needs and ➤ Assessment processes may need to be modified to reach conditions for shock response new beneficiaries. The instruments and techniques for assessing needs and In an emergency, a modification of the routine instruments conditions for regular cash transfer programs are determined may be more appropriate to assess needs. For horizontal ex- by the program’s nonemergency design parameters, using pansions (or piggybacked programs) aiming to reach new ben- information collected at the intake and registration stage. eficiaries, the procedures and underlying data management For programs targeting the chronically poor and vulnerable, systems for assessing needs and conditions can be adapted. for example, this can include assessing if the potential ben- If intake and registration processes have been adapted to re- eficiaries’ incomes or their PMT poverty scores are under a duce barriers to registering new beneficiaries by collecting particular threshold or whether they have been ranked as vul- less information, for example, then the process for assessing nerable according to particular criteria by community mem- needs and conditions must similarly be modified. Meanwhile, bers or committees. For categorical programs, it will include if a cash transfer program assesses eligibility based on a pov- assessing into which demographic groups the applicants fit. erty threshold, this threshold can be set higher during shock For many programs, either poverty targeted or categorical, response to capture the less poor who are also affected. whether individuals or households live in particular districts Defining assessment tools in advance of an emergency can or communities (geographical targeting) will also be assessed. enable more rapid response. For accurate targeting, however, In contrast, emergency assistance is often targeted based on program implementers must be confident any preset methods data collected in post-disaster household assessments during will accurately reflect the needs and conditions of the popula- the intake and registration phase, which uses set criteria in- tion following a crisis. The ability to do this may vary, depend- dicating vulnerability to the shock. Household-level assess- ing on the context and severity of the emergency. Recent ments review disaster impacts and the needs of affected attempts at climate-aware targeting have been designed to households to determine priorities for response. Depending address this. Kenya’s Hunger and Safety Net Program (HSNP), on the shock, the criteria may be similar to those used for discussed in box 1.11, is commonly cited as the prime exam- nonemergency situations (such as socioeconomic indicators ple of a program that collects operationally relevant data in or demographic groups) but may include different indicators advance. On the one hand, this has allowed rapid expansion (such as level of damage to property, in rapid-onset disasters), in times of need; on the other, the assessment process for de- as well. The existing cash transfer assessment process may, termining eligibility for shock response is not well understood therefore, not be a “good fit” for assessing needs and vulner- by affected communities. :: 24 :: The delivery chain and shock response Box 1.11: Predefining modified rules and processes for assessing needs and conditions during horizontal expansion: Kenya In Kenya on the HSNP, the scaling up of assistance to new beneficiaries at times of drought is first based on a monthly assessment of early warning data—the Vegetation Condition Index (VCI)—which is derived from remote-sensed satellite imagery and provides infor- mation for geographical targeting. Within the counties where the HSNP operates new beneficiary households for emergency support are then assessed based on their wealth ranking scores from the program’s social registry. These households have scores above (that is, they are “less poor than”) but are otherwise similar to those enrolled in the routine HSNP. The key advantage of targeting emergency assistance based on existing wealth ranking data has been the speed with which it can be done; beneficiary household lists can be generated within days of the VCI data’s being received. Evaluation has, however, highlighted the difficulty of targeting both routine and emergency assistance to poor households where poverty rates are high and uniform across communities. A 2016 assessment of the program’s efforts to do so found both inclusion and exclusion errors very high. This has implications for the use of this targeting approach in emergencies. Moreover, a study of the scale ups also stressed the difficulty of communicating the approach to communities or local leaders because of the complicated technical processes involved. Sources: Farhat et al. 2017; O’Brien et al. 2018. ➤ When modifying assessments, new criteria and screening of household-level data on climate exposure and livelihoods processes can be included to assess household needs post to inform targeting and distinguish the temporarily from the shock. chronically poor (Kuriakose et al. 2012). These databases can also contain georeferenced data on households or even ad- Where a program aims to reach new beneficiaries who are dresses or locations, which can be combined with DRM tools “most vulnerable” to a shock, the process for assessing needs and data to provide a more detailed assessment of the popu- and conditions can be modified for routine cash transfer lation exposed to different risks. programs by incorporating into it additional indicators of vulnerability. The process can make use of data already held If the same government department is managing the scale in the beneficiary registry or social registry underpinning the up, these activities don’t present issues for data sharing. Of- cash transfer program, or it can use new data collected from tentimes, though, such scaling up may require collaboration households post shock that may include shock-specific as- across government departments (for example, with a ministry sessment criteria, such as degree of destruction to a dwelling responsible for emergencies) or with nongovernmental ac- (see section 3, Intake and Registration). tors. Box 1.12 details emerging good practices and challenges based on experiences in several countries of using social reg- ➤ Existing cash transfer data can be used to run a new istries to target emergency assistance. Highlighted by these assessment of needs and conditions for shock response. experiences is that the regulations around ownership and Social protection information systems can provide infor- use of the data must be clear and may need to be adapted to mation against which to assess households’ vulnerability to ensure sufficient flexibility for shock response. Furthermore, shocks. Since social registries include data on all registrants, clear procedures and preestablished agreements should out- they provide a better platform for identifying vulnerable line how data can be shared and with which parties—for ex- households than beneficiary registries. Both kinds of regis- ample, whether other departments or organizations should tries can play two roles: first, they can be used as a basis for be granted access to the data management system, or if data providing immediate support to everyone in affected areas; will be extracted and shared in some other form (Barca and and, second, they can inform rapid assessment of households O’Brien 2017). most vulnerable to the crisis. This can be through integration ADAPTIVE SOCIAL PROTECTION :: 25 :: Box 1.12: Using social registries for targeting emergency assistance: Dominican Republic, Malawi, Mauritania, Pakistan, and the Philippines In the Dominican Republic, the Índice de Vulnerabilidad ante Choques Climáticos (Index of Vulnerability to Climate Shocks, or IVACC), which is part of the Sistema Único de Beneficiarios (SIUBEN, a georeferenced social registry covering 85 percent of the population) calculates the probability that a given household may be affected by certain climate shocks. The IVACC index includes three dimensions: housing characteristics (such as walls and ceilings); estimated income; and proximity to a hazardous natural element (such as a river, stream, or ravine). Using it, the government can map the vulnerability of households in the social registry. Overlain with data from risk and vulnerability assessments and hazard risk mappings, IVACC could become a powerful tool for answering questions on who should be supported and where. Malawi has had consecutive climatic shocks in recent years. Accordingly, emergency responses of international humanitarian actors have increased steadily, operating in parallel to cash transfer systems in the country. Each year humanitarian actors re-register and assess the affected population using community-based targeting (CBT), implemented by different actors without the support of any “central” database. This makes it impossible to track which households are receiving assistance from one year to the next. A trial was set up to assess the potential operational benefits of using the Unified Beneficiary Registry (UBR)—developed as a social registry to serve the needs of routine social protection programs—in the targeting of beneficiaries for emergency responses. In one district this tested using the ranked UBR list (based on PMT) within communities for endorsement using standard criteria used on the emergency responses. It enabled the collection before an emergency of some information for these households, which humanitarian partners could verify and update. Households not in the UBR could also be added (and data “pushed” to the UBR). Using the UBR as an initial list to focus community targeting minimized the influence and bias of traditional authorities. Using the pre-populated data also saved time at critical stages of the response (for example, at registration and data entry), while enhancing coordination among partners across the humanitarian- development nexus. While demographic indicators were helpful for identifying the vulnerable, other indicators from the UBR could not be used for pre- targeting households for emergency assistance, as they were too out of date. The PMT used in the UBR to rank households was found to be unsuitable for targeting emergency response, as it was not highly predictive of vulnerability to food insecurity. The questionnaire of the Unified Beneficiary Registry has been modified to identify household vulnerability to annual predictable food gaps and climate shocks. The study stressed that, for better future performance, a digitized registry should be used consistently across humanitarian actors and which pushes data back to the UBR. Moreover, coverage of the UBR could be usefully expanded. In Mauritania, the Elmaouna shock response program has been running since 2017. Managed by the CSA, it complements the government’s regular cash transfer program, Tekavoul, managed by the Tadamoun agency. Elmaouna uses the core instruments of the safety net system (in particular, the social registry and the payment platform). WFP, the CSA, and Oxfam first piloted the social registry for the targeting of lean-season interventions, which demonstrated the technical feasibility of using it for shock response, while also raising practical and methodological challenges around targeting accuracy. This operational engagement with the social registry and subsequent constructive feedback has allowed the registry to be modified and gradually improved to reflect better the requirements of seasonal shock response. In Pakistan, the International Rescue Committee (IRC) conducted a small pilot research project in Sindh province to compare the efficiency of targeting emergency assistance through CBT versus using data in the National Socioeconomic Electronic Registry. Results showed that using the prepositioned data resulted in a higher level of operational efficiency, with assistance administered in 16 days—more than twice as fast as the usual 35 days—provided relevant memoranda of understanding for data sharing with the social registry administrators were already in place. Benazir Income Support Programme (BISP) can share NSER data with other organizations upon request, though the process to date has been ad hoc and not particularly clear to external agencies. In 2016, IRC successfully accessed and made use of the NSER data for targeting an emergency program, and evaluation showed this led to efficiency and effectiveness gains, though the emerging recommendation was to create data-sharing agreements in advance to maximize timeliness (without such an agreement the wait to receive the data was 44 days). Using poverty scores from NSER to select the most vulnerable flood-affected households resulted in targeting accuracy similar to that of the CBT approach. In the Philippines, the roadmap for institutionalizing shock-responsive social protection highlights the need to introduce climate- and disaster-related vulnerability criteria into the Listahanan to strengthen shock-responsive targeting of beneficiaries. The Listahanan is the social registry used by DSWD to target cash transfer program beneficiaries and is the one registry authorized by Cabinet to be used during shocks. It is, however, a static list of households. It is not spatially referenced to hazard-prone areas, nor does it contain a hazard or a disaster variable in the PMT formula. Improving the sensitivity of the Listahanan to hazards will improve the targeting of those households during responses to shocks. Sources: Beazley (2017). King and Tranchini (2017); Holmes and Costella (2017), Hobson (2018); World Bank and WFP (2019); IRC (2016). :: 26 :: The delivery chain and shock response Social registries need good coverage, and household data need to be updated regularly to maintain their relevance and Box 1.13: Why the coverage of a social registry accuracy for use in shock response. Contrasting experienc- matters: Kenya and Lesotho es of using social registries for shock response in Kenya and Lesotho (box 1.13) demonstrate why coverage matters. This In Kenya, prior to the rollout of phase 2 of the Hunger Safe- is further illustrated by figure 1.1, which shows how cover- ty Net Programme (HSNP2), a census was completed of all households in the drought-affected counties where the age by the social registry can influence its utility for shock HSNP was implemented, and participating households were response. The accuracy of these data is also important. Over registered in the program’s MIS (and pre-enrolled). This in- time, wealth categorizations in the social registry are likely tegral feature of the HSNP’s design resulted in a database of to become outdated, and the registry may also not reflect most households in northern Kenya, containing a range of population movements or changes in contact details. More household characteristics along with poverty scores, which dynamic registration systems, with processes built in for the meant households in the social registry could be wealth ranked. The approach resulted in the registration of an addi- regular updating of records, may help to retain the relevance tional 180,000 poor and vulnerable households, who poten- of the data for targeting emergency assistance. Investment in tially could be reached with periodic emergency payments. staff will be needed to support such activities and to record In contrast, in Lesotho, despite its high coverage of house- the new data in the MIS, as will procedures allowing particu- holds, the National Information System for Social Assistance lar staff to access the MIS to ensure records are actively up- (NISSA) social registry contained no data for 28 out of 64 dated. Also needed will be procedures and infrastructure for community councils that were affected by the 2016 El Niño, receiving updated information from beneficiaries and regis- which prevented the use of NISSA data for targeting emer- trants—for example, through hotlines or visits to administra- gency assistance. tive offices (Barca and O’Brien 2017). Sources: Otulana et al. 2016; Kardan et al. 2017. Figure 1.1: Social registry coverage and utility for shock response a. Lesotho b. Mozambique CGP beneficiaries 22% 100% Beneficiaries 100% NISSA social 5% a. Lesotho across all 8% Mozambique a. population registry population program b. c. c. c. Pakistan d. The Philippines 85% 20% 100% 75% 20% 100% NSER social BISP Paksitan Listahanan Pantawid Philippines registry beneficiaries population social registry beneficiaries population a. a. b. b. c. c. Social registry data Beneficiary data for noncontributory social protection(s) Households potentially affected by shock Source: Barca and Beazley 2019. ADAPTIVE SOCIAL PROTECTION :: 27 :: CASE STUDY: The “assess” phases in Turkey In 2017, 3.3 million refugees were registered in Turkey, of quickly to the millions in need. The government therefore whom approximately 3.1 million were Syrian. In May 2017, modified the assessment process for the ESSN program, lim- an estimated 64.2 percent of refugees were estimated to iting it to six readily verifiable demographic indicators. This be living below the Turkish poverty line. In response to this means that during registration, refugee applicants are only protracted crisis, donors have supported the Facility for the required to complete their basic information and 19 of the Refugees in Turkey (FRiT), to address both the immediate usual 49 questions in the application form. In addition, ESSN emergency and longer-term development needs of refugees applicants do not receive a household verification visit until a and host communities and forge links with national systems. year after enrollment. Under the FRiT initiative, the Emergency Social Safety Net Integration of systems in support of intake and registration: (ESSN) for refugees was initiated in 2017. Its first phase ran Under the social assistance regulations for refugees, appli- until 2019, implemented through a partnership of the Minis- cants must be formally registered with the DGMM and is- try of Family and Social Policy, the World Food Programme sued with a temporary protection ID. The data management (WFP), and Turkish Red Crescent, and provided cash transfers system for social assistance (the Integrated Social Assistance to meet the basic needs of over 1.3 million refugees. Phase System, or ISAS) and the DGMM refugee registry have been 2 began in 2020. The ESSN “piggybacks” on and makes use of the operational systems and processes of the Turkish cash integrated so the social assistance foundation staff can use transfer system, which are highly advanced. Successes so far the ISAS to screen automatically for proof of this registration. have entailed the following adaptations to business process- Ensuring intake and registration processes are accessible es for outreach, intake and registration, and assessing needs for refugees: Early in the ESSN program, monitoring showed and conditions: some refugees encountered barriers to intake and registra- Ensuring outreach processes are accessible to refugees: tion. Under the social assistance regulations for refugees, Communication channels and materials used for outreach by for example, applicants must have formally registered their the routine Turkish cash transfers system are not accessible residential addresses with the population department office. to the Syrian population. The ESSN uses printed communi- The varied living arrangements of refugee families (in shared cation materials in Arabic and other languages spoken by the accommodations or nonresidential buildings) made this chal- refugee population in Turkey and distributes them through lenging. This issue was addressed through intergovernmen- the Turkish Social Assistance Foundations, as well as Turkish tal cooperation. The population office amended its protocols Red Crescent service centers, offices of the government’s to support the registration of multiple families at the same Directorate General of Migration Management (DGMM), address and those living in nonresidential accommodation. community centers, temporary education centers, and NGOs Monitoring also revealed that vulnerable families (those with working with refugees. New communication channels are mobility constraints or poor literacy) were struggling with the also used to expand outreach to the dispersed refugee pop- on-demand registration process. International humanitarian ulation. Among these are a helpline providing information in actors such as NGOs, funded by international donors, pro- five languages and a Facebook page (CaLP 2018). The inclu- vided complementary support by, for example, covering the sion of social media in the outreach strategy was found to costs of transporting applicants to social solidarity foundation be effective because the primary target population of Syrian (SASF) offices or helping them complete the application form refugees was highly technologically literate and already using (Smith 2016; CaLP 2018). social media to maintain links with friends and family (EUD Ensuring capacity of institutions to implement assessment 2017b). processes: The ESSN’s processes for outreach, intake and Adapting processes for intake and registration and assess- registration, and assessing needs and conditions are managed ment of needs and conditions for the refugee population: In- by the existing network of SASFs used in the Turkish cash take and registration for cash transfers under the Turkish sys- transfer system. These are supported by the establishment tem is an “on-demand” process. Since assessment of needs of Turkish Red Crescent service centers in areas with large and conditions is based on a proxy means test, registration populations of refugees, to bolster the capacity of SASFs and involves collecting an extensive range of socioeconomic data ensure the foundations do not become overburdened. The from households through visits. In the shock response for ref- ESSN program also covers the costs of recruiting interpreters ugees, verifiable socioeconomic data on refugees was lacking, for the SASFs to reduce language barriers during registration and the government needed to be able to scale up assistance (Smith 2016; CaLP 2018). :: 28 :: The delivery chain and shock response SECTION 2: ENROLL ASSESS ENROLL PROVIDE MANAGE Eligibility Determination Provision of Beneficiaries Exit decissions, Assessment and of benefits Notification benefits compliance, notifications, Intake of of needs and enrollment and service and and/or updating, and and case Outreach registration conditions decissions package onboarding services grievances outcomes RECURRING CYCLE PERIODIC REASSESSMENT Photo: MediaNation ADAPTIVE SOCIAL PROTECTION :: 29 :: Eligibility and Enrollment Decisions Based on assessment results, who most requires support after a shock, and how will they be enrolled into the program? D uring the fourth phase of the delivery chain, reg- viders, enrollment includes opening any accounts required istrants become beneficiaries if they are deemed for delivery of the benefits and procurement of the requisite eligible and enrolled in a program. This phase payment instruments. The beneficiary operations manage- makes use of data from the profiles of regis- ment system manages much of this process and supports trants generated by the assessment of needs and conditions the generation of a list of beneficiaries (a beneficiary regis- phase. When programs have insufficient capacity to cover try) (Lindert et al. 2020). all needs, not everyone considered eligible can be enrolled. This phase also involves finalizing an updated beneficiary list Programs use a variety of methods to manage this demand before each payment and benefits delivery. This is normal- in the face of budget constraints. Examples include ranking ly an automated back-office process that combines results households from poorest to least poor; the use of waiting from any new enrollment with data on the previous cohort, lists (primarily with on-demand approaches); random se- taking into account results from any reassessment of eligi- lection of beneficiaries from among all eligible participants bility and monitoring of program compliance and adherence (with census-based registration); enrollment based on the to conditions (phase 8) (Lindert et al. 2020). order in which applicants applied (with on-demand registra- tion); caseworkers’ discretion based on profiling (for exam- These systems and processes for eligibility for and en- ple, for services); and prioritization of households based on rollment in cash transfer programs can support shock re- categorical vulnerabilities (such as households with persons sponse—for example, by supporting vertical expansion with with disabilities or children). Depending on the registration existing enrollment data and beneficiary accounts. Notable process, additional personal data from eligible applicants constraints to be aware of include accessibility of enroll- may be collected at this phase for inclusion within the pro- ment processes for new beneficiaries who are to be reached gram’s MIS. Some programs also conduct community meet- through horizontal expansion post shock, piggybacking, or ings or household visits to validate new beneficiaries. Where standalone emergency programming. These strengths and payments are to be delivered through financial service pro- constraints are outlined in table 2.1. :: 30 :: The delivery chain and shock response Table 2.1: Strengths and constraints of eligibility and enrollment processes for shock response Strengths Constraints to be aware of • Where systems for determining eligibility are automated • Some social protection programs do not automatically notify through the program’s underlying MIS, this provides po- applicants in cases of ineligibility, which can create confusion tential to manage effectively responses “at scale” for those and tensions during the enrollment of new applicants for whose data are in these systems. shock response. • They provide an established database of program beneficia- • Some households or individuals affected by the shock and ries, their identification details, and account details for facili- in need of assistance may not be included in national ID sys- tating rapid payments during shock response. tems or legal identification systems or may struggle to pro- vide the needed documentation for the provision of bank • Where beneficiary identification is based on a nationally accounts and bank cards (in line with Know Your Customer recognized legal identity, a common ID system can easily be (KYC) regulations). This will certainly be a challenge in the used for broader enrollment (and deduplication) during shock cases of noncitizens and refugees. response. Where this doesn’t exist, functional ID systems set up for the program (for example, bank cards, social securi- • The more verification factors involved, the higher degree of ty numbers, etc.) can also provide a recognized system for confidence that transfers are reaching the “right” person but broader identification purposes. Biometric data are increas- also the greater potential for bureaucracy’s creating barriers ingly being used in enrollment systems. to and delaying enrollment and, subsequently, payments. In the case of digital enrollment systems based on biomet- • They provide established systems for notifying applicants of ric data, the technology can have an error rate of around 10 eligibility decisions for post-shock benefits. percent. Ensuring continuity of enrollment and eligibility Following a shock, affected beneficiaries may have lost the during or after a shock forms of identification (foundational or functional) with which they were enrolled in the beneficiary database and/ Programs can implement procedures to relax routine pro- or any related payment instruments—such as program ID cesses for validation of eligibility, such as home visits or com- cards, bank books, bank cards, sim card numbers, or personal munity validation, during crisis periods. Doing so can reduce identification numbers (PINs)—that were issued during enroll- the burden on implementing staff and ensure continuity for ment. This is likely to be a common problem for beneficiaries those already enrolled. Such measures were successfully ap- where severe disruption follows a natural disaster or forced plied in Turkey, for example (box 2.1). Having these waivers displacement. built into program SOPs and a digital MIS that automatically applies the modified program rules will ensure a smoother In such circumstances, programs can implement protocols for transition. Where social assistance laws or regulations stip- helping beneficiaries recover their IDs and payment tokens. ulate the requirement for such activities, it may be necessary Where programs make use of program-specific forms of ID, to insert clauses that specify relaxation under specific condi- such as bank cards or program ID cards, administrative staff tions of disaster. and payment service providers can mobilize to issue replace- ments. In large-scale disasters, this may present a significant administrative task that must be factored into staff’s work- Box 2.1: Waiving home visits: Turkey ing hours. Lead time may also be required for issuance of re- In Turkey, the protocols of the social assistance system mean placement bank cards if the cards need to be procured from that information provided by applicants during registration suppliers. Where costs are usually incurred for replacement at Social Assistance and Solidarity Foundations (SASF) of lost cards, program administrators can negotiate with pay- offices must be validated through home visits by SASF staff. ment service providers to waive them following a disaster. In When social assistance was expanded to include refugees, a rapid scale up was needed. The government agreed that the case where legal forms of ID (such as national ID cards) ESSN and Conditional Cash Transfer for Education (CCTE) are used, the process for replacing them is outside the social beneficiary households would receive a visit within one year protection program’s jurisdiction and will be subject to the of enrollment into the program rather than as a prerequisite administrative processes and bureaucracy of the providing for eligibility decisions to be finalized. institution. With effective intergovernmental coordination, Source: CaLP 2018. however, it may be possible to help cash transfer beneficia- ries navigate this process or be “fast tracked” through the sys- tem in some way. ADAPTIVE SOCIAL PROTECTION :: 31 :: Programs based on digital verification of biometric data will In the case of a standalone emergency program or a separate have less of an issue in terms of lost IDs and payment tokens. program piggybacking on administrative systems, introducing For cases in which beneficiaries have been injured and their a program-specific ID accessible to the target group may be biometric data are no longer verifiable, having a procedure for more appropriate. The type of ID or token selected has very accepting alternative forms of ID will be important. different implications for the time needed for this process and for the workload of program administrators. Moreover, In the interim, defining procedures for relaxing “proof of eli- program staff and the target group must be fully informed of gibility” requirements at payout points will allow alternative the changes. Experiences in Yemen are described in box 2.2. forms of identification to be used, ensuring that loss of ID doesn’t contribute to significant delays in disbursing pay- ments. Where programs use a financial service provider, the Box 2.2: Modifying proof of ID requirements during national financial regulations will influence the temporary enrollment: Yemen ID requirements. If bank cards have been lost, beneficiaries In Yemen, the wife, female household head, or primary could provide some other form of acceptable ID, such as a female caretaker was the designated recipient for the temporary program ID card or attestation of identification Emergency Cash Transfer program in response to the civil from program administrative offices or local government au- war. Many Muhamasheen, particularly women, do not have thorities, for “over the counter payments.” To run smoothly, national IDs. The payment provider Amal Bank therefore verifies beneficiary identity through one of several IDs, these procedures should be established ahead of an emer- including national ID card, passport, family card, voter card, gency, and any activation must be well communicated to or Social Welfare Fund (SWF) ID or through attestation by both the beneficiaries and the staff of the payment service the traditional leaders (aqel). To provide a common ID for providers. use at payout points for the emergency program, photos are taken of the designated recipient and program ID cards Eligibility and enrollment considerations produced, incorporating biometric data and a unique family for shock response code linked to the beneficiary registry. Source: EUD 2019c. In the cases of horizontal expansion, piggybacking, or stand- alone emergency programs, eligibility needs to be assessed and new beneficiaries enrolled. It is important that enrollment ➤ Adapt procedures for enrollment to ensure vulnerable mechanisms can be implemented quickly and efficiently fol- groups can get access to them. lowing a shock, and that procedures do not create barriers to enrollment. Identified beneficiaries belonging to vulnerable groups may be provided with extra support to ensure they can complete ➤ Modify proof of identity requirements for enrollment of new the enrollment process. This might include, for example, as- beneficiaries. sistance in filling out forms, translation services, or the provi- The requirement to provide documentation to establish proof sion of “last-mile” enrollment services in accessible locations of identity during enrollment and to open accounts with ser- close to communities. Most such measures have been highly vice providers can impede access to the scaled-up program beneficial in shock response programs in Turkey and Yemen for particular vulnerable groups (as also discussed for regis- (box 2.3). This additional support will require sufficient human tration in section 3).6 This may be because the required doc- resource and logistical capacity on the part of administrative uments have been lost in the shock or because certain popu- staff and the staff of payment service providers. Insufficient lations do not hold the requisite legal proof of identity (such capacity will contribute to bottlenecks and delays in the pro- cess. Where operating procedures allow, the enrollment of a as national ID cards) or proof of address, and it is likely to be third party who is authorized to collect payments on behalf of more of a problem in cases where noncitizens, refugees, or the designated recipient can be a way of ensuring vulnerable internally displaced persons (IDPs) require assistance. families can receive their emergency payments in situations It may be possible to relax these documentation requirements where particular vulnerable groups —especially older people, in the interest of a speedy and inclusive response. Where pro- people living with disabilities, and, in some cases, women— gram regulations or Know Your Customer regulations need to may struggle to get physically to payout points. Authorizing a be modified, this will require approval from the appropriate third party in such cases is standard practice for many routine authorities within government. social protection programs. Depending on program design, formal identification may be conducted at either the registration or enrollment phase. 6 :: 32 :: The delivery chain and shock response Box 2.3: Adapting enrollment to ensure accessibility for vulnerable groups: Turkey and Yemen In Turkey, applicants eligible for support through the ESSN can be enrolled by the SASF staff, and also by staff of Turkish Red Crescent (TRC) service centers. An eligible family is issued an ATM card and PIN from Halk Bank, receiving a text message when the card is ready and must be collected from a bank branch in their district. Beneficiaries must also sign a consent form that authorizes TRC to have access to their accounts. Support is provided in the form of transportation for those who struggle to get to bank branches, translation services at the branches, and payment of notary fees so someone else can sign on a beneficiary’s behalf. These measures have been set up within the ESSN project procedures, managed by TRC, and also through complementary donor-funded activities run by NGOs where TRC service centers don’t exist. In Yemen, when responding to the civil war in 2017, SWF staff were supported during enrollment in the Emergency Cash Transfer program by a local implementing partner contracted by UNICEF. The payment service provider, Amal Bank, conducted household visits to enroll beneficiaries, after which the program ID cards were distributed to households through a doorstep service implemented by SWF staff, staff of the local implementing partner, and Amal Bank. Women were the designated recipients but could nominate other household members to collect the cash assistance on their behalf. This proved useful in areas affected by conflict and where women would struggle to get to payout points. Sources: Smith 2017; EUD 2019b, 2019c; CaLP 2018. ➤ Pre-enroll vulnerable households where shocks are more predictable and recurrent. Box 2.4: Pre-enrollment of vulnerable An alternative to modifying document requirements and en- households: Kenya rollment procedures is to pre-enroll vulnerable households. In Phase 2 of the HSNP in Kenya, the scalable emergen- This approach involves collecting all pertinent operational cy cash transfer response to droughts is based on the prior data and pre-issuing required documentation for use when identification of 470,000 households, in addition to its core needed and is most relevant for slow-onset and recurrent beneficiaries, who are considered vulnerable to drought in shocks, such as droughts, in contexts where a cohort of the respective districts as part of emergency preparedness. Each of these households has been enrolled in the program households that is likely to be badly affected and in need of MIS, with bank accounts opened for them and bank cards assistance can be confidently identified ahead of time. Expe- distributed in advance. Although resource intensive to put riences of pre-enrollment in Kenya are presented in box 2.4. into place, this preparation is facilitating rapid transfers post One issue to reflect on is whether pre-enrollment will equate disaster. to automatic inclusion in any future scaling up or whether Not all these pre-enrolled households, however, are assured further assessment of eligibility according to the scale of the of assistance post disaster. This is determined by the severi- emergency and level of vulnerability will need to be under- ty of the drought status of each sublocation and the poverty score of each household, according to a predesigned alloca- taken post disaster (which could also be designed ex ante). If tion formula. Evaluation of the HSNP Phase 2 has shown additional assessments in the future re expected before eligi- this has not been not well understood by the population or bility will be confirmed and payment made, this will need to local leaders, leading to questions about why some pre-en- be clearly communicated to minimize confusion, disappoint- rolled households have been left out. ment and complaint. Sources: Riungu et al. 2017; Fitzgibbon 2016. The objective of phase 5 of the delivery chain is to determine the benefits package that will be provided to each beneficiary ADAPTIVE SOCIAL PROTECTION :: 33 :: Determine Benefits Package: Once enrolled, what kind of benefits will best address beneficiaries’ post-shock needs? during that payment cycle. In cash transfers, the determina- ● The frequency with which payments are to be made tion of benefits is normally an automated back-end process The determination of the benefits provided for a shock re- based on the updated beneficiary list, which determines sponse will depend on the systems and processes of the so- transfer amounts and calculates the payroll (World Bank cial protection programs involved. In particular, automated 2017a). Depending on the program, decision making here can systems for setting and adjusting transfer values will enable include the following: simple adjustments to be made to the existing benefits pack- ● The modality of payment—that is, whether the benefit is to age to reflect post-shock needs (as assessed in phase 3). As be provided wholly in cash or by combining cash with other noted in table 2.2, however, the actual adjustment of benefits types of support, such as vouchers or in-kind transfers packages can be complicated in many cases in the absence ● The value of the transfer to be provided, and whether this of such automated systems and by a number of other fac- will vary depending on characteristics of the household, tors, including regulations related to social assistance, politi- such as its size, or characteristics of its members, such as cal pressures, and payment service provider constraints (see gender, age, and compliance with conditionalities section 7). Table 2.2: Strengths and constraints of processes for determining benefits packages for shock response Strengths Constraints to be aware of • Where systems for setting the • The package of benefits and services is designed with chronic poverty objectives in mind transfer value are automated rather than emergency objectives. These will have been set partly by technical design but through the program’s underly- also perhaps to conform to social assistance regulations, and taking into account political ing MIS, this provides potential pressures. These factors may restrict the modification of the benefits and services pack- to manage responses effectively age to meet emergency needs. at scale. • Any underlying MIS will still need the capability to select from and apply various benefits • The process offers potential to packages and to include specific data fields that will affect the benefit level (e.g., number vary the transfer size according of dependents/household size). Unless a program was designed with shock response in to specific needs. mind, its existing MIS is unlikely to be able to manage all such fields and functions imme- diately. Reprogramming will take time and come at a cost. • Where programs are not underpinned with a digital MIS, the complexity of such changes to payroll and the time needed to manage them manually will limit the ability of systems to offer variation in the benefits package. • While varying the transfer value and switching between conditional and unconditional transfers may be possible, the vast majority of cash transfer programs provide assistance in a specific form (cash or food). Existing systems and processes for establishing the ben- efits package are unlikely to allow for the flexibility to switch between these forms of assistance. Where food transfers might be useful post disaster, these will need to be provided through parallel or complementary programs. • Transfer values for emergency response will also be driven not only by need, but in many instances mostly by budget constraints and access to additional resources. It will be im- portant to prepare a comprehensive risk financing strategy and link to prepositioned financing instruments to better address these constraints. :: 34 :: The delivery chain and shock response Key benefits package determination considerations only a portion (often a small portion) of consumption (World for shock response Bank et al. 2018c). In emergency contexts, where needs rap- idly become acute, these values are likely to be inadequate to When working with an existing cash transfer program, atten- meet either basic needs or recovery objectives. In the case of tion must be given to the following: both vertical and horizontal expansion of an existing program, ● Whether the current design of the benefits and services transfer values in the routine benefits package will need to be appraised and most likely increased. This adjustment has package will effectively meet needs after a shock implications for program systems and processes, as demon- ● Whether adapting the benefits and services will make strated from experiences in several countries (box 2.6). them more relevant, efficient, or effective for meeting post-shock needs The expertise required for calculating the shock response benefits package is likely to be different from that needed for ● Whether making such adaptations to the existing system a routine cash transfer program. Program managers need to is feasible take into account various parameters, including recognized ➤ Adjust the size and frequency of the benefits package to minimum standards for meeting particular emergency needs meet emergency needs. (such as minimum calorie intake per person in the case of food); the costs of goods and services acquired locally; and The value of cash transfers and the frequency with which what households can provide for themselves and the gap in payments are made in existing programs may not be ade- meeting these needs. The government department responsi- quate to meet the heightened need of a post-shock context. ble for the design and management of the cash transfer pro- Values of routine cash transfers for social protection are typi- gram should build the expertise to calculate a “minimum ex- cally set with reference to the national poverty line and cover penditure basket” (MEB; see box 2.5 and emergency transfer Box 2.5: Setting transfer values: Lesotho, Mozambique, Nepal, the Philippines, and Tonga In Lesotho, the transfer value top-up of the response to the 2016 El Niño through the social protection system (Child Grants Programme, or CGP) was much lower than that of WFP’s parallel cash response. This was partly because coordination of transfer values was lacking between actors and partly because of different program objectives. WFP’s program was purely emergency oriented and aimed to meet acute food needs resulting from the drought in line with humanitarian standards. In contrast, the CGP top-up was intended to sustain the efforts of the long-term program to reduce chronic vulnerabilities, despite its being designed as part of the humanitarian response. The top-up value was not based on any calculation of emergency need. As households did not receive assistance from both programs, CGP beneficiaries (some of the most vulnerable to the drought) received less. In Mozambique, the operational manual that guides implementation of the government’s emergency cash assistance (PASDE) linked to the social protection system sets a standard value for the emergency transfer of Mt 2,500 per household per month. While this is equiv- alent to the minimum expenditure basket (MEB) for meeting a household’s food needs, it does not take into account to the severity of the disaster or real food prices (though it does say the amount may be adjusted according to the nature of the disaster and the country’s socioeconomic conditions). The value is some six times higher than the basic rate of social assistance. In the vertical expansion of the government’s cash transfer programs in response to the 2015 earthquake in Nepal, the size of the top-ups was not based on calculation of an emergency needs gap and was smaller than cash transfers provided by other humanitarian actors. The expansion of the social protection program was intended to support basic consumption smoothing and was not envisaged as a replacement for all humanitarian action; rather, households’ needs would be met through a range of assistance provided (as cash and noncash) through interventions by government and nongovernmental actors. In the Philippines, the DSWD’s vertical expansion of the Pantawid program during the 2013 Typhoon Haiyan response was financially and technically supported by WFP. The top-up transfer value was based on the cost of the MEB and estimated as being sufficient for meeting households’ food needs. Households had financial needs beyond food, however. Limited coordination between clusters and sectors meant that multipurpose cash grants to meet this variety of needs were not considered, and parallel interventions of humani- tarian agencies were necessary. In Tonga, in the social protection shock response to Tropical Cyclone Gita, all cash transfer beneficiaries received a standard top-up to meet emergency needs. Evaluation showed most of the benefits were used to ensure food security and meet health needs. While around 80 percent of the elderly who received the top-up were able to cover the costs of repairing damage from the cyclone, the mon- ey was insufficient for those who still had repairs to make. The evaluation recommended a multisectoral and coordinated approach to setting benefits levels in the future, according to needs. Sources: Smith et al. 2017; Parsons 2018; Maunder et al. 2015; Kardan et al. 2017; Merttens et al. 2017; Government of Mozambique 2019. ADAPTIVE SOCIAL PROTECTION :: 35 :: values, or it may rely on input from the governmental depart- to meet all the needs of some households. To address the ments or international humanitarian agencies that routinely shortfall, the program introduced additional quarterly top-up undertake these calculations. grants that vary according to household size; these were fac- tored into the program MIS for payments (CaLP 2018). The government should coordinate with international hu- manitarian actors to harmonize the value of assistance that The frequency of transfers can also be modified to meet is provided through government with that from other cash emergency needs better. In the case of horizontal and verti- transfer programs. As noted in the key messages section, ac- cal expansion of an existing program, it will be important to tors involved in the social protection response (government consider whether the program’s regular payment schedule is authorities and any supporting humanitarian agencies) should a good fit for achieving the emergency objective of meeting engage in the humanitarian coordination forums (most com- immediate basic needs. Routine payments from the program monly the cluster system and cash working groups) and joint- may be relatively infrequent, occurring every two months to ly set transfer standards. quarterly, while assistance following a shock is often month- Finalization of the transfer value will need to take into ac- ly. The department responsible can temporarily adapt the count any regulations or legislation that indicate the size of cash transfer program’s payment schedule to provide more the assistance that can be provided through cash transfers, as regular assistance to affected households, as well as to har- well as national poverty lines and minimum wages. This is a monize the frequency of these payments with those of other negotiation process, finely balancing government constraints cash programs implemented within the humanitarian system. and emergency needs. Where refugees are involved, they The decision to make such changes must be based on care- may be the greatest issue, since they are likely to have great- ful consideration of the affect they will have on the work- er economic needs than citizens; but governments may be loads of program staff and payment service providers (this is worried about creating social tensions by assisting them. Fi- discussed further in section 8 on benefit provision). Existing nite financial resources will also constrain transfer size. Such beneficiaries (some of whom may be displaced) must also be issues have been faced in several contexts, as highlighted in informed of any changes to the payment schedule (see sec- box 2.6. If the government, or the program regulations, can- tion 6, Notification and Onboarding). not allow sufficient variation in or changing of the transfer Finally, the government will need to adapt the program MIS value to meet the full gap in emergency needs, the additional to incorporate the revised transfer values. Programs without needs can be met through a complementary aligned program, a sophisticated MIS or where payrolls are developed manu- where one exists, and vertical expansions financed by non- ally may lack the capacity to offer this flexibility in benefits governmental actors. packages. A related issue will be negotiating the changes and Transfer values can be standardized for all beneficiaries or tai- modifying agreements with the program’s contracted pay- lored according to differing needs. On the one hand, adjust- ment service providers (see phase 7). These actions can be ing the amount of the transfer to align it more closely with time consuming and may delay the response. They do not re- a household’s real needs—which can vary according to, for quire huge workloads for administrative staff, however, and example, the size of the household or the location, season, or will be undertaken only once. Nevertheless, actors must con- phase of the response—may be more effective. On the oth- sider whether the program and its systems have the capaci- er hand, such tailoring adds a layer of complexity to program ty to deal with these complexities effectively and in a timely administration. This dilemma is exemplified by Turkey. There, fashion or whether such capacities can be built. If not, then the ESSN transfer value was standardized for ease of imple- a standard transfer value to allow delivery at scale and speed mentation, but monitoring data confirmed it was insufficient may be more useful, at least in the initial stages of a response. :: 36 :: The delivery chain and shock response Box 2.6: Political and legal constraints to setting transfer values: Lesotho, Nepal, and Turkey In Lesotho, following the 2016 El Niño, the post-shock transfer value was set lower than needed to meet household emergency needs. This was because of political concern that increasing the CGP transfer levels would damage the public’s perception of the program and lead them to question the value of the lower routine social protection transfer. In Nepal, in response to the 2015 earthquake, the government’s partner on the vertical expansion program, UNICEF, proposed a transfer value of NPR 3,000 per month, for two months, for all households in the 11 worst-affected districts. The government insisted the coverage should include all 19 affected districts, which effectively halved the value of the assistance provided per household (and was less than half the value of the unrestricted cash assistance being provided by nongovernmental humanitarian actors). This transfer amount was not sufficient to address the needs of whole families. In Turkey, setting the transfer value for the ESSN began with the government and its international humanitarian partners calculating the MEB and undertaking a gap analysis. During these consultations, the government expressed reluctance to cover the full basic needs gap for refugees because it did not want the transfer value to exceed the benefits provided to poor Turkish citizens through the national social assistance system. The initial ESSN transfer was, therefore, based on the calculation of needs, plus broader concerns around sustainability and social cohesion. Monitoring data demonstrated this amount was insufficient to achieve the ESSN objective of meeting basic needs. The government of Turkey, WFP, and TRC therefore negotiated an increase in the transfer value (from TRY 100 to 120 per person), as well as quarterly top-ups for households, to close the gap. As the Conditional Cash Transfer for Education (CCTE) for refugees is an extension of the national CCTE, the Turkish government want- ed the program to make use of the same design features for the modality, frequency, duration, and value of the transfer. The monthly transfer value for refugees is the same as that provided through the national CCTE and varies according to the gender and age of the recipient. Some of these design parameters may not be optimal, however, as the value is insufficient to cover the income gaps refugee families face in meeting the needs of their children. Alignment with the ESSN and significant overlap of beneficiaries in the ESSN and CCTE for refugees means the CCTE effectively serves as additional top-up assistance for education over and above the basic needs assistance provided by ESSN. Sources: CaLP 2018; Merttens et al. 2017; Kardan et al. 2017. ADAPTIVE SOCIAL PROTECTION :: 37 :: Notification and Onboarding What is the best way to let the selected beneficiaries know they will receive support? O nce decisions on eligibility and enrollment have existing beneficiaries, this phase can also include new noti- been made and the benefits package determined, fications as needed—for example, communicating changes to households need to be notified of the decision their benefits packages or to the regular functioning of the and new beneficiaries oriented to the program program. and issued any IDs or payment instruments they will need at For regular programming, enrollment often includes an on- payout points. This takes place in phase 6, with the actions boarding (orientation) session with beneficiaries to explain managed in person, by phone or SMS, by mail, or online, de- the details of the program, co-responsibilities, payment pro- pending on the outreach and communication channels used cesses, and so on. Beneficiaries are often asked to bring sup- by the program (World Bank 2017a). During these meetings porting documentation to these sessions, including payment staff may also collect any additional information needed for and bank account details, national identification, proof of ad- program operations. This may include a photo (for the pro- dress, and so on, and they will often need to sign contracts or gram ID), a cell phone number, or signed consent forms. agreements to formalize their entry to the program. Programs For new beneficiaries, notifications should indicate what they typically have handbooks or leaflets they distribute during will receive and when, where, and how they will receive it; onboarding sessions to help with this process and to serve rights and responsibilities; contact points and information; later as reference for key program parameters, co-responsi- and next steps. Registrants who are waitlisted or deemed in- bilities, and other pertinent topics. eligible should be informed of the basis for the decision and The systems and processes for notification and onboarding provided with instructions for filing grievances (see section used by cash transfer programs have potential to support 9), although this is often skipped, and only beneficiaries are shock response. As table 2.3 suggests, the strengths in this formally notified. This is not desirable, as it leaves registrants regard come from preexisting networks of implementation who do not become beneficiaries wondering about their staff that are close to communities, as well as digital commu- status, and it undermines credibility and transparency. For nication systems, in some cases. Table 2.3: Strengths and constraints of notification and onboarding processes for shock response Strengths Constraints to be aware of • Where cash transfer programs employ • Physically onboarding beneficiaries in centralized locations where a number of digital communications as part of their them must gather may be challenging amid impacts of a destructive disaster. Sim- notification processes, this can be ef- plified onboarding processes and decentralized approaches will be needed—for fective for mass communication with example, holding community sessions in affected communities. large numbers of beneficiaries over a • Some social protection programs do not automatically notify applicants in cases dispersed area, provided they have not of ineligibility, which can create confusion and tensions during the enrollment of been affected by the shock. new applicants for shock response. • In disaster-affected areas with existing • The selection of existing communication channels and media to be used by a program staff, the institutions managing cash transfer program to support onboarding is based on their accessibility and face to face onboarding sessions for cash perceived trustworthiness for the program’s usual target group. They may not be transfer programs can provide ready “go as accessible to other population groups affected by the shock or appropriate in to” points for notifying new beneficiaries the post-shock operating environment. of the shock response. • Face to face notification depends on human resources, which may limit the ability to scale up activities. :: 38 :: The delivery chain and shock response Ensuring notification and onboarding continuity tification and can lead to some confusion. For any type of during or after a shock shock response (whether vertical, horizontal, or piggyback- ing), notifying beneficiaries about the details of the payment Notification and onboarding processes must continue to schedule and how long any changes will be in effect is essen- operate post disaster to ensure new beneficiaries (and non- tial. Existing beneficiaries must be told about the expected beneficiaries) can be notified about enrollment decisions and duration of any top-ups and new beneficiaries that they will adequately informed to participate in the response program. not be permanently enrolled in the program (O’Brien et al. As in phase 1, the communication channels used can be mod- 2018). The case of Turkey, related in box 2.7, illustrates the ified, or “tweaked,” to ensure they continue to be effective importance of simple program rules that are easy to commu- (that is, accessible to beneficiaries) post disaster. Where a nicate. shock has temporarily disrupted digital communication chan- nels, for example, other channels can be used for notification, such as word of mouth. A program that uses a range of com- Box 2.7: Modifying key messages: Turkey munication channels in normal times will be strongly placed to manage the changes needed. In addition, the location or In Turkey, ESSN transfers to refugees are made on the last timing of onboarding meetings may need to be changed to day of every month. In contrast, the payment schedule for take any post-shock disruption into account. the CCTE for Refugees mirrors that of the CCTE for Turk- ish citizens, with cash transfers delivered every two months Also important is that staff advise beneficiaries of any tem- and only for the corresponding 10 months of the school porary changes made to the usual administrative processes year. While this is good from the perspective of aligning the two programs, the different payment schedules have creat- to ensure the program can continue to operate post disaster. ed some challenges for their harmonization in practice, re- Beneficiaries may need to be informed of the following, for quiring careful communication between programs and with example: beneficiaries. ● What to do in the case of displacement (how to register for Sources: Smith et al. 2017; CaLP 2018. receipt of transfer in the new location) ● What to do in the case of lost identification documents, bank cards, or sim cards (payment process and process for ➤ Adapt mechanisms needed to communicate eligibility replacing or using alternative identification) decisions and onboard new beneficiaries. ● Any changes to the enforcement of conditions, payment schedule, or location of payment points, and how long the Existing mechanisms for communicating eligibility decisions changes will last to and onboarding beneficiaries should be assessed for their accessibility to the new intended beneficiaries, and, where Where programs implement procedures to suspend new reg- necessary, new communication channels should be added. istration during crisis periods to avoid overburdening staff As already mentioned in section 2, this may imply changes and ensure continuity for those already enrolled, these notifi- to staff workloads and budgets that must be accounted for. cation and onboarding activities will also be also suspended. In cash transfer programs that do not automatically notify Key notification and onboarding considerations for ineligible applicants of the eligibility decision, procedures shock response will need to be modified to ensure notifications go out during a shock response to minimize confusion and com- ➤ Adapt messages to be communicated to the beneficiaries plaints. In an emergency, procedures that ordinarily involve related to the shock response program parameters. gathering large groups of beneficiaries may not be feasible. When the size and frequency of the benefits package must Adaptations to the processes should be defined ex ante and be adapted to meet emergency needs, beneficiaries must be include decentralized onboarding in smaller groups within informed of the changes. This increases the complexity of no- the community. ADAPTIVE SOCIAL PROTECTION :: 39 :: CASE STUDY: The “Enroll” phases in Kenya In the arid and semiarid lands of northeastern Kenya, 80 Challenges in notifying beneficiaries of eligibility for shock percent of the land mass and one-third of the people are response: Not all households “pre-enrolled” in Group 2 are routinely exposed to drought. Drought cycles are increas- assured of assistance in the event of a post-shock expansion ing in frequency and intensity, meaning households cannot of the program. The actual scale of any scale up is deter- adequately recover livelihoods between them. This has led mined by the severity of the drought in each area and the to food insecurity, negative coping strategies, and erosion poverty score of each household, according to a previously of assets. The government’s Hunger Safety Net Program designed allocation formula. While evaluation has shown that (HSNP) was established to provide an alternative, predictable the pre-enrollment of Group 2 has helped ensure a rapid re- response to this seasonal shock for chronically vulnerable sponse to drought, it has also indicated the post-shock de- households. Since 2013, a shock response component incor- termination of actual eligibility is not well understood by the porated into it has enabled the program to expand assistance population or local leaders, leading to questions as to why temporarily at times of drought to include additional house- some pre-enrolled households have been left out (Riungu et holds in drought-affected areas. (These are known as Group 2 al. 2017; Fitzgibbon 2016). households, while routine long-term beneficiaries are known as Group 1 households.) The shock response component was Pre-defining the benefits package for shock response: built into the design of the program ex ante to enable a rap- During expansion of the HSNP to Group 2 drought-affected idly scalable, cash-based response to early signs of drought. households, the value of the transfer provided to them is pre- The following elements were among the considerations: determined under the program’s SOPs. It is the same value as the routine monthly payment provided to Group 1 long-term “Pre-enrollment” to facilitate rapid response to shock: To beneficiaries (that is, the payment they receive in nonemer- maximize timeliness of response, the program identified ex gency times), which is calculated as 46 percent of the food ante those households that may be eligible for support under basket. While this approach was intended to prevent confu- Group 2 and enrolled them in the program. This was achieved sion and tension between Group 1 and Group 2 beneficiaries at preceding phases of the delivery chain, and it included un- and speed up processes through a simple automation, its ap- dertaking a census of all households in the drought-affected propriateness and effectiveness has been questioned by non- counties, registering them into the program’s MIS, and run- governmental humanitarian actors. The value does not reflect ning a proxy means test. The result was a database of most the gap households encounter in meeting their basic needs households in northern Kenya, along with poverty scores, during drought emergencies or account for the high prices which allowed the households to be ranked by wealth. The of commodities in remote locations. The result, a review has poorest were eligible for Group 1 (longer) assistance and the found, is that the HSNP emergency payments have general- slightly less poor Group 2 households for the shock response. ly only been used for immediate consumption and have not A further 470,000 Group 2 households (comprising more prevented the affected population from resorting to negative than 80 percent of the population in the four counties) had coping strategies, such as the depletion of productive assets accounts opened with the payment service provider, were (Farhat et al. 2017). notified, and were provided with their cards. :: 40 :: The delivery chain and shock response SECTION 3: PROVIDE ASSESS ENROLL PROVIDE MANAGE Eligibility Determination Provision of Beneficiaries Exit decissions, Assessment and of benefits Notification benefits compliance, notifications, Intake of of needs and enrollment and service and and/or updating, and and case Outreach registration conditions decissions package onboarding services grievances outcomes RECURRING CYCLE PERIODIC REASSESSMENT Photo: MediaNation ADAPTIVE SOCIAL PROTECTION :: 41 :: Benefits provision How is the assistance physically delivered to the selected beneficiaries in the post-shock setting? T he objective of phase 7 of the delivery chain is to Ensuring continuity of benefit provision during distribute the correct amount of benefits to the right or after a shock people, at the right time and with the right frequency. Payments take two main forms: cash and electronic. To ensure the continuation of regular cash transfer payments The first involves delivering the benefit amount by manual and the system’s effectiveness as a payment channel for means, directly to the beneficiary, and the second through any shock response, payment systems and processes must a bank account or mobile financial service. The process of continue to operate and to reach people in a timely fashion electronic transfers can be divided into two steps: cash in— during or immediately following shocks. Natural disasters and that is, the crediting of the funds to the beneficiary’s bank conflict can damage payment infrastructure, restrict access account—and cash out, when the beneficiary collects the to offices and payment sites, limit liquidity in affected areas, benefit (World Bank 2017a). Depending on the institutional and displace staff and businesses. This may disrupt pay- arrangements, payments can be made by the implementing ment schedules, especially soon after a crisis (Bastagli 2014; agency, decentralized to a local government, or outsourced to O’Brien et al. 2018). a financial service provider (which may be a private or state- Where disruption and damage have affected the infrastruc- owned company). ture of the payment system—for example, the mobile phone Cash transfer programs that already have payment systems network, ATM network, offices of payment service providers, have several attributes that equip them to support shock re- payment sites, or liquidity (O’Brien et al. 2018)—a priority will sponse. As summarized in table 3.1, these include strong part- be to restore services and to provide flexibility so payments nerships and an existing payment network, although the ca- can be received through alternative means if one payment pacities of these systems and networks must be considered. channel cannot be restored and used. Actions on the part of Table 3.1: Strengths and constraints of routine payment systems for shock response Strengths Constraints to be aware of • Most large-scale cash transfer programs will contract with a • Where payments are not outsourced to a payment service payment service provider to lead the payment process, en- provider, the task will often overburden social protection gaging organizations with the requisite systems and expertise program staff and reduce the time available for wider admin- to reach vulnerable groups and manage financial transactions istrative tasks. This can affect the quality of the programming quickly and safely while minimizing risks. These preestab- and is likely to limit the ability of the program, or that of the lished relationships can also save time in undertaking due underlying payments system, to flex and scale. diligence and tendering and in establishing procedures. • Where services are outsourced, the terms of contracts or ser- • National coverage and good coverage in areas with high vice agreements may add rigidity to the process and limit the vulnerability to disasters provide programs with a preestab- capacity of program management to add flexibility, such as to lished network of payment agents and processes, allowing shorten payment cycles. for administrative efficiency and economies of scale. • Manual payment processes are more time consuming and • Social protection payment processes are increasingly being labor intensive to administer than digitized ones, which can digitized through the use of card-based and mobile technol- limit the opportunity to scale up or provide flexibility in pay- ogy. This provides opportunities for real-time transactions to ment schedules and amounts. large, dispersed, and inaccessible populations, with increased • Digital payment services are still emerging and expanding potential for last-mile delivery and for supporting mobile and do not cover all the population, especially in rural areas. and displaced populations. Digitization introduces greater The financial and digital illiteracy of a population also can lim- transparency, with an audit trail from source to beneficiary. it the effectiveness of digital payment mechanisms. It also facilitates varying the distribution schedule and trans- fer values according to changing needs, through automated • Complex and bureaucratic payment processes can contrib- processes. ute to bottlenecks and delays. :: 42 :: The delivery chain and shock response the payment service provider and the government may in- process in the delivery chain is automated or desk based, the clude the following: mechanics will be relatively unaffected by the shock. Where such changes imply a greater workload for payment service ● Digital payment channels could be temporarily replaced providers or program staff, however, their capacities to imple- with manual payments until services come back online. ment them should be taken into account, and, where possible, This will be contingent on having sufficient human resourc- a commitment to preposition capacity to fill the gap should es to manage a manual payment process. be written into payment service provider contracts ex ante. ● Where necessary, payment service providers and the cen- tral bank could undertake special measures to move phys- Key considerations for benefits provision in shock ical currency from national or regional headquarters to re- response gional offices and agents in the disaster-affected areas to ➤ Preestablish the capacity needed to pay adjusted benefits ensure liquidity for an increase in manual cash payments. packages (vertical expansions) ● Payment service providers could set up payout points in In the cases of horizontal and vertical expansion of an existing more accessible and secure areas. program, payments may have to be more frequent to meet ● Payment service providers and government and humani- immediate post-shock needs. An example of this took place in tarian actors could introduce measures to ensure security Ecuador in response to the 2016 earthquake (box 3.2). These at payout points. decisions (which are made in phase 6, benefit determination) may increase the workload of payment service providers and Where payment service provider staff are directly affected administrative staff, especially if payments are made manual- and incapacitated by the disaster, or where a switch to manu- ly, which could hamper the program’s day to day operations. al payments is required, support staff can be brought in from Indeed, the schedule may have been set up with less frequent other regions to address this gap. payments to account for the capacities of those making them. In line with decisions made in phase 5, determine benefits Government or other humanitarian actors could support the package, the payment schedule of the existing program may requisite capacity building needed, although it might be bet- need to be modified to ensure beneficiaries receive their reg- ter to retain the existing payment schedule if changing it risks ular payments in a manner that best supports their resilience doing harm. to the shock. This may include bringing scheduled payments forward to coincide with the period of greatest need (for in- stance, in relationship to lean or rainy seasons) or making pay- Box 3.2: Changing payment schedule: Ecuador ments more frequently to support consumption smoothing, In Ecuador, in response to an earthquake in 2016, cash as was done in Mozambique (see box 3.1). Since much of this transfers were delivered through public and private banks and cooperatives—a solid mechanism with the advantages of having large coverage and low costs, along with being Box 3.1: Modifying the payment schedule: administratively simple and user friendly. The one hurdle that Mozambique delayed disbursal of transfers was the digital platform for delivering the payments, which was not designed to enable In Kenya, prior to the rollout of phase 2 of the Hunger In multiple top-up transfers to the same recipients. Once the Mozambique, the main safety net program—the Basic issue was resolved, payments to affected populations were Social Subsidy Program—regularly experiences delays in the smooth and timely, thanks to the systems in place and the disbursement of payments at the start of each new financial people’s trust in them. year in January. This coincides with the period of greatest risk of climate shocks, such as cyclones, when beneficiaries Source: Beazley 2017. are most in need of financial assistance. The Department for International Development’s (DFID) research on the potential of shock responsive social protection in ➤ Adapting payment processes to provide top-ups to existing Mozambique recommended a simple modification to the administrative procedures for payments that would beneficiaries or assistance to newly enrolled ones introduce a double payment in December in place of the Before additional payments can be started—whether top- one in January, to help ensure households receive support ups for existing beneficiaries or payments to more beneficia- when they are most vulnerable. ries—revising underlying financial procedures may be neces- Source: Kardan et al. 2016. sary to outline the source of and the flow of funds and any reconciliation requirements. This will be of great importance ADAPTIVE SOCIAL PROTECTION :: 43 :: when shock response activities are to be funded by separate programs to noncitizens or refugees, specific capacity issues budget envelopes from development or humanitarian part- may need to be addressed—for example, language barriers, as ners, as their financial regulations may require reconciliation seen in Turkey (box 3.1). processes different from those of the existing program. This Something else to consider is the workload involved if trans- process can be time consuming to complete ex post, as illus- fer values vary—according to household size or level of need, trated by experiences in the 2016 El Niño response in Leso- for instance. Variation adds an additional layer of complexity tho (box 30). The design of the top-up was only elaborated with implications for the time needed to make and reconcile at the time of the crisis, rather than being an integral part of the social protection administrative processes. This hindered the payments, especially where these are made manually. smooth implementation of the shock response. Top-up assis- Digital payment systems are advantageous here, as process- tance was only disbursed six months after drought was de- es are more automated and additional workload is minimized. clared—four months after the implementation of standalone Their usefulness, however, relies on the ability of any underly- cash assistance programs. ing MIS or IT platform for payments to manage these chang- es. Box 3.3 details some of the challenges faced in Nepal in The introduction of top-up payments or payments to new this regard. beneficiaries can overburden staff or systems. This can delay emergency payments and also undermine long-term social Also important to bear in mind is the need for additional li- protection programs. Providing top-ups to existing beneficia- quidity, especially at the local level. Even programs that are ries as part of the existing payment rather than as an addition- implemented through digital payment systems generally need al payment will minimize additional workload. Governments to have a “cash out” function at ATMs or through agents. The and their contracted payment service providers should assess payment service provider (and/or its affiliated agencies) will the added workload for program administrative staff and for need to compare the expected volume of transactions to the frontline staff and affiliated agents of the service provider and usual volume of currency that “cash out” services (cash desks, whether they have the capacity (in terms of systems, office ATMs, or their registered agents) will handle and increase pro- space, and personnel) to take on the additional payments, visions accordingly. This may mean developing plans to move while ensuring reconciliation and audit. When expanding currency from headquarters and regional hubs to provincial Box 3.3: Importance of ensuring capacity for payments: Nepal and Turkey In Nepal, social protection payments are made by Village Development Committee (VDC) and ward secretaries, who were enlisted to take on the administration of the vertical expansion in response to the 2015 earthquake. These officials, whose institutions suffered from high turnover and a shortage of staff in general, were involved in coordinating other projects of humanitarian actors in their communities, as well. The scaling up of the social protection programs placed additional responsibilities on this already overworked government staff, who were frustrated that national government actors and UNICEF had not adequately considered their capacity to deliver the additional funds. An evaluation concluded that assessing the real capacity of administrative processes and staff and providing the necessary support to ensure successful delivery of these payments would have improved effectiveness significantly. More generally, in Nepal the intention had been to synchronize the top-up payment schedule with the payment of the regular transfers in June, so as not to burden further the actors involved. This happened for only around half the beneficiaries, however, because of delays in VDCs’ receiving the funds, which, in turn, were the result of errors in transfers to the VDCs brought about by the complexity of the administrative processes to transfer funds to District Development Committees (DDCs) and the low capacity of the DDCs and banks to deliver the payments. In several districts, payments for the Emergency Cash Transfer Program took place after the regular social assistance payments had already been made. Some VDCs made payments as additional one-off transfers, meaning staff incurred additional workload. Others waited until the next round of regular payments (in October) to disburse the emergency top-up, meaning beneficiaries didn’t receive their emergency payments until five months after the earthquake. In Turkey, problems arose in making payments to Syrian refugees from the ESSN and CCTE, as the language barrier make it difficult for the staff of the payment service provider, Halk Bank, to communicate with the beneficiaries. When the issue came to light, Turkish Red Crescent placed dedicated program staff for the ESSN and CCTE directly in bank branches to help the bank staff make the payments. Source: EUD 2019b; Merttens et al. 2017. :: 44 :: The delivery chain and shock response and district distribution networks. The larger the scale of the disaster and size of the response, the more important this be- Box 3.5: Modifications to payment interface for new comes. Regardless of their frequency, setting a regular day of beneficiaries: Turkey and Yemen the month on which emergency payments are to be made will be useful to assist both communication to beneficiaries and In Turkey, to facilitate payments to refugees enrolled in the the planning activities of payment service providers and their ESSN program, the payment service provider has updated all ATMs to include an Arabic language function. Although agents (see the experiences of Kenya in box 3.4). this change has improved accessibility for Syrian refugees, the new language function was not set as an additional op- tion, for selection by the user, but rather as an automatic function of the service, applied for all ESSN cardholders. Box 3.4: Ensuring liquidity for shock response This has created problems for non-Syrian refugees who do payments: Kenya not speak Arabic. In 2015, the HSNP in Kenya horizontally expanded in re- In Yemen, when transfers were made to vulnerable and sponse to drought. Two rounds of emergency payments food-insecure households in enclaved areas affected by the provided a proof of concept of HSNP’s ability to scale up civil war, the (private sector) payment service provider for coverage rapidly. They also showed, however, that signifi- the Social Welfare Fund was able to move money discrete- cant capacity on the part of the payment service provider ly into and within the enclaved areas. Staff selected payout and its agents is required to deliver emergency payments points that were accessible to the affected communities over a wide and remote area. Payment agents outside of (especially women) and set up temporary payout points in county capitals had significant problems maintaining liquid- more secure community spaces. They also conducted home ity during the scale up. An evaluation of the HSNP con- visits for those unable to go to the payout points. ducted by Oxford Policy Management (OPM) found that Sources: EUD 2019c; CaLP 2018. advance planning by bank branches was required to ensure enough cash would be available ahead of payment disburse- ment dates. A review of the emergency expansion of the HSNP in Kenya recommended that a fixed payment date per tion) is in place. Again, remuneration agreements may need month be agreed on for all emergency payments. updating to reflect this additional work. Source: OPM 2015. ● Where new beneficiaries are unfamiliar with the payment system (for example, the use of ATMs or of mobile devices to make transfers), governments should ensure sufficient ➤ Modify the payment interface to ensure new beneficiaries support is provided at the point of transaction, perhaps by can conveniently and safely access payments. personnel of the payment service provider, program staff, Depending on the characteristics of the newly targeted ben- or both. Again, this has implications for staff time and bud- eficiaries, the existing payment processes and systems may gets that must be thought through and for which capacities need to be modified in the case of standalone emergency must be built. programs, piggybacking, and horizontal expansion to ensure ● Where the payment interface (for example, the mobile they can easily receive their transfers. Examples of how this money operating system, the ATM menu, or SMS notifica- can be done are provided in box 3.5 and include the follow- tions of payments) presents language or technological bar- ing: riers to new beneficiaries, governments can ask payment ● If new beneficiaries are located in new geographical areas service providers to provide ex ante training or services in (potentially far away from existing payout points) or where an additional language. This may take a little time to put mobility is restricted, payment service providers may need into place but is a one-off investment. Such services should to establish new payout points in convenient, secure loca- be designed with caution to ensure the changes for one tions or provide doorstep services. Service providers must beneficiary cohort do not reduce service accessibility for consider the workload implications of doing so and ensure another. requisite capacity (in terms of personnel and transporta- ADAPTIVE SOCIAL PROTECTION :: 45 :: CASE STUDY: The “Provide” phases in the Philippines The Philippines has a relatively advanced social protec- eas. The capacities of some service provider branches was tion system that has been used several times to respond to also reduced, since personnel and infrastructure had been shocks. The flagship cash transfer program is the Pantawid affected. Pamilya Pilipino Program (or Pantawid), a nationwide con- Challenges in adapting the payment schedule: The Pantaw- ditional cash transfer whose aim is to alleviate poverty and id program usually provides payments every two months. At improve the health, nutrition, and education of poor children. WFP’s request, this was changed to every month during the The program is implemented by the Department for Social shock response to align the schedule with that for emergency Welfare and Development (DSWD), and it reached over assistance being provided to other households through the 4.4 million households in 2015. In 2013–14, Typhoon Hai- international humanitarian system. The increased frequency yan affected some 16 million people across nine provinces. in payments, however, created additional work for the pay- The DSWD, in partnership with the World Food Programme ment agents, who had to prepare, implement, and reconcile (WFP), delivered a vertical expansion to the Pantawid pro- the extra distribution at a time when staff capacity was al- gram, providing cash top-up payments to Pantawid benefi- ready stretched. This and the direct damage to the busi- ciaries in 60 “worst-affected” municipalities. The intervention ness from the typhoon put PHLPost (the Philippines Postal required adaptations to the business process for providing Corporation) behind on reconciliation reporting, leading to benefits and generated several lessons in the following areas temporary suspension of its payment conduit license. A les- (Smith et al. 2017): sons-learned study concluded that sticking with the original Addressing disruption to the payment system from the payment plan of the Pantawid program would have reduced shock: Pantawid program beneficiaries can receive their pay- the burden on the social protection system. ments either by using ATM cards or as cash over the count- The importance of adapting financial procedures where er. After the typhoon, power outages prevented the use of needed: The reconciliation procedure for benefit payment in the ATM payment channel for several weeks. Some benefi- the Pantawid program includes an acknowledgment receipt ciaries also lost their ATM cards, and replacing them would that each beneficiary signs. These are produced in triplicate, have taken several months. The lead payment service pro- with one copy retained by the beneficiary, one by the pay- vider, Land Bank, provided three mobile ATMs to help dis- ment service provider, and one by the social welfare depart- burse cash payments to beneficiaries in affected areas and ment for submission to the Commission on Audit. Upon the switched from ATM payments to cash payments over the vertical expansion of the Pantawid program, WFP’s financial counter while ATM services were reduced. Payment service procedures required that it also receive a copy of the acknowl- providers managing the over the counter cash payments also edgment receipt. This was not made clear to the government changed the location of some payout points to ensure they at the beginning, meaning paperwork was already filed with were still accessible to households post shock. This ensured the Commission on Audit. It took administrative staff over a payments were received without significant delays. As these year to address this retroactively. adaptations were not defined in the Pantawid program pro- Revising the remuneration structure for payment service cedures, however, they took some time to put into place, providers: Although the top-up payments made during the making the delivery of shock response payments later than shock response entailed considerably more work than regular originally planned. payments, the payment service providers managing them for Dealing with capacity challenges: The payment service pro- the Pantawid program still received the same transaction fee. viders managing the cash payments faced some challenges. The memorandum of understanding between DSWD and Ensuring the availability of physical cash in the first months Land Bank was later amended to reflect this lesson and adjust was difficult until Central Bank directed it to the affected ar- their fee structure for future shock responses. :: 46 :: The delivery chain and shock response SECTION 4: MANAGE ASSESS ENROLL PROVIDE MANAGE Eligibility Determination Provision of Beneficiaries Exit decissions, Assessment and of benefits Notification benefits compliance, notifications, Intake of of needs and enrollment and service and and/or updating, and and case Outreach registration conditions decissions package onboarding services grievances outcomes RECURRING CYCLE PERIODIC REASSESSMENT Photo: Mohammad Al-Arief/The World Bank ADAPTIVE SOCIAL PROTECTION :: 47 :: Beneficiaries Compliance, Updating, and Grievances Is the program functioning effectively in response to post-shock needs, or are adjustments needed? O verall, a program aims to monitor each individual management systems, which require a cadre of trained case, ensuring a beneficiary’s needs and situation caseworkers or social workers operating across a country)7 are assessed continuously and addressed appro- Monitoring and compliance data collected during this phase priately. This eighth phase, sometimes referred to feed up and down the delivery chain. They are used down broadly as “beneficiary management” or “case management,” the chain to inform periodic reassessment of needs and con- comprises three functions: ditions and subsequent eligibility, as well as for updating the 1. Beneficiary data management (including monitoring ben- benefits package based on changing needs and conditions. eficiaries’ receipt of benefits, progress, and continued fit They feed up the chain to trigger exit decisions based on with eligibility criteria and any changes in circumstances changes to eligibility status. Data from grievance mechanisms that may render them ineligible, which may require period- also feed down the chain to inform the design and implemen- ic collection of new household profiling data) tation of delivery processes. 2. Monitoring compliance with program conditions (if appli- These systems and processes of cash transfer programs have cable) and conditions the potential to support shock response. Table 4.1 summa- 3. Collecting, assessing, and reporting on appeals and griev- rizes the benefits, which include established networks of ances raised through grievance mechanisms (may also in- staff and predefined grievance mechanisms, and limitations, clude providing tailored guidance and support, though few among them capacity and infrastructure constraints that may LICs and MICs have fully established and functional case affect staff networks and grievance mechanisms post shock. Table 4.1: Strengths and constraints of routine beneficiary data management, compliance verification, and grievance redress systems and processes for shock response Strengths Constraints to be aware of • Household-level monitoring is carried out through an • Where administrative budgets are limited, cash transfer programs established network of social workers and other ad- can be understaffed, and time available for beneficiary operations ministrative staff, often based close to and trusted by management tasks can suffer as a result. This is a bigger issue communities. This provides a vital human resource to where social welfare staff are tasked with managing cash transfer support program follow up and monitoring of shock re- payment and reconciliation processes and will limit ability to take sponse activities. on additional tasks during shock response. • Grievance mechanisms often provide multiple commu- • Channels for communicating complaints are designed for the nication channels for receiving and responding to que- needs of routine cash transfer programs in normal times. They ries and complaints from beneficiaries and the wider may not be accessible to or trusted by those affected by a shock. community. Where available for grievance redress, dig- • Monitoring compliance with program conditions can impose a ital communication channels can ensure reach across heavy administrative burden on the cash transfer administrative dispersed and inaccessible areas, while person to person staff and related institutions (schools, clinics), especially where communication channels are generally trusted by com- data have to be collected and entered manually. These institu- munities. These mechanisms provide a means of under- tions are likely to become further overburdened at times of shock. standing targeting errors, as well as issues with wider program processes. This phase and phase 9 are sometimes referred to broadly as “case management.” Together they form part of the beneficiary operations manage- 7 ment stage. Beneficiary operations management denotes the activity of continuously engaging and collecting information from the field or other sources (such as other databases), which is then processed through a set of protocols, recorded, and used to make decisions. :: 48 :: The delivery chain and shock response Ensuring continuity of the beneficiary compliance, Key considerations for the beneficiary compliance, updates, and grievance processes updates, and grievance processes Shocks can disrupt program infrastructure, damaging offices ➤ Where appropriate, temporarily suspending monitoring of and communication lines, affecting administrative staff, and conditions on existing programs rendering communities inaccessible, all of which can affect Cash transfer programs usually include conditions or co-re- the ability to continue household-level monitoring and com- sponsibilities relating to attendance of beneficiaries at school, pliance activities, which are particularly staff intensive. Where their use of health services, and similar activities to support staff and offices are directly affected by a disaster, measures long-term human development objectives. Adherence to for service recovery include bringing in support staff from these conditions is often not feasible or appropriate in the other regions and relocating operations to temporary sites. If aftermath of the crisis. Schools may be closed, for instance; any telecommunication services used in monitoring or griev- health offices only attending to emergencies; and behavioral ance processes are disrupted or damaged, the communica- change sessions within the community difficult or undesirable tion service provider and the government will need to act to to arrange. Additionally, administrative capacity to monitor restore them. conditions is constrained. Where new beneficiaries are add- Where more than one communication channel is used for ed, they may also face different, higher barriers to access to beneficiary monitoring and grievance processes, programs such services. This will certainly be the case for refugees and can focus on that which is most accessible post shock. Where noncitizens. It is advisable to waive the enforcement of the hotlines cannot function, priority can be placed on using face cash transfer conditions whenever these factors exist during to face channels, provided staff capacity for this task can be crisis periods. assured and any negative impacts on other aspects of pro- Programs can implement procedures to suspend the usual gram implementation can be avoided. Where communities processes for continual reassessment of eligibility, or adher- and administrative offices are inaccessible, phone communi- ence with conditions, during crisis periods for several reasons cation may be a useful means of continuing monitoring ac- (Bastagli 2014; Beazley et al. 2016; OPM 2017; O’Brien et tivities. This will require having up to date contact details of al. 2018): beneficiaries in the MIS. ● Where households are displaced, infrastructure damaged, Alternatively, it may make sense to waive certain activities or service providers affected by the shock, the enforce- temporarily. These include the monitoring of conditions or ment of conditions may no longer be appropriate. household eligibility. Likewise, if a hotline is the sole chan- nel for beneficiaries to provide feedback or raise queries and ● Suspension will ensure beneficiaries in the affected areas complaints, program managers may need to suspend the are not penalized or made to exit the program but will con- grievance mechanism until digital communications are re- tinue to receive their regular payments, so as to ensure re- stored. Beneficiaries should be notified of any changes and silience to the shock. informed of alternative, temporary mechanisms for commu- ● It can reduce the burden of labor for program administra- nicating grievances. Key considerations for beneficiary data tors and social service providers involved in monitoring management, compliance verification, and grievance redress adherence to conditions and continued eligibility at a time for shock response. when capacities can be overstretched, meaning more time can be given to post-shock program recovery. Having these modifications built into program SOPs, as is done in Mexico (box 4.1), and a digital MIS that automatically applies the modified rules, will help ensure a smoother tran- sition. Where social protection laws or regulations stipulate the use of conditions, it may be necessary to insert clauses that allow their relaxation under specific conditions of disas- ter, as was the case in the Philippines (box 4.1). ADAPTIVE SOCIAL PROTECTION :: 49 :: Box 4.1: Waiving conditionality following a shock: Mexico, the Philippines, and Turkey In Mexico, the former Prospera program guidelines allowed for conditions in a locality or municipality to be temporarily waived in the event of a natural disaster, social or epidemiological contingencies, or any declared emergencies that prevented health and educational services from being provided. The waiver is triggered by the Home Ministry’s declaration of a state of emergency or when the program’s state technical committee makes a request to the Prospera National Coordination Office. In the Philippines, transfers from the Pantawid program are usually conditioned on regular school attendance and health checks for children and pregnant women and attendance of parents at monthly family development sessions. In 2013, DSWD passed a resolution that when a state of calamity was declared, program conditions would be waived for three months. This was included in the program operational procedures and was activated following Typhoon Haiyan to ensure affected families would not be penalized and would receive assistance during this time. In Turkey, for the Conditional Cash Transfer for Education for Refugees, the government decided the condition of 80 percent school attendance should continue to be implemented for both refugee beneficiaries and Turkish citizens. Refugee children, however, face greater barriers to gaining access to education. The Turkish government’s partner, UNICEF, is addressing this through complementary case management activities alongside the cash transfer program, to identify and support those at risk of dropping out of school. Sources: Smith et al. (2017); Beazley et al. (2019); SEDESOL (2017c). Emergency programs can also apply conditions to beneficia- ed to support and monitor enforcement, and access must be ries. These are less common for programs that aim to meet provided to any underlying MIS for the recording and sharing food and other basic needs but quite common for others, of data. This was the case with the CCTE in Turkey, where ver- such as shelter rehabilitation and livelihoods recovery. Here, ifying compliance with the educational attendance condition further installments of the benefits package are conditioned required schools to input attendance data into the Ministry on the beneficiary’s having undertaken certain activities (for of Education’s MIS. Since this MIS links with the Integrated example, having completed a specific construction task or Social Assistance System (ISAS), the government could auto- attended relevant training). Adapting existing cash transfer matically screen for CCTE beneficiaries who did not meet the program processes and systems to accommodate verification attendance condition before each payroll. The same proce- of such new conditions ex post is likely to be difficult, as they dure was followed for the CCTE for Refugees with refugee will not be part of the existing MIS or program rules and will children who attended the Turkish public schools. The staff require time and effort to modify. of the temporary education centers where refugee children were enrolled had to support this monitoring activity. These While these changes could be incorporated ex ante, the po- schools use a separate MIS, called “YOBIS”—a standalone da- tential benefits of linking scaled-up transfers to such condi- tabase for refugees that needed to be integrated with ISAS tions need to be weighed against the costs of making these before school attendance could be verified and payments changes and of monitoring and enforcing the conditions. made (EUD 2019b). Where existing cash transfer programs are to be expanded in emergency times, therefore, the introduction of conditions ➤ Prepare the grievance redress mechanism to manage an should be avoided. If the nature of the emergency support increased volume of complaints and enquiries. really requires their use, such needs may be best met through In the cases of horizontal expansion and piggybacking, the a standalone program (which could piggybacked on other sys- number of beneficiaries will increase. This may increase the tems of the cash transfer program). volume of grievances and calls for sufficient staff and data management and redressal systems that can effectively take ➤ Ensure new beneficiaries can be effectively included in on this additional work without creating backlogs. In any monitoring and grievance activities. scale up for emergency response, the nature of the grievanc- In the case of horizontal expansion of a program, or where a es will also change (although they will still include appeals new program is piggybacking on an existing program’s pro- against exclusion and complaints about implementation), and cesses and systems, these may need to be modified to moni- staff must be prepared to handle these. The capacity of the tor and capture data on new beneficiaries effectively. Where existing grievance redress mechanism (GRM) could be im- conditions are to be enforced, for example, new institutional proved by hiring additional staff, providing training, providing partnerships with links to the new population may be need- support staff through humanitarian partners, and introducing :: 50 :: The delivery chain and shock response data management systems to streamline the process. In Ye- ➤ Adapting grievance mechanisms to ensure they are men, for example, for the Emergency Cash Transfer Program accessible to new, shock-affected beneficiaries linked to the Social Welfare Fund, UNICEF supported the es- In the case of horizontal expansion or piggybacking, the com- tablishment of a new grievance mechanism on a digital MIS munication channels used in the grievance mechanism must and trained SWF staff in how to use and manage it (Smith be accessible to the new beneficiaries. Barriers to access in- 2017). In Turkey, a new grievance mechanism was set up to clude the following: manage queries and complaints of refugees in the CCTE and ESSN programs, staffed by Turkish Red Crescent to reduce ● Lack of physical access to the offices of staff managing the the burden of labor on the cash transfer program staff (EUD grievance mechanism 2019b). Such measures have budget implications and may ● Language barriers (especially relevant in the case of refu- take time to establish. gees) Similar issues will arise for other aspects of monitoring: where ● Marginalization of or discrimination toward new beneficia- new beneficiaries are added, workloads will increase for the ries by staff managing the grievance mechanism or lack of staff involved. The capacity of the institutions involved within trust on the part of the beneficiaries the social protection system and their staffs to take on these As shown in box 4.3, several modifications to improve access further responsibilities must be considered. Capacity can be can be made, including setting up program staff in locations built through support provided by the government or its hu- close to new beneficiaries; establishing phone hotlines to manitarian partners—for example, through additional human reach dispersed and isolated populations, such as in Yemen resources, training, and administrative support and vehicles. (EUD 2019c); translating feedback forms; providing staff or If capacity cannot be assured, it may be better instead to lim- call center operators who are fluent in relevant languages, as it the requirement for monitoring activities—by temporarily is done in Turkey (CaLP 2018); and recruiting new, trusted waiving conditions and eligibility screening, for instance. The organizations to participate in grievance mechanism opera- importance of addressing monitoring capacity gaps is illus- tions. Any new channels must be understood and trusted by trated in box 4.2. and convenient for beneficiaries. Box 4.2: Challenges of monitoring after shocks: Box 4.3: Adapting grievance mechanisms to ensure Kyrgyzstan and the Philippines they are accessible: Nepal and the Philippines In Kyrgyzstan, social welfare officers of the State Agency In Nepal, the grievance redressal system of the national for Social Welfare (SASW) did not practice a “case manage- cash transfer system has people communicate their com- ment” approach prior to the conflict of 2010. As part of its plaints directly to VDC or ward secretaries. During the hori- support during the crisis, UNICEF provided skills and meth- zontal expansion of these programs in response to the 2015 ods training and coaching to social protection managers and earthquake, a toll-free phone number and SMS platform social workers on additional outreach measures to ensure were also introduced; however, beneficiaries generally pre- family welfare. They introduced new documentation—a care ferred to use the traditional and familiar approach of com- and support plan for the family—for monitoring needs, re- municating with VDC and ward secretaries. ferrals to services, and progress. This monitoring approach In the Philippines, during implementation of the national was subsequently adopted by the government. Emergency Cash Transfer following disasters in 2016 and In the Philippines, additional responsibilities taken on by so- 2017, grievances could be reported through a range of cial welfare officers to implement the emergency cash trans- channels, including through a hotline, social welfare staff, fer placed extra strain on personnel and other social welfare and local government officials. Grievance desks set up activities. The DSWD’s provincial offices were not provided during the distribution of cash cards and at the DSWD re- with extra equipment or operational budgets to fulfill the gional office were staffed by assigned grievance officers. All administrative requirements of the emergency cash transfer grievances were recorded and urgent ones elevated to the and had to cover these costs from existing budgets. program management team for action. Grievance resolu- tion included home visits and mass information broadcasts Sources: EUD 2019a; Smith et al. 2017. through various media. Sources: Merttens et al. 2017; Government of the Philippines 2019. ADAPTIVE SOCIAL PROTECTION :: 51 :: Exit decisions, notifications, and closing cases When should the shock response program be wound down? Who needs ongoing long-term support? A major decision in beneficiary data management is beneficiary. Exit triggers are changes in a beneficiary’s basic when to move beneficiaries out of the cash transfer information or compliance metrics that indicate the benefi- program. Beneficiaries may exit from a program be- ciary has fallen out of the program’s eligibility parameters. cause they have completed it (where the program Triggers can be built into the MIS as “red flags” that initiate has a predefined and timebound duration); had a change in the subsequent steps. They may be identified automatically the condition required for entry to the program—for exam- during scheduled updating of the MIS (for example, the age ple, in poverty status, employment, disability or marital status, of the beneficiary); through interfaces with other government residential location, or life cycle status (age/death)—meaning databases (for example, civil registration databases); through they no longer meet the program’s criteria; or have not com- new data collection under program beneficiary management plied with program rules and conditions. Well-functioning processes (including reassessment); or through on-demand beneficiary registries are dynamic tools, regularly incorporat- updates initiated by beneficiaries. ing new beneficiaries throughout the enrollment phase and These systems and processes for social protection programs moving out those who need to leave the program. have the potential to support the application of shock re- The steps in the process of moving beneficiaries out of the sponse. As summarized in table 4.2, there are also constraints beneficiary list and stopping assistance include identifying to be aware of concerning staff workloads and the need to an exit trigger, followed by reassessing the beneficiary’s el- revise exit rules with relationship to shock response. igibility criteria, making an exit decision, and notifying the Table 4.2: Strengths and constraints of routine exit and notification systems and processes for shock response Strengths Constraints to be aware of • Routine systems and processes can immediately flag and • Periodic verification of eligibility (unless the criterion is one that notify ineligible beneficiaries of exit from the program to can be automatically verified by the MIS, such as age) is likely to ensure the response focuses on the most urgent needs. require some level of new monitoring and data gathering. This has implications for staff time and workload in the compliance verification, updating, and grievance phase, as well as program budgets. • The rules governing assessment and exit are set up according to the requirements of a long-term cash transfer program (per its objectives), which may mean beneficiaries are moved out of the program while still in need of emergency assistance. :: 52 :: The delivery chain and shock response Ensuring continuity of processes for exit decisions, notifications, and closing cases during or after a Box 4.4: Standardizing terms of exit and scale down: shock Mozambique, Nepal, and the Philippines The communication channels used to notify beneficiaries of In Mozambique, the operational manual guiding imple- exit can be modified, if necessary, to ensure they continue to mentation of the government’s emergency cash assis- be accessible to the beneficiaries post disaster. This involves tance Post Emergency Direct Cash Transfers Program the same considerations set out earlier under the outreach includes standard guidance on criteria for beneficiary exit. The program provides emergency assistance to all phase and the notification and onboarding phase. Where beneficiaries for six to twelve months, with the specific programs implement procedures to suspend elements of eli- duration defined at the start of the intervention based gibility monitoring during crisis periods so as not to overbur- on severity of the disaster. Beneficiaries can also be den staff and ensure continuity of support for those already moved out of the program during this period if they have enrolled, these exit and notification activities will also be sus- a change of residence (to places outside the program pended. implementation area) or do not collect two consecutive payments. Key considerations for exit decisions, notifications, In Nepal and the Philippines, all beneficiaries of the ver- and closing cases for shock response tical expansions (for the 2015 earthquake and 2013 Ty- phoon Yolanda, respectively) received assistance for the ➤ Adapting the terms of exit or scale down following the same length of time for ease of implementation and to shock response avoid creating tensions. For the more vulnerable house- holds (for example, those with more dependents), some For a routine cash transfer program, assistance is generally needs remained when this assistance ended. provided until beneficiaries no longer meet the eligibility cri- Sources: Smith et al. 2017; Merttens et al. 2017; Government of teria for it. Eligibility may be lost because of a change in a Mozambique 2019. household’s income or poverty status or a beneficiary’s de- mographic characteristics (such as age or disability status). Change in eligibility is sometimes linked to duration of partic- have recovered equally at the point of exit. Box 4.4 highlights ipation in the program, with a reassessment scheduled after experiences of implementing standardized exit processes in a specific interval. Mozambique, Nepal, and the Philippines. Similarly, emergency assistance should be provided until An alternative, more resource-intensive approach is to move households have been able to meet their immediate basic shock response beneficiaries out of the program after veri- needs in the short term and/or have sufficiently recovered fying some change in their vulnerability status. In some con- from the crisis in the medium to long terms. Depending texts, the underlying program MIS and registry will already on the purpose of the emergency assistance or short-term contain relevant indicators of this. If the status can be regu- relief and/or the objective for medium- to long-term re- larly updated—for example, through data exchange with oth- covery, beneficiaries should remain in the program until er government departments, as in Turkey (box 39)—this will they are considered to have met these criteria. In practice, serve to screen and verify continued eligibility automatically however, adequacy of funding will often limit the compre- and periodically. Such a process could be feasible in countries hensiveness of the intervention. This means eligibility veri- where these underlying data management systems are well fication checks and criteria for exit from a social protection developed. Alternatively, rather than verifying vulnerability shock response program will differ from those of a routine at the household level, transfer duration could be linked to cash transfer program. changes in geographical vulnerability indicators—for exam- The simplest approach for moving beneficiaries out of a re- ple, changes in environmental indicators or according to the sponse program is to implement the program for a predefined severity of the disaster in different areas, as has been done period of time that is communicated at the outset. This ap- in Kenya (box 4.5). This is less time and labor intensive, pro- proach is easiest to communicate to beneficiaries. Where vided systems are in place within or outside of government financial resources for shock response are limited, this can to make these assessments and data can be entered into the also be a pragmatic approach in practice, as opposed to un- program’s MIS to inform calculation of the benefits package. dertaking reassessment of needs and conditions. Since dif- To avoid confusion or social tensions, there must be confi- ferent households will follow different trajectories for recov- dence in the accuracy of the data and their usefulness as an ery, this approach means not all vulnerable households will indicator for household vulnerability. ADAPTIVE SOCIAL PROTECTION :: 53 :: ➤ Moving the shock response beneficiaries to long-term social protection programs if they match the routine eligibility Box 4.5: Verifying changes in vulnerability status for criteria exit or scale down: Kenya and Turkey In the case of horizontal expansion, what should happen with During the horizontal expansion of the HSNP in Ken- newly enrolled beneficiaries who also match the regular el- ya at times of drought, the government of Kenya’s Na- tional Drought Management Agency (NDMA) monitors igibility criteria for the long-term cash transfer program? If drought conditions by satellite. Cash transfers are paid the government has the will and necessary financial capaci- monthly in the month after a drought is forecast via this ty, these cases could be enrolled permanently in the regular satellite data, according to agreed-on vegetation condi- program. This has been done in several shock responses to tions being reached. The duration of the scaled-up pay- date (box 4.6). Any such cohort would need to be flagged to ment is directly linked to monthly data provided from the Vegetation Condition Index (VCI) which builds from ensure its members stay enrolled when the shock response these satellite data. This means emergency payments payments end. Such design decisions would also need to be continue to eligible households as long as at least one communicated carefully to all new beneficiaries so it is clear subcounty has reached “severe” or “extreme” VCI status. who will be moved out of the program and who will receive In Turkey, the eligibility of routine social assistance longer-term assistance. beneficiaries is verified through the Integrated Social Assistance System (ISAS). This links to data on citizens held in various government databases and automatical- Box 4.6: Incorporating New Beneficiaries: Ethiopia, ly screens for changes in households’ income, access to job-related social security, and assets. The DGMM da- Kyrgyzstan, and Mozambique tabase of refugees was integrated into ISAS to enable In Ethiopia, the expansion strategy of the Productive verification for beneficiaries of the ESSN program. Safety Net Program is directly linked to previous shock Sources: EUD 2019b; Farhat et al. 2017. relief assistance. As defined in the 2014 manual, for ex- ample, new woredas (districts) are added every five years if they have received shock relief assistance in three of In the case of vertical or horizontal expansion, the usual el- the five years preceding, while new kebeles (wards) are added if they have received recurrent food assistance for igibility reverification processes may need to be waived for at least three of the past five years. the period of the response to ensure no vulnerable household In Kyrgyzstan, beneficiaries enrolled as a result of the gets exited from the regular program during the post-shock 2010 conflict met the eligibility criteria of the regular period and loses continued access to assistance at a critical cash transfer programs and were therefore enrolled time. With poverty-targeted programs for which verification permanently by the government. The fact that the gov- involves an infrequent poverty survey, this will be less of an ernment was financially responsible from the outset for issue. It will be more important to address with programs providing the transfers from the national budget contrib- uted to this sustainability. for which such screening is automated within the MIS and periodically carried out (for example, those that base reas- In Mozambique, the operational manual guiding imple- sessment and exit on age). In the latter case, any automatic mentation of the government’s Post Emergency Direct Cash Transfers Program states that beneficiaries of the screening and exit procedures in the program MIS will need emergency assistance can be integrated into other long- to be turned off for the post-shock period. If such adaptations term social protection programs at time of exit, providing are not possible within the program regulations, coordination they meet the criteria and depending on the availability and referral procedures must be put into place so that any of these programs in the beneficiaries’ area of residence. households removed from the program mid-response (for For this purpose, the National Institute for Social Action (INAS) must complete the registration of household data example, when a child reaches the maximum age) will be di- as per the INAS program forms and assess needs and rected to other assistance, such as programs implemented by conditions by applying the PMT. humanitarian partners. Sources: EUD 2019a; Bowen 2015; Government of Ethiopia 2014; Government of Mozambique 2019. :: 54 :: The delivery chain and shock response Summary: Key Considerations In summary, table S.1 highlights the key considerations for effective shock response at each stage of the delivery chain, based on which table S.2 provides a checklist of key questions to consider when planning to use or adapt the cash transfer delivery chain for shock response. Table S.1: Effective shock response: key considerations across the delivery chain Issue to address at this Stage Phase phase of the chain Key considerations for effective shock response Outreach How will shock-affected ● Routine messages must be modified for shock response out- households be informed they reach activities. may be eligible for support? ● Outreach mechanisms for delivering modified messaging must be accessible to new, shock-affected beneficiaries. ● Internal communication with and training for those involved in delivering shock-response outreach messaging will be needed. Intake and How should information on ● Registration processes and tools may need to be adapted to col- registration shock-affected households lect new data on disaster vulnerability. be gathered to assess their ● Registration processes may need to be modified to speed up needs and eligibility? registration of new shock-affected beneficiaries and ensure reg- istration is accessible to vulnerable groups. ● Coordination with other actors implementing emergency trans- ASSESS fers may be necessary to fill gaps in intake and registration. Assess needs and How will the needs of those ● In the interest of a rapid response, assessment processes may conditions who are registered be be kept the same as for routine cash transfer programs. assessed to determine their ● Alternatively, assessing eligibility for assistance post shock eligibility for post-shock may require the incorporation of new criteria and screening support? processes. ● New criteria and screening processes can make use of existing data in cash transfer registries, for which the screening process- es can be automated. ● Data collected after a shock with post-disaster household needs assessments can fill gaps in information ● Newly collected data should be integrated into social protection information systems, primarily social registries. Eligibility and Based on assessment results, ● Proof of identity requirements may need to be modified and enrollment who most requires support enrollment procedures adapted to reduce barriers to the enroll- after a shock, and how will ment of new beneficiaries. they be enrolled into the ● Pre-enrolling households that are vulnerable to a shock is an- program? other option to improve timeliness of assistance. Determine Once enrolled, what kind of ● It’s important to adapt the size of the benefits package to meet benefits package benefits will best address emergency needs. beneficiaries’ post-shock The frequency and duration of benefits distribution can also be ENROLL needs? ● adapted in line with the objectives of the program (relief, recov- ery). Notification and What is the best way to ● The mechanisms for communicating eligibility decisions to new onboarding let the selected beneficia- beneficiaries and those not selected for assistance may need ries know they will receive to be adapted to ensure they will be received in a post-shock support? environment. ● Adjust messaging to inform beneficiaries of the parameters of the shock response program. ADAPTIVE SOCIAL PROTECTION :: 55 :: Table S.1: Effective shock response: key considerations across the delivery chain (cont.) Issue to address at this Stage Phase phase of the chain Key considerations for effective shock response Benefits provision How is this assistance ● Capacity of the payment provider to pay adjusted benefits pack- physically delivered to the ages, potentially to caseloads that are significantly increased, selected beneficiaries in the should be preestablished. post-shock setting? ● The beneficiary payment interface can be modified to ensure PROVIDE beneficiaries can conveniently and safely get access to pay- ments post shock. ● Alternative payment processes may be needed if certain routine channels cannot be used in the post-shock context. Beneficiaries Is the program functioning ● Conditions for the program can be temporarily waived to ac- compliance, effectively in response to commodate urgent needs and limited access to services post updating, and post-shock needs, or are shock. grievances adjustments needed? ● Processes and systems can be adapted to ensure the effective inclusion of new beneficiaries in data management, compliance verification, and grievance redress activities. MANAGE Exit decisions, When should the shock ● Terms of exit and scale down can be adapted to ensure no notifications, and response program be wound households moved out of the program while still requiring assis- closing cases down? Who needs ongoing tance to meet post-shock needs. long-term support? ● New beneficiaries who match the eligibility criteria for routine long-term cash transfer can be enrolled in the routine program when the shock response ends. Table S.2: Checklist for using or adapting the cash transfer delivery chain for shock response Phase in Checklist: Questions to consider when planning to use or adapt the cash transfer delivery chain delivery chain for shock response. ● Which key messages that are communicated during program outreach activities need to be modified (for example, with regard to objectives; who is eligible; how to apply) to inform communities about the shock response program? Outreach ● Will outreach mechanisms be accessible to (new) targeted beneficiaries, and what modifications (language; location; media) will be needed to ensure they are? ● Must internal communication be adapted and/or training provided for those involved in outreach during a shock response to ensure accurate information is conveyed to communities? ● For identifying new beneficiaries through a demand-led registration system, must systems be mod- ified (for example, with regard to mobile registration; additional staff; relaxation of procedures) to identify eligible individuals rapidly? ● Will demand-led registration processes be accessible to vulnerable groups after a shock, and what modifications (language; location; hand holding; transportation; financial assistance) are needed to Intake and registration ensure they are? ● For identifying new beneficiaries through a census-based registration system, what is the coverage of the affected population; how up to date/accurate are the data; and how will gaps in coverage be managed (completion of new census activities; implementation of parallel assistance to fill gaps)? ● If needed, how will collection of new data on shock-related vulnerability be managed (budget; staff; procedures; data collection and management systems; partnerships and coordination)? ● Must assessment processes (criteria or thresholds) be modified to ensure assessment captures vul- nerability post shock? Assess needs and conditions ● How will assessment processes need to be modified (procedures; data management systems; part- nerships and coordination) to incorporate new data collected post shock? :: 56 :: The delivery chain and shock response Table S.2: Checklist for using or adapting the cash transfer delivery chain for shock response (cont.) Phase in Checklist: Questions to consider when planning to use or adapt the cash transfer delivery chain delivery chain for shock response. ● Must mechanisms for communicating eligibility decisions be adapted? ● Will modifying the proof of identity requirements for program enrollment (for example, with regard to waiving documentation requirements; support for document recovery; support for getting access to required documentation) allow for more rapid provision of assistance? Eligibility and enrollment ● Will enrollment processes be accessible to vulnerable groups after the shock, and what modifica- tions (identification requirements of financial service providers; transportation/financial assistance to get to banks) will be needed to ensure this? ● Can program systems and procedures (the data management system; staffing; opening of accounts; provision of payment tokens) support a pre-enrollment of households that are vulnerable to shock? ● Will flexibility in the form of the cash transfer benefits package (that is, cash versus food) better enable the program to meet emergency needs? Determining benefits ● Can procedures be adapted to modify the size or frequency of the benefits package to meet emer- gency needs? ● Which key messages that are communicated during notification activities (for example, with regard to payment schedule; relaxed conditions) need to be modified to inform beneficiaries of changes in Notification and onboarding processes for the shock response? ● Must notification processes be adapted to ensure they are accessible to new beneficiaries and to existing ones post shock? ● Are business continuity plans in place to ensure payment channels can continue to function after a shock? ● If the existing payment schedule is modified to serve emergency needs better, how will this affect the workload of staff and payment conduits? ● Should payment processes be adapted (for example, with regard to provision of additional staff; plans for movement of funds and increased liquidity; changes to reconciliation processes) to provide Benefits provision payments to newly enrolled beneficiaries or top-up payments to existing ones? ● Are modifications to the beneficiary payment interface needed (payout point locations close to communities; mobile ATMs; temporary over the counter payments where e-payment systems are knocked out; security at payout points) to ensure beneficiaries can conveniently and safely get ac- cess to payments after a shock? ● If payment channels of the national system cannot be used, can a suitable alternative payment ser- vice provider be identified and contracted with? ● Do procedures contain a clause to relax program conditions, or can they be adapted to relax them, where these are a barrier to meeting emergency needs? ● Must processes and systems be adapted (for example, with regard to post-distribution monitoring locations; use of technology; language of questionnaires and social welfare officers; partnerships Beneficiaries compliance, with community service organizations with links to beneficiaries) to ensure new beneficiaries can be updating, and grievances more effectively included in monitoring activities? ● Can grievance mechanisms be adapted (new communication channels; use of technology) to ensure they are accessible to new beneficiaries? ● Can procedures be put in place to create referrals and linkages to other long-term social programs for longer-term recovery? ● Can procedures be adapted to modify the duration of and/or exit from the program to meet emer- Exit decisions, notifications, gency needs? and closing cases ADAPTIVE SOCIAL PROTECTION :: 57 :: References Asfaw, S., and B. Davis. 2018. “Can Cash Transfer Programmes Bowen, T., C. del Ninno, C. Andrews, S. Coll-Black, U. Gentilini, Promote Household Resilience? Evidence from Sub-Saha- K. Johnson, Y. Kawasoe, A. Kryeziu, B. Maher, and A. Wil- ran Africa.” In Climate Smart Agriculture, edited by L. Lipper, liams. 2020. Adaptive Social Protection: Building Resilience to N. McCarthy, D. Zilberman, S. Asfaw, and G. Braca, 227–50. Shocks. International Development in Focus. Washington, DC: Rome: Food and Agriculture Organization of the United Na- World Bank. tions. Coll-Black, S., C. Holmemo, J. Sandford, and W. Soer. Forthcom- Barca, V. 2017. Integrating Data and Information Management for ing. From Programs to Systems: A Decade of Lessons Learned Social Protection: Social Registries and Integrated Beneficiary from Designing and Implementing Ethiopia’s Productive Safety Registries. Canberra: Commonwealth of Australia, Depart- Net Program. Washington, DC: World Bank. ment of Foreign Affairs and Trade. CaLP (Cash Learning Partnership). 2018. The State of the World’s Barca, V., and C. O’Brien. 2017. “Factors Affecting the Usefulness Cash Report: Cash Transfer Programming in Humanitarian Aid. of Existing Social Protection Databases in Disaster Prepared- Accenture and CaLP. https:/ /www.calpnetwork.org/wp-con- ness and Response.” Policy brief. Oxford Policy Management, tent/uploads/2020/03/calp-sowc-report.pdf. Oxford, UK. Del Ninno, C., F. Pierre, and S. Coll-Black. 2016. Social Protection Barca, V., and R. Beazley. 2019. Building on Government Systems Programs for Africa’s Drylands. Washington, DC: World Bank. for Shock Preparedness and Response: The Role of Social As- Dercon, S. 2005. “Risk, Poverty and Vulnerability in Africa.” Jour- sistance Data and Information Systems. Canberra: Common- nal of African Economies 14 (4): 483–88. wealth of Australia, Department of Foreign Affairs and Trade. EM-DAT. 2020. The Emergency Events Database. Université Barrett, S., and S. Kidd. 2015. “The Design and Management of Catholique de Louvain (UCL)—CRED, www.emdat .be, ac- Cash Transfer Programmes: An Overview.” KfW Development cessed May 2019. Bank Materials on Development Financing, no. 3. Frankfurt am Main. Ethiopia, Government of. 2014. Productive Safety Net Program, Phase IV Program Implementation Manual. Bastagli, F. 2014. “Responding to a Crisis: The Design and Deliv- ery of Social Protection”. ODI. EUD. 2019a. Guidance Package on Social Protection across the Hu- manitarian-Development Nexus: Kyrgyzstan Case Study. EUD. Bastagli, F., and L. Harman. 2015. “The Role of Index-Based Trig- gers in Social Protection Shock Response.” ODI. ———. 2019b. Guidance Package on Social Protection across the Humanitarian-Development Nexus: Turkey Case Study. EUD. Beazley, R. 2017. Study on Shock-Responsive Social Protection in Latin America and the Caribbean: Ecuador Case Study. WFP and EUD. 2019c. Guidance Package on Social Protection across the Hu- Oxford Policy Management, Oxford, UK. manitarian-Development Nexus: Yemen Case Study. EUD. ———. 2018. Strengthening the Capacity of ASEAN Member States Farhat, M., F. Merttens, and C. Riungu. 2017. Evaluation of the to Design and Implement Risk-Informed and Shock-Responsive Kenya Hunger Safety Net Program Phase 2: Emergency Pay- Social Protection Systems for Resilience: Thailand Case Study. ments Deep Dive Study. Oxford Policy Management. Oxford Policy Management, Oxford, UK. Fitzgibbon, C. 2016. “Shock-Responsive Social Protection in Beazley, R., A. Solórzano, and K. Sossouvi. 2016. Study on Practice: Kenya’s Experience in Scaling Up Cash Transfers.” Shock-Responsive Social Protection in Latin America and the Ca- Blog on the HSNP, February 23. https:/ /odihpn.org/blog/ ribbean: Theoretical Framework and Literature Review. WFP and shock-responsive-social-protection-in-practice-kenyas-ex- Oxford Policy Management, Oxford, UK. perience-in-scaling-up-cash-transfers/. Beazley, R., A. Solórzano, and V. Barca. 2019. Study on Shock-Re- Gentilini, U., M. Almenfi, I. Orton, and P. Dale. 2020. “Social Pro- sponsive Social Protection in Latin America and the Caribbean: tection and Jobs Responses to COVID-19: A Real-Time Re- Summary of Key Findings and Policy Recommendations. WFP view of Country Measures.” Living paper version 11. https:/ / and Oxford Policy Management, Oxford, UK. openknowledge.worldbank.org/handle/10986/33635 (ac- cessed June 12, 2020). Bowen, T. 2015. “Social Protection and Disaster Risk Manage- ment in the Philippines: The Case of Typhoon Yolanda (Hai- Grand Bargain. 2016. “The Grand Bargain: A Shared Commit- yan).” Policy Research Working Paper 7482. Background Pa- ment to Better Serve People in Need.” World Humanitarian per for Shockwaves: Managing the Impacts of Climate Change /reliefweb.int/sites/reliefweb.int / Summit, Istanbul. https:/ on Poverty. World Bank Group, Washington DC. files/resources/Grand_Bargain_final_22_May_FINAL-2.pdf. :: 58 :: The delivery chain and shock response Hallegatte, S., M. Bangalore, L. Bonzanigo, M. Fay, T. Kane, U. Mahler, D., C. Lakner, A. Aguilar, and H. Wu. 2020. “The Im- Narloch, J. Rozenberg, D. Treguer, and A. Vogt-Schilb. 2016. pact of COVID-19 (Coronavirus) on Global Poverty: Why Shock Waves: Managing the Impacts of Climate Change on Pov- Sub-Saharan Africa Might Be the Region Hardest Hit.” erty. Washington, DC: World Bank. World Bank Blogs. https:/ /blogs.worldbank.org/opendata/ impact-covid-19-coronavirus-global-poverty-why-sub-saha- Hallegatte, S., A. Vogt-Schilb, M. Bangalore, and J. Rozenberg. ran-africa-might-be-region-hardest. 2017. Unbreakable: Building the Resilience of the Poor in the Face of Natural Disasters. Washington, DC: World Bank. Maunder, N., N. Dillon, G. Smith, S. Truelove, and V. De Bauw. 2015. Evaluation of the Use of Different Transfer Modalities in Harvey, P., and S. Bailey. 2011. “Cash Transfer Programming in ECHO Humanitarian Aid Actions 2011–2014: Final Report. A Emergencies: Good Practice Review.” Humanitarian Practice report for ECHO by ADE. Network. Issue 11. ODI. Merttens, F., J. Upadhyay, N. Kukrety, S. Karki, and Z. Majeed. Hidrobo, M., J. Hoddinott, J. Kumar, and M. Oliver. 2018. “Social 2017. Evaluation of the Nepal Emergency Cash Transfer Program Protection, Food Security, and Asset Formation.” World Devel- through Social Assistance: Final Report. Oxford Policy Manage- opment 101: 88–103. ment, Oxford, UK. Holmes, R., and C. Costella, with M. Bailey, A. Kruczkiewicz, R. Mexico, Government of. 2018. Reglas de operación de prospera Poulter, K. Sharp, and L. Scott. 2017. Towards a Shock Sensi- programa de inclusión social [Rules of operation: Social inclu- tive Social Protection System for Malawi. London: ODI and The sion program PROSPERA]. Government of Mexico: Mexico Hague: Red Cross Climate Centre. City. Hobson, M. 2018. “Road Map for Establishing a Risk-Informed Mozambique, Government of. 2019. Programa apoio social directo and Shock-Responsive Social Protection System in the Philip- post-emergencia: Manual de operações. Draft. October. pines.” Food Economy Group. O’Brien, C., Z. Scott, G. Smith, V. Barca, A. Kardan, R. Holmes, IRC (International Rescue Committee). 2016. Cash Research and C. Watson, and J. Congrave. 2018. Shock-Responsive Social Development Pilots—Pakistan 2015–2016. International Res- Protection Systems Research: Synthesis Report. Oxford Policy cue Committee. Management, Oxford, UK. Islamic Republic of Pakistan. n.d. FATA Temporarily Displaced Per- OPM (Oxford Policy Management). 2015. “HSNP Evaluation: sons Emergency Recovery Project (FATA TDP-ERP) Operations Drought Emergency Scale-Up Payments Process Review.” Manual, version 1.9. Power point presentation by OPM. Kardan, A., C. O’Brien, and M Masasa. 2017. Shock-Responsive ———. 2017. Shock-Responsive Social Protection Systems Research: Social Protection Systems Research: Case Study—Lesotho. Ox- Literature Review. 2nd ed. Oxford, UK: OPM. ford Policy Management, Oxford, UK. Otulana, S., C. Hearle, R. Attah, M. Merttens, and J. Wallin. 2016. King and Tranchini. 2017. “UBR MVAC Trial: Assessing the Feasi- Evaluation of the Kenya Hunger Safety Net Program Phase 2: bility of Using a Social Registry for Targeting a Humanitarian Impact Evaluation Qualitative Research Study—Round 1. OPM. Response in Malawi” (unpublished). Ovadiya, M., and C. Costella. 2013. Building Resilience to Disaster Knippenberg, E., and J. Hoddinott. 2017. “Shocks, Social Protec- and Climate Change through Social Protection. Washington DC; tion, and Resilience: Evidence from Ethiopia.” Ethiopia Strat- World Bank. egy Support Program Working Paper 109. International Food Policy Research Institute, Washington, DC. Parsons, K. 2018. Towards Adaptive Social Protection: Elderly Ben- efit and Disability Scheme Top-Ups in the Aftermath of Tropical Kuriakose, A., R. Heltberg, W. Wiseman, C. Costella, R. Cipryk, Cyclone Gita in Tonga. World Bank, Washington, DC. and S. Cornelius. “Climate-Responsive Social Protection.” So- cial Protection & Labor Discussion Paper No. 1210. World Philippines, Government of. Department of Social Welfare and Bank, Washington, DC. Development. 2019. “Guidelines in the Implementation of the Emergency Cash Transfer During Disasters.” Memoran- Leite, P., T. George, C. Sun, T. Jones, and K. Lindert. 2017. “Social dum Circular No. 17, Series of 2019. Registries for Social Assistance and Beyond: A Guidance Note and Assessment Tool.” Social Protection & Labor Discussion Riungu, C., C. O’Brien, and A. Haynes. 2017. HSNP Phase 2 Eval- Paper No. 1704. World Bank, Washington, DC. uations: Special Themes Report - Emergency Payments, January 2017. Oxford Policy Management, Oxford, UK. Lindert, K., T. G. Karippacheril, I. Rodriguez Caillava, and K. Nishi- kawa Chavez. 2020. Sourcebook on the Foundations of Social SEDESOL (Ministry of Social Development). 2017a. Implemen- Protection Delivery Systems. Washington, DC: World Bank. tation Guidelines for the Temporary Employment Program (PET). Mexico City. ADAPTIVE SOCIAL PROTECTION :: 59 :: ———. 2017b. Operation Rules for the Temporary Employment Pro- Ulrichs, M., and R. Slater. 2016. “How Can Social Protection Build gram for the Fiscal Year 2018 (PET). Mexico City. Resilience? Insights from Ethiopia, Kenya and Uganda.” Work- ing paper, Building Resilience and Adaptation to Climate Ex- ———2017c. Operation Rules for the Social Inclusion Program tremes and Disasters (BRACED) Series, London. PROSPERA for the Fiscal Year 2018. Mexico City. UNICEF. 2019. “Kyrgyzstan: Supporting National Social Protec- Seyfert, K., V. Barca, U. Gentilini, M. Luthria, and S. Abbady. tion Systems to Respond in Times of Crisis.” UNICEF. 2019. “Unbundled: A Framework for Connecting Safety Nets and Humanitarian Assistance in Refugee Settings.” Social UNISDR (United Nations International Strategy for Disaster Re- Protection and Labor Discussion Paper 1935. World Bank, duction). 2015. Making Development Sustainable: The Future of Washington, DC. Disaster Risk Management: Global Assessment Report on Disas- ter Risk Reduction. Geneva: UNISDR. Skoufias, E. 2005. “PROGRESA and Its Impacts on the Welfare of Rural Households in Mexico.” Research Reports 139. Interna- Williams, A. 2020. “Post-Disaster Household Assessments and tional Food Policy Research Institute, Washington, DC. Eligibility Determination for Post-Disaster Social Protection Benefits.” Guidance note. World Bank, Washington, DC. Smith, G. 2016. Developing Strategic Options for Building Social Assistance for Syrian Refugees in Turkey: Findings and Policy Op- World Bank. 2017a. “SPL Delivery Systems Brief” World Bank, tions. A report for UNICEF Turkey. Washington, DC. ———. 2017. “Linking Humanitarian Cash Transfers with Nation- ———. 2017b. Strategic Collaboration between World Bank Group al Social Protection Systems in the MENA Region.” A lessons and World Food Program: Bridging Humanitarian Assistance and learned case study for the UNICEF Middle East and North Social Protection Systems—The Case of Yemen. World Bank, Africa Regional Office. Washington, DC. Smith, G., and L. Mohiddin. 2015. “A Review of Evidence of Hu- ———. 2018a. Post-Disaster Damage Assessment and Targeting manitarian Cash Transfer Programming in Urban Areas.” IIED. Mechanisms in Jamaica. World Bank, Washington, DC. Smith, G., Z. Scott, E. Luna, and T. Lone. 2017. Shock-Responsive ———. 2018b. Process Evaluation of Top 225 for Tropical Cyclone Social Protection Systems Research: Case Study—Post-Haiyan Gita Response and Recovery Package. World Bank Group, Cash Transfers in the Philippines. Oxford Policy Management, Washington, DC. Oxford, UK. ———. 2018c. The State of Social Safety Nets 2018. Washington, TRANSFORM. 2017. Administration of Non-Contributory So- DC: World Bank Group. cial Protection: Delivery Systems—Manual for a Leadership and World Bank and World Food Programme. 2019. “The World Transformation Curriculum on Building and Managing Social Pro- Bank and WFP Partnership in the Sahel: Mauritania.” Summa- tection Floors in Africa. Authored by V. Barca, Oxford Policy ry note. Draft, November. Management. World Economic Forum. 2017. The Global Risks Report 2017. Uganda, Government of. 2016. Third Northern Uganda Social Ac- 12th ed. Geneva. WEF. tion Fund Project (Nusaf III): Disaster Risk Financing Sub-Compo- nent Handbook. Draft. June.