9/16/2021 Malawi COVID-19 Urban Cash Intervention Process Evaluation Report Boban Varghese Paul;Chipo Msowoya;Archibald, Edward;Massimo Sichinga;Alejandra Campero Peredo;Muhammad Abdullah Ali Malik © 2021 International Bank for Reconstruction and Development / The World Bank 1818 H Street NW Washington DC 20433 Telephone: 202-473-1000 Internet: www.worldbank.org This work is a product of the staff of The World Bank with external contributions. The findings, interpretations, and conclusions expressed in this work do not necessarily reflect the views of The World Bank, its Board of Executive Directors, or the governments they represent. The World Bank does not guarantee the accuracy, completeness, or currency of the data included in this work and does not assume responsibility for any errors, omissions, or discrepancies in the information, or liability with respect to the use of or failure to use the information, methods, processes, or conclusions set forth. The boundaries, colors, denominations, and other information shown on any map in this work do not imply any judgment on the part of The World Bank concerning the legal status of any territory or the endorsement or acceptance of such boundaries. Nothing herein shall constitute or be construed or considered to be a limitation upon or waiver of the privileges and immunities of The World Bank, all of which are specifically reserved. Rights and Permissions The material in this work is subject to copyright. Because The World Bank encourages dissemination of its knowledge, this work may be reproduced, in whole or in part, for noncommercial purposes as long as full attribution to this work is given. Any queries on rights and licenses, including subsidiary rights, should be addressed to World Bank Publications, The World Bank Group, 1818 H Street NW, Washington, DC 20433, USA; fax: 202-522- 2625; e-mail: pubrights@worldbank.org. Cover photo: © Jane Chidengu/ Ministry of Gender, Community Development and Social Welfare. Used with the permission of Jane Chidengu/ Ministry of Gender, Community Development and Social Welfare. Further permission required for reuse. 1 Table of Contents Acknowledgements....................................................................................................................................... 5 Executive Summary....................................................................................................................................... 7 1. Introduction ........................................................................................................................................ 11 Global and Sub-Saharan Africa ........................................................................................................... 11 Impacts of COVID-19 ........................................................................................................................... 11 Social protection responses to COVID-19 ........................................................................................... 13 Malawi ................................................................................................................................................ 18 Impacts of COVID-19 ........................................................................................................................... 18 Responses to COVID-19 ...................................................................................................................... 19 2. COVID-19 Urban Cash Intervention .................................................................................................... 21 Implementation Context..................................................................................................................... 21 Key Design Considerations.................................................................................................................. 21 3. The Process Evaluation: Analytical Framework and Methodology .................................................... 23 Analytical Framework ......................................................................................................................... 23 Methodology ...................................................................................................................................... 24 Setting the Reference Points .............................................................................................................. 24 Data Collection Framework ................................................................................................................ 25 Data Collection Approaches ................................................................................................................ 25 Sampling Method ................................................................................................................................ 26 Data Quality and Assurance ................................................................................................................ 27 Data Analysis ....................................................................................................................................... 27 Evaluation Timeline............................................................................................................................. 28 Limitations of the Process Evaluation ................................................................................................. 28 4. Overview of findings ........................................................................................................................... 28 Outreach and Sensitization ................................................................................................................. 29 Intake and Registration ....................................................................................................................... 30 Assessment of Needs .......................................................................................................................... 32 Eligibility and enrollment decisions .................................................................................................... 35 Determination of Benefits and Services Package ............................................................................... 37 Notification and Onboarding .............................................................................................................. 38 Provision of benefits ........................................................................................................................... 40 2 Beneficiary Compliance, Updating and Grievance Redress ................................................................ 42 Exit Decisions, Notifications, and Case Outcomes .............................................................................. 43 5. Lessons learnt ..................................................................................................................................... 44 Outreach and registration................................................................................................................... 44 Intake and Registration ....................................................................................................................... 45 Assessment of Needs .......................................................................................................................... 46 Eligibility and enrollment decisions .................................................................................................... 46 Determination of Benefits and Services Package ............................................................................... 47 Notification and Onboarding .............................................................................................................. 48 Provision of benefits ........................................................................................................................... 48 Beneficiary Compliance, Updating and Grievance Redress ................................................................ 49 6. Conclusions and Recommendations ................................................................................................... 51 Elements of the Delivery Chain ........................................................................................................... 51 Intake and registration........................................................................................................................ 51 Setting of transfer values .................................................................................................................... 52 Eligibility criteria and enrolment process ........................................................................................... 53 E-payments and KYC verification ........................................................................................................ 53 Notification and communication to beneficiaries and nonbeneficiaries ........................................... 54 Grievance Redress Mechanism ........................................................................................................... 54 Monitoring framework........................................................................................................................ 55 Building blocks .................................................................................................................................... 55 Coordination and leadership .............................................................................................................. 55 Capacity building of local actors ......................................................................................................... 56 Enabling environment ......................................................................................................................... 56 Financing ............................................................................................................................................. 56 Data and the national social registry .................................................................................................. 57 References .................................................................................................................................................. 58 Appendix 1: Data Collection Framework .................................................................................................... 62 Appendix 2: CUCI Process Evaluation Reference Points ............................................................................. 80 Appendix 3: Summary of CUCI Process Evaluation KII and FGDs................................................................ 84 Appendix 4: Calculation of Sample Size ...................................................................................................... 88 LIST OF FIGURES 3 Figure 1. Evolution in number of countries/territories and social protection measures ........................... 14 Figure 2. Delivery chain for cash transfers.................................................................................................. 23 Figure 3. Four Building Blocks of Social Protection Delivery Systems ........................................................ 24 Figure 4. Respondents’ recollection of the content of the CUCI sensitization ............. Error! Bookmark not defined. Figure 5. Share of respondents that thought the registration process was transparent, inclusive and not influenced by community leaders.................................................................. Error! Bookmark not defined. Figure 6. Channels for received notifications pertaining to CUCI targeting .. Error! Bookmark not defined. Figure 7. Multiple preventive measures were enforced by MNO agents ..... Error! Bookmark not defined. Figure 8. How did beneficiaries use CUCI money? ........................................ Error! Bookmark not defined. LIST OF TABLES Table 1. Cash transfer initiatives/social protection programs introduced in Sub-Saharan Africa during the pandemic .................................................................................................................................................... 15 Table 2. Sample size distribution across the four cities .............................................................................. 26 Table 3. Noneligibility test for households .................................................... Error! Bookmark not defined. Table 4. Economic Activity Score ................................................................... Error! Bookmark not defined. Table 5. Beneficiaries’ perception of the adequacy of transfers ................... Error! Bookmark not defined. Table 6. Respondents knowledge on the basis of transfer level ................... Error! Bookmark not defined. Table 7. Progress on disbursement of CUCI funds to beneficiaries as of June 30, 2021 .... Error! Bookmark not defined. 4 Acknowledgements The process evaluation was undertaken by a team led by Boban Varghese Paul (Economist, Social Protection and Jobs) and Chipo Msowoya (Social Protection Specialist) and, comprised of Edward Archibald (Social Protection Senior Consultant), Massimo Sichinga (Social Protection Consultant), Alejandra Campero (Financial Sector Consultant) and Muhammad Abdullah Ali Malik (Social Protection Consultant). Ivan Drabek (Senior Social Protection Specialist), Peter Pojarski (Social Protection Senior Consultant) and Evie Calcutt (Financial Sector Specialist) provided useful technical inputs into the study. The team hired a Malawi based socioeconomic research and development (R&D) consulting firm, Green Innovation Center (GIC), to design the surveys and undertake the interviews, focus group discussions, and key informant interviews. Deepika Davidar (Consultant) ably edited the report. The report was prepared under the supervision of Mara K. Warwick (Country Director, Malawi, Zambia, Tanzania, and Zimbabwe), Hugh Riddell (Country Manager, Malawi), and Robert Chase (Practice Manager, Social Protection and Jobs). Inaam Ul Haq (Program Leader) and Preeti Arora (Manager, Operations) provided guidance to this work at the concept note stage. Excellent comments were received from peer reviewers Thomas Bowen (Social Protection Specialist) and Luis Alton (Senior Financial Sector Specialist). Additional comments were received from Ruth Hill (Lead Poverty Economist) and Sumati Rajput (Financial Sector Specialist) during the preparation stage of the evaluation. The team is thankful for the numerous discussions and comments received from government officials on the scoping, findings, and recommendations. This includes Mrs. Patricia Zimpita (Director Poverty Reduction and Social Protection), Mr. Dalitso Kalimba (Deputy Director); and Mrs. Bessie Msusa (Chief Economist); Mr. Geoffrey Kumwenda Jnr (Economist) from the Ministry of Economic Planning and Development and Public Sector Reform; Mr. Laurent Kansinjiro (Deputy Director); Mr. Brighton Mndambo (Principal Social Welfare Officer); Mrs. Jane Chidengu (Principal Social Welfare Officer) from the Ministry of Gender, Community Development and Social Welfare; Mr. Alifeyo Banda (Executive Director); Mr. Paul Chipeta (Director of Infrastructure and Economic Development); Stanley Chuthi (Community Driven Development Specialist) and Mr. Benjamin Kayala (Monitoring and Evaluation Officer) from the National Local Government Finance Committee. The team is also thankful to the feedback and insights provided by social protection donor partners from UNICEF Malawi; World Food Programme (WFP) Malawi; (Kreditanstalt für Wiederaufbau) KFW; Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) Malawi, Irish Aid and International Labour Organization (ILO) Malawi. The study was co-financed by the Global Risk Financing Facility (GRiF). Cover photo credit: Covid 19 Urban Cash Intervention (CUCI) Beneficiaries receiving cash during the launch of the intervention. Photo by Jane Chidengu 5 ACRONYMS COVID-19 Coronavirus Disease 2019 CUCI COVID-19 Urban Cash Initiative EPD Economic Planning and Development FAO Food and Agriculture Organization FGD Focus Group Discussions GOM Government of Malawi GRM Grievance Redressal Mechanisms G2P Government to Persons KII Key Informant Interviews KYC Know Your Customer MGCDSW Ministry of Gender, Community Development and Social Welfare MNO Mobile Network Operators MNSSP Malawi National Social Support Program MSME Micro, Small and Medium Enterprises NSSP National Social Support Policy PSR Public Sector Reforms SAP Social Amelioration Program SCT Social Cash Transfer SCTP Social Cash Transfer Program TWG Technical Working Group UBR Unified Beneficiary Registry VHR Very High Resolution 6 Executive Summary Background As part of the 2020 National COVID-19 Preparedness and Response Plan, the Government of Malawi initiated the COVID-19 Urban Cash Initiative (CUCI) to mitigate the adverse impacts of the pandemic on livelihoods, human capital accumulation and basic consumption for poor populations in the cities of Lilongwe, Blantyre, Zomba and Mzuzu. The intervention targeted 35 percent of each city’s population and was implemented in the poorest hotspots, enrolling over 199,000 households in total. The monthly transfer level was equivalent to the minimum wage at the time – MK35,000 (approx. US$45) – for a total of three months. The design and implementation of the CUCI was led by the Government of Malawi with support of Development Partners including the World Bank, Kreditanstalt für Wiederaufbau (KFW), European Union Delegation to Malawi (EU), Deutsche Gesellschaft für Internationale Zusammenarbeit GmbH (GIZ), UNICEF, WFP, and ILO. Being the country’s first major shock responsive cash transfer program to focus on urban areas, the CUCI provides a conducive and fertile learning ground for shaping future policy and programming. This process evaluation report is a key analytical piece on the CUCI. Using a mixed methods approach, the process evaluation explores key questions of what, how and why the CUCI activities were implemented, and what worked well and did not work well along the CUCI’s delivery chain. To obtain an overview of the situation, 23 focus group discussions; 55 key informant interviews; and door-to-door interviews of 1,313 households (985 beneficiaries and 328 nonbeneficiaries) were conducted between May and June 2021. Performance was assessed according to the social protection delivery chain, focusing on four key phases (assessment, enrollment, provision of benefits, and management) and nine discrete stages (outreach; intake and registration; assessment of needs and conditions; determination of eligibility and enrolment; decisions on the benefits package; notification and onboarding; payments of benefits; beneficiaries’ compliance and updating and grievances; and exit decisions, notifications, and case outcomes). The process evaluation also integrated the four “building blocks� of adaptive social protection—programs; data and information systems; finance; and institutional arrangements and partnerships—across all stages of the delivery chain. Main findings Overall, the results show that CUCI implementation has largely been satisfactory across the delivery chain, and the survey data points to a range of positive highlights of the program. For instance, most beneficiaries perceived the hotspot identification and household registration processes to be transparent, inclusive, and correctly done, with little political influence from community leaders. Beneficiaries also felt that payments were delivered at a time when their livelihoods were most impacted, and that they received the full amount to which they were entitled. The implementation phase nonetheless experienced some challenges. For example, one of the main shortcomings of the program was that there have been substantial payment delays due to significant data quality issues at the registration stage. The data constraints reduced the overall pace of program implementation and affected the scope of coverage. 7 These and other challenges faced by the CUCI do not, however, undermine the overall satisfactory implementation of processes across the delivery chain. Moreover, the CUCI has also generated many lessons learnt, and the program has laid a strong foundation for future shock responsive social protection initiatives and urban safety nets – both in Malawi and beyond. Below we summarize the major takeaways from the report: Intake and registration: The challenging context of COVID-19 notwithstanding, different stakeholders were able to collaborate effectively and complete a large-scale registration within a short period of time. City Councils adapted to the process despite having no previous social protection experience, and the involvement of independent enumerators minimized the risk of perceived bias and political capture. Significant data quality challenges, especially mismatch of beneficiary names and identification (ID) relative to e-wallet Know Your Customer (KYC) details, undermined the overall efficiency of registration (and other processes in the delivery chain, such as enrolment). It is recommended that similar future interventions ensure participation of MNOs during intake and registration processes, as appropriate, with a view to eliminating any KYC errors at source. In addition, the government should also explore innovative mechanisms of ascertaining eligibility through existing data (e.g. complementing basic UBR data with satellite data), and layering this with an enumeration exercise to confirm and validate key information. Determination of benefit packages: While leveraging some data analysis to provide a degree of rigor to the proposed transfer value of MK35,000 (approx. US$45), the setting of transfer values was not undertaken based on a needs assessment – as would often be done for social protection programs. It is recommended that, in future, a more detailed analysis of needs be undertaken to facilitate a more rigorous and objective approach to determining the transfer value for an urban program. This lays the foundation for any potential variation in transfer levels according to intra-household characteristics but also helps inform decisions on the inevitable trade-off between coverage and adequacy. Eligibility criteria and enrolment: Benefitting from multistakeholder experience and collaboration during design, the CUCI’s eligibility criteria included both financial and structural criteria. Experiences from regular social safety net programs such as the Malawi Social Cash Transfer Program (SCTP) were leveraged, and existing digital investments in delivery systems facilitated implementation of the criteria. Survey data shows that previous experience with information systems, such as the SCTP Management Information System (MIS), shaped the design of the CUCI MIS and ultimately helped the CUCI eligibility process to operate efficiently. However, an oversight in jointly working with MNOs to validate beneficiary information for eligible households, including matching beneficiary ID and phone/e-wallet information, led to substantial confusion and delays. It is recommended that government implementing agencies jointly work with MNOs, including in the design of intake data collection tools, as well as during enumeration, to ensure that potential beneficiaries validate their IDs and phone/e-wallet information at the intake and registration stage of the delivery chain. Payments of benefits: Survey data showed that the use of an electronic payment method was highly appreciated by beneficiaries, who felt that it reduced the prospect for theft and/or corruption when compared to receiving cash in hand. Beneficiaries also indicated their preference for cash over in-kind assistance, especially for vulnerable groups. Survey findings also showed that CUCI was able to leverage the lessons from the use of e-payments in the SCTP. It is recommended that any urban interventions and shock responsive safety net interventions continue to leverage lessons from the SCTP and broader cash- 8 based humanitarian responses, and seriously consider the use of e-payments in delivering cash-based emergency response interventions across Malawi. E-payments: Despite supporting market development for mobile money-based e-payments among poor households in urban areas, liquidity challenges by most MNO agents affected CUCI’s first payment. This suggests that, despite prior knowledge, the MNOs and their agents’ network were inadequately prepared for the large influx of cash in hotspots and likely instant withdrawal of the same by beneficiaries. Broadly, this might suggest low adoption levels for digital payments across Malawi, requiring more investment in digital payments awareness and usage. This will require multisectoral collaboration beyond social protection and investment by both public and private sectors. Besides increased multistakeholder digital payment engagement and dialogue to address adoption and usage concerns, it is recommended that future social safety net payments take full advantage of the interoperability of Malawi’s payment systems to consider other payment points options such as banks. Notification and communication to beneficiaries and nonbeneficiaries : While efforts were made to sensitize eligible households, survey data showed that the notification process was unsatisfactory, both for beneficiaries and nonbeneficiaries. This created unnecessary expectations and raised unjustified fears of elite and political capture. It is recommended that communication methods be adapted to an urban context by building on rigorous evidence from national surveys on access and usage of ICT services. More innovative means, such as digital technologies, should be considered to reach urban communities. Grievance Redress Mechanisms: Survey data showed that the CUCI Grievance Redress Mechanisms (GRM) faced challenges. Major concerns included lack of timeliness in setting up GRM committees in three of the four cities, and limited functionality of the Call Centre. It is recommended that greater efforts be made to develop appropriate GRM structures for urban areas, and that this be planned ahead of another shock. There could be complementary lessons here from the humanitarian sector, which regularly sets up accountability mechanisms at short notice. A harmonized approach to grievance redress across the humanitarian and social protection sectors may help to alleviate such issues in future. Monitoring framework: Evaluation findings showed that monitoring of CUCI was another area that faced challenges in its design and execution. That said, it is recognized that there were additional challenges created by both the COVID-19 context and the move into new “uncharted� geographical areas that were not familiar with social safety nets implementation. It is recommended that the Malawi National Social Support Program (MNSSP II) monitoring and evaluation framework be enhanced to account for urban measures. This may either be through adding an urban-specific component to the framework or mainstreaming an urban dimension throughout the existing framework. The enhancement should also consider the increased human resource needs that may arise in the context of a shock response. Coordination and leadership: One of the qualified successes of CUCI has been the coordination across different levels, sectors and geographic areas, and the leadership shown by central government agencies in a challenging context. Survey data showed positive feedback on structures such as the Technical Working Group (TWG) and the way in which the division of labor considered stakeholders’ comparative advantage and experience. Nevertheless, coordination was more complex in the hotspots. It is recommended that efforts be directed to establishing coordination structures and relationships between central government agencies and decentralized urban structures (including City Councils and Ward Committees). 9 Enabling environment: The findings of the process evaluation point to a number of opportunities to strengthen the social protection policy framework for Malawi. Firstly, insights from “learning by doing� approaches to urban programming such as the CUCI should provide a fertile ground for sectoral dialogue, amid existing priorities and emerging issues, to enhance the urban dimensions of the social protection enabling environment. A second dimension is the principle that routine safety net programs should continue to operate through a shock and not be compromised by scalability efforts, including new programs. This reflects on one of the GoM’s under-recognized achievements of 2020-21, which is ensuring that the rural SCTP continued to make regular and predictable cash transfers to its approximately 292,000 beneficiary households, notwithstanding the pandemic and the demands of designing and implementing the CUCI. It is recommended that both these issues be incorporated into the reviews of the National Social Support Policy (NSSP) and the MNSSP II. Data and the national social registry: Malawi’s Unified Beneficiary Registry (UBR) has the potential to be a strong tool for future urban responses. The data that have already been collected through CUCI may be of significant value in the event of a future shock, although basic updating may be necessary. Going forward, it is recommended that the government consider an expansion of the UBR to pre-register a larger proportion of the urban population. Further, stronger links need to be explored and established where feasible, between the UBR, the national ID, and digital payment systems for improved delivery efficiency. 10 1. Introduction This report presents the findings of a process evaluation for the COVID-19 Urban Cash Intervention (CUCI) program. The CUCI was initiated by the Government of Malawi in 2020 to help mitigate the adverse health and economic effects of COVID-19 on urban poor populations in Lilongwe, Blantyre, Zomba and Mzuzu. The design and implementation of CUCI was led by the government with the support of development partners including the World Bank, KFW, EU, GIZ, UNICEF, WFP, and ILO. CUCI is the first major cash transfer program in Malawi to focus on urban areas. It therefore provides a conducive and fertile learning ground for shaping future policy and programming, both in terms of urban cash transfers and shock-responsive interventions. The goal of this process evaluation is to understand the quality of implementation of CUCI at each stage of the delivery chain, with a view to informing future policies and programs. The study looks at what, how and why CUCI activities were implemented, what worked well, and what did not work well along the nine stages of CUCI’s delivery chain. The assessment also integrates the four “building blocks� of adaptive social protection1 —i.e. programs; data and information systems; finance; and institutional arrangements and partnerships—across all stages of the delivery chain. The process evaluation plans to draw lessons that can inform medium- to long-term considerations regarding the design and implementation of both urban safety nets and shock sensitive social protection policies and programs. The remainder of this section lays out the context in which the CUCI was designed and implemented. It does this through examining the impacts of COVID-19 and social protection responses by governments – firstly at a global and regional level, and then with regard to Malawi. Global and Sub-Saharan Africa Impacts of COVID-19 Globally, the COVID-19 pandemic has led to a colossal loss of human life and caused unparalleled challenges to food security and systems, public health, and the jobs-employment sector. The pandemic has disproportionally affected the poor and vulnerable. Recent analysis predicts that COVID-19 may have pushed an additional 88 million people into extreme poverty – and this is just a baseline number. In a worst-case setting, the figure could go as high as 115 million people (World Bank 2021). Despite rates of infection not being as high as in other regions, the economic consequences of COVID- 19 in Sub-Saharan Africa have been severe due to the combination of declining global demand and local efforts to contain the disease. The World Bank forecast a recession in the region in 2020 for the first time in more than 25 years (IFC 2020). The 2020 cost in output losses alone was projected to be between US$37 billion and US$79 billion. The continent is also expected to experience its first increase in the proportion of people in extreme poverty in two decades as the informal sector, a driving source of income and employment, is one of the sectors which has been hit hardest by COVID-19 (IFC 2020). Men, women, and children alike, from both urban and rural settings, are vulnerable to the worldwide surge in food insecurity. According to an early assessment by the UN Food and Agriculture Organization (FAO), the number of malnourished people, then projected at nearly 690 million, could increase by 83 million to 132 million people by the end of 2020 (Blake and Wadhwa 2020; FAO and WFP 2021). According 1 Bowen and Smith (2020). Adaptive Social Protection: The delivery chain and shock response. 11 to WFP (2020) estimates, 270 million people in its client countries were or were at risk of being acutely food-insecure in 2020. It has been suggested that food security concerns stem from loss of incomes owing to strict lockdowns and restrictions to mobility, and are aggravated by disruptions to global and domestic markets and food value chains (Aggarwal et al., 2020; Amjath-Babu et al., 2020; Khan et al., 2021; Mahajan and Tomar, 2021). Food insecurity has been particularly acute in Sub-Saharan Africa, with levels tripling in Malawi, Nigeria, Ethiopia, and Uganda, compared to the previous year. In Malawi, Nigeria, Kenya, South Africa, and Sierra Leone, more than 50 percent of households ran out of food in the 30 days prior to the Rapid Response Phone Survey conducted by the World Bank in 2020, with urban households being disproportionally affected. Closure of schools across the region exacerbated the problem by limiting children’s access to school feeding programs (Pierella Paci 2021). Today, millions of businesses and livelihoods are experiencing an existential threat. Companies, especially the micro, small and medium enterprises (MSMEs) in the developing world are under crushing pressure; more than half are either in arrears or about to fall into arrears. Nearly 50 percent of the world’s 3.3 billion workforce is close to losing their livelihoods (Apedo-Amah et al. 2020). Khamis et al. (2020) state that a large share of workers stopped working in all countries. They also confirm large disruption in the labor market with a substantial number of employees experiencing partial or no payments for their respective work. Kugler et al. (2021) find that the brunt of livelihood losses was borne by women, young, less educated, and urban segments of the workforce. Informal workers are particularly susceptible to losing livelihoods as the majority lack social protection and access to quality health care in addition to losing access to productive assets (Blake and Wadhwa 2020; Apedo-Amah et al. 2020). Between April and August 2020, these special groups experienced disproportionate employment gains; the gains, however, were not sufficient to nullify the size of the initial losses, thereby possibly reflecting an evolution towards a lower-level equilibrium (Kugler et al. 2021). The pandemic has also affected the private sector in Sub-Saharan Africa, putting livelihoods at risk. Disruption to trade and the value chain has impacted commodity exporters and companies dependent on global value chains to source materials and sell their products. Containment measures have put pressures on small businesses, and the decrease in foreign financing flows from remittances, tourism, and foreign direct investment have threatened business survival and employment (IFC 2020). Agricultural income and revenues from nonfarm businesses have fallen since the onset of COVID-19. Job losses are widespread and common among urban workers and female workers. Kenya reported 62 percent job losses as a percentage of the employed population before COVID-19, followed by Gabon (61 percent) and Democratic Republic of Congo (42 percent) (World Bank COVID-19 High-Frequency Monitoring Dashboard, n.d.). The disruption of the job market is also noticeable from the high share of workers changing jobs during COVID-19. In the Sub-Saharan Africa region, job changing ranged from 2 percent to 21 percent (Khamis et al. 2020). Widening gender gaps during and after the pandemic could stand to reverse achievements in human capital, economic empowerment, and voice and agency made by women and girls over the past decade. The gender dimensions of COVID-19 are unlike any experienced before and is largely due to within-sector differentials rather than across sector differentials (Kugler et al. 2021). Women are facing job loss at a faster rate than men, with the latest estimates suggesting that the loss of women’s jobs could reach the 64 million mark worldwide, with 86 percent moving completely into inactivity (UNDP 2021). Within the 12 sub-Saharan region, a large share of households witnessed a drop in household income in both rural and urban settings as a result of COVID-19, which may further reduce the opportunities for women due to prevailing social norms regarding control of the allocation of household resources by men in many parts of sub-Saharan Africa. Since the onset of COVID-19, domestic violence against women has been exacerbated due to lack of employment, low social support, substance abuse, increased stress, and poor mental health (Peterman and O’Donnell, 2020). It is also more challenging for rural women to support their children’s learning at home as they already spend more time caring for other household members compared to both men and urban women (six hours vs. three hours for men and four hours for urban women) (Aoyagi 2021). Facing increasing expectations to take on care-related tasks at home, girls may find it harder to stay engaged in schools over the longer term. It is estimated that 11 million girls may never return to school post COVID-19 (Blake and Wadhwa 2020). Social protection responses to COVID-19 The global response to COVID-19 has been significant, with a dramatic expansion of social protection programs following the onset of the pandemic. More specifically, 222 countries had planned or implemented 3,333 social protection measures between March 20, 2020, and May 14, 2021 (Figure 1). This represents an increase of nearly 148 percent since December 2020. Social assistance and social insurance grew by about 120 percent and 110 percent, respectively, while active labor market interventions increased by nearly 330 percent. Of the social protection measures, social assistance accounted for the highest proportion of global responses (55 percent); cash transfers were the most used form of social protection worldwide (23 percent) and accounted for 42 percent of total social assistance (Gentilini et al. 2021). Almost 17 percent of the world’s population has been covered by at least one COVID-19-related cash transfer payment between 2020 and 2021 (Gentilini et al. 2021). The rate at which cash transfers have been scaled up in comparison to pre-COVID-19 coverage levels remains high. For instance, 15 countries, mostly low- income and fragile countries, had scale-up rates of 1,000 percent due to low pre-COVID coverage rates. 13 Figure 1. Evolution in number of countries/territories and social protection measures 250 3600 222 212 215 3100 Number of countries/territories (bars) 200 3333 190 195 200 181 2600 Number of measures (line) 171 159 151 2100 150 133 126 1600 106 100 84 1414 1100 1179 1024 1055 870 937 803 600 45 685 732 50 564 505 418 100 283 103 0 -400 1. May (2020) 8. May (2020) 15. May (2020) 22. May (2020) 12. June (2020) 14. May (2020) 20. Mar (2020) 27. Mar (2020) 10. July (2020) 11. Dec (2020) 3. April (2020) 10. April (2020) 17. April (2020) 24. April (2020) 18. Sept (2020) Source: Gentilini et al. 2021 The composition of social protection programs differed greatly by country income groups and regions. Nearly three quarters (73 percent) of all social protection measures in low-income countries were social assistance programs, substantially higher than the 45 percent observed in high-income settings. The proportion in in Sub-Saharan Africa was 65 percent. The extent to which social insurance programs were deployed in response to COVID-19 has been broadly aligned with income levels, ranging from 9 percent in low-income countries to 26 percent in high-income countries. The generosity of responses has also varied greatly. In a sample of 125 countries, the average cash transfer represented 31 percent of monthly GDP per capita, varying between 18 percent in North America to 52 percent in sub-Saharan Africa (Gentilini et al 2021). The duration of cash transfers remains short, but longer than before. Information on the duration of cash transfer programs exists for 142 schemes. Among these, the duration range goes between 1 and 12 months, for an average of 4 months. For a subset of 79 cash transfer programs (or 10.4 percent of the cash transfers sample in 54 countries), information was recorded on program extension. For 49 out of those 79 programs, information was available on exact duration of such extension, which was for an average 3.1 months (Gentilini et al. 2021). With regard to overall spending, there were vast differences across regions and countries. Social assistance expenses ranged from US$1.7 billion in low-income countries to US$515.3 billion in high- income countries, and from US$6.1 billion in Africa to US$290 billion in North America (Gentilini et al. 2020). Global social protection spending is largely driven by the Unites States, which accounts for 65 percent of global spending in the sector (Gentilini et al. 2021). Generally, the bulk of social protection investment has been accounted for by high-income countries, which have spent US$2.6 trillion (87 percent of the total global investment) (UNDP 2021). To finance current scale-up efforts, the most 14 prevalent choice among countries was restructuring or re-prioritizing budget lines followed by domestic debt and deficit spending and tapping of state reserves/contingent funds. The ability to respond to COVID- 19 was not only lower among poorer countries, but in addition, not all of them were able to offer any income support to mitigate the short-term effects on income losses (Gentilini et al. 2020). Only a minority of cash transfer programs are being scaled up based on existing schemes . Looking at vertical expansions, 47 countries have increased benefit amounts of 61 existing cash transfer schemes. Simultaneously, 31 countries have provided additional payments to beneficiaries in 54 pre-existing programs. With regard to horizontal expansion, 35 programs in 26 countries scaled up their existing cash transfers. More noticeably, the number of new programs has been significant, i.e. 509 schemes in 166 countries. (Table 1 provides an overview of selected cash transfer initiatives/social protection programs introduced in Sub-Saharan Africa in response to COVID-19.) Finally, while 48 schemes in 38 countries did not expand transfers either vertically or horizontally, they still benefited from administrative simplifications and/or the bringing forward of payments which were due in the future. Overall, the positive implementation performance across vertical and horizontal scale up can be attributed a range of factors, including countries’: pre-existing social registries, delivery capabilities (Gentilini et al. 2021), legal framework, funding sources, and ability to quickly gather relevant information (Beazley et al. 2021). Table 1. Selected cash transfer initiatives/social protection programs introduced in Sub-Saharan Africa in response to the COVID-19 pandemic Country Pillar Program Description Comoros Social Unconditional Target beneficiaries: 20,000 households Assistance cash transfers Geographic scope: Urban and sub-urban areas (UCTs) Kenya Social Oxfam transfer Transfer target beneficiaries: 40,000 people Assistance Transfer geographic scope: Mombasa and Nairobi informal settlements Transfer program duration: 3 months Transfer distribution method: MPESA Madagascar Social Tosika Fameno Target population: 368,000 (revised, initial target was Assistance Unconditional 150,000) poor and vulnerable households Cash transfers Benefit amount: 100,000 Aviary per household (TS-UCTs) Geographic scope: Antananarivo and Toamasina city. Roll-out timeline: 2 transfers between April and May 2020. Total project budget: US$15,000,000 Mauritius Social Employment Target beneficiaries: 11,000 employees Assistance Support Scheme Benefit amount: Rs. 10,200 per capita for SMEs Benefit frequency: Monthly Mozambique Social Post Emergency Horizontal expansion: 1,102,825 new households Assistance Direct Cash Benefit amount: 1,500 per month per family Transfers Benefit payment frequency: Every two months Program (PASD- Program duration: 6 months PE COVID) Geographic scope: Urban, peri-urban and border areas Targeting methodology: Combine multi- poverty index, geographical and community-based targeting to reach intended population Total cost: US$50 million (Phase 1) and US$140.5 million (Phase 2) 15 Rwanda Social Vision 2020 Existing beneficiaries: 310,000 families Assistance Umurenge Horizontal expansion: 56,000 families Programme (VUP) Vertical expansion of Nutrition Sensitive Co- responsibility Cash Transfers: Include more poorer households in the existing 17 out of 30 districts where the program is functional Somalia Social Baxnano Cash Program cost: US$65 million Assistance transfers Target population: 200,000 poor and vulnerable households, approx. 1.3 million individuals Beneficiaries served: 73,478 households by January 2021 Benefit amount: US$60 Benefit payment frequency: Quarterly Geographic scope: National including fragility areas due to widespread poverty, acute drought and long-drawn- out conflicts and insecurity. Uganda Social Child-Sensitive Target population: 56,500 women (59% refugees and Assistance Social Protection 41% citizens) Program in Benefit payment frequency: One-off payments Refugee-Hosting Geographic scope: West Nile Region Districts of West Nile Region Zambia Social Cash transfers Target population: Vulnerable homes of the elderly, Assistance women and their children Benefit: Cash as well as food hampers Program duration: 6 months. Zimbabwe Social Emergency Target population: 25,000 vulnerable households Assistance Harmonized Benefit amount: US$13 per eligible individual Social Cash Geographic scope: 6 urban domains Transfer (HSCT) Total cost: US$10,104,000 Source: Gentilini et al. 2021. Against the backdrop of the extensive global social protection response, there is a growing body of evidence and analysis about the effects of social assistance in response to pandemics. Since the onset of COVID-19, a small number of studies have investigated the effects of cash assistance in mitigating the short-term consequences of pandemic-induced shocks. Some of these are highlighted below in Box 1. Box 1: Selected studies investigating the effects of cash assistance in mitigating the short-term consequences of pandemic-induced shocks Younger et al. (2020) simulate the poverty impacts of COVID-19 in Uganda in the presence of temporary transfers. Without social assistance, they estimate that income loss is equivalent to 9.1 percent of monthly GDP and poverty-rate increases by 7.9 percentage points. Their calculations suggest that expanding the two largest existing social programs would lessen poverty by 1.6 percentage points, and that to neutralize about two-thirds of the increase in poverty, the government would need to transfer about 0.7 percent of their monthly GDP to targeted poor and vulnerable households. Blofield et al. (2021) perform simulations to study the impact of the economic contraction on poverty and income equality, alongside the poverty and inequality reduction effects of the social assistance 16 programs initiated during the COVID-19 crisis in Argentina, Brazil, Colombia, and Mexico. Aside from Mexico, where there was practically no expansion of emergency social assistance, the results suggest that the introduction of cash-based benefits might have more than nullified the negative effect on income in all countries, particularly Argentina and Brazil, due to their large expansion in emergency social assistance. Egger et al. (2021) use telephone-based surveys, collected across nine countries in Africa, Asia, and Latin America, to assess the impact of the pandemic on income losses and to investigate the effect of social assistance in reducing food security. Their findings show that although a sizeable share of respondents benefited from social support (at a median of 11 percent), it was still insufficient to maintain pre-COVID-19 living standards and, hence, to prevent food insecurity. The slightly pessimistic findings can be attributed to the fact that the surveys were conducted over the first three months of the pandemic, April-June 2020, well before the bulk of social assistance and cash transfers were implemented globally. Paul et al. (2021) undertakes a microsimulation of the potential impact of a fully financed Social Cash Transfer (SCT) Program2 in Zambia. Due to historic debt issues, the SCT program was only partially financed, i.e., a sizeable number of SCT beneficiaries would not receive their regular transfers. Simulations were, therefore, conducted for three scenarios: i) status quo SCT that pays its full caseload benefits; ii) an enhanced SCT that pays its full caseload benefits at a higher transfer amount, equivalent to the 2017 value of the transfer; and iii) a further enhanced SCT that pays its full caseload benefits at a higher transfer amount, equivalent to a target of 15 percent of the poverty line. Simulations estimate that if all those enrolled in the SCT were paid benefits, this would bring down poverty by 3.7 percentage points. This implies that, in 2020, 0.7 million less people are expected to be in poverty if SCT pays benefits to its full caseload, compared to the case if only the current partial caseload is paid. Enhanced SCTs could potentially reduce poverty incidence by up to 6 percentage points. Factoring for a COVID- 19-induced increase in poverty in 2020, this translates to a net poverty reduction of 4.2 percentage points. Further increasing the transfer value to 15 percent of the national poverty line would have an even larger poverty impact. Cho et al. (2021) undertook a process evaluation of the second tranche of the Social Amelioration Program (SAP2)3 in the Philippines to document lessons learned from its implementation. Implementation of SAP2 outlined several prerequisite areas that should be enhanced for digital G2P payments. The chief area for improvement is the quality of beneficiary data. The outmoded social registry was unreliable for identification and verification of beneficiaries, forcing government officials to rely on paper-based manual data collection during the pandemic for SAP1. This not only resulted in increasing the risk of infection, but also created a large amount of invalid or duplicative data which was inherited by the SAP2 process and greatly hindered digital delivery of payments. Other important areas 2 The Harmonized Social Cash Transfer (SCT) Program is Zambia’s flagship national social assistance program that aims to reduce extreme poverty and intergenerational transfer of poverty among poor households with vulnerable individuals. Piloted in 2003 with only 159 households in Kalomo district, the program has undergone considerable evolution since its inception. By 2020, SCT covered 616,464 beneficiary households representing about 18 percent of the population. Owing to data limitations, a microsimulation method was used for SCT to estimate the ‘potential’ impact on poverty, poverty gap, and inequality of a fully operational SCT program vis-à-vis the current underfunded program. 3 The Social Amelioration Program (SAP) introduced under the Bayanihan To Heal as One Act in March 2020 and supervised by the Department of Social Welfare and Development (DSWD), aimed to provide cash-based assistance to 18 million households (i.e. 75 percent of total households) in the country. The program intended to provide a benefit of Php 5,000 to Php 8,000 (about US$100 to US$160) two times, contingent on the minimum wage of workers and households’ subsistence expenditure in the region. The first payment of SAP (SAP1) suffered from significant delivery issues, leading to several changes in the operation of the second tranche of SAP (SAP2). Some notable changes included extension of beneficiaries, online application form to digitize beneficiary records, digital G2P (government to persons) payment delivery, and incorporation of customer choice in G2P payments. 17 include institutional arrangements and coordination challenges, systematic and efficient grievance redressal mechanisms (GRM), and communications with beneficiaries. Malawi Impacts of COVID-19 The COVID-19 pandemic reached Malawi in March-April 2020. Malawi declared a State of Disaster on March 20, 2020 and registered its first confirmed coronavirus case on April 2, 2020 (Mzumara et al., 2021). On April 4, 2020, the Government of Malawi recommended a series of restrictive measures including virtual working arrangements and staff shifts for workplaces, especially for nonessential staff. Public gatherings were limited to a maximum of 100 people. Although public transport was not completely closed, it was recommended to reduce the seating capacity by 60 percent. All commercial flights and entry into Malawi by foreign nationals were suspended, but land borders remained open for cargo and returning residents (who were required to undergo mandatory 14 days of quarantine). Other preventative measures included closure of schools, restrictions on travel of nonessential personnel within and outside the country, and the promotion of mandatory hand washing, mask wearing and social distancing in supermarkets, open markets and other public spaces. Unlike many African nations, there was no lockdown in Malawi. On April 14, 2020, the President of Malawi announced a national lockdown slated to begin on April 18 and to last for 21 days. However, the lockdown was challenged in court, and a court injunction on April 18 prevented the lockdown from going ahead. The High Court later upheld the challenge to the lockdown, with a decision on September 10 finding that the lockdown restrictions were unconstitutional as they posed potential negative consequences to the socioeconomic status of a majority of Malawians. Although there was no local lockdown, Malawi’s economy has been heavily affected by the pandemic due to both domestic and international factors. Macroeconomic factors include a) disruption in global commodity value chains and trade and logistics; b) decrease in tourism; and c) decrease in remittances. Economic growth in Malawi was projected to decrease to 1 percent in 2020, down from earlier projections of 4.8 percent and representing a 2 percent downturn in per capita GDP when measured against population growth of around 3 percent (World Bank, 2020).4 The economic downturn has also weakened the government’s ability to generate much needed revenue. A large proportion of the urban poor in Malawi have experienced disrupted or impaired economic welfare. At the macro level, some of the key services and industry sectors that offer employment opportunities have been hard hit, leading to a heavier impact in urban areas among populations that traditionally suffer from unstable income, high unemployment, fragile markets, and unstable asset-base (World Bank, 2020). At the household level, the ability to survive and rebuild or protect their livelihoods was severely affected through loss of income. Urban and rural poverty were projected to increase by 4.9 4 An early assessment of potential containment measures in Malawi forecast an economic downturn and increasing poverty. The International Food Policy Research Institute (IFPRI) modelled the impact of two potential scenarios: two months of social distancing and a hypothetical 21- day urban lockdown. IFPRI projected that Malawi would face a national GDP loss of 11.6 percent from the social distancing scenario and a loss of 22.3 percent under the lockdown scenario. These equate to GDP losses of approximately US$26 million and US$49 million per week, respectively, during the periods the restrictions were enforced (Baulch et al. 2020). Between 1.1 and 2.2 million people were projected to fall into poverty under the two scenarios. It was also suggested that Malawi’s GDP would decline by between 4 and 5.2 percent in 2020 under the social distancing scenario and by between 6.2 and 9.1 percent under the lockdown scenario. (On the upside, the GDP projections for Malawi were considerably better than those calculated for Ghana, Nigeria, and South Africa (Baulch et al. 2020).) 18 percent and 2.2 percent respectively. A study conducted in May 2020 by the Institute of Public Opinion and Research revealed that 82 percent of Malawians feared going hungry during the COVID-19 pandemic more than they feared being infected by the virus itself (GLD & Ipor, 2020). Survey data suggests that more than two-thirds of households have lost income since the onset of COVID-19. An IFPRI survey from August 2020 found that 69 percent of respondents affirmed that their household’s total income had decreased during July 2020. The decrease was noticeable across the three main sources of income: agriculture, nonfarm enterprises, and wage employment. However, the share of households reporting a decline in income was the highest for nonfamily businesses (67 percent), relative to agriculture (60 percent) and wage employment (48 percent) (World Bank 2020). The survey also sheds light on the reasons behind difficulties in accessing food and the negative coping mechanisms during the pandemic. 56 percent of respondents said that the foods that they usually bought were not available in the nearby markets. Around 50 percent of respondents suggested that prices for their regular food purchases had increased post COVID-19. Suggestive of a high level of vulnerability, 50 percent of respondents said that they had decided to limit the size of their meals while 45 percent reduced the number of meals consumed in a week (Ambler et al. 2020). The negative impacts of COVID-19 on employment in Malawi have been greater in urban areas. A phone survey conducted by the World Bank between May and August 2020 suggested that the crisis had translated into job losses for 12 percent of the employed population, with income losses being more significant in urban areas (77 percent of households) in comparison with rural areas (72 percent). This reflects the levels of employment in the services and industry sectors, which have been more heavily affected by the pandemic (World Bank 2020). That said, while the impact is significant, the labor market in Malawi has been more resilient than in other countries, perhaps reflecting Malawi’s relatively high reliance on the agriculture sector, which has been less affected by the crisis. Overall, the proportion of survey respondents who are working has continued to increase, although there is evidence that people might have changed jobs, indicating job market volatility (World Bank 2020b). Numerous challenges have affected household farms and businesses. According to the IFPRI survey, conducted in August 2020, more than 50 percent of respondents mentioned lower than normal prices for outputs, difficulty accessing credit, and difficulty accessing inputs in the three months prior to the survey. Other reported challenges included lower than normal demand for outputs, higher than normal prices for inputs, and difficulty traveling. However, while this data is informative, the pandemic is the only possible cause for these issues. There may have been other reasons at play such as good harvests across the country driving down output prices (Ambler et al. 2020). With regard to the agriculture sector, the overall proportion of households involved in crop farming increased between the 2019 and 2020 dry seasons, although there were minimal changes to the proportion of households keeping livestock. Households reported a greater impact of COVID on livestock sales compared to crop sales (50 percent vs. 44 percent) (World Bank 2020b). Government responses to COVID-19 Malawi employed a multisectoral approach to address both the health and socioeconomic aspects of the pandemic. On April 8, 2020, Malawi launched an ambitious national COVID-19 Preparedness and Response plan with a total budget of US$213 million. The plan involved targeted interventions across the health, education, economic and social sectors, and was developed in coordination with the humanitarian cluster system. The multisectoral approach is reported to have successfully increased the capacity for 19 COVID-19 testing and also heightened policy responsiveness to the unintended socioeconomic consequences of the pandemic that were anticipated to follow the proposed lockdown (Mzumara et al. 2021). The breadth of policy responses included introduction of tax waivers on some medical equipment (e.g., personal protective equipment), credit facilities for SMEs, and increasing the health workforce by hiring 2,000 health care workers (Mzumara et al. 2021). The most significant social protection measure proposed by the Government of Malawi (GoM) in response to COVID-19 was an urban cash initiative which is the focus of this report – the CUCI. Chapter 2 provides further detail on the initiative, but some of the key characteristics are as follows. The overall objective of CUCI was to help the urban poor population cope with the health and economic effects of the pandemic. The intervention was implemented in the poorest hotspots of four cities—Lilongwe, Blantyre, Zomba and Mzuzu—and targeting 35 percent of the population in these cities (over 199,000 households in total). The transfer level was MK35,000 (approx. US$45) per household per month for a total of three months. The design and implementation of the intervention led by the GoM with the support of development partners in the social protection sector (including World Bank, KFW, EU, GIZ, UNICEF, WFP, and ILO). There were also other social protection and related interventions introduced in Malawi in response to COVID-19, including bringing forward future payments for SCTP beneficiaries and otherwise continuing to provide routine SCTP payments throughout 2020 and the first half of 2021. It is beyond the scope of this report to elaborate further on those interventions. 20 2. COVID-19 Urban Cash Intervention Implementation Context A new work environment was ushered in by COVID-19 in a context already challenged by complex institutional and development partner relations. CUCI was the GoM’s flagship response to support at- risk and vulnerable urban populations, and there was much public and political interest in the initiative and its technical design issues such as transfer levels. Nevertheless, the government and donors were dealing with innumerable challenges induced by the pandemic, including a complete change in working arrangements. For instance, almost all officials were working from home or going to the office in shifts during much of the preparation phase. Meetings related to the design of CUCI were mostly conducted virtually. An overall atmosphere of fear and anxiety pervaded the period during which the CUCI was designed and implemented. Moreover, this was the first time the government was designing and implementing any kind of safety net intervention in urban areas of Malawi. It had to build new expertise in a completely novel context and at a rapid pace given the shock responsive nature of the intervention. The government was able to draw on its substantial rural safety net experiences in designing CUCI. Yet the dynamics of urban poverty, livelihoods and vulnerability in Malawi are typically quite different from those in rural areas. Officials had to balance an array of competing priorities, including day-to-day responsibility for pre- existing and ongoing social protection programs. Central government staff – particularly from MGCDSW – were already heavily committed with ensuring the rural safety net program, the SCTP, continued to deliver routine and predictable payments to its approximately 200,000 beneficiary households. Their contributions to the design and implementation of the CUCI ran in parallel to their responsibilities for routine service delivery. To further complicate implementation, the design period of CUCI coincided with the campaigning period for the Presidential Election re-run in June 2020. This was the result of the Malawian High Court having nullified the results of the May 2019 Presidential Elections. This was preceded by multiple demonstrations across the country led by civil society, which sometimes became violent5. GoM officials and development partners had to focus on the immediate response to the crisis while simultaneously dealing with the risks associated with an intensely politically charged period. Beyond politicization and potential tampering of technical design considerations of CUCI during implementation, security issues dogged frontline staff – including enumerators – who would be interacting with communities during a politically volatile period. Key Design Considerations Against the backdrop of a highly challenging context, the GoM opted to build on the foundations of the existing rural-focused Malawi Social Cash Transfer Program (SCTP) to provide cash transfers to households in urban areas. The CUCI has five objectives that aim to reduce the adverse impact of COVID- 19 on livelihoods, human capital accumulation and basic consumption. Specifically, these objectives are: a) Prevent the poor/vulnerable from falling further into poverty and assist them to cover their basic 5 https://www.bloomberg.com/news/articles/2019-08-06/election-dispute-in-malawi-heats-up-as-fresh-protests-erupt 21 consumption needs such as food; b) Support the continuous uptake of nutritious meals to help prevent the outbreak of opportunistic diseases during the pandemic; c) Protect the poor and vulnerable from engaging in negative coping mechanisms and selling productive assets as a result of loss of livelihood sources due to strict crowd control measures aimed at controlling spread of the disease; d) Promote health-seeking behaviors of the poor and vulnerable during the COVID-19 outbreak; e) Assist the poor and vulnerable in reconstructing their livelihoods post-pandemic. The intervention used a mix of geographic and household level characteristics to identify its beneficiaries, with geographic eligibility based on a government assessment that identified a number of COVID-19 hotspots. The hotspots were determined according to the estimated vulnerability of poor urban populations due to the socioeconomic effects of COVID-19. The identified hotspots are densely populated areas with households living in very small, shaky houses built with mud or burnt bricks, earning low income and/or relying on casual labor, petty trade, and other such poor livelihood sources, and with high unemployment and/or job insecurity and thereby persistently not meeting basic needs (food, hygiene, shelter, etc.) (UBR/CUCI, 2020). A total of 76 wards were selected across the four cities as follows: Mzuzu (15), Lilongwe (27), Zomba (11) and Blantyre (23). At the household level, the eligibility criteria focused on a household’s livelihood vulnerability . The criteria considered which households in the hotspots were likely to lose their livelihoods due to COVID- 19-related economic slowdowns and social distancing measures. This was determined by gathering information to understand household level livelihood characteristics and sources of income, household composition, and job and food insecurity status. The CUCI eligibility criteria also considered the structural vulnerability of households. It was decided that the CUCI caseload would be equivalent to 35 percent of the urban population, which was expected to total about 185,000 households. Ultimately, however, the urban population was higher than anticipated, and over 199,000 households (or 37 percent of households in the four cities) were determined to be eligible. Cash transfers were provided on an unconditional basis for a duration of three months. The monthly transfer level of MK35,000 (approx. US$45) per household was equivalent to the prevailing minimum wage at the time of design in 2020. The transfer level was determined on the basis that most urban poor in Malawi are engaged in some form of informal employment (such as vending, ganyu/piecework and other less reliable types of employment) that provides a monthly income of at least the minimum wage. The duration was determined by an initial assessment of the duration for which COVID-19 adverse impacts on livelihood sources would last and, critically, by availability of financing. The intervention cost about 0.3 percent equivalent of 2020 GDP. CUCI monthly transfers commenced in February 2021. 22 3. The Process Evaluation: Analytical Framework and Methodology Analytical Framework The key analytical framework for the CUCI process evaluation is the delivery chain for social protection as conceptualized in Lindert et al. (2020) and applied to shock responsive social protection by Bowen and Smith (2020). As represented in Figure 2, the delivery chain presents the operational processes for social protection with reference to four phases (assess, enroll, provide, and manage) and nine discrete stages (outreach; intake and registration; assessment of needs and conditions; determination of eligibility and enrolment; decisions on the benefits package; notification and onboarding; payments of benefits; beneficiaries’ compliance, updating and grievances; and exit decisions, notifications, and case outcomes). (The references in Figure 2 to “recurring cycle� and “periodic reassessment� were not relevant to the CUCI.) The CUCI process evaluation collected data against each of the nine stages to help understand what was implemented through the CUCI, how it was implemented, and the successes and challenges experienced. Figure 2. Delivery chain for cash transfers Figure 1: Delivery chain for cash transfers Eligibility Determination Provision of Beneficiaries Exit decisions, Assessment and of benefits Notification benefits compliance, notifications, Source: Lindert et al. 2020. The CUCI process evaluation also incorporated the four main building blocks of adaptive social protection: programs; data and information systems; finance; and institutional arrangements and partnerships (Bowen et al, 2020). This conceptualization – as depicted in Figure 3 – represents the second part of the analytical framework for the CUCI process evaluation. To understand the extent to which they have influenced the design and implementation of the CUCI, the process evaluation collected data on each of these building blocks at all nine stages of the delivery chain. 23 Figure 3. Four Building Blocks of Social Protection Delivery Systems Source: Bowen et al (2020) Methodology Setting the Reference Points CUCI was unique in that, without prior experience of a similar shock response in urban Malawi, and given the rapidly evolving context, the implementation had to rely heavily on iterative adaptation. Broadly, the process was as follows: a manual for implementation of the CUCI was developed and agreed on by the government and development partners prior to implementing CUCI. Yet, as implementation took off, the government quickly adopted a method of iterative adaptation to adjust the intervention to emerging realities on the ground. While this worked well in some instances, it did not work out well in others, and this report discusses these issues in Chapters 4 and 5. Critically, though, this iterative adaptation meant that the process evaluation did not have a clear reference point against which to evaluate implementation. A process evaluation is about whether an initiative was implemented as intended, and it was therefore necessary to have consensus on the precise design of each stage of the CUCI delivery chain – prior to the commencement of implementation. The design parameters of each stage are referred to in this report as the “reference points�. Based on document reviews and consultations with the CUCI Process Evaluation Reference Group, and the Ministry of Economic Planning and Development and Public Sector Reforms (MEPD&PSR) and the Ministry of Gender, Community Development, and Social Welfare (MGCDSW), the broad timeline of the design is as follows. The CUCI program was initially designed in April/May 2020, as outlined primarily in the CUCI Guidelines and the CUCI Master Plan program documents. The initiative was paused around June 2020 due to the “fresh� Presidential Elections, with a plan to resume in August 2020. Between August and November 2020, prior to the rollout of the program, several changes were made to the design of the CUCI to reflect emerging information and developments in what was a very fluid context. These adjustments were not necessarily reflected in revised program documentation. The first step in the process evaluation was therefore to develop and reach a consensus with government on the reference points which represent the final design of the CUCI at each stage of the delivery chain. This was done through: (a) a series of discussions held by the Green Innovation Center 24 (GIC) with central government officials, city councils, and development partners; and (b) reference to documentation, including the CUCI Master Plan (2020) and CUCI Technical Working Group minutes of August, 2020. These reference points were then validated by the CUCI Process Evaluation Reference Group on March 18, 2021, with subsequent review and clarifications provided by officials from MEPD&PSR and MGCDSW on April 10, 2021. Appendix 1 presents the validated reference points. These will also be described before delving deep into the findings of the process evaluation for each stage. As highlighted below, some of the stages of CUCI do not align exactly with that of the analytical framework, and the differences are highlighted wherever relevant. Data Collection Framework A detailed Data Collection Framework was developed to summarize – across all stages of implementation – the key areas of inquiry, the types of inquiry methods to be used, and relevant stakeholders. This was done in the form of a matrix, and its purpose was to provide a synopsis of the entire evaluation and to stimulate cross-validation checks. For each of the nine stages of the delivery chain, the matrix included the following information with regard to data collection: reference points, questions to be asked in surveys, the type of tools to be used, and potential respondents. A copy of the Data Collection Framework is attached as Appendix 2. Data Collection Approaches The process evaluation adopted a mixed method approach involving a combination of qualitative and quantitative evaluation approaches. The mixed method approach helped to enhance the process evaluation findings through triangulation, complementarity, and expansion of breadth/coverage. Qualitative methods were used to strengthen the quantitative findings from the individual interviews, providing a basis for in-depth analysis and insights into the experiences of CUCI implementers, beneficiaries, and nonbeneficiaries. The mixed method approach involved the following: • Document review: A literature review was conducted of Malawi’s social protection strategic and program documents and relevant global reports, including on shock responsive social protection. The documents included the MNSSP II, the Master Plan for the implementation of CUCI and its related guidelines and implementation progress reports. The review also included collection of data from local resources including the UBR, SCTP-MIS and the COVID-19 response plan, which were used to complement the findings from the qualitative analysis. The document review helped to develop evaluation themes and questions as well as identifying data points, and possible interviewees for key informant interviews and consultations. • Focus Group Discussions (FGD): The FGDs were held to understand all stages of the delivery chain and shed further light on information collected through other tools, including how the program affected social dynamics. FGD checklists were developed to guide data collection from beneficiaries and nonbeneficiaries, and community, GRM, and ward committees. A total of 23 FGDs were carried out (Appendix 3). • Key Informant Interviews (KII): Key informant interviews were conducted to identify successes, challenges, and lessons learnt in relation to the implementation of the CUCI. KIIs were carried out at the community, district, and national levels. The KIIs included development partners, private sector actors (Airtel/TNM), block leaders, Ward Development Committees, government officials, City Councils, ward councilors, and implementing partner staff. The interviews were conducted individually to encourage active participation and questions were tailored to the interviewee's role and area(s) of expertise. In most cases, the interviews were done virtually to minimize contact in view 25 of the prevailing COVID-19 situation. A total of 55 KIIs were conducted (15 national and 40 City Council/community levels) (Appendix 3). • Survey for beneficiaries and nonbeneficiaries: A semi-structured questionnaire was developed to collect individual data from both beneficiary and nonbeneficiary households. The beneficiaries were identified through the CUCI MIS. The survey questions included information on household demographics, household economy, local economy, and community social networks, as well as impressions of the operational implementation of the CUCI across the delivery chain. In total, 1,313 door-to-door interviews were completed (985 beneficiaries and 328 non-beneficiaries). The household data was collected through KoBo Toolbox6-enabled Android smartphones by 16 enumerators. The recruitment of the enumerators was rigorously conducted: an advertisement was placed in the newspaper and a panel of GIC team members selected the most suitable candidates based on academic qualifications and experience in data collection and social protection programs. Sampling Method The survey used statistical computation to determine a sample size for beneficiaries and nonbeneficiaries. A minimum sample size for the beneficiary caseload (of 199,613 households) was determined as 799 households, at 95 percent confidence level with 3 percent precision (refer to Appendix 4 for sample size calculations). However, to make up for nonresponse or nonavailability, a 20 percent margin was factored in, increasing the sample to 959 households. This target was exceeded, as a total of 985 beneficiary households were reached. The survey also collected data on nonbeneficiary households to cross reference and triangulate relevant findings. The nonbeneficiary group was selected from households who were initially registered for the CUCI but later deemed ineligible and not enrolled in the program. (Such households were identifiable as their information was entered in the UBR.) Both beneficiary and nonbeneficiary populations lived in the same hotspots. The sample size for the nonbeneficiaries (ineligible households) was determined to be 288 households (using the rule of thumb by White II (2018) of 25 percent to 30 percent of the beneficiary sample size). The target was exceeded, as a total of 328 nonbeneficiary households were interviewed. Probability to Population Sample Size distribution was applied in calculating or allocating sample size per city and per ward within each city. Table 2 presents the distribution of sample sizes from 23 sampled wards across the four cities. The process evaluation employed systematic random sampling to select both the beneficiary and nonbeneficiary households. Specifically, every fifth household for both eligible and ineligible households was selected for interviews. In cases where selected households were not present, a replacement was made using the same selection approach; in practice, however, cases requiring replacement were very few. Table 2. Sample size distribution across the four cities City Enrolled HHs Sample Wards Total Sample HHs Control group HHs Mzuzu 13, 364.00 4 107 39 Lilongwe 82, 191.00 6 444 155 Zomba 23, 939.00 5 49 15 Blantyre 80, 119.00 8 385 119 Total 199, 613.00 23 985 328 6 https://www.kobotoolbox.org/ 26 Data Quality and Assurance A number of data quality assurance efforts were instituted and maintained throughout the evaluation. This was particularly critical given that much of the work was being undertaken by a research team that was working across time zones, and in a context in which physical visits were not possible during the entirety of the survey firm’s work. The World Bank research team held weekly briefings and monitored progress with the survey firm and undertook further analysis of the survey data to ensure data quality and findings. Where necessary, the World Bank team research team also joined meetings with government colleagues to ensure that the survey firm was up to speed and accurately understood the details of the intervention. The reference points and data collection framework were jointly developed with the survey firm, with constant feedback from government. Prior to the commencement of data collection, the evaluation team ensured the following: random sampling of beneficiaries and nonbeneficiaries, engagement of 16 qualified and experienced enumerators, adequate training of enumerators prior to data collection, pre-testing of the study tools and logistics, digitalized tools with necessary validation checks, and increased sample size to compensate for non-responses/errors. The survey firm team was separated into two groups, each led by two consultants and one supervisor, responsible for two cities each. All field team members applied uniform criteria throughout the data collection period. Supervisors provided feedback on the performance of the research assistants immediately after observation, thus helping to ensure early detection of any anomalies. During data collection, supervisors and consultants: (a) made frequent spot checks on the RAs; (b) ensured that pictures and videos were uploaded to ensure data originality; (c) made objective observations of respondents and the environment, (d) ensured that each data collection tool was checked for completeness and labeled daily by supervisors; (e) held daily and summative debriefing meetings with the enumerators to reflect on their experience in the field; and (f) required field supervisors and consultants to re-interview a sample of respondents. The evaluation concluded that there were no major issues with the interviews. Data Analysis The quantitative data collected through the individual beneficiary questionnaire and other quantitative tools was first cleaned before analysis. The cleaning process was to remove outliers, inconsistences, and exaggerated patterns of responses. Corrections, as necessary, were added to the database while the original data were retained in a separate field for reference. Data were largely analyzed using SPSS Data software and Excel computer software. Data outputs included frequency counts, percentages, charts, cross tabulations, means and standard deviations, which provided a distribution of respondents across the parameters. Once the data and analysis were handed over by the survey team, the World Bank team undertook another round of analysis to confirm the findings and further elaborate them, as necessary, using STATA software. The qualitative data from FGDs and KIIs were organized and reduced through a process of coding. Before coding the data, the team analyzed the participant information, clarified shorthand interview notes, reviewed transcripts for any inaudible comments and made necessary corrections to the file. The team developed a coding scheme that included a set of high-level codes geared toward separating the raw data into large buckets by theme, with lower level subcodes to identify data that addressed specific subtopics. The study team also conducted spot-checks for consistency of the coded information. The idea was to understand the prevalence of responses to deduce common themes. The team also examined the remaining data to determine whether comments from minority participants were outliers or reflected a 27 larger issue faced by a smaller group. The data were arranged into thematic-content narratives and embedded in the report to complement the quantitative findings and provide a reflective voice of whether the CUCI program was implemented as intended. Here again, once handed over, the World Bank team undertook another round of analysis to confirm the findings and further elaborate them, as necessary. Evaluation Timeline The process evaluation was undertaken over a period of 8-9 months starting in December 2020. The evaluation commenced with an inception period that concluded at the end of January 2021. Detailed documentation for the survey, including reference points, data collection framework, evaluation framework, survey tools and training and data quality plans, were developed and finalized between February and May 2021. Fieldwork was conducted in May-June, including training of enumerators and pretesting of study tools (17-18 May), field data collection (19-29 May) and national level KIIs (7-18 June). Data analysis and the draft report were completed by the end of June. There were various touch points with government and the CUCI Process Evaluation Reference Group that were incorporated into the evaluation process. This report was finalized in August-September 2021. Limitations of the Process Evaluation As with any process evaluation, this one had various limitations particularly as a result of the implementation context. First, the process evaluation was undertaken under the restrictions provided by the Malawi Health Rules on prevention, containment, and management of COVID 19 (GoM, 2020). Some participants were unwilling to attend physical meetings considering the prevailing COVID-19 situation. Thus, many KIIs, particularly for national level stakeholders, were conducted virtually through videoconference, phone calls and emails. Unfortunately, this often meant an inability to read body language and facial expressions, which are important ingredients for effective communication. To mitigate these challenges, the survey team used probing questions to understand the interviewees’ perspectives more deeply. Second, relatedly, the World Bank research team was unable to physically be available to monitor and participate in the evaluation. Weekly progress meetings were held with the survey firm and a thorough review of all survey-related documentation was undertaken. Third, there may be potential respondent bias as CUCI implementers and beneficiaries might have felt compelled to answer interview questions more positively than is actually the case to avoid a negative portrayal of the program in a report or for fear of being removed from the program. The survey team was careful to ensure anonymity of all respondents and to underscore to interviewees that the purpose of the process evaluation was to learn about program implementation and ultimately to improve service delivery, and so all types of feedback (positive and negative) were useful. Further, prior to interviewing any beneficiary, enumerators obtained consent from the respondents and were informed that their responses had no bearing on their enrolment and entitlement in the CUCI. Fourthly, the condensed intervention timeline for CUCI implies that it was challenging to identify the nine stages of the social protection delivery chain distinctly (Figure 2). While having a benchmark analytical framework helped to ensure the process evaluation was methodical, this also made it complex at times to clearly identify the successes and challenges of the CUCI. 4. Overview of findings This chapter provides an overview of the findings of the process evaluation by each of the nine stages of the social protection delivery chain. Each sub-section commences with a short definition of the relevant stage, before articulating what occurred and how the stage was implemented. 28 Outreach and Sensitization Outreach is the first stage in the social protection delivery chain (note that the language used to describe this stage of the CUCI is generally “sensitization� rather than “outreach�). The objective of this stage was to build awareness about the CUCI with the public, with a focus on urban hotspots where potential beneficiaries were living. It was intended to build awareness on a range of issues including a broad description of the program, its intended population, requirements and eligibility, and the entire process of on- boarding. Implementation The outreach stage of the CUCI was guided by a communication and sensitization strategy, which was developed by the Ministry of Gender, Community Development and Social Welfare (MGCDSW). The communication strategy was intended to inform target beneficiary communities about CUCI including information on benefit levels, GRMs, duration of the interventions and targeting mechanism. However, the Ministry reported that the strategy also helped to disseminate the messages more broadly to stakeholders at national, City Council and community levels. Primarily, two formal channels of sensitization were used, namely, cascading briefings, and radio jingles and press releases. In terms of cascading briefings, MGCDSW held a two-day physical meeting to orient all four City Councils, including Members of Parliament (MPs), Ward Councilors and the Council Secretariats, about the proposed design and implementation of the CUCI. Emphasis was placed on key elements of the delivery chain such as the transfer amount, the duration, the eligibility criteria for beneficiaries, GRMs, and the implementation modality. This also included an orientation of the City Council Secretariat and Ward Councilors about their expected role and responsibilities, such as carrying out community orientation and monitoring of the CUCI. The Ward Councilors were expected to hold similar briefing meetings with the block leaders, who were in turn supposed to share the information to the general public. Accordingly, they collaborated in disseminating the information to the wider public using funeral ceremonies and town criers as the main communication modes. MGCDSW also organized radio jingles and press releases to disseminate information on CUCI. However, the frequency and effectiveness of radio jingles and press releases were limited, and the CUCI registration process itself served as the most prominent sensitization mechanism. The radio jingles and press release were undertaken only once due to funding delays from the supporting development partner. Yet the survey data from the process evaluation shows that radio jingles were cited more often than the cascading briefings (19.6 percent versus 8 percent from block leaders and 6.8 percent from ward councilors) as the channel through which households became aware of the CUCI. On the other hand, the press release was mentioned by only 0.4 percent of households; through FGDs, the communities highlighted that newspapers were too expensive for them to access. However, about 80 percent of households highlighted that they found out about CUCI only during the registration process (stage 2 of the delivery chain). The depth of understanding of the CUCI’s parameters was limited. Beneficiaries and nonbeneficiaries were asked to recollect the content of the CUCI sensitization (Figure 4). Around two-thirds of households had learned of the CUCI transfer size, and approximately one-third were clear about the eligibility criteria, but very few of them had learned how to raise complaints or how to update information collected through the registration process. This points to the need for continuous and detailed sensitization to ensure that beneficiaries and the broader community are adequately informed of the program and its parameters. 29 Figure 4. Respondents’ recollection of the content of the CUCI sensitization Respondents were told what amount HHs Respondents were clear about the eligibility would be receiving criteria for the program 45% 40% 90% 86% 79% 40% 37% 80% 72% 34% 67% 69% 35% 32% 31% 67% 70% 60% 30% 60% 24% 49% 25% 22% 50% 20% 17% 40% 30% 15% 20% 10% 10% 5% 0% 0% Blantyre Lilongwe Mzuzu Zomba Blantyre Lilongwe Mzuzu Zomba % of total beneficiaries % of total non-beneficiaries % of total beneficiaries % of total non-beneficiaries Respondents learned about raising Respondents learned about updating complaints information collected during registration 30% 29% 18% 17% 16% 14% 25% 14% 13% 20% 12% 10%10% 13% 10% 15% 8% 10% 8% 8% 8% 6% 7% 4% 4% 4% 3% 5% 2% 1% 0% 0% 0% Blantyre Lilongwe Mzuzu Zomba Blantyre Lilongwe Mzuzu Zomba % of total beneficiaries % of total non-beneficiaries % of total beneficiaries % of total non-beneficiaries Nevertheless, about 65 percent of households that received some type of sensitization were satisfied with the outreach and sensitization process. Overall, 73 percent of beneficiaries and 41 percent of nonbeneficiaries that received sensitization in one or the other form rated the sensitization process to be good or very good. This implies that while many households were not reached by sensitization activities prior to the registration process, the quality of sensitization was perceived as satisfactory. Intake and Registration The second stage of the CUCI delivery chain is “intake and registration�, which refers to the process of collecting information on households in the eligible geographic areas (hotspots). (Information collected at this stage was used in the subsequent stages to assess the registrants’ needs and conditions (Stage 3) and determine eligibility (Stage 4).) 30 Implementation Due to the perceived risk of political capture and bias, external enumerators were engaged to independently carry out the registration process. 520 enumerators were engaged to register households for a period of 21 days (during September-October 2020). The enumerators were divided into teams of four with each team led by a supervisor and an official from the central government. Three days’ training was arranged for enumerators to be oriented on the CUCI program design, targeting approaches and the Rapid UBR Tool—which was a brief questionnaire developed by the CUCI Technical Working Committee to conduct door-to-door interviews. The Rapid UBR Tool was pretested for one day, after which there were no major changes. Registration was limited to households that displayed characteristics of less well-off households within selected hotspots. While all households in an identified hotspot were originally envisaged to be registered, due to time limitations (enumerators had targets of registering 30 people per day and they spent only two days in each hotspot), only those households that displayed specific characteristics were registered. These characteristics included reliance on casual labor, residing in densely populated areas, low income levels, loss of income source, engaged in petty trade, and persistent inability to meet basic needs i.e. food, clothes, and shelter. This process relied heavily on the enumerators’ judgement of household characteristics , potentially leading to inclusion and exclusion errors. The FGDs found that enumerators were required to make a quick judgement of whether a household was fit to be registered or not. While there were criteria for this, as highlighted in the paragraph above, there was still a level of subjectivity in the process. Additionally, there were no measures to cross-validate this mechanism, leaving room for targeting errors. The report explores these issues below in the analysis of the eligibility and enrollment stage (Stage 4). Nevertheless, a large share of beneficiary and nonbeneficiary household reported that the registration process was transparent and inclusive, with little political influence from community leaders (Figure 5). A large majority of beneficiaries (74 percent) and nonbeneficiaries (61 percent) were of the view that the process was transparent. They were also of the view that despite not all households being registered in targeted hotspots, almost all relevant groups in the area that were the focus of the intervention – namely older persons, unemployed, people with disability, women and the poor – were included in the program. Indeed, about 75 percent of both beneficiaries and nonbeneficiaries thought the process was inclusive. Similarly, about 80 percent of both beneficiaries and nonbeneficiaries thought that the process was not influenced by local politics and dynamics given the engagment of external enumerators. Such trust in the process, despite potential targeting errors, are crucial in successful implementation of time-sensitive shock responsive programs. Figure 5. Share of respondents that thought the registration process was transparent, inclusive, and not influenced by community leaders 31 Was the reigstration process in this Was the registration process in this community inclusive enough? community transparent? 100% 80% 74% 80% 61% 80% 75% 60% 60% 40% 30% 20% 40% 20% 16% 6% 9% 20% 11% 9% 9% 0% 0% No Not sure Yes No Not sure Yes % of total beneficiaries % of total beneficiaries % of total non-beneficiaries % of total non-beneficiaries Did community leaders have an influence on who should be targeted for CUCI after registration? 100% 82% 79% 80% 60% 40% 20% 10% 13% 8% 8% 0% No Not sure Yes % of total beneficiaries % of total non-beneficiaries There were, however, some significant data quality issues that delayed the overall pace of implementation of the program. Key information which was captured by enumerators included the name of household head and/or alternative, phone number, national ID number, and occupation. Some of the errors reported to have occurred during the registration process include collection of incomplete or inaccurate data. As a result of this, MNOs had to undertake a comprehensive Know Your Customer (KYC) exercise to match up the names of beneficiaries with phone numbers and National Registration IDs. This process caused a substantial delay in the implementation of the CUCI. As of June 2021 (6 months after the KYC process commenced), MNOs reported that only 53 percent of the registered households were corrected in the KYC system. The delay in completing the KYC demonstrates the degree and complexity of errors originating from the registration exercise. Assessment of Needs The third stage of the delivery chain focuses on assessing the needs and conditions of affected communities or households. The purpose is to identify the appropriate support to be provided, and this stage of the social protection delivery chain commonly encompasses a systematic process and detailed methodology for conducting the assessment. 32 Implementation Assessments of needs and conditions were reportedly done at two levels. The first assessment was meant to facilitate the identifcation of hotspots and was done prior to intake and registration of potential beneficiaries; the second assessment was aimed at identifying the households within the hotspots and was done during the intake and registration exercise. Hotspots were identified using a unique method of triangulating across multiple data sources. First, the Government of Malawi used City Councils’ socioeconomic profiles to help with classification of areas/wards into high, medium, and low-income areas based on the economic status of residents in that particular ward/area. Second, the classified areas were triangulated with other data sources both on income classification and population density, namely: UN Habitat reports (2011); WFP reports classifying areas of the four cities in terms of income-earning opportunities and characteristics of dwelling units (2020); and the Survey of Urban Poor Settlements in Lilongwe and the Ministry of Lands, Housing and Urban Development Report (2015). Despite the income classification of areas, the reality is that most areas had residents with mixed economic status and dwelling units. To narrow down to specific hotspots for the CUCI, further criteria were used: (a) areas with a population living in very small, shaky houses; (b) houses built with mud or burnt bricks; (c) densely populated areas; (d) low-income areas, with residents relying on casual labor, petty trade, and others with poor or no livelihoods sources and with high unemployment and/or job insecurity; and (e) areas with households who are persistently not meeting basic needs (food, hygiene, shelter, etc.). WFP and ILO played an active role in verifying and confirming the hotspots identified by the government in consultation with the City Councils. See Box 2 for more details. 33 Box 2: Hotspot verification exercise conducted by the World Food Programme and the International Labour Organization The main objective of the verification exercise was to independently demonstrate that the hotspots identified by the government met the targeting criteria. The exercise was done at the request of the Government of Malawi to show a proof of concept i.e., a proxy that shows that the process of identifying the hotspots was impartial and transparent. The exercise was carried out in Blantyre and Lilongwe because these districts had up-to-date data readily available. The exercise involved overlaying the identified hotspots with income classifications, Integrated Household Survey (IHS) 4 data and Very High Resolution (VHR) satellite imagery data as highlighted by the following steps: • Step 1: Income Classifications: Produced maps for the cities by ward, delineating and visualizing the income classifications (low to high, including mixed areas). Low-income classification characteristics included spatially concentrated residential dwelling and smaller size housing, limited green space between properties and irregular and unpaved road/street network. • Step 2: Disaggregated IHS4 data and identified households classified as informal labor sector and urban poor. Geotags for these households were mapped against the income classifications. • Step 3: Demarcated hotspots identified by the government using Google Earth. • Step 4: Used VHR satellite imagery to independently identify areas with poor quality houses and densely populated areas. The verification exercise confirmed that hotspots identified by the government did indeed represent settlements of high informality with low incomes as defined by the hotspot identification criteria. The maps shows that government-identified hotspots met the hotspot identification criteria with no inclusion errors (area of 96 km2 for Blantyre and 163km2 for Lilongwe). This is shown by the map below whereby the pink shaded areas representing settlements independently classified as low incomes by the verification exercise match with hotspots identified by the government shown by a mix of yellow and blue dotted lines for Lilongwe and blue lines for Blantyre. The exercise also found very small areas of exclusion errors i.e., informal settlements with low income, not identified by government as hotspots (area of 3.5km2 for Blantyre and 12km2 for Lilongwe), represented by the purple lines. Source: WFP (2020) 34 A majority of households reported that the hotspots were accurately identified. When the communities were asked about their perception of the selected hotspots, about 78 percent of respondents indicated that the selection was satisfactory. There was some variation across the cities, however, with Blantyre recording the lowest levels of satisfaction. The proportional responses per cities were Blantyre (67.6 percent), Zomba (82.8 percent), Lilongwe (80.8 percent) and Mzuzu (79.5 percent). Assessment of needs at the household level was based on only a few household characteristics that aimed at assessing the conditions rather than the actual needs of households. In other words, the assssment was based on the minimum expenditure for a household to survive, not on the actual needs of the households – which would have been generally high. This choice was based on the availability of time and resources. Assessment of needs, in addition to assessment of conditions, can be quite useful in determining an appropriate and varying benefit package (transfer level and duration) by beneficiary needs. In the case of CUCI, there was an interface with humanitarian response platforms such as the Cash Working Group, which had recently undetaken market assessments to determine benefit levels for the humanitarian sector. The needs assessment was therefore not largely used because the transfer level had already been predetermined at a flat amount for all eligible households, and all CUCI beneficiaries were to be paid the same amount over the same period, irrespective of diversity in needs as well as differences in household composition and financial/structural vulnerability. However, several national stakeholders underscored the importance of a needs assessment, and about 12 percent of households interviewed shared this view. Some FGDs and KIIs reasoned that the needs assessment could have allowed for the transfer levels to vary according to food prices in the different cities (as was originally planned). Eligibility and enrollment decisions The fourth stage relates to making decisions about whether registrants meet the eligibility criteria and should become enrolled in a social protection programme. Decisions are informed by the data collected during stages 2 and 3, and the decision involves producing or finalizing an updated beneficiary list before delivery of payment (Smith & Bowen, 2020; Lindert et al, 2020). Implementation The CUCI eligibility and enrollment process combined a proxy for income/ consumption to filter out ineligible households, and a community validation process to authenticate them on the ground. There were three stages to doing this, as follows. First, the registered households were subjected to an eligibility test and had to answer the three questions outlined in Table 3: If the household was considered not poor and as having of formal emplyment as a main and secondary livelihood, it was classified as ineligible and therefore excluded from the intervention. Table 3. Noneligibility test for households Question Answer Result Is the household poor? No Ineligible What was main livelihood source of the Formal Employment household during the last 12 months What was the secondary Livelihood source of Formal Employment the household during the last 12 months 35 Second, the households were ranked and prioritised based on financial vulnerability as indicated by one of the primary objectives of the CUCI. This was the first level of prioritisation. Three questions were used, out of which the highest level of importance was given to question 1, indicating the primary livelihood source of the household and the lowest to question 3, with home ownership status (Table 4). The financial vulnerability score of the households was calculated as the summation of questions 1, 2 and 3. The households were arranged according to the score, starting with the highest (where the score would equal 18 when the household answers, “the main source was begging, the secondary source was no activity, and home ownership was rented.�). Table 4. Economic Activity Score Question Answer Score 1. What is the main livelihood source of the household Begging 10 during the last 12 months Ganyu ( Duckerman) 8 Petty Trading 6 Informal Employment 4 2. What is the secondary livelihood source of the No activity 5 household during the last 12 months Begging 4 Ganyu 3 (Duckerman/woman) Petty Trading 2 Informal Employment 1 3. Home ownership status Rented 3 Free, Not Authorized 2 Free, Authorized 1 Employer Provides 0 Owned 0 Being Purchased 0 A second level of prioritisation was undertaken, based on the structural vulnerability of the household. In this case, three variables were used: (a) Household size; (b) Number of elderly above 60 years of age; and (c) Number of children under 18 years of age. The most structurally vulnerable households were those with the highest score, which was calculated by adding the three numbers (household size + elderly members + children). The households were arranged in a list according to the score, starting with the highest. The highest vulnerability score was provided to households with large households with elderly and children as members, while the lowest score was allocated to small households with neither elderly nor children as members. Households were selected for the program on the application of both the first level of prioritization (under financial vulnerability) and the second level of prioritization (under structural vulnerability). Households had to meet both criteria in order to be eligible. A total of 199,460 registered households were deemed eligible as CUCI beneficiaries, representing 90.2 percent of the registered households. About 9.8 percent were deemed ineligible because they did not meet the eligibility requirements of the program. Broadly speaking, the ineligible housholds included those whose houses were owned, or had neither elderly nor children as members. 36 Third, the list of eligible household names was sent to the four city councils, where the Ward Councilors disseminated the list to the communities through Ward Development Committees. During this process, the Ward Councilors and City Councils convened meetings, normally in schools, churches, or ward meeting halls, to verify and authenticate the existence of these households. This was a critical exercise as it helped to match names with National ID, phone numbers and payment service providers (TNM Mpamba or Airtel Money). Determination of Benefits and Services Package The objective of stage 5 of the delivery chain is to determine the benefits package (transfer value) that will be provided to each beneficiary during the payment cycle. Implementation This stage is usually closely linked with the assessment of needs and conditions (stage 3) discussed earlier, as information from the latter is used to inform the benefit levels in programs that have variable levels of benefits to help meet needs. In CUCI, however, the transfer value was determined prior to commencement of the program, and it was also decided that the same transfer level would apply across all beneficiary households given the time-sensitive nature of the CUCI and the consequent potential for errors and confusion in implementation. The process evaluation therefore does not assess whether this stage was implemented as intended. Instead, it describes how the transfer level was determined, followed by lessons learned. The CUCI transfer level was determined so as to meet the basic needs of a household whose members were unemployed as a result of the pandemic. As early as April 2020, it was envisaged that the transfer value for the CUCI would align with the minimum wage in Malawi during 2020, which was MK35,000 (approx. US$45) per month. The decision was taken after considering a range of transfer levels, including: (i) indexation of transfer levels to market prices across different urban areas, (ii) aligning with the SCTP transfer level, and (iii) adjusting for household size. In-kind food was also considered in case markets became inaccessible due to continued restrictions and/or the pandemic continued unabated. Some of these options came with complications, however, such as the need to conduct market assessment surveys in each city. It was also recognized that the payroll system should not be overly complex given the time- sensitivity. While data availability to guide this process was initially limited, a survey by WFP helped ascertain the transfer level. The most recent nationally representative data was the Integrated Household Survey 4 (IHS4), which showed a very wide variation in income levels—an extremely poor household with an average of size of 4.4 members lived on income ranging from MK20,000 (approx. US$26) to MK90,000 (approx. US$116) per month. To help move beyond the broad parameters of the IHS4, WFP conducted a Minimum Expenditure Basket survey7 in the four cities in July 2020. The result was that estimated food needs in the sampled hotspots were just under MK35,000 (approx. US$45) per month (MK32,584 – approx. US$42 – for food items only). Further analysis of household surveys by WFP and ILO suggested that a transfer equivalent to the minimum wage to 35 percent of urban households would both offset the anticipated impacts of a lockdown and cover anticipated food security requirements (Roelen et al, 2021). Ultimately, it was decided to confirm that the transfer level would be the minimum wage. 7 https://reliefweb.int/sites/reliefweb.int/files/resources/MEBReport_23July_ROUND%206.pdf 37 A little over half (54 percent) of beneficiaries considered the transfer value to be adequate, with nonbeneficiaries suggesting a reduction in transfer value to help expand the CUCI caseload (Table 6). The remaining 46 percent of beneficiaries proposed an increase in the CUCI transfer level to between MK70,000 (approx. US$90) and MK200,000 (approx. US$257) per month. They also suggested that the transfer value should have considered the following factors: size of the household, child-headed households, households with school-going children, and the cost of living. Conversely, nonbeneficiaries proposed a reduction in the transfer value from MK35,000 (approx. US$45) to MK10,000 (approx. US$13) so that CUCI could reach every household in the selected hotspots or all those that registered. And despite the legislative change to raise the minimum wage to MK50,000 (approx. US$64) in June 2020, the actual transfer value was not adjusted. Table 6. Beneficiaries’ perception of the adequacy of transfers City Yes No Total Yes No Blantyre 228.00 276.00 504.00 45.2% 54.8% Zomba 38.00 26.00 64.00 59.4% 40.6% Lilongwe 315.00 284.00 599.00 52.6% 47.4% Mzuzu 85.00 61.00 146.00 58.2% 41.8% Total 666.00 647.00 1,313.00 53.9% 46.1% Note: Yes = adequate; No = not adequate Unsurprisingly, households’ understanding of the rationale for the current transfer level was low (Table 7). When beneficiaries were asked if they had any idea of how the transfer level was determined, only 9 percent expressed some knowledge and were able to point out that the MK35,000 (approx. US$45) was based on the minimum wage. There was some variation across the cities, with households in Blantyre and Zomba showing significant stronger understanding than those in Lilongwe and Mzuzu (Table 6). Table 7. Respondents’ knowledge of the basis of transfer level City Percentage of respondents who were aware Blantyre 18.4% Zomba 18.4% Lilongwe 0.8% Mzuzu 3.7% Total 9.0% Notification and Onboarding This sixth stage of the delivery chain is about notifying beneficiaries that they have been enrolled in CUCI and providing them with information about the program. Implementation The program envisaged a dual notification mechanism to inform the beneficiaries and communities of the status of enrollment in the program and the schedule of payments. This included: (a) publication of names of enrolled heads of household by Ward Committees and block leaders, either through public meetings or noticeboards in schools (CUCI Guidelines, June 2020); and (b) text messages being sent through the mobile money operators to enrolled households notifying them about: their beneficiary 38 status, the amount to be received during the first transfer and subsequent months, and the mode of transferring the funds (in this case through mobile money). In practice, while the process of notification aligned with the design of the program, it does not appear to have been fully or effectively implemented. Less than half of surveyed beneficiaries confirmed they had been notified. Figure 6 shows that 48 percent of beneficiaries confirmed receiving notification primarily through SMS (48 percent) and Ward Councilors (27 percent). Only 21 percent of nonbeneficiaries confirmed receiving notification about their ineligibility, mainly from ward councilors (47 percent). The evaluation also established that the channels and intensity of notification were not consistent across the cities. For instance, in certain hotspots, some community members were not aware that a list of eligible households had been sent to their respective areas. Onboarding and notifications were intertwined and, often, during this process beneficiaries were also required to do their KYC with MNOs. Figure 6. Channels for received notifications pertaining to CUCI targeting 60% 48% % of total beneficiaries (N = 983) 50% 40% 30% 27% 20% 15% 9% 9% 10% 0% SMS Block leader visit Public place listing Ward councilor Ward committee Data gathered by the evaluation suggests that the information received by beneficiaries was not clear enough, including detail on the next stages of CUCI and other relevant information. Only a small proportion of households reported that the notification was clear and comprehensive8. An average of just 1 in 10 beneficiaries (11 percent) reported acquiring clear and comprehensive information about their eligibility. Zomba had the highest proportion of households who reported having received clear information, compared to Blantyre (8 percent), Lilongwe (13.1 percent) and Mzuzu (9.8 percent). There was also a particular lack of information for the households that were ineligible for CUCI. The evaluation found that a majority of beneficiaries were not aware of the payment schedules. On average, only 18.8 percent of the beneficiaries knew about the payment schedule through the notification/onboarding process, and mainly via SMS sent by MNOs. It appears that TNM Mpamba, which covered Blantyre and Zomba, was more efficient than Airtel Money in sending notifications to 8 Comprehensiveness here means that the notification clear enough about all relevant information for the next stages of CUCI. 39 beneficiaries about the impending payments as 26 percent in Blantyre and 35 percent in Zomba reported knowing the payment schedule. In Lilongwe and Mzuzu, only about 10 percent of beneficiaries knew of the payment schedule. Given these findings, it is not surprising that a majority of households felt that the notification and onboarding process needed to be improved. Between 50 percent (Blantyre and Zomba) and 75 percent (Lilongwe and Mzuzu) of beneficiary households across different cities felt that this process needed to be improved. Provision of benefits The focus of stage 7 of the delivery chain is providing transfers to beneficiaries. This means distributing the correct amount of benefit to the right people, at the right time, and with the right frequency. Implementation As urban economies usually have functioning markets and relatively more robust financial systems, the program decided to provide the cash transfers through Mobile Network Operators (MNOs). The MNOs used in the program were Airtel Money and TNM Mpamba. This builds on the experience of using e- wallets for payments in the SCTP in Machinga, Mchinji and Balaka districts of Malawi. The assumption was that most households in urban areas have handsets registered with Airtel Money or TNM Mpamba (the registration questionnaire was tailor-made to collect such data). The payment process was designed to be seamless to reduce the burden on beneficiaries in terms of logistics and the cost of accessing the transfers. MGCDSW was responsible for preparing the payroll for beneficiaries, using the CUCI MIS. When all authentication processes were completed, the payroll was sent to MNOs for payments. The first payments were intended to be made in January 2021. The service charge and cash out charges were to be absorbed by CUCI and not borne by beneficiaries. Beneficiaries were expected to access funds from MNOs’ agents located in various townships in the four cities. For each transfer made, the MEPD&PSR was expected to make a press release in the media. This action was meant to meet the transparency requirement to the general public. However, in practice, the first payment was delayed beyond January 2021 primarily due to issues with the KYC requirements of eligible households. As of June 30, 2021, all three payments had been made to 32,000 households, representing 16 percent of all enrolled households. About 33 percent of beneficiaries received only the first payment, 22 percent of households received two payments, and 29 percent of enrolled households hadn’t received any payments at all (Table 8). As described earlier in the report, the issues with KYC verification were the result of inaccurate data collected during registration (e.g., mismatching names, SIM card details and phone numbers, and mixing up of names). The design phase had assumed that urban households would have up-to-date KYC on the MNO systems. Unfortunately, this was not the case, and MNOs had to undertake KYC verification for a majority of eligible households. Table 8. Progress on disbursement of CUCI funds to beneficiaries as of June 30, 2021 Eligible 194,762.00 Frequency (percent) First Payment 63,624 32.7 Second payment 42,148 21.6 Third Payment 32,000 16.4 Not paid even once 56,990 29.3 40 Despite delays, beneficiaries were of the opinion that payments were done in a timely manner, at a time when their livelihoods were most impacted, and that they received the full amount to which they were entitled. Except for Mzuzu, over 65 percent of the beneficiaries who knew the schedule agreed that CUCI payments were delivered on schedule. Moreover, 93 percent of beneficiaries were of the view that payments were made when they were suffering the most from the negative consequences of COVID-19. Nearly all (96 percent) beneficiaries thought they received the amount to which they were entitled. As payments are the most critical aspect of the program in terms of achieving its objectives, this is a strong sign of the program’s success. Two-thirds of beneficiaries reported not facing liquidity issues at MNO agents and only a small minority of beneficiaries reported being charged more than they expected to cash out. Approximately 65 percent of beneficiaries reported that their MNO agents had sufficient liquidity to pay their cash entitlements. Where liquidity issues were experienced, beneficiaries were referred to super MNO agents located in strategic places within the townships but generally outside the hotspots. Special arrangements were made to extend their normal operating hours to serve CUCI beneficiaries. Nevertheless, there were some issues with mobility, especially for the elderly and people living with disability to access super MNOs located further away from the hotspots. Only about 8 percent of beneficiaries reported being charged more than they expected to cash out their transfers. Overall, more than half of surveyed beneficiaries reported that MNO agents adopted appropriate COVID-19 measures. Use of masks by agents was the most widespread COVID-19 measure used across all cities. There was a large variation in terms of other measures across the cities. Agents in Blantyre and Zomba performed strongly in terms of not admitting customers without masks and having handwashing facilities in place. On the other hand, whereas agents in Lilongwe and Mzuzu did not perform well on those two measures, they were stronger at observing and enforcing social distancing (Figure 7). Figure 7. Multiple preventive measures were enforced by MNO agents % of total beneficiaries (N = 551) 120% 99% 96% 91% 89% 100% 61% 80% 55% 60% 39% 31% 27% 25% 25% 25% 22% 40% 17% 14% 12% 12% 12% 10% 9% 9% 7% 7% 7% 7% 6% 20% 5% 2% 0% Agent wore Agent did not Agent Agent Agent Agent had Agent mask admit sanitized observed enforced handwashing enforced customers hands physical physical facility at handwashing without masks distance distance by his/her place before service clients Blantyre Lilongwe Mzuzu Zomba 41 A large majority of beneficiaries fully cashed out the transfers at the MNO agents, driven by need and cultural beliefs. Eighty-five percent of beneficiaries cashed out the money at once. This was commonly justified on the basis that beneficiaries feared losing access to the funds, particularly those who used other people’s handsets. The fear was intensified by the perceived randomization of the time for receiving the funds (which corresponds with the lack of knowledge of the payment schedule). In fact, many community members referred to the transfer as “satanic money which would vanish anytime, if not withdrawn�. For those who had to travel a distance to cash out from MNO agents, avoiding the cost of transportation was a factor. Transfers were overwhelmingly used to purchase foods; less than half of beneficiaries used it to purchase household assets or shared with friends/relatives in need; and fewer still used it to meet education expenses or to boost their existing businesses (Figure 8). The uses highlighted here strongly point to the success of the program in meeting some of the basic needs of households arising from the economic impacts of COVID-19. That some households shared money with relatives and friends points to the widespread needs prevalent in urban Malawi at the time of the CUCI. At the same time, it is encouraging to see the potential human capital and productive effects these transfers will have on beneficiaries who allocated the funds to non-food purposes. Figure 8. How did beneficiaries use CUCI money? Share of all households (percentage) Not spent yet 1% Repaid loans 5% Medical expenses 7% Building materials/ House reconstruction 15% Boost business 17% Education expenses 27% Shared with friend/relative 39% Purchased household items/assets 41% Purchased food 87% 0% 20% 40% 60% 80% 100% Nearly every beneficiary household was satisfied with the e-payment modality used to transfer payments. About 97 percent of beneficiary households were either very satisfied or satisfied with the system. They highlighted that they perceive mobile money payments as secure and safe, and promoting the dignity of the household. Beneficiary Compliance, Updating and Grievance Redress For the purposes of the CUCI, the key focus of stage 8 is the GRM and program monitoring . In other contexts, this stage of the delivery chain also considers issues such as beneficiary data management and compliance with conditions. 42 Implementation The GRM was conceived as a platform to handle grievances through two channels, namely, a call center and local committees. The purpose of the GRM (both committees and call center) would be to ensure that beneficiaries and nonbeneficiaries from the hotspots had access to appropriate channels through which they could voice any concerns about CUCI. The referral system would ensure that responsible institutions duly resolve all registered complaints. It was intended that there would be a CUCI-specific GRM committee established in each ward of the four city councils. Each GRM committee would be mandated to handle any complaints arising from implementation of the CUCI. In practice, the GRM was established only around the time of field work for this process evaluation, and so was not widely known. Only 10 percent and 8 percent of all respondents (beneficiaries and nonbeneficiaries) were aware of the GRM committee in their ward and the Call Center, respectively. Moreover, GRM committees were not active in all wards. The Call Center was established two months later than expected and was housed in MGCDSW, with support from UNICEF. While M&E frameworks and protocols were in place during implementation of CUCI, lack of coordination between central ministries and local institutions hampered these activities from being undertaken in a structured way. At the central level, the TWG, MoG and MEPD&PSR had weekly meetings where all aspects of program implementation were shared, including transfer levels, financial commitments, and KYC, among others. Central agencies (MEPD&PSR and MGCDSW) also conducted spot checks in the hotspots across the four cities to check on agent liquidity, ensure that beneficiaries were receiving the correct amount, and deal with any other issues. These spot checks were relatively light, however. Monitoring activities were also undertaken by City Council members but their roles and responsibilities with respect to these activities was less clear. As a result, they produced no M&E reports or other form of outputs. The M&E framework for CUCI was not fully known at the City Council level, and Council staff believed that MEPD&PSR and MGCDSW would be responsible for writing progress reports on the implementation of CUCI. The evaluation found that City Council members had no access to the M&E tools designed by central ministries. Not surprisingly, only 3 percent of beneficiaries reported that some form of monitoring was done by the government. The Call Center was overwhelmed with calls, many of which went unanswered due to capacity constraints. On some days, each of the four handsets at the Call Center would receive 31 calls per hour. Exit Decisions, Notifications, and Case Outcomes For the purposes of CUCI, the ninth stage of the delivery chain is about the conclusion of the program and how that process is managed with beneficiaries. The program had not yet completed this stage and so could not be evaluated in-depth. Yet, the evaluation finds that all beneficiaries surveyed were aware that the duration of CUCI was only three months, i.e., three transfers. Only one-fifth (21.9%) were aware of the exact month when CUCI would come to end, but this is unsurprising as many of them were not aware of the payment schedule. 43 5. Lessons learnt This chapter describes the lessons learnt with regard to each stage of the CUCI delivery chain, based on analysis of the overall findings. Outreach and registration Equipping CUCI registration/enumeration teams with information on CUCI was an effective way to safeguard against the limited execution of the formal channels of outreach and sensitization. While the use of multiple channels of outreach and sensitization provided a potentially wider reach to stakeholders at national and community level, its limited execution implied constrained reach in practice. The radio jingle and press release fitted well for people who had access to radio and newspapers, especially given the prevailing COVID-19 related restrictions on gathering and in-person meetings. At the same time, the physical interactions through community-based meetings (by block leaders and Ward Councilors) were more suitable for people with limited access to information technologies. Yet it was the CUCI enumerators that had the greatest reach in terms of informing households about the intervention. In any future intervention, the government could pay attention to this layer of staff and their understanding of the program to ensure quality of sensitization. However, doing so has its own drawbacks in terms of creating unnecessary expectations of the enumerators and inadvertently reducing their effectiveness at the registration exercise. The outreach process helped to establish a critical foundation for cooperation between central government ministries and urban governance structures which can be useful for future interventions in urban areas. The outreach activity involved several stakeholders at national, City Council and community levels. As described above, MGCDSW developed the communication and sensitization strategy and oriented the City Council. In turn, the City Council through the Ward Councilors briefed the block leaders who passed on the information about CUCI to the communities on the ground. Such a cascade model is effective in terms of resource use and maximization of reach. Given that this was the first time that central government ministries had collaborated on a social protection program with these urban structures, this is an important success. Sufficient and timely availability of resources is an important aspect of successful sensitization efforts, especially in time-sensitive interventions. One of the reasons cited by the government for insufficient execution of these activities was the delay in disbursement of funds – especially for radio jingles. At the same time, government officials acknowledged that the stipulated awareness/outreach meetings at Ward and block level might have overstretched the staffing and equipment of the City Councils. The City Councils testified that the staffing and funds allocated to them were inadequate to effectively cover all the community awareness meetings (using COVID-19-friendly tools i.e., cars with loudspeakers) across the targeted Wards. MGCDSW admitted that the sensitization process could have been done better if more resources were availed on time. Going forward, shock sensitive interventions need to pre-plan for such activities and have financing ready for implementation, to improve project execution. Further engagement of local sensitization methods may yield greater returns in terms of creating awareness of the program. The process evaluation found that, in some instances, the block leaders did not use their usual means of sensitization but instead relied on representatives who went door-to-door 44 to inform households of the program. Some households emphasized the use of local methods such as ‘whistleblowing at late hours’ to improve the reach of key program information. Continuity of sensitization efforts is key to how well beneficiaries and communities will understand a new intervention. Given the lack of similar programs in urban areas, it seems critical that efforts be made to continually sensitize the relevant stakeholders of the objectives of CUCI and its parameters. While CUCI has a relatively short lifespan compared to other more longer-term social protection programs in Malawi, it is a relatively complex and novel intervention. Continuous sensitization would ensure that information on the program is made available to stakeholders over a longer period of time. This could be through providing the information in smaller chunks and/or through refresher initiatives which restate and reinforce the key information – rather than providing all the information on one occasion only, which can impede understanding and overload the communication channels. Intake and Registration Leveraging the experience of Malawi’s rural social registry (the UBR) provided the registration process with a significant head start. The MEPD&PSR and MGCDSW – as key stakeholders of both the UBR and SCTP respectively – adapted and employed existing programmatic linkages between the SCTP MIS and the UBR. They innovated and developed an abridged version of the UBR’s harmonized data collection tool, referred to as a Rapid Data Collection Tool, which helped to facilitate intake and registration of potential CUCI beneficiary households. Using existing institutional structures further facilitated its rapid implementation. The registration exercise was undertaken in a short period of 21 days. The City Councils, despite never having done such an exercise before, proactively used their existing local structures such as Community Development Committees (CDC), Ward committees and block leaders to facilitate the registration process. The use of external enumerators to undertake the data collection helped build trust with the communities. Cash transfer programs in Malawi usually use local government structures, working closely with community structures, to carry out household registration. However, due to the emergency nature of the program, MGCDSW engaged independent enumerators for the task. This helped to reduce the risk of perceived bias and political capture which might otherwise have adversely affected the implementation of the program. The mechanisms for quality control were insufficiently strong to optimize the overall efficiency of program implementation. The evaluation found that enumerator training was lacking. For example, in many instances they were not aware that the phone numbers and SIM cards were critical inputs to the payment process. At the same time, many enumerators did not take the time to verify the details of households being registered. For example, some households used the ID numbers of other households as they had forgotten their own. While these issues are part of any data collection exercise, oversight usually provides a quality control or early check on implementation. However, COVID-19 restrictions implied that in-person monitoring, and oversight was limited on the part of the central ministries. City Council members were not required to play this role either. As a result, monitoring and oversight of the registration process was weak, leading to data quality issues. Improved planning and preparation for similar shock responsive programs is critical for success. Despite the support provided by government and partners to the registration process, it is likely that there were some targeting errors arising from the design of the process itself, time sensitivity, and the scale of need. 45 For example, the evaluation found multiple instances wherein poor elderly women were not registered on the program. It is recognized that every program has targeting errors, and that there may also be valid explanations for the anecdotal examples of exclusion encountered by the evaluation. Nonetheless, it must also be noted that – in light of the procedural weaknesses highlighted above – it is possible to reduce these errors through better planning and preparation for such programs. Assessment of Needs Given the need for agility and speed, use of multiple datasets can provide a robust targeting mechanism. As highlighted above, the identification of hotspots was done by overlaying and triangulating multiple datasets, which proved to be quite successful both in terms of accurately identifying hotspots but also building trust with communities. Furthermore, the process demonstrated the benefits of working with several stakeholders such as ILO, WFP, Ministry of Lands, UN Habitat and City Councils as part of identifying the hotspots. Planning for shocks can provide an opportunity to collect a greater amount of data on each household to accurately assess needs and cater to those needs to the extent possible when shocks hit. The data collected through the UBR Rapid Data Collection Tool was limited to a few variables only, and did not allow for a more accurate measure of household needs. Going forward, the social registry (UBR) could be expanded to urban areas and other geographies or subgeographies not currently covered by it, to allow for a more robust household identification mechanism. Eligibility and enrollment decisions Clearer verification and monitoring mechanisms need to be in place to reduce the inclusion errors which are suggested by both qualitative and quantitative data. Of the beneficiary households surveyed for this process evaluation, 17.2 percent had formal employment as their main livelihood strategy while a total of 43.3 percent either owned their own house (42.5 percent) or resided in a house which they were in the process of purchasing (0.8 percent). According to the eligibility criteria for the CUCI, it would be expected that households with such characteristics would not be eligible. However, these households were found eligible and included in the beneficiary list. FGDs and KIIs across the cities confirmed the inclusion errors as evidenced by inclusion of some households who were landlords, employed, and/or who owned prosperous businesses. The lesson here is that errors must have occurred during the registration process. Qualitiative evidence suggests that this was likely due to the fact that verification mechanisms were not clearly established in the registration process. In addition, as highlighted earlier, oversight and monitoring during registration were weak and as a result led to these inclusion errors. Leveraging existing local structures in the verification and oversight process could potentially reduce targeting errors. Such a community validation and use of local institutional structures is part of rural social protection programs in Malawi. Given the scale of need in the country and financing constraints, exclusion errors are likely; however, strengthening the targeting criteria and quality of implementation can reduce exclusion errors. Table 5 looks at the livelihoods of beneficiaries and non-beneficiaries. It finds that, on average, there is little difference between the two. While these characteristics do not comprehensively cover the targeting criteria used (e.g., financial and structural vulnerability indicators are not included here), this evidence does suggest that the targeting criteria used may not be clear enough to distinguish beneficiaries. In a time-sensitive situation, it may be worth considering if universal targeting of specific hotspots is a better targeting approach to reduce exclusion, even though it may involve inclusion errors. 46 Table 5: Livelihood types by treatment status Livelihood type Benficiairy Non-beneficiary Main livelihood: Formal employment 12.29% 13.72% Main second livelihood: Formal employment 5.18% 2.44% Main livelihood: Begging - Main livelihood: Piecework/Duckerman 1.73% 2.13% Main livelihood: Petty trading 13.92% 10.37% Main livelihood: Informal employment 4.07% 2.13% Main livelihood: None of the above 0.91% 1.83% Main second livelihood: Begging - - Main second livelihood: Piecework/Duckerman 9.76% 3.05% Main second livelihood: Petty trading 8.63% 5.49% Main second livelihood: Informal employment 3.86% 1.52% Main second livelihood: None of the above 60.87% 76.52% Determination of Benefits and Services Package Multistakeholder collaboration across the humanitarian and social protection sectors was effective. The coming together of humanitarian and social protection partners to help determine the transfer value of CUCI was applauded by both government and development partners. There has been increasing interaction and engagement between humanitarian and social protection actors in Malawi in recent years, through sustained efforts to advance the shock-responsive social protection agenda. While the sectors have often focused on their respective priorities, this was not the case in the design of the CUCI transfer level, and the pre-existing collaboration bore fruit. Using multiple sources of data provided robustness to the analysis. As highlighted above, using multiple sources of data, including a new quick survey using the expertise of humanitarian actors, was crucial in providing a clear rationale and robustness to the transfer value. This is critical as shocks come unannounced and require that implementing agencies undertake quick and dirty analyses using the best available data to make decisions regarding program parameters. With sufficient preparation and planning, it may be possible to implement a shock responsive program that responds to household level needs. The CUCI transfer package was “one size fits all� in the sense that it did not cater to the complex family structure, needs and household-specific situations such as size, presence of lactating mothers or persons with disabilities, number of school-going children, and other factors that can have an impact on the adequacy of a transfer. That said, the complexity of administering such a scheme is likely to be significant. Nevertheless, by building a strong enough database of households prior to a shock, it may be possible to address household and spatial differences in needs induced by a shock. Benefit levels tend to generate strong opinions in the community regarding the fairness of the program, and a robust communication mechanism can help inform some of those opinions. A lesson for future social protection programs is that it would be beneficial for beneficiaries to fully understand the basis on which a package or transfer level is decided. In the context of CUCI, such information would have helped in guiding the CUCI beneficiaries to understand the rationale and intended uses of the money. This is 47 particularly important if the program decided to customize the benefit levels to household and spatial differences in needs. Notification and Onboarding Stakeholder collaboration was critical, particularly, the engagement of local institutional mechanisms. It was encouraging that many stakeholders were involved in working to improve the effectiveness of this stage of the CUCI delivery chain. The GRM committee members and City Council officials and, in some cases, members of the TWG, worked in conjunction with Airtel and TNM to expedite the process of onboarding. The evaluation found that the use of traditional institutional structures (e.g., local structures and the City Council) was key to the implementation of this stage. Regular and clear information to communities and beneficiaries/nonbeneficiaries is key to building trust in the program and for beneficiaries to start factoring information about eligibility into their household budgeting processes. Despite payments having started, the evaluation found instances of nonbeneficiaries not knowing whether or not they had been selected for the program. One female nonbeneficiary said, “We just hope that maybe after this 3-month cohort which our friends are benefiting from right now, we may be considered in the next cohort. The blue card which we received during registration gives us hope.� On the other hand, some beneficiaries reported that between registration and payments there was a huge silence, and they got to know of their eligibility only when they received a text message about having received the payment. Beyond the fact that lack of communication on eligibility can lead to confusion at a time when the shock may have already created a sense of insecurity, this issue also has consequences on the financial decisions taken by beneficiaries. For example, beneficiaries may be able to avoid negative coping mechanisms if they are aware that they will be paid in the immediate future. For beneficiaries, the notification should include information such as: the amounts they would receive, the date of payment, the locations where the transfer could be cashed out, and information regarding beneficiary rights and a GRM contact should they encounter any hurdles. Good communication with beneficiaries and nonbeneficiaries will also help fortify local structures. The work of councilors and GRM Committee members was made difficult by incomplete or, at best, partial information. These structures spent a substantial amount of time handling cases of nonbeneficiaries. In some instances, disgruntled households organized demonstrations against Ward Councilors and GRM Committee members. Provision of benefits E-wallets are generally trusted by beneficiaries. The modality of transfers was a clear success of the program. A large number of respondents viewed the modality of e-wallet as a highly preferred option because of the reduced risk of theft and robbery compared to physical cash. Cash (as opposed to in-kind) was also noted as an appropriate means of support for vulnerable groups such as older persons and women, as there was no need for them to carry heavy in-kind items (such as food). A few opportunistic theft cases were noted, but these were ably handled by GRM committees and, where necessary, the police were engaged. It was also reported that the e-wallet helped to reduce the chance of local leaders being accused of bribery. The creation of proxies for those beneficiaries who did not have a handset or whose National IDs had expired was seen by many beneficiaries as an excellent innovation. That said, it might not have always been successful. At the time of data collection, a case was being handled by the police where a close relative of an elderly beneficiary in Zomba tricked her and cashed out all the money from her e-wallet account. There were also other anecdotal cases of family feuds, but these emanated from 48 differences on how to use CUCI money as a family. Such cases were handled by GRM committees and Victim Support Units within the communities. E-wallets also help to manage fiduciary issues efficiently. On a programmatic level, electronic payments make tracking of expenditures much easier than cash-based transactions. An audit trail of transactions can be easily identified within a reasonable timeframe, and the data is also secure and not easily tampered with in electronic devices. An electronic payment system also promotes accountability and easier reconciliation of transactions resulting in reduced costs and less time spent in auditing the program. Liquidity issues should be factored in when scheduling payments and in performance management of the payment service providers. The most significant challenge for this stage of the delivery was liquidity. The low liquidity levels experienced at local MNO agents meant that beneficiaries needed to travel to super agents to cash out. This was particularly so during the initial cash transfer payments. The need to cash out from super agents may have led to increased barriers and risks to vulnerable groups including people with disabilities, women-headed households, or older persons. Managing KYC requirements is key for the timely delivery of benefits to those affected by shocks. The CUCI design assumed that urban residents already had their KYC up to date with MNOs. Unfortunately, this was not the case and together with errors in the registration, has resulted in significant delays in payments. The delays caused by the verification exercise led to many beneficiaries losing hope that they would ever receive the CUCI transfers. Planning this into the preparatory activities for shock response interventions will be crucial. Relatedly, planning for early engagement of payment service providers, even before the shock occurs, would be advantagous. Good coordination and collaboration in addressing the KYC issues helped facilitate CUCI payments sooner rather than later – at least for a proportion of beneficiaries (as at the time of fieldwork for this evaluation, the KYC process was not yet fully complete). Strong and consistent collaboration efforts between central and local institutions (city officials, transfer agencies, ward councillors and GRM committees) and payment service providers helped to facilitate the process of updating KYC information. The MNOs have showed a strong commitment to ensure that beneficiaries’ records in their system are updated and correctly documented. To fast track the KYC process, the MNOs have provided free SIM cards to those whose records were not accurately presented. MNOs, as key stakeholders in the CUCI program, have been forthcoming with solutions to help beneficiaries with technical hitches. Beneficiary Compliance, Updating and Grievance Redress Resources in terms of both financing and time are critical to design and successfully implement GRM and M&E systems. CUCI’s GRM and M&E system suffered from lack of financing as well as time to get them up and running. This was partly due to the time-sensitive nature of CUCI and the lack of an existing social protection program in urban areas. Nevertheless, these are important aspects of short-term shock responses and the window of opportunity to refine implementation issues is very small. Significant resources must be devoted to aligning and clarifying roles and responsibilities across all levels of government with respect to GRM and M&E. The central ministries had assumed that the urban institutional set up would necessarily follow their obligations and mandate under the Decentralization Act to undertake M&E activities. The CUCI experience showed that this may not be sufficient, and that central 49 ministries must devote resources to aligning expectations and, as required, devote greater financial resources to ensure these activities are undertaken. 50 6. Conclusions and Recommendations This chapter sets out the conclusions and recommendations of this evaluation, which have been structured as follows. After some initial overarching remarks, we focus in the first instance on certain aspects of the delivery chain. Some of these are grouped together under broader headings, for convenience and ease of use. Second, we focus on elements that fall within the “building blocks� of shock sensitive social protection. For every topic under each of these headings, we commence with observations/conclusions, before proposing recommendations. Overarching comments Overall, the results show that CUCI implementation has largely been satisfactory across the stages of the delivery chain, and the survey data points to a range of positive highlights of the program. For instance, most beneficiaries perceived the hotspot identification and household registration processes to be transparent, inclusive, and correctly done, with little political influence from community leaders. Beneficiaries also felt that payments were delivered at a time when their livelihoods were most impacted, and that they received the full amount to which they were entitled. The implementation phase nonetheless experienced some challenges. For example, one of the main shortcomings of the program was that there have been substantial payment delays due to significant data quality issues at the registration stage. The data constraints reduced the overall pace of program implementation and affected the scope of coverage. These and other challenges faced by the CUCI do not, however, undermine the overall satisfactory implementation of processes across the delivery chain. Moreover, the CUCI has also generated many lessons learnt, and the program has laid a strong foundation for future shock responsive social protection initiatives and urban safety nets – both in Malawi and beyond. Elements of the Delivery Chain Intake and registration There were several encouraging aspects to the registration process, including the fact that many different stakeholders were able to collaborate effectively and complete a large-scale registration within a short period of time. The challenging context of COVID-19 placed additional demands on the actors involved. Other positive elements include the fact that City Councils adapted to the process despite having no previous experience in such exercises, and that independent enumerators were engaged to reduce the risk of perceived bias and political capture. That said, the quality of data collected during the registration process created significant challenges at later stages of the delivery chain and undermined the overall efficiency of registration. There are two aspects in this regard: first, the fact that about 10 percent of registered households were later found to be ineligible due to duplicate information or living outside a hotspot; and second, the mismatch of names and IDs relative to information provided by respondents in their e-wallet database. We recommend that the government investigate the source of the inaccurate data and put measures in place to address those concerns. The process evaluation has been unable to determine the source/s of 51 the errors that led to almost 10 percent of registered households being found ineligible. It appears that the registration exercise was not adequately monitored for data quality and that errors were not identified and corrected at the data source. This is not a KYC issue. It is likely to be a question of the standard operating procedures for data collection, checking, and storage, and adherence to those agreed processes. It may also have been a question – in whole or in part – of the extent of supervision by local structures (such City Councils, Ward Councilors and block leaders). This aspect is addressed below under the recommendations relating to capacity building. We also recommend that measures be taken to reduce the potential for a future mismatch between information collected at registration and that held by the e-wallet database. MNOs could participate in future registration processes, as appropriate, with a view to eliminating any KYC errors at source. All potential beneficiaries could be asked about which phone number they would like to use to receive the money, if they are found eligible. While there is a risk that this sets up an expectation of receiving benefits (which may not occur, as it is quite possible that the number of registered households exceeds the number of potential beneficiaries), good communication can mitigate that risk to some extent. Another mechanism could be to undertake another enumeration process once eligibility is ascertained to confirm phone numbers and any other basic logistical information. Through innovative mechanisms, the government may be able to ascertain eligibility through existing data (e.g., in rural areas, the UBR may already have basic data that could be complemented with satellite data) and undertake an enumeration exercise to confirm key information. The CUCI, unfortunately, didn’t have the benefit of such a database. Setting of transfer values One of the most commendable aspects of CUCI was the extent to which there was collaboration between actors across the social protection and humanitarian sectors. It could not be expected that the implementation of a new urban initiative such as CUCI would be seamless in this regard. But it is important to highlight the setting of transfer values as one of the areas in which recent years of collaboration on adaptive social protection paid dividends. The leveraging and deeper analysis of data provided a degree of rigor to the proposed transfer value of MK35,000 (approx. US$45). That said, the setting of transfer values was not undertaken on the basis of a needs assessment, as would often be the case in a social protection program. The original proposal for a transfer value of MK35,000 (approx. US$45) was proposed ahead of the analysis being undertaken which subsequently gave credibility to that figure. It is recommended that, in future, a more detailed analysis of needs be undertaken to facilitate a more mature and reasoned approach to determining the transfer value for an urban program. Ideally this would be undertaken in advance of a proposed transfer value being discussed in public, so as to reduce the risk of expectations being created. This is not to say that more information should have been collected by enumerators during the registration process. This evaluation recognizes that additional questions on a survey can have a significant impact on survey time, and we have not evaluated the UBR Rapid Data Collection Tool with this in mind. During a time of emergency, a more comphensive survey, such as one that captures a fuller picture of a household’s livelihood and financial abilities, is unlikely to stay current for very long. There are many moving parts and the risk of collecting and relying on data that may quickly become out-of-date should not be overlooked. Our suggestion is therefore focused on a stronger upfront assessment of the 52 relationship between existing needs and the proposed transfer level, based on the information that was collected during registration. A stronger link between the transfer value and an assessment of needs would also provide the foundation for any potential variation in transfer levels according to intrahousehold composition such as young children, pregnant and lactating women, persons with disabilities, or older persons. While this evaluation understands that there may been a range of reasons for setting a flat transfer level for all households, including what was administratively feasible and also perceived by the public to be reasonable, the lack of an assessment of needs meant that there was no analytical foundation on which any such decisions could have been made. Such analysis would also help inform the inevitable trade-off that governments need to make between coverage and adequacy. The evidence from the evaluation in this regard is ambiguous. Nonbeneficiaries thought the transfer level should be lower, whereas beneficiaries thought it adequate or even insufficient. This evaluation does not pass judgement on whether it should have been lower or higher or the same, but recommends that in future there should be more analysis to inform these difficult decisions. As an aside to the broader issue of how this part of the delivery chain was implemented, we note that the evaluation found CUCI funds to be beneficial to the urban poor community. The funds facilitated investment and opening of new businesses (e.g. through purchasing a fridge or commencing operation as an MNO agent), resuscitation of closed businesses, allowed households to buy food and items such as iron sheets and mattresses. Eligibility criteria and enrolment process The process of setting the eligibility criteria also likely benefited from collaboration from multiple stakeholders, as evidenced by the inclusion of both financial and structural criteria. The recognition of households with vulnerable groups, such as children under 18 and older persons, is suggestive of a mature social protection sector that is leveraging the lessons of other programs such as the SCTP. Another successful aspect of this stage of the delivery chain was the determination of household eligibility by the CUCI MIS. This appeared to function smoothly, and it is likely that central government officials, including those in the MGCDSW, have benefited from their previous experience with information systems such as the SCTP MIS. It seems probable that their familiarity with the SCTP MIS helped the CUCI eligibility process to operate efficiently. However, there was also a major oversight in this stage of the delivery chain which led to significant delays. Not asking MNOs to validate the information regarding eligible households led to substantial confusion and delays. While the delays may have been inevitable given that the registration period had already been completed, and in light of the fact that the errors would have needed to be corrected at any point, more advance warning would have helped to manage expectations and implementation planning. It is recommended that action be taken at the registration stage, as noted above, such as asking potential beneficiaries on which phone number they would like to receive the money – if they are found eligible. E-payments and KYC verification The use of an electronic payment method was highly appreciated by beneficiaries who felt it reduced the opportunity for theft and/or corruption when compared with cash in hand. Beneficiaries in FGDs also indicated their preference for cash over in-kind, especially for vulnerable groups. 53 The evaluation also found that CUCI was able to leverage the lessons from SCTP regarding the use of e- payments. It is recommended that any urban intervention continue to leverage lessons from SCTP, but also that there be consideration of what can be learned from the use of e-payments by other programs in Malawi – such as those in the humanitarian sector. It may therefore be worthwhile seeking to understand whether there are insights or evidence from humanitarian responses that could be used in the social protection sector. This could also be an opportunity to forge further collaboration between the two sectors. One of the major shortcomings of the CUCI payment system was that it was overloaded at the time of the first payment and there was insufficient liquidity in many hotspots. This suggests that the MNOs were inadequately prepared for what was already known to them—that there would be a large influx of cash in certain geographic areas, and it was likely that much of this would be withdrawn instantly. There may be lessons from the humanitarian sector on how it deals with large amounts of cash transfers. (The SCTP transfer level is much lower, and the payment method is through banks rather than MNOs, meaning that liquidity of MNO agents is not as much of a concern.) One of the most significant risks which we have not been able to quantify is the risk posed to vulnerable groups who are forced to access their funds from a super MNO agent located outside the hotspot. We recommend that future urban cash transfer initiatives consider other payment points such as banks, post offices and other points beyond MNO agents. Notification and communication to beneficiaries and nonbeneficiaries While efforts were made to sensitize eligible households, we recommend this be an area for substantially better planning and execution in future initiatives. The bottlenecks that occurred in the CUCI process – such as financing – should be identified and rectified in future initiatives. We also recommend that communication methods be adapted to an urban context. CUCI used channels of communication that are commonly used in rural areas, but it appears that these are not well-suited to an urban environment. Consideration could also have been given to the use of more innovative means of reaching urban communities, such as the use of digital technologies. The notification process was also unsatisfactory, both for beneficiaries and non-beneficiaries. We recommend there should be adequate and timely communciation regarding eligibility and enrollment decisions to all those who were registered, irrespective of whether they were found eligible or not. High quality notification would help to avoid creating unnecessary expectations from ineligible households. And it would also help insulate local structures, particularly the GRM Committees and Ward Councilors, from being accused of meddling with the system of selecting eligible households. Grievance Redress Mechanism On the basis of the evidence collected by this evaluation, the CUCI GRM was disappointing in many respects. The lack of timeliness in establishing GRM Committees in three of the four cities, and the limited functionality of the Call Centre, are causes for concern. The evaluation was not able discern the precise nature of the bottlenecks, although there appears to have been disagreement or lack of clarity on the topics of finances and roles and responsibilities in the case of the GRM committees. We recommend that greater efforts are made to develop appropriate GRM structures for urban areas, and that this be planned in advance of another shock. Again, there could be lessons here from the humanitarian sector, which regularly sets up accountability mechanisms at short notice. A harmonized approach to grievance redress across the humanitarian and social protection sectors may help to alleviate such issues in future. This 54 would include the thorny and difficult issue of finances, so as to ensure there are operational costs available—irrespective of the source—for local councils and Ward structures. Monitoring framework The monitoring of the CUCI was another area that the evaluation found to be disappointing in its design and execution. That said, it is recognized that the COVID-19 context added a significant additional challenge. Further, this was an urban initiative and the central government officials had previous experience with rural areas only. We recommend that these obstacles be addressed through various means. First, the MNSSP II monitoring and evaluation framework should be adjusted to account for urban measures. This may be through adding an urban-specific component to the framework, or through mainstreaming an urban dimension throughout the existing framework (including the shock response component). To some extent this depends on what the GoM sees as the likely future trajectory for urban measures – as part of routine social protection, or only as shock responses, or both. It is important to note that this adjustment would extend beyond issues such as selecting indicators that are appropriate for the urban context. Crucially, the roles and responsibilities of urban actors should be defined and agreed. It appears that this may have been a shortcoming of the CUCI. We also recommend that the monitoring and evaluation framework should consider the increased staffing needs that may arise in the context of a shock response. We note that this is not specific to urban areas. In the event of shock responses such as a vertical expansion, or a new caseload (e.g., through horizontal expansion or a new program), it is crucial that existing and routine social protection programs continue to function as normal. This means that the design, implementation, and monitoring of emergency programs will create an extra burden on staff. Preparing for such eventualities, through measures such as hiring additional staff, sub-contracting staff, or reassigning existing staff, will lead to a timelier and effective response, and higher quality monitoring. Building blocks Coordination and leadership As has been noted above and throughout the evaluation, one of the qualified successes of CUCI has been the coordination across different levels, sectors and geographic areas, and the leadership shown by central government agencies in a very challenging context. The evaluation received positive feedback on structures such as the TWG and the way in which many tasks were allocated to actors with a comparative advantage. Deciding on what household-level interventions are appropriate and feasible during an emergency can be a daunting task, especially in a complicated urban environment. The setting up of the TWG was an enabler; it facilitated sharing of experiences and information which guided the programming of CUCI. It appears that the efforts of the past 5-6 years to strengthen the shock-responsiveness of social protection in Malawi have borne fruit in this regard. That said, coordination was more complex in the hotspots, and we recommend that efforts be directed to establishing coordination structures and relationships between central government agencies and urban structures (including City Councils and Ward Committees). There is a need to establish stronger implementation arrangements for urban social protection interventions, whether shock responsive or routine. This may also include having clearer roles and responsibilities across the delivery chain. 55 Relatedly, and building on experiences of the SCTP, it would be valuable for central government and local structures to spend some time building relationships and an awareness of each other’s priorities and ways of working. Capacity building of local actors While there has been much to commend regarding the work of local actors in the implementation of CUCI, we recommend that consideration be given to building their capacity with regard to social protection. Similar efforts have been undertaken for a number of years with regard to district level officials who are involved in the SCTP. Given the likelihood that urban social protection programming will occur again at some point in the future, either in response to a shock or potentially as part of some routine provision, we see strong merit in building the capacity of local actors. Peer to peer learning can be effective and efficient, and consideration should be given to working with district councils that are already experienced in the SCTP to share lessons and experiences with City Councils. Enabling environment The current policy framework for social protection in Malawi does not yet include an urban dimension. This evaluation does not necessarily propose that there is an urgent need to prioritize a new policy for urban social protection. It is recognized that there is a long list of recommendations, and that adding a new policy to that list could result in significant time and resources being absorbed for a lengthy process. This is not to say that a policy or strategy is ill-placed, but rather that it should be considered at a later time and perhaps when there has been more urban programming and more lessons learnt. What we would recommend at this point is the incorporation of urban dimensions into current thinking on shock responsive social protection in Malawi, and that urban programming also be incorporated into the review of the MNSSP II (which currently expires in 2023) and its successor. One of the under-recognised successes of 2020-21 for the GoM is that it was able to fulfil a crucial responsibility in the midst of a shock, namely: ensuring that the rural SCTP continued to make regular and predictable cash transfers to its approximately 200,000 beneficiary households. This achievement should not be underestimated. Existing programs can sometimes collapse in the face of unexpected widescale shocks, and efforts to ensure the resilience of a routine social protection program represent an active policy choice by governments (Barca and Archibald, 2020). It is therefore highly encouraging that the GoM was able to maintain effective delivery of the SCTP throughout the design and implementation of the CUCI and in the midst of a global pandemic. We recommend that, when the time comes to review and revise Malawi’s social protection policy framework, that it be updated to reflect this important dimension of shock response. While the MNSSP II includes a significant focus on shock sensitive social protection, it does not yet reflect – e.g. as a guiding principle – that routine safety net programs should continue to operate through a shock and not be compromised by scalability efforts, including new programs. Financing The evaluation found that financing commonly arose as a bottleneck at various stages of the delivery chain. We are not aware of broader discussions regarding the financing of social protection – including shock responsive social protection – but we would see merit in taking a more coherent and comprehensive 56 approach. There is a range of recent guidance and literature on the financing of shock responsive social protection. Data and the national social registry Malawi’s UBR has the potential to be a strong tool for future urban responses. The data that has already been collected through CUCI may be of significant value in the event of a future shock, although a light touch updating may be necessary. More broadly, and looking ahead, we recommend that strong consideration be given to an expansion of the UBR to pre-register a larger proportion of the urban population in Malawi. Globally, there are heightened risks to urban populations from disasters such as flooding and storms, and the COVID-19 pandemic has laid bare the vulnerability of urban residents in the event of covariate shocks. In this regard, attention should be given to registering those with heightened vulnerability to various shocks, including geographic locations (e.g., vulnerable to natural disasters) and also members of vulnerable groups (e.g., older persons, people with disabilities). Given pockets of urban poverty, a more detailed understanding of urban income poverty may also be useful and would underscore the magnitude of the challenge at hand. Investment in such preparedness actions could greatly improve the timeliness and accuracy of future urban responses. 57 References Aggarwal, Shilpa, Jeong, Dahyeon, Kumar, Naresh, Park, David Sungho, Robinson, Jonathan, Spearot, Alan. 2020. Did COVID-19 Market Disruptions Disrupt Food Security? Evidence from Households in Rural Liberia and Malawi. CEGA Working Paper Series No. WPS-141. Center for Effective Global Action. University of California, Berkeley. https://doi.org/10.26085/C3JS39 Ambler, Kate; Herskowitz, Sylvan; Maredia, Mywish K.; and Mockshell, Jonathan. 2020. COVID-19 in rural Malawi: Perceived risks and economic impacts. Project Note November 2020. Washington, DC: International Food Policy Research Institute (IFPRI). https://doi.org/10.2499/p15738coll2.134156 Amjath-Babu, T.S., Krupnik, T.J., Thilsted, S.H. et al. 2020. Key indicators for monitoring food system disruptions caused by the COVID-19 pandemic: Insights from Bangladesh towards effective response. Food Sec. 12, 761–768. https://doi.org/10.1007/s12571-020-01083-2 Apedo-Amah, Marie Christine; Avdiu, Besart; Cirera, Xavier; Cruz, Marcio; Davies, Elwyn; Grover, Arti; Iacovone, Leonardo; Kilinc, Umut; Medvedev, Denis; Maduko, Franklin Okechukwu; Poupakis, Stavros; Torres, Jesica; Tran, Trang Thu. 2020. Unmasking the Impact of COVID-19 on Businesses: Firm Level Evidence from Across the World. Policy Research Working Paper; No. 9434. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/34626 Aoyagi, Chie. 2020. Effects of COVID-19 on Regional and Gender Equality in Sub-Saharan Africa: Evidence from Nigeria and Ethiopia. International Monetary Fund. https://www.imf.org/en/Publications/WP/Issues/2021/06/18/Effects-of-COVID-19-on-Regional-and- Gender-Equality-in-Sub-Saharan-Africa-Evidence-from-460912 Barca, V. and Archibald, E. 2021. Social Protection Interventions as Medium- and Long-Term Responses to the Pandemic, in Subramaniam, R., Perdiguero, A., Rush, J. and Asis-Layugan, P. 2021. Policy Actions for COVID-19 Economic Recovery: A Compendium of Policy Briefs. Asian Development Bank. Manila, Philippines. https://www.adb.org/sites/default/files/publication/715196/policy-actions-covid-19- economic-recovery.pdf Baulch, Bob; Botha, Rosemary; and Pauw, Karl. 2020. Short-term impacts of COVID-19 on the Malawian economy: Initial results. MaSSP Report June 2020. Washington, DC: International Food Policy Research Institute (IFPRI). https://doi.org/10.2499/p15738coll2.133788 Beazley, R., Marzi, M., Steller, R. 2021. Drivers of Timely and Large-Scale Cash Responses to COVID19: what does the data say? Social Protection Approaches to COVID-19 Expert Advice Service (SPACE), DAI Global UK Ltd, United Kingdom. https://socialprotection.org/discover/publications/space-drivers-timely- and-large-scale-cash-responses-covid-19-what-does-data Blake, Paul, and Divyanshi Wadhwa. 2020. 2020 Year in Review: The Impact of COVID-19 in 12 Charts. Accessed August 23, 2021. https://blogs.worldbank.org/voices/2020-year-review-impact-covid-19-12- charts. Bowen, T. and Smith, G. 2020. Adaptive Social Protection – Delivery Chain and Shock response. World Bank. Chikoti, Lizzie; Caruso, German Daniel; Vundru, Wilbert Drazi; Ilukor, John; Kanyanda, Shelton Sofiel Elisa; Kanyuka, Mercy; Kilic, Talip; Mleme, Tiope; Moylan, Heather G.; Mvula, Bright; Mwale, Isaac; Mwalwanda, Twika; Yoshida,Nobuo.2020. Monitoring COVID-19 Impacts on Households in Malawi: 58 Findings from the Fourth Round of the High-Frequency Phone Survey (English). Monitoring COVID-19 Impacts on Households in Malawi Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/784171607320182211/Findings-from-the-Fourth-Round- of-the-High-Frequency-Phone-Survey Cho, Yoonyoung; Kawasoe, Yasuhiro; Rodriguez, Ruth Reyes; Valenzuela, Myra. 2021. COVID-19 G2P Cash-Transfer Payments: Case Study: Philippines (English). Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/368351625163110870/COVID-19-G2P-Cash-Transfer- Payments-Case-Study-Philippines CUCI Technical Working Group (2020): Minutes of Technical Working Group, Lilongwe Gentilini, Ugo; Almenfi, Mohamed Bubaker Alsafi; Blomquist, John D.; Dale, Pamela; De La Flor Giuffra, Luciana; Desai, Vyjayanti Tharmaratnam; Fontenez, Maria Belen; Galicia Rabadan, Guillermo Alfonso; Lopez, Veronica; Marin Espinosa, Ana Georgina; Natarajan, Harish; Newhouse, David Locke; Palacios, Robert J.; Quiroz, Ana Patricia; Rodriguez Alas, Claudia P; Sabharwal, Gayatri; Weber,Michael.2021. Social Protection and Jobs Responses to COVID-19: A Real-Time Review of Country Measures (May 14, 2021) (English). COVID-19 Living Paper Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/281531621024684216/Social-Protection-and-Jobs- Responses-to-COVID-19-A-Real-Time-Review-of-Country-Measures-May-14-2021 GoM, 2020. Government notice no.48. public health act (cap.34:01) public health (corona virus and covid-19) (prevention, containment and management) rules IFC. 2020. Sub-Saharan Africa: COVID Economic Impact. https://www.ifc.org/wps/wcm/connect/05008fd5-a427-4fae-8521-fa6c73d82d88/2020-09-09-SSA- Regional-Brief-COVID-19.pdf?MOD=AJPERES&CVID=nhO.iFv. Khamis, Melanie; Prinz, Daniel; Newhouse, David; Palacios-Lopez, Amparo; Pape, Utz; Weber, Michael. 2021. The Early Labor Market Impacts of COVID-19 in Developing Countries: Evidence from High- Frequency Phone Surveys. Policy Research Working Paper; No. 9510. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/35025 Khan, Syed Abdul Rehman, Asif Razzaq, Zhang Yu, Adeel Shah, Arshian Sharif, and Laeeq Janjua. 2021. Disruption in Food Supply Chain and Undernourishment Challenges: An Empirical Study in the Context of Asian Countries. Socio-Economic Planning Sciences, February 13, 2021, 101033. https://doi.org/10.1016/j.seps.2021.101033. Kugler,Maurice David; Viollaz,Mariana; Vasconcellos Archer Duque,Daniel; Gaddis,Isis; Newhouse,David Locke; Palacios-Lopez,Amparo; Weber,Michael.2021. How Did the COVID-19 Crisis Affect Different Types of Workers in the Developing World? (English). Policy Research working paper; no. WPS 9703; COVID-19 (Coronavirus) Washington, D.C.: World Bank Group. http://documents.worldbank.org/curated/en/409921624030877958/How-Did-the-COVID-19-Crisis- Affect-Different-Types-of-Workers-in-the-Developing-World Lindert, C et al (2020): Sourcebook on Foundations of Social Protection Delivery Systems, World Bank Group Mahajan, Kanika, and Shekhar Tomar. 2021. COVID-19 and supply chain disruption: evidence from food markets in India. American Journal of Agricultural Economics, vol. 103, No. 1, pp. 35–52. Malawi Government (2016): National Resilience Strategy – 2018 – 2030 59 Ministry of Finance Economic Planning and Development (2020): Proposal for the Implementation of Crisis Interventions to address Covid-19 Effects (CUCI), Lilongwe. Ministry of Gender, community development and Social Welfare (June 2020): Technical Document- COVID 19 Urban Cash Transfer (CUCI), Lilongwe Ministry of Local Government and Rural Development (2013): A Guide Book on Local Government Systems in Malawi, Lilongwe Mzumara GW, Chawani M, Sakala M., (2021): The health policy response to COVID-19 in Malawi. BMJ Global Health 2021. http://dx.doi.org/10.1136/bmjgh-2021-006035 NSO (2010): Enumerator Manual for the Household Questionnaire, Zomba Paci, Pierella. 2021. How Livelihoods Deteriorated in Sub-Saharan Africa Due to COVID-19. https://blogs.worldbank.org/africacan/how-livelihoods-deteriorated-sub-saharan-africa-due-covid-19. Paul, Boban Varghese, Arden Finn, Sarang Chaudhary, Renata Mayer Gukovas and Ramya Sundaram. 2021. “COVID-19, Poverty and Social Safety Net Response in Zambia�, World Bank Policy Research Working Paper 9571. Peterman, Amber, and Megan O’Donnell (2020). COVID-19 and violence against women and children: a third research round up for the 16 days of activism. CGD Notes. Washington, D.C Rubio, G (2012): Building Results Frameworks for Safety Net Projects: Discussion Paper 1218. World Bank. Smith G and Bowen T (2020): Adaptive Social Protection. The delivery chain and shock response, World Bank The World Bank (2020): Malawi Economic Monitor: Doing more with less; Improving Service Delivery in Energy and Water Sectors, Lilongwe. Tiare Cross and Andrew Johnson (2011): Cash Transfer Programming in Urban Emergencies: A tool kit for practitioners. Oxford UK. TRANSFORM (2020): Shock Responsive Social Protection: Manual for Leadership and Transformation United States Agency for International Development, (2009): An evaluation framework for USAID- funded TIP prevention and victim protection programs. Social Transition Team, Bureau for Europe and Eurasia. https://riselearningnetwork.org/wp-content/uploads/2015/09/An-evaluation-framework-for- USAID-funded-TIP-Governance-and-Social-Transition-of-the-USAID-2009.pdf. Vaessen Jos et al (2020): Evaluations of International Development Interventions. An Overview of Approaches and Methods, The World Bank. UNDP. 2021. Mitigating Poverty: Global Estimates of the Impact of Income Support during the Pandemic. WHO. 2020. Impact of COVID-19 on People’s Livelihoods, Their Health and Our Food Systems. https://www.who.int/news/item/13-10-2020-impact-of-covid-19-on-people's-livelihoods-their-health- and-our-food-systems. World Bank. 2020. Malawi Economic Monitor, December 2020: Doing More with Less - Improving Service Delivery in Energy and Water. https://openknowledge.worldbank.org/handle/10986/34931 60 World Bank. 2020b. COVID-19 Impact Monitoring: Round 4. Lilongwe, Malawi. https://microdata.worldbank.org/index.php/catalog/3766/download/49198 World Bank. 2021. Global Economic Prospects, January 2021. Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/34710 World Bank. “COVID-19 Household Monitoring Dashboard.� Accessed August 23, 2021. https://www.worldbank.org/en/data/interactive/2020/11/11/covid-19-high-frequency-monitoring- dashboard. World Food Programme (July 2020): Minimum Expenditure Basket Report, Round 6, Lilongwe World Food Programme (September 2020): Integrated Food Security Phase Classification, Lilongwe World Food Programme/ILO (September2020): Hotspot Verification Exercise-CUCI, Lilongwe 61 Appendix 1: Data Collection Framework Table 1.A provides the Data Collection Framework which was used to develop the survey tools for the process evaluation. The table maps the stages of the delivery chain with the specific questions, respondents and tools. Table 1.A: Summary of key areas for process evaluation for CUCI Stage Steps What is known so far Specific questions for process Potential respondents Tools/methods (to be updated) assessment ASSESS Outreach Aim: To ensure that · What communication Beneficiaries/ Non HH Qtr. all target methods were used to inform beneficiaries FGD checklist communities are well you/target communities about MoGCDSW EPD City informed about the the CUCI cash transfer KII checklist council, NLGFC, wards objective of CUCI program? (was it mobile van, program. town criers, press release, meetings with block leaders, radios) Planned outreach · How were the MGCDSW, EPD, UNICEF, KII Checklist message: communication methods GIZ, NLGFC, City · Duration of cash identified/ chosen? councils, wards transfer, possible · What was the content of the Beneficiaries, Non-Ben. HH Qtr., FGD eligibility criteria and message you received during checklist registration process the sensitization meetings? · What gender issues were Beneficiaries/ Non HH Qtr. Planned Activities: integrated in the outreach beneficiaries FGD checklist design? Were gender issues · Stakeholders’ MoGCDSW EPD City also integrated into the KII checklist sensitization council, Wards outreach activities that were · Identification of delivered? partners at all levels Tip: Timing, venues of · Community meetings, child friendliness, mobilizations target criteria, exploitation etc. · Resource · Why did you/communicators Min GCDSW, EPD, City KII checklist mobilization choose to use such Councils, Wards communication methods? · Full district council Were there different forms of meetings communication methods at · Production of different levels of hierarchy? posters, short videos, Explain. 62 Radio jingles, mobile · Which communication Beneficiaries, Non-Ben. HH Qtr. vans, leaflets, social methods were more effective Block leaders, Min KII checklist media – digital and less effective and why? GCDSW, City Councils, platforms, bulk SMSs What other methods could UBR, Transfer agencies, FGD checklist have been more appropriate Block leaders, Ward · Use of influential than what was used and why? committee members people – faith-based groups, local leaders · How many visits were made Min GCDSW, HH Qtr., KII for outreach. to the wards (by Councils & beneficiaries, Non-Ben. checklist central level team) to inform City Councils, Block FGDs communities about the CUCI leaders, Ward cash transfer program? committees · How were the local Block leaders, City KII checklist structures involved in Councils, ward community sensitizations about councilors, ward the CUCI cash transfer committee program? · Were all population groups Min GCDSW, City KII checklist reached, such as the elderly, Councils, Beneficiaries, FGDs those lacking literacy, PWD? If Non-Ben. City, Wards so, how was this achieved? If not so, how and why? · Was any data used to Min GCDSW, EP&D, KII checklist determine what population WFP and City Councils, groups lived within the wards? Wards · Did you adapt the messages Min GCDSW, City FGD checklist for particular needs such as Councils, EP&D, UNICEF, KII checklist PWD (e.g. sign language), NLGFC, Communication illiteracy, those without access team to media (e.g. radio or TV)? If so, how was this attained? If not, why? · How much financial and Min GCDSW, EPD; FGD checklist human resources did you NLGFC, City Councils KII checklist allocate, or were allocated, to and development the outreach activity? partners · To what extent were the Min GCDSW, EPD; KII checklist resources adequate to meet NLGFC, City Councils the needs of the intended outreach activities? Were there any innovations developed to improve the speed or timeliness of planned activities? What are some of the 63 innovative processes that are worth learning about or documenting for future reference in similar programs? · How effective was the Min GCDSW, EPD, FGD checklist coordination between central NLGFC, beneficiaries, KII checklist and local government levels for City Councils, Block the outreach phase of the CUCI leaders, Ward cash transfer program? committees · Were any outreach / Min GCDSW, EPD, FGD checklist communication methods of beneficiaries, City KII checklist existing programs, such as the Councils, Block leaders, SCTP, used or leveraged for the Ward committees CUCI? · Based on your reflection Min GCDSW, EPD, FGD checklist what do you consider as having beneficiaries, City KII checklist gone well or not during the Councils, UBR team, outreach activities? And why? Block leaders, PSPs [explain your answer] Were there any disruptions to Min GCDSW, EPD, City FGD checklist the outreach activities i.e. due Councils, Wards KII checklist to 2020 elections, COVID 19 pandemic, political interference? Intake and · What key criteria and tools City Council, GCDSW, KII Checklist registration were used to define the EP&D, city council, ward What was planned by eligibility of the wards councilors & WFP EPD and Min GCDSW: (hotspots) and their sub- · Register all sections? What data sources HHs in hotspot areas were used? · The · Do you think there could have City Council, GCDSW, KII Checklist identification of been improvements in the EP&D, city council, ward FDG hotspots will be done process of selecting hotspots? If councilors, beneficiaries by the City Council so, what? & block leaders/ward Technical Team committees · Door to door · Which levels of government, City Council, Wards KII Checklist data collection via and which MDAs, were involved GCDSW UBR collection tool in deciding the process for Central and local govt, eligibility criteria of geographic DPs areas (hotspots). How were they involved? Was there an effective coordination process 64 · Use of for involving all the different tablets, ODK forms to government actors? collect the data Note: It is possible for · Recruit 520 respondents at this stage to enumerators for start also talking about HH household eligibility criteria. Notes must registration be taken to ensure HH eligibility criteria question is rephrased in · Collect Hhd subsequent sections based on data within 21 days response in the question. · Target · Who ultimately decided City Council, GCDSW, KII Checklist 185,247 HHs which geographic locations EP&D, city council, ward · Exercise to were eligible as hotspots? councilors/committee be funded by GIZ, · How effectively were City Council, GCDSW, KII Checklist ILO,WB and WFP enumerators trained to EP&D, city council, UBR administer the Rapid UBR Tool? team How many days were enumerators trained for? · When reflecting on the City Council, GCDSW, KII Checklist training of enumerators, could EP&D, city council, FGD you identify any gaps or areas beneficiaries for improvement? · Were all HHs registered in Min GCDSW, EPD, City FGD checklist each hotspot area? If not, how Councils, Block leaders, KII checklist were HHs to be registered in a beneficiaries and non- particular area selected? beneficiaries, Ward committees · The RUT (Rapid UBR Tool) City Council, GCDSW, KII Checklist required that enumerators EP&D, city council, indicate the vulnerability status of each registered HHs based on their observation. How accurate was this process? · Were the relevant protocols EPD, UBR MU KII Checklist for data collection, sharing and storage in place prior to the process commencing? Were previous protocols leveraged for the CUCI? · How long did it take to collect EPD, Councils KII Checklist the Hhd data? Did you manage 65 to collect all the necessary data? If not, why not? · Who financed the EPD, MGCDSW, City KII Checklist registration of households (eg. Councils govt, development partners)? Was the financing of registration process adequate? Effective? Efficient? What could have been improved? · How well was the EPD, UBR MU, KII Checklist registration process MGCDSW, City councils coordinated across the different actors, including different levels of government? · How was gender EPD, City Councils, KII Checklist integrated in the design and Wards FDG Checklist implementation of the Beneficiaries & GRM registration of households? committees Tip: Check respondents are including issues of timing in the community, sex of enumerators, safeguarding · To what extent did the Central (EPD & Gender KII Checklist registration process involve a and local Councils, DPs leveraging of existing social protection programs (such as the SCTP), or previous lessons learnt from these programs and responses to shocks. Tip: consider responses given above avoid repetition · What challenges were GCDSW, EP&D, UBR mgt KII Checklist faced during registration unit, city council, Block FGD process and how were they leaders, Ward addressed/mitigated? committee beneficiaries & non-beneficiaries. · How was the EPD, UBR MU KII Checklist household data stored? Who was/is responsible? Were there any challenges with the management and usage of the data – e.g. in terms of data security, infrastructure 66 capacity, human resource capacity? · Did politicians, local GCDSW, EPD, City KII Checklist leadership or anyone else have Council, beneficiaries / Hhd Qtr., FGD any influence on the intake and Non beneficiaries registration of the HHs? Explain [How and Whys]. · What went well or did GCDSW, EPD, City KII Checklist not in the registration of Council, UBR mgt unit households? Are there any key innovations or lessons learnt? Assess needs Aim: Among those · What key criteria and City Council, GCDSW, KII Checklist and that are registered, tools were used to define which EP&D, city council, ward conditions how will their needs HHs should be eligible for CUCI? councilors & WFP, UBR be assessed (to Were any data sources used to MU determine their develop the eligibility criteria? potential eligibility) · Do you think there City Council, GCDSW, KII Checklist What was planned: could have been improvements EP&D, city council, ward FDG or variations in the design of councilors, beneficiaries · Determine the eligibility criteria? If so, & block leaders/ward HH eligibility criteria what? committees · Which levels of City Council, GCDSW KII Checklist government, and which MDAs, Central gvt. and DPs were involved in deciding the process for household eligibility criteria? How were they involved? Was there an effective coordination process for involving all the different government actors? Note: It is possible for respondents at this stage to start also talking about HH eligibility criteria. Notes must be taken to ensure HH eligibility criteria question is rephrased in subsequent sections based on response in the question. · To what extent were EP&D, Local Councils, KII Checklist the household eligibility criteria Min of Gender, central based on existing social and local govt, DPs protection programs (such as the SCTP) or systems (such as 67 the UBR), or previous lessons learnt, including responses to shocks. Note: Observe if some answers were not already given when responding similar question above · Do you think the Central and local govt, KII Checklist eligibility criteria are sufficiently DPs GCDSW city council, gender sensitive? Block leaders, Ward committees · What challenges were GCDSW, EP & D, DPs KII Checklist faced during the process to city council, Block FGD determine the eligibility leaders, Ward criteria? committee · What lessons would Central and local govt, KII Checklist you share for future DPs GCDSW programming regarding eligibility criteria for urban responses to a shock in Malawi? ENROLL Eligibility & Aim: To (i) identify · Tip Questions to Head Beneficiaries HH Qtr. enrolment most appropriate HHs of HH: How old are you? What to benefit from the is your main source of income? CUCI cash transfer What is the ownership status of program and (ii) the house residing in? At the enroll them in the time of the registration, did you CUCI have children under the age of 18, a member above 60 years? Will also capture some HH basic What was planned demographic data such as marital status, sex, and HH size, · Transfer etc. registered HHs from the UBR to the CUCI · How were the total Gender, EPD, UBR MU KII Checklist MIS beneficiaries per city per ward and per zone/area determined? · Enroll HHs Was a quota system applied in based on source of the process and who was income, living with involved in making these Elderly people and decisions? Was there any children with 18 years correlation with the vulnerability ranking? 68 · Use the · Who undertook the GCDSW, EPD, NLGFC, KII Checklist Rapid Tool to assess process of identifying eligible City Councils households that are households? eligible · How was data used to Gender, UBR MU, DP, KII Checklist · Run the identify and enroll eligible HHs? City, eligibility criteria in How did you screen the the CUCI MIS to registered HHs to arrive at the determine eligible final beneficiaries list? Was this HHs done at once or in phases? · Carry out Were any geographical and GCDSW, EPD, UBR MU KII Checklist KYC to determine gender trends noticed during whether selected the screening process? beneficiaries have TNM/Airtel SIM cards · How transparent was the Gender KII Checklist or not. HHs without screening process? Which , EPD, UBR MU, DP, City, SIM cards to be given stakeholders were involved? for free · Was the screening/eligibility EP&D and GCDSW, City KII checklist process well-coordinated across council the different actors, including different levels of government? · How were the vulnerability Gender, EPD, UBR MU KII Checklist criteria applied among eligible HHs? What was the outcome geographically, was there gender trend that was noticed in the vulnerability ranking? · Why was your household not Non beneficiaries HH Qtr. selected for CUCI program? In FGD checklist your opinion was the enrolment process transparent – i.e. the selection of eligible households? Explain What were the key EPD/Gender, UBR MU KII Checklist findings/issues that arose after data registered from Hot spots was transferred from UBR to CUCI MIS for the first time? Tip: Proportion of HHs with available National IDs (NID), level duplicates, proportion of HHs with or without phone numbers 69 How long did it take for the Gender, EPD, PSPs KII Checklist eligibility data to be sent to the mobile money service providers or payment service providers (PSPs) for validation? What key issues emerged during phone validation by transfer agency? How long did the PSPs take to provide report of validation to Ministry? Tip: Proportion of HHs whose phone numbers were valid or not, proportion of HHs whose phone numbers KYC verified or not and proportion of HHs with E-wallet verified or not, without E-wallet · How was the transfer agency phone validation report used and by who, after how long? · When did the CUCI MIS Gender, EPD, UBR MU, KII Checklist custodians apply the eligibility criteria, was it done after transfer agency report or before? What were the outcomes of the eligibility application? What gender issues emerged at this stage? Were there any differences with what emerged during transfer agency validation as well as first CUCI data cleaning process Tip: Proportion of HHs eligible and ineligible · Did the transfer agency, city, Gender, EPD, UBR MU, KII Checklist CUCI management team visit transfer agency, city, FGD the Hotspots to issue SIM cards beneficiaries, block to selected beneficiaries with leaders, ward phone numbers, create E- committee members wallets for those who didn’t have or KYC for HHs that had not done so by them? How many SIM cards were issued? Were they bought or free? 70 Which months were the visits made and what other activities were done during these visits? How was gender and diversity integrated during these visits? · What were the challenges Gender, EPD, UBR MU, KII Checklist encountered in the process of transfer agency, city, FGD determining the enrollment of beneficiaries, block the beneficiaries? leaders, ward committee · How long did the enrollment EP&D and GCDSW, City KII checklist process take to be completed? council, Ward What were the key challenges committees and how were these resolved? Explain · Was the enrolment process EP&D and GCDSW, City KII checklist well-coordinated across the council different actors, including different levels of government? · How often were all HHs that EP&D and GCDSW, City KII checklist were registered updated about council, beneficiaries, FGD the progress after data Ward committees collection up to this stage? (Enumerators had urged registered HHs to be checking their phone CUCI updates) · How much funding was EP&D and GCDSW plus KII Checklist allocated to the facilitation of TNM & Airtel, NLGFC the enrollment process? Was this amount adequate? Did the private sector i.e., TNM/Airtel offer any corporate social responsibility [CSR] to leverage their role in CUCI program? Explain Determine Aim: Smoothen HH · What was the proposed Beneficiaries Hhd Qtr. benefits consumption and monthly transfer level for package cushion unemployed beneficiaries? category. · What was the process for EP&D and GCDSW KII checklist What was planned: determining the transfer level? FGD 71 · Make TIP: Reflect on previous monthly cash transfer responses if they did not include of MK35,000 to this validated beneficiaries (value · What data was used to EP&D and GCDSW KII Checklist based minimum determine the transfer level? wage) TIP: Reflect on previous responses if they did not include this · Which levels of government, EP&D and GCDSW plus KII Checklist and which MDAs, were involved TNM & Airtel in determining the transfer level? Note: Observe if some answers were not already given when responding similar question above · Did politicians, Govt, donors EPD and GCDSW KII Checklist or anyone else influence the transfer level? Explain Note: Observe if some answers were not already given when responding similar question above · What were the factors in EP&D and GCDSW KII Checklist NOT varying the transfer value according to size of the household or other considerations? Note: Observe if some answers were not already given when responding similar question above · How much did you receive Beneficiaries Hhd Qtre per month? Was the amount adequate to meet the needs of your household? How much EPD & GCDSW, COMSIP, KII Checklist would you consider as Case workers adequate for your household? What have you used or intend to us the money for? List/Mention 72 Notification Aim: To notify Hhds · How were you notified about Beneficiaries Hhd Qtr. and and general public your enrolment in the CUCI onboarding about the details of program, including the monthly the monthly package transfer value? Did it include use of SMS, press release, word Planned messaging: of mouth? · Transfer · How were beneficiaries EP&D, beneficiaries, KII checklist Hhd value of MK35,000 for notified about their enrolment block leaders, ward Qtr. 3 months in the program? How many committees · Date of days, and how many staff, were payments for allocated to notify and onboard transfers. beneficiaries into the program? · Funds will be · How much did it cost to notify EP&D, Councils & Min. KII checklist Hhd transferred through and onboard beneficiaries for of Gender Qtr. Mobile money the program? · Use of 153 as · Was there financing available EPD & Min GCDSW KII Checklist Toll free line in case for the notification and of questions onboarding? Was it adequate? · Messages to · Was the notification and Beneficiaries Hhd Qtr., FGD be done through SMS onboarding process well- Checklist (TNM/Airtel) and coordinated across the Press releases different actors, including · Caution the different levels of government? beneficiaries about · When were the beneficiaries Beneficiaries Hhd Qtr., FGD possible tricksters notified before commencement Checklist who may hijack the of the payment? How many SMS facility times were the notification messages passed on? · Did the notification also Beneficiaries Hhd Qtr. include cautioning of the beneficiaries about the tricksters? How were the FGD Checklist beneficiaries determining the authenticity of the notification messages? · Did the notification include any message about COVID-19 preventive measures? · In your view how effective was the notification process to different groups of people i.e. PWD etc.? 73 · Were any key structural stakeholders that you think were missed or excluded, or misplaced in CUCI? List PROVIDE Benefits Aim: Pay the agreed · Who was responsible for Min GCDSW (+EP&D) KII Checklist provision transfer value to the preparation of the enrolment right people at the worksheet and preparation of right time payroll for beneficiaries? Planned payment · What was the process for EPD & Min GCDSW, KII Checklist procedures by EPD determining the payment Development Partners mechanism for the CUCI? What (especially UNICEF) · Prepare actors, and what levels of payroll of selected government, and which MDAs, beneficiaries were involved in this? · Service Tip: Be self-aware if this was charge and cash out not already answered charges to be included in the What data or other sources of GCDSW, EP & D, DPs KII Checklist transfer value information were used to determine the payment · Payroll to be mechanism? sent to TNM and Airtel who will send · To what extent was the Central and local govt, KII Checklist back to EPD for payment mechanism based on DPs review and existing social protection authorization programs (such as the SCTP), or · EPD to send previous lessons learnt, the authorized payroll including responses to shocks. to MNOs for transfer Note: Observe if some answers to beneficiaries were not already given when · Beneficiaries responding similar question to access funds from above MNOs agents located · Do you think the payment Central and local govt, KII Checklist in the 4 cities mechanism is sufficiently DPs GCDSW city council, · Press release sensitive to the needs of Block leaders, ward to be made about vulnerable groups, such as the committee each cash transfers elderly, or women? done · What challenges were faced GCDSW, EP & D, DPs & KII Checklist during the actual establishment Councils (procurement) of the payment mechanism? 74 · How has the payment EPD & Min GCDSW, DPs KII Checklist mechanism been financed? Tip: Be self-aware if this was not already answered · Was the payment mechanism EPD & Min GCDSW KII Checklist – both its design and execution – well-coordinated across the different actors, including different levels of government? · Were the beneficiaries GCDSW, beneficiaries, KII Checklist, HH guided on what to use the Qtr., FDG money for? Explain your answer · What gender issues EPD, city KII Checklist emerged during the provision of payments? (GBV, which phone numbers were mostly registered by main receivers, proportion of women in the first blast, pattern of withdrawing money between women and men) · What specific roles did the EPD, TNM & Airtel KII Checklist mobile service providers and EPD play in the design and implementation of the payment mechanism? · What challenges were there MoGCDSW, EPD, TNM & KII Checklist for beneficiaries to cash out the Airtel, FGD, MNO FGD MNOs agents? i.e. in terms of Agents, beneficiaries geographical location of the Airtel and TNM agents? Did the agents experience any liquidity challenges? First day, second day or a week? If there were any liquidity challenges, how were these resolved? · What evidence (data) is there EPD, TNM & Airtel, Document that the right amount of FGDs, MNO Agents review, KII transfer value was received, Checklists and right beneficiaries were reached? Are there systems to capture use of transferred 75 funds? What would you say are the key lessons i.e. system based strengths and weaknesses gathered so far? In your view, are there any other better ways that service provision would have implemented/handled? · How much did you receive Beneficiaries Hhd Qtr. per month? Was the amount EPD & GCDSW, COMSIP, adequate to meet the needs of Case workers your household? How much KII would consider as adequate for your household? What have you used or intend to us the money for? List/Mention · Were the payments delivered Beneficiaries HH Qtr. and in a timely manner – i.e. were FGDs they delivered on schedule? · If the payments were EPD, GDSW KII Checklist delayed, why did this occur and how was it managed? · Do you think the payments Beneficiaries, EPD & HH Qtr. and were delivered at a time when Min GCDSW, City FGDs livelihoods were most at risk? Councils, DPs, Ward Would it have been preferable committees to provide the payments in a different month? · It was planned that EP&D and SW, COMSIP, KII checklist Hhd beneficiaries will be provided Beneficiaries, Qtr. with financial literacy. Was this service provided? If so, how. And is there any evidence as to how effective the service has been? Was there any system to capture feedback from beneficiaries on the quality of this service? · What is your perception on EP&D, GCDSW, City KII Checklist whether using Councils, Ward Hhd Qtr. MNOs/TNM/Airtel was an Committee, NLGFC effective and efficiency means of delivering payments compared to other potential 76 modalities? What other Beneficiaries, City models of cash transfer would Councils, Ward you have preferred? i.e. banks, committee e-wallet)? MANAGE Beneficiaries Aim: Conduct · What is the GRM process for EPD, GCDSW, UNICEF, KII & FDG compliance, beneficiary the CUCI? GIZ, TNM & Airtel, Checklists updating, & monitoring, beneficiaries grievances management of admin and grievance · Which levels of government, EPD & GCDSW, KII Checklist, issues related to cash and which MDAs, were involved transfer. in designing and deciding the GRM process? · Was the GRM process – both EPD & GCDSW, Call KII Checklist, What was planned: its design and execution – well- Centers, Councils, FGDs · Call Centre coordinated across the records, beneficiaries, Desk review to be in place by June, different actors, including ward committees, 2020. different levels of government? · GRM to be · Did you set up a call center? GCDSW, TNM & Airtel, KII & FDG operational in 4 city When and why? What beneficiaries Checklists councils messages are commonly submitted to the call center? · EPD and Any evidence to share on the Gender officers’ call center messages? conduct spot checks in MNOs agents and · Are there any GRM GCDSW KII Checklist sites committees? At what level? · EPD to have When were they formed and a taskforce for how useful have they been in managing grievances the whole delivery chain? How and complaints does the GRM committee address the grievances? Which stakeholders are involved in the management of the grievances? · Are there any spot checks EPD, GCDSW, City KII Checklist that were conducted to Councils, CUCI Donors, monitor the performance of the Ward committees MNOs’ agents? Who conducted the spot checks? · How were different EPD & GCDSW KII Checklist stakeholders of CUCI program involved in the spot checks? Any reports to share? 77 · Have there been cases of GCDSW, EPD, KII Checklist Ghost beneficiaries or anyone beneficiaries / Non Hhd Qtr. registering a relative or friends? beneficiaries If any, how did you deal with such cases? · Are you aware of the Beneficiaries HH Qtr. grievance redress mechanism for the CUCI? Can you describe it? Have you used it? · What data is used for the EPD & GCDSW, KII Checklist, HH GRM processes? What are the beneficiaries, ward Q’re, FGDs processes involved in sharing committees, councils and storing this data? What and councilors, wards other uses is the CUCI data for? · How is the GRM process EPD & GCDSW, Transfer KII Checklist, financed? agencies · What were the common EPD & GCDSW, KII Checklist, HH complaints, who raised most beneficiaries, ward Q’re, FGDs complaints between men and committees, councils women, how were gender and councilors, wards considerations incorporated in the GRM? · What worked well and didn’t EPD & GCDSW, KII Checklist, HH with regard to grievance beneficiaries, ward Q’re, FGDs redress mechanism? Do you committees, councils still have outstanding issues and councilors, wards Exit Aim: Manage · Are beneficiaries [you] aware Beneficiary, EPD, Min Hhd Qtr., FGDs decisions, expectations and exit that the cash transfer program GCDSW notifications, of the CUCI program will or was to end on 31st and closing beyond 31st March March? Of what value has the cases 2021. CUCI program been to you? Is the program still necessary? What was planned: · CUCI · How are beneficiaries EPD & GCDSW, KII Checklist, HH notified that they will be exiting beneficiaries, Q’re, FGDs program will close by the program? 31st March 2021 · No plan for · What happens to the EPD & GCDSW, KII Checklist, long term support beneficiary data when a other than SCTP beneficiary exits CUCI? · Was the amount of financing EPD & GCDSW, KII Checklist, a factor in determining the 78 duration of CUCI? What were other factors? · Has the exit process been EPD & GCDSW, KII Checklist, well-coordinated across the different actors, including different levels of government? · When CUCI concludes, and a EPD & GCDSW, DPs KII Checklist beneficiary exits the program, are there any arrangements in place to link the beneficiaries with other programs that could complement the outcomes of CUCI? If not, what recommendations would you have in this regard? Are there any possible sustainable program design approaches you would recommend or suggest for future similar shock response programs? 79 Appendix 2: CUCI Process Evaluation Reference Points Following document review, scoping with City Councils, consultations with government officials, coupled with a validation meeting that CUCI TWG members attended on 18 March 2021, Reference Points were established for the program, as shown in this table. Pillar / Stage CUCI design in May/June CUCI design just prior Reason for Reference point for 2020 to implementation changes designing the questionnaires (~Nov 2020) between (November 2020) May/June and ~Nov. 2020 Pillar 1: Planned activities Planned Activities Using Posters, Actual Implementation Plan ASSESS Short videos, • Using Posters, Short • Stakeholders’ Radio jingles • Stakeholders’ Stage 1: videos, Radio jingles sensitization were sensitization Outreach considered • Social media – digital • Identification of • Identification of partners costly and time platform and bulk SMS partners at all at all levels consuming levels • Community platforms – • Community yellow vans, mobile audio • Community mobilizations systems mobilizations • Resource mobilization • Influential people – faith- • Resource based groups, local mobilization • Full district council leaders meetings • Full district council • Leaflets distribution meetings • Use of influential people – faith-based groups, • Use of influential local leaders for people – faith- outreach. based groups, local leaders for • Radio announcements outreach. • Community platforms – • Radio yellow vans, mobile announcements audio systems • Town Criers (local boys announcing CUCI in the streets) Pillar 1: Planned activities Planned activities Call Centre was Actual Implementation Plan Assess planned to be • All HHs within hotspots • Register all HHs in in place by • Not all HHs within Stage 2: hotspot areas June 2020 but hotspots were registered Intake & • Target 185,246 shifted to Jan registration • The identification of • Registered 270,000 • Use of call Centre if Covid hotspots will be 2021 How was case identified within done by the City • 520 enumerators were information hotspot or use councilors Council Technical used collected on Team Not all • Use of enumerators • Case management used potential vulnerable HHs GRM beneficiaries? • Case management in some 80 • Pockets of vulnerable HHs • Door to door data hotspots were • Engage ward committees in other areas included collection via UBR included due and police collection tool to time • Rapid tool include under constraints • Use tablets to collect data 18 • Use of tablets, ODK forms to collect the • Collect HHs data within 21 • Engage ward committees data days and police Police was not • Recruit 520 used at stage • Target 199,460 • Use tablets to collect data enumerators for • Exercise funded by GIZ, household ILO, WB, UNICEF and WFP registration Original target of 185,240 • Rapid Tool included under • Collect HHs data changed 5 and age over 64 within 21 days • Door to door data • Target 199,460 collection via UBR Vulnerability collection tool • Exercise to be criteria funded by GIZ, ILO, changed from • Use of tablets, ODK forms WB and WFP under 18 to to collect the data • Rapid Tool include under 5 under 5. Pillar 1: Planned Activities Planned Activities As an Actual Implementation Plan Assess emergency • Map out social and • Map out social and program, TWG • Map out social and Stage 3: economic aspects of 4 economic aspects considered the economic aspects of 4 Needs cities with assistance of of 4 cities with needs cities with assistance of assessment City Council assistance of City assessment as City Council and WFP Council and WFP time Use of the • Assess structural and • Clearly define urban registration financial vulnerability of • Clearly define urban consuming, vulnerability data to households: livelihood vulnerability hence identify characteristics/sources, dropped. • Develop Rapid tools for eligible HHs job and food insecurities. • Develop Rapid tools assessment for assessment • Assessment of Needs to • Rank hotspots based on be conducted by National • Rank hotspots poverty profile and Technical Team jointly based on poverty poverty mapping with City Council levels • Assess structural and Technical Team. financial vulnerability of households: livelihood characteristics/sources, job and food insecurities. Pillar 2: Planned Activities Planned Activities Actual Implementation Plan ENROLL: • CUCI MIS to facilitate • Transfer registered • Transfer registered HHs Stage 4: determination of eligibility HHs from the UBR from the UBR to the CUCI Eligibility & to the CUCI MIS MIS enrolment • Enroll HHs based on source of income, living • Enroll HHs based on • Run the eligibility criteria Identifying with Elderly people and source of income, in the CUCI MIS to eligible HHs children under 18 years living with Elderly determine eligible HHs 81 based on • Run the eligibility criteria people and children • MoGCDSW selects criteria in the CUCI MIS to under 5 years registrants to become determine eligible HHs beneficiaries • Use the Rapid Tool • CUCI Guide proposes to assess • Enroll HHs based on prioritization of HHs into 3 households that are source of income, living categories; CUCI manual eligible with Elderly people and refers to 2 categories children under 5 years • Run the eligibility • Eligibility status assigned criteria in the CUCI • Use the Rapid Tool to by MoGCDSW MIS to determine assess households that eligible HHs are eligible • MoGCDSW selects HHs to become beneficiaries • Carry out KYC to • Run the eligibility criteria determine whether in the CUCI MIS to selected determine eligible HHs beneficiaries have TNM/Airtel SIM • Carry out KYC to cards or not. HHs determine whether without SIM cards selected beneficiaries to be given for free have TNM/Airtel SIM cards or not. HHs without SIM cards to be given for free SIM cards Pillar 2: Enroll Planned Activities Planned Activities Timeline Actual Implementation Plan changed from Stage 5: • Duration of the • HHs vulnerable to June – • HHs vulnerable to Determine intervention to be three economic shock of November, economic shock of the transfer level to six months (June – Nov the pandemic to be 2020 to Jan – pandemic to be enrolled 2020). enrolled March 2021 • Structural vulnerability of due to June • Exact length to be • CUCI-MIS will under 5 and above 64 age determined on needs and facilitate the 2020 Fresh resources prioritization of General • CUCI-MIS will facilitate the households Elections - to prioritization of • Monthly cash transfer to avoid political households be MK35,000 (based • Duration of interference minimum wage) intervention to be 3 donors advised • Duration of intervention months (Jan-March government to to be 3 months (Jan- • Transfer value to vary 2021) suspend the March 2021) based on HHs size, needs, program market prices • Monthly cash • Monthly cash transfer to transfer to be be MK35,000 (based on MK35,000 (based minimum wage) on minimum wage) Pillar 2: Enroll Planned messaging Planned messaging: Transition plan Actual Messaging for Stage 6: • transfer value and • Transfer value of beneficiaries • Transfer value to be Notification guaranteed months of MK35,000 for 3 not feasible MK35,000 and assistance at that transfer months due to low onboarding value • Cash transfer to be done resource • Date of payments for 3 months envelope Notifying • adaptive nature of the for transfers. beneficiaries intervention • Date of payments for that they will transfers. 82 benefit from • Tailoring needs to HHs • Funds will be • Funds will be transferred the program size transferred through through mobile money mobile money • how to transition • Use of 351 as Toll free line households onto new • Use of 153 as Toll in case of questions social protection free line in case of programs in urban areas questions • Messages to be done (e.g. social insurance, an through SMS (TNM/Airtel) urban Social Cash Transfer • Messages to be and press releases Program etc.). done through SMS (TNM/Airtel) and • Caution the beneficiaries press releases about possibility of SMS facility being • Caution the compromised beneficiaries about possibility of SMS facility being compromised Pillar 3: Planned Activities Planned Activities Decision that Actual Implementation Plan Provide Give Direct • CUCI management runs • Prepare payroll of would not • CUCI management runs Stage 7: pre-calculations i.e. no. of selected handle pre-calculations i.e. no. of Benefits HHs to receive transfers beneficiaries transfers of HHs to receive assistance provision funds as an • Prepare payroll of • Service charge and • Prepare payroll for the Payment of selected beneficiaries agent as its beneficiaries cash out charges to transfers be included in the services are • Service charge and cash transfer value only with Airtel • Service charge and cash out charges to be included out charges to be included in the transfer value • Payroll to be sent to in the transfer value TNM and Airtel who • Payroll to be sent to TNM will send back to • Payroll to be sent to TNM and Airtel who will send EPD for review and and Airtel who will send back to EPD for review authorization back to EPD for review and authorization and authorization • EPD to send the • EPD to send the authorized payroll • EPD to send the authorized payroll to to MNOs for authorized payroll to MNOs for transfer to transfer to MNOs for transfer to beneficiaries beneficiaries beneficiaries • Beneficiaries to access • Beneficiaries to • Beneficiaries to access funds from MNOs agents access funds from funds from MNOs agents located in the 4 cities MNOs agents located in the 4 cities located in the 4 • Press release to be made • Press release to be made cities about each cash transfers about each cash transfers done • Press release to be done made about each • Give Direct to be cash transfers done contracted as Transfer Agent 83 Pillar 4: Planned Activities Planned Activities Actual Implementation Plan Manage • Call Centre to be in place • GRM to be • GRM to be operational in Stage 7: by June, 2020. operational in 4 city 4 city councils Beneficiaries councils compliance, • GRM to be operational in • EP&D and Gender officers updating, & 4 city councils • EPD and Gender to conduct spot checks in grievances officers’ conduct MNOs agents and the • Cash Working Group will spot checks in hotspots in the 4 cities monitor household’s MNOs agents and needs sites • Call Centre hosted by Conduct Gender to employ officers beneficiary • Cash Working Group will • EPD to have a at the Call Centre to monitoring, issue Monthly Advice to taskforce for record grievances management increase/decrease managing of admin and transfer value grievances and • Technical Working Group grievance complaints to manage grievances and issues related complaints to cash • Reconciliation report to be • Reconciliation report to transfer. produced by Airtel be produced by Airtel and and TNM TNM on monthly basis to EP&D • Cash Working Group will monitor household’s needs Pillar 4: Planned Activities Planned Activities No plan to Actual Implementation Plan Manage transition • HHs renounce • CUCI program will beneficiaries • HHs renounce Stage 9: Exit close by 31st March into SCTP decisions, • HH members died 2021 • HH members died notifications, • HHs benefiting from other • No plan for long • HHs benefiting from other & closing interventions term support The plan was interventions cases/exit to close CUCI strategies • CUCI exit rationale and • CUCI program will close by by 31st March procedures will be 31 March 2021 Manage 2021 but it is designed in line with expectations reported that prevailing conditions. and exit the program process of will extend to beneficiaries April 2021. Appendix 3: Summary of CUCI Process Evaluation KII and FGDs Focus Group Discussion (FGD) City Target Conducted Description of FDGs per City Mzuzu 4 3 • 1 FGD at Mchengautuba Ward • 1 FGD at Katawa Ward 84 • 1 FGD at Msongwe Ward Lilongwe 5 7 • 2 FGDs at Chinsapo 1 Ward • 3 FGDs at Maria Ward • 1 FGD at Mvunguti Ward • 1 FGD at Ngwenya Blantyre 5 8 • 1 FGD at Blantryre City Centre ward • 1 FGD at Chilomoni Ward • 1 FGD at Misesa Ward • 1 FGD at Chigumula Ward • 1 FGD at Ndirande Namalimwe • 1 FGD at Ndirande Makata • 1 FGD at Mbayani Ward • 1 FGD at South Lunzu Ward Zomba 4 5 • 1 FGD at Masongola Ward • 1 FGD at Chinamwali Ward • 1 FGD at Chirunga Ward • 1 FGD at Mbeza Ward • 1 FGD at Ntiya Ward Total 18 23 Key Informant Interviews (KII) City Target Conducted Description of KII per City Mzuzu 3 6 • KII with Mchengautuba West Councilor • KII with Chimdima Block Leader, Zolozolo East Ward • KII with Katawa Ward Councilor • KII with City Council Official, the Public Relations Officer • KII with Zolozolo Councilor • KII with Msongwe Councilor Lilongwe 4 5 • Chinsapo 1 councilor, Temwanani Muhango • Maria Ward Councilor, Patrick Makumba 85 • 1 KII with 2 Airtel money operators, Pilirani Halmiton and Madalo Thewe at Mvunguti Ward • GRM Chair for Mvunguti ward, Mr. Mazumba • 1 KII with Magirinjala Block Leader, Ngwenya Ward Blantyre 4 19 • 3 in Blantyre City Centre (Councilor, Block Leader and Airtel Agent) • 2 in Chilomoni (Councilor and Block Leader) • 3 in Misesa (Councillor and Block Leader and TNM Agent) • 2 in Chigumula (Councilor and Block Leader) • 2 in Mbayani (Councilor and Block Leader) • 3 in Makata (Councilor and Block Leader and Airtel Agent) • 3 in Namalimwe (Councilor, Block Leader and TNM Agent) • 3 in South Lunzu (Councilor and Block Leader and Airtel) Zomba 3 10 • 1 in Masongola (Councillor) • 2 in Chinamwali (Councillor and Block Leader) • 1 in Chirunga (Mayor) • 1 in Mbeza (Councilor) • 1 in Ntiya (Councillor) • 4 in Zomba City Council (CUCI Coordinator, Enumerator, Director of M&E and Social Welfare Officer) Central level 8 15 • EPD-PRSP agencies • Min of Gender • National Local Government Finance Committee • TNM • Airtel • World Bank • Irish Aid • UNICEF • WFP • GIZ • ILO 86 • UBR MU • COMSIP • KFW • Give Directly Total 55 87 Appendix 4: Calculation of Sample Size The sample size for the households/individuals that were interviewed in this study was determined by the following statistical equations (Health and Social Care Information Centre, 2009): Where: Proportion of CUCI beneficiaries located in the hotspots; is the number of respondent who complete the household questionnaires (which was unknown); is the size of the population being surveyed (sampling universe); is the number of beneficiaries sampled out (the sample size); is the response rate i.e. how many people out of those sampled agreed to be interviewed and complete the interview (in our case assumed it was set at 75%) The study calculated the sample size based on the target population planned to benefit from the project. CUCI enrolled a total of 199,613 beneficiaries. The sample size was therefore calculated as follows: 1.96²*0.5*199,613*(1-0.5) 0.04²*199,613+1.96²*0.5(1-0.5) n= 191,708.33/320.3118 = 599 c = 599/0.75 = 799 Thus 799 is the minimum sample size for the 199,613 beneficiaries adequate at 95 percent confidence level and 3 percent precision. However, to account for the possibility of spoilt questionnaires, an additional 20 percent was factored in, making the final sample size approximately 959 households. 88