SouthSudan-PWP-May 23-2019 Final With additional learnings from the Central African Republic, the Democratic Republic of Congo, Uganda, and Yemen May 2019 Authors i List of Abbreviations AAH-I Action Africa Help International AFD Agence Française de Développement (French Development Agency) CAR Central African Republic DRC Democratic Republic of Congo DRF Disaster Risk Financing ECRP Emergency Crisis Response Project ERP Eastern Recovery Project FAO Food and Agriculture Organization of the United Nations FCV Fragility, Conflict, and Violence FSRDC Fonds Social de la République démocratique du Congo (Social Fund of the Democratic Republic of Congo) GBV Gender-based Violence GDP Gross Domestic Product GIS Geographic Information System GPS Global Positioning System GRM Grievance Redress Mechanism ICR Implementation Completion and Results Report IDA International Development Association IDP Internally Displaced Person IHISP Improved Household Investment Support Program IOM International Organization for Migration IP Implementing Partner ISR Implementation Status and Results Report JCC Juba City Council LIPW Labor Intensive Public Works M&E Monitoring and Evaluation MAFS Ministry of Agriculture and Food Security MEPCI Minister of Economy, Planning, and International Cooperation MFI Microfinance Institution MINUSCA United Nations Multidimensional Integrated Stabilization Mission in the Central African Republic MIS Management Information System MSEs Medium and Small Enterprises NGO Nongovernmental Organization NUSAF Northern Uganda Social Action Fund OPM Office of the Prime Minister PAD Project Appraisal Document PID Project Information Document PIU Project Implementation Unit PWP Public Works Project QC Quarter Council R-ARCISS Revitalized Agreement on the Resolution of the Conflict in South Sudan SFD Social Fund for Development SLP Sustainable Livelihood Pilot SNSDP Safety Net and Skills Development Project SWF Social Welfare Fund TAAC Transparency, Accountability, and Anti-Corruption TST Technical Support Team ii UN United Nations UNDP United Nations Development Program UNHCR United Nations High Committee for Refugees UNICEF United Nations Children’s Fund UNMISS United Nations Mission in South Sudan UNOPS United Nations Office for Project Services WASH Water, Sanitation, and Hygiene WFP World Food Program iii Table of contents List of Abbreviations .................................................................................................................................... ii Executive Summary ...................................................................................................................................... 1 Introduction .................................................................................................................................................. 5 Methodology ................................................................................................................................................ 7 Selection of countries / case studies ........................................................................................................... 8 1. Project Description and Results ......................................................................................................... 10 Project design ..................................................................................................................................... 10 Project Implementation ..................................................................................................................... 10 Environmental and Social Safeguards ............................................................................................... 14 Fiduciary safeguard ............................................................................................................................ 16 Monitoring and Evaluation ................................................................................................................ 17 Capacity Building ................................................................................................................................ 18 Results................................................................................................................................................. 18 Challenges and Lessons Learned ........................................................................................................ 21 2. Evidence from Other Countries ......................................................................................................... 25 2.1. Strategic Objectives / models for PWPs ........................................................................................ 25 Implications for public works objectives in South Sudan ..................................................................... 27 2.2. Design and Implementation Features ........................................................................................... 28 Targeting ................................................................................................................................................. 28 Targeting ................................................................................................ Error! Bookmark not defined. Implications for targeting in South Sudan ......................................................................................... 32 Wages ..................................................................................................................................................... 33 Implications for wage rates in South Sudan ...................................................................................... 34 Public works activities / subproject selection ................................................................................... 35 Gender Issues ..................................................................................................................................... 36 Implications for gender-sensitive public works in South Sudan ....................................................... 39 Implications for extending public works in South Sudan ................................................................. 41 Conflict-sensitivity .............................................................................................................................. 42 Implications for conflict-sensitive public works in South Sudan ...................................................... 45 Summary and Conclusions ................................................................................................................. 46 Bibliography................................................................................................................................................ 48 Annexes ...................................................................................................................................................... 50 List of Boxes, Figures, and Tables Box 1 : South Sudan: Recovery and Need Deficits ........................................................................................ 6 Box 2: The Grievance Redress Mechanism ................................................................................................. 14 Box 3 Models of public work programs applicable to South Sudan ........................................................... 25 Box 4 Using a Force Account to delivery public works in the Central African Republic…………………………….34 Figure 1: Before and after public works...................................................................................................... 19 Figure 1 Cartoon explaining the lottery system for LONDO in CAR ............................................................ 30 Table A.1: Countries, Public Works Case Studies, and Sources of Information.......................................... 50 Table A.2: Project Development Objective-level Results Indicators by Project and Country .................... 51 Executive Summary 1. The long history of conflict and on-going insecurity in South Sudan has resulted in the deaths of thousands of people, and collapse of the local economy, socioeconomic infrastructures, and access to basic services like education and health, with incidence of poverty worsening from 47 percent in 2011 to 66 percent in 2015, and falling to the current rate of 86.5 percent in 2017. 1 The Government of Republic of South Sudan has identified a national social protection system as a priority to help the poorest and the most vulnerable persons to build their resilience to shocks and stresses. In response, the World Bank supported the Government through the Safety Net and Skills Development Project (SNSDP) to implement Public Works (PW) activities to address immediate needs of the poorest and the most vulnerable households, while also putting in place the building blocks of a national social protection system for delivery of predictable and reliable safety net in the longer term. This report intends to build on the experiences in South Sudan and other countries implementing public works programs, to inform specific and innovative design choices for public works interventions. 2. Despite the volatile environment, the project has achieved important results in South Sudan:  53,163 poor and vulnerable households (355,871 individuals)2 were provided with about US$ 11 million in wages for participation in 5173 public works activities.4  Biometric and Global Positioning System (GPS) based enrolment and payment verification tools were developed and payment were made by commercial banks, facilitating transparent and accountable payments.  Local level oversight and coordination structures, such as development committees and the Grievance Redress Mechanism (GRM), were established and supported participatory management of the project and improved relationships between community members and local leadership. 1 This estimate does not include the conflict-affected Greater Upper Nile region due to on-going insecurity which limited the access to conduct household survey. Therefore, it is expected that national poverty levels are likely even worse than the estimated 86.5 percent. See World Bank. 2019. “South Sudan Economic Brief,� Washington, DC. 2 This exceeded the original target of 34,000 households by about 57 percent. About 74 percent of the beneficiaries working on public works activities were female, against the target of 50 percent female beneficiaries. In some locations, female participation was over 80 percent. In addition, about 30 percent were internally displaced persons (IDPs) 3 The original target was 150. 4 Project Completion Report. February 2019 1  The PW platform allowed community members from different ethnicities who did not usually interact to come and work together, contributing to strengthened community cohesion and social unity. Nine out of 10 respondents in Juba felt that PW united people of different ethnicities.5  Implementation by the United Nations Office for Project Services (UNOPS) on behalf of the Government resulted in deepening its understanding of safety net interventions and strengthening humanitarian-development linkages, as it enabled UNOPS to address most immediate needs and vulnerabilities while building community capacity necessary for longer term resilience. 3. However, the project faced the following challenges:  The existing program failed to cover most of the poor people given the limited resources and insecurity in the country, resulting in higher risk of conflict between targeted and non- targeted households and communities. Government was unable to reach the neediest people in opposition held areas.  Due to hyperinflation, insecurity, and lack of infrastructure (reliable road networks, functional markets etc.), the real purchasing power of the transfers reduced greatly, endangering impact. Moreover, the project had to deal with high implementation costs and delays.  Weaknesses in the Management Information System (MIS) and the biometric payment system led to errors in payrolls and delayed payments in some locations, risking escalated tension and violence between project staff and community members in a few project locations.  In certain locations, weak capacity of community and local government structures meant that community mobilization and sensitization were not conducted adequately, leading to communication gap among some beneficiaries and increasing the risk of misunderstanding about the project. This also contributed to risk of violence against project staff. 4. Given these challenges, the following lessons could be drawn from the projects implemented in South Sudan, Central African Republic (CAR), Democratic Republic of Congo (DRC), Uganda, and Yemen for PW design:  In fragile and conflict affected environments like South Sudan, public work design needs to be pragmatic and focused on consumption smoothing and delivering public works in ways that 5 Hub Consulting. 2018. “Juba Public Works Project. End of Project Evaluation,� Juba, South Sudan. 2 allow greater participation and engagement of local communities in decision-making and give a voice to local grievances may provide social cohesion dividends over time.  For targeting, if poverty rates are near universal (like in South Sudan), which could lead to oversubscription for the program, it is important to use a targeting method that gives a fair chance to all. Experience form case studies show that in an environment of high poverty rate, setting public work wages below market rate for self-targeting is not an effective targeting method, as too many people want to work even when the wage rate is far below the market rate. In this regard, in addition to community targeting, using lotteries for targeting beneficiaries or providing support on rotation might be considered. If there are barriers for women to enter, dedicated lottery draws should be adopted for women.  Options should be built into a public work design to allow for transfers to poorest and most vulnerable households without able-bodied adult members to participate on the public works.  Community mobilization and sensitization is essential to (a) ensure community ownership and buy-in to support project implementation and sustainability; (b) minimize risk of conflict between communities and project staff due to communication gap and misunderstanding of project objectives, targeting approach and eligibility, implementation processes and procedures etc.; and (c) ensure that male members of households view income from public works as a household, rather than individual, entitlement to maximize spending aimed at improved household welfare.  Regarding women’s participation in public works, it is important to ensure that public work program design and implementation pay more attention to (a) support elderly, pregnant women and nursing mothers to identify lighter work options or different activities that allow women to combine public works with their reproductive/domestic roles, (b) provide child care support at work site for mothers with small children, and (c) protect women from gender based violence (GBV) at public work sites and at home.  While cash transfers through public works can improve women’s welfare and contribute to enhanced gender empowerment, it can also heighten risks of GBV within households due to shifting power dynamics between men and women. These can be mitigated through efforts aimed at changing mindsets of communities and local level leadership (i.e. through increased awareness raising exercises and in-depth community engagement), ensuring GBV focal points and GBV sensitive reporting channels within GRMs, availing information on existing referral 3 services and systems in target locations; and establishing gender sensitive work norms and PW designs (i.e. work sites within communities, security at payment sites etc.) among others.  In situations characterized by fragility, conflict, and violence (FCV), project implementation should rely on neutral institutions accountable to the communities, the government, and donors, to the extent possible. To this end, close collaboration and coordination with a range of government and non-government (UN and NGO) partners is important. Related, it would be important to invest in the relationships between these institutions and communities and build their capacities to deliver social protection programs, should conflict worsen or renew. Hence, the objective of the SNSDP in South Sudan to take advantage of pockets of stability to provide reliable and predictable safety net is relevant.  Use of information technology such as biometric and GPS based enrolment and payment verification tools in collaboration with the private sector (for example commercial banks) can ensure transparent and accountable management of program resources, mitigating risks of error, fraud, and corruption.  Effective coordination and collaboration with other organizations involved in social protection is key to avoid duplication and contradiction and implement income-generating activities to deliver more integrated package of services.  Despite all these innovative design features, achieving a fully effective social protection in South Sudan will not be possible without peace. 4 Introduction 1. After decades of conflict, South Sudan gained its independence from Sudan on July 9, 2011. Despite many historical challenges of peace building and institutional and socioeconomic deficits, prospects for development were high at the time of independence given the groundswell of popular citizen support, government control over significant oil revenues, and substantial international assistance.6 Renewed conflicts in December 2013 and July 2016 have however highlighted weaknesses in the state-building agenda in terms of dissensions among the elites who failed to agree on an “inclusive- enough� system for sharing power and resources. 2. Lack of effective and accountable institutions and democratic processes associated with citizen’s self-determination to hold elected authorities to account, as well as a common ‘national identity’ to bind diverse communities are generally cited as key reasons for the outbreak of the conflict in 2013. 7 The protracted situation of violence has resulted in a severe humanitarian crisis, exacerbating an already fragile situation. With a population estimated at 13 million in 2017, South Sudan counts almost 2 million of internally displaced persons (IDPs) while 2.3 million had fled the country entirely. 8 Food insecurity also remains a critical challenge, with more than half the population classified as severely food insecure between May and July 2019.9 3. Furthermore, the country displays all the signs of a war economy with contraction of real gross domestic product (GDP) by 6.9 percent in 2017,10 hyperinflation, and depreciation of the South Sudanese pound in the parallel exchange rate market. The country is highly oil-dependent (60 percent of GDP), and lower oil revenues combined with higher spending on defense and security operations at the expense of poverty reduction, have worsened the fiscal deficit. For instance, security and accountability/public administration and rule of law spending have accounted for a little more than 70 percent of the total budget over the past three fiscal years. By contrast, combined expenditures on health and education are estimated to make up around 6 percent of total government spending. Outside the oil sector, livelihoods 6 Nearly 99 percent of voters elected for independence. In the first two-and-a-half years of independence, South Sudan received US$4.3 billion in humanitarian and development assistance, with oil revenues averaging US$1.8 billion per year between 2009 and 2012 (World Bank 2017). 7 There are about 63 different ethnic/language groupings in South Sudan. 8 http://macropovertyoutlook.worldbank.org/mpo_files/mpo/mpo-am18-ssd.pdf. 9 WFP. 2019. “WFP South Sudan Crisis – Situation Report #241,� Rome. 10 Dihel and Pape 2017. 5 are concentrated in low-productive, unpaid agriculture and pastoralists’ work. Consequently, extreme poverty has also increased from 47 percent in 2011 to 86.5 percent in 2017.11 4. South Sudan has some of the worst health and education outcomes in the world. The under-five mortality rate is about 106 per 1,000 live births12 while primary completion rate represents 29 percent, with only 19 percent of girls completing school.13 Moreover, according to the Human Capital Index of the World Bank released in October 2018, South Sudan ranks 156 out of 157 countries. 5. Access to basic infrastructure is also limited, affecting economic productivity and growth prospects. In addition to being a landlocked country, South Sudan also has the lowest road density in Africa with less than 2 percent of the population having access to electricity and 30 percent of the population lacking access to safe water. Box 1 provides some key development indicators. Box 1 : South Sudan: Recovery and Need Deficits High levels of vulnerability  86.5 percent of the population live in poverty in 2017.  92 percent suffered some form of shock in the last five years.  53 percent are food insecure between May and July 2019. Majority lack education or skills with high levels of unemployment, especially among youth.  Adult illiteracy is about 61 percent.  Primary school completion rate is 29 percent.  Only about 19 percent of girls complete school. Low health indicators  Maternal mortality: 2,054 per 100,000 births  Under-five mortality: 106 per 1,000 births  High levels of malnutrition and food insecurity Low basic infrastructure base  Lowest road density in Africa and less than 2 percent of the roads paved.  Less than 2 percent of the population has access to electricity.  Three-fourths of the population lack access to sanitation.  30 percent of the population lacks access to safe water. Source: World Bank (2018), Country Engagement Note - South Sudan; World Bank (2019) South Sudan Economic Update. 6. While a revitalized peace agreement was signed in September 2018, the long-term prospects for a peaceful resolution to the on-going conflict appear tenuous. The main challenge going forward is to ensure all parties to the conflict remain committed to implementing the new peace agreement. Subsequently, the government will be expected to tackle the underlying causes of the country’s current macroeconomic crisis, improve food production, boost employment, build infrastructure, and diversify 11 World Bank. 2019. “South Sudan Economic Update,: Washington, DC. 12 South Sudan Health Survey 2016. 13 World Bank 2015c. 6 the economy, among many other priorities that need to be addressed for sustained peace, stability, recovery, and longer term development. 7. Overall, the situation in South Sudan represents an exceptional development challenge in terms of security and state capacity building as well as human capital. In response to the crisis, many partners, including the World Bank, are actively working to support peace and development, relying on strong coordination and complementarity between development and humanitarian efforts. 8. The World Bank Group operation in South Sudan is guided by the Country Engagement Note endorsed by the Board of Directors in January 2018. It has two objectives: (a) support to basic service provision for vulnerable populations; and (b) support to livelihoods, food security, and basic economic recovery. The Public Works component of the SNSDP corresponds to the second objective. 9. The purpose of this report is to build on the experiences of the World Bank in South Sudan and other countries implementing public works programs to inform specific and innovative design choices in challenging FCV environments. The focus will be on core design features, including targeting, benefit structure (i.e., transfer level), assets, and institutional arrangements (i.e., partnerships with the United Nations [UN] and other actors). The contribution of this report is twofold. First, it informs options for public works implementation approaches in South Sudan to better align with the emerging volatile and uncertain conditions on the ground. Second, the lessons learned supports enhanced understanding among different actors involved in public works in South Sudan —government ministries, World Bank, UNOPS, NGOs, development partners, and others— about the direction and implementation practices of the SNSDP public works. 10. The rest of the report proceeds as follows. After a presentation of the methodology used, section 1 presents the Public Works Project in South Sudan along with the different results and challenges. In Section 2, we discuss the evidences from other countries and lessons learned. Finally, the last section concludes by analyzing the implications for designing public works in South Sudan. Methodology 11. Robust evidence about development initiatives in fragile situations are rare, and evidence about PWs even more so. In a review of evidence about aid in FCV countries, Gisselquist (2015) comments that despite substantial efforts, there are significant gaps in our knowledge about what has worked, why and the transferability of lessons from one context to another. 12. This report’s analysis is mainly based on the SNSDP and other World Bank project documents – especially Project Appraisal Documents, Implementation Status and Results Reports, Implementation 7 Mission aide memoires, and Completion and Results Reports. In practice, this means that while the analysis can explore the rationale for specific interventions and their implementation mechanisms and the outputs of projects, it can say less about the outcomes and impacts of the PWs. Nonetheless, there are advantages to this methodology and the sources of evidence used – exploring why specific instrument and design choices were made in particular contexts allows us to move beyond the internal validity of robust evaluations to a focus on external validity and the extent to which specific design features from other countries might be appropriate and replicable in the context of South Sudan. Selection of countries / case studies 13. In selecting case studies, priority was given to experiences gained in World Bank public works programs, as the aim of this report is to specifically support future activities of the World Bank’s safety net interventions in South Sudan. The rationale was that, whilst there are important lessons to be learned from the activities of other development partners and NGOs, most of these initiatives cover fewer beneficiaries, are ad hoc and have short implementation time frames, and therefore, would be less relevant to the objectives of the report. 14. In each country, the analysis was generally drawn from one specific program. In some cases, the program has numerous phases (e.g. Uganda’s Northern Uganda Social Action Fund - NUSAF) or related programs – particularly where the delivery mechanisms depended on the institutional arrangements in another, long standing project (e.g. DRC and Yemen). 15. The different countries (CAR, DRC, Uganda and Yemen) were also selected based on their similarities with South Sudan in terms of context. For instance, DRC is likely to provide greater insights given existing levels of insecurity and the challenges of working in remote environments within an evolving conflict environment. In Yemen, like in South Sudan, World Bank Group staff have challenges of access to project sites, as they are no longer in the country and operations are carried out using third party implementing partners, namely UN agencies. 16. The labor-intensive public work programs selected for this study are as diverse as the countries in which they are implemented, following different implementation modalities. For example, the Yemen program used UN and parastatal organization to implement the program, while CAR and Uganda used the governments’ staff and systems and local government structures, respectively. However, all of the programs in the selected countries share two key features that are central to this study: (i) the building or repair of productive community asset; and (ii) the provision of a cash to beneficiaries in exchange of work 8 on any such community infrastructure. The lessons from these different implementation arrangements to achieve same objective would be important to inform future designs of similar programs in FCV context. 9 1. Project Description and Results Project design 17. The World Bank’s SNSDP in South Sudan consists of a US$21 million IDA credit . Approved in June 2013 and effective in November 2014, the project included three components: (a) Systems Building and Project Management, (b) Public Works, and (c) Skills Development, which was later restructured to only include components (a) and (b). The SNSDP’s development objective was “to provide access to income opportunities and temporary employment to the poor and vulnerable and put in place building blocks for a social protection system in South Sudan.� The project closed in February 28, 2019. 18. The project was designed in a context of very limited institutional and implementation capacity and experience at all levels of government, civil society organizations and the private sector. The operational tools for targeting, identification, and registry of beneficiaries, and payment transfer mechanisms were rudimentary at best and nonexistent in most cases. In addition, prevailing socio- economic fragility in South Sudan and poor governance, weak accountability and lack of clarity around functional assignments and responsibilities for social protection activities were among the major risks identified during the project design. 19. The recruitment of implementing partner NGOs and private sector expertise to implement the public works component of SNSDP was a key design features of the project to mitigate the risks identified during the project design. As such, while overall project coordination and management was the responsibility of the Ministry of Agriculture and Food Security (MAFS), project implementation was undertaken by Implementing Partner NGOs. To augment MAFS capacity, a Project Implementation Unit (PIU) was established and staffed with consultants, supported by government-assigned focal points from the following ministries – MAFS, Ministry of Gender, Child and Social Welfare (MGCSW), Ministry of Labor, Public Service and Human Resource Development (MLPSHRD), and Ministry of Finance and Planning (MFP). At the grassroots level, the planning and day-to-day implementation activities were monitored in a participatory manner by local government institutions and community structures, including county, Payam and Boma level government staff, and the communities at large. Project Implementation 20. Recognizing implementation capacity limitation at the local level, two NGOs (SNV and Action Africa Help-International) were initially contracted to support in the planning and implementation of the 10 public work component and capacity building support for the local institutions. The project also allocated adequate resources which supported local level joint monitoring and implementation by appropriate government institutions and community structures. 21. However, due to the increasing conflict and insecurity, worsening economy and escalating needs, the challenges of implementation increased over time and the project needed further adjustments. The outbreak of conflict in July 201614 escalated insecurity in the country and the two NGOs implementing the project evacuated from South Sudan, pausing implementation of the project for a year. The conflict also displaced millions of people, further destabilized the macroeconomic condition of the country and exacerbated poverty and vulnerability. When the conflict subsided only one of the NGO returned to South Sudan and was available to restart the implementation. 22. In response, the project was restructured in June 2017 to effectively address the changing socio- economic dynamics in the country and better align the project design to the emerging needs on the ground. The restructuring adapted the project to focus primarily on public works in order to more effectively respond to the alarming humanitarian and food insecurity needs on the ground. The restructuring brought a number of additional adjustments in the implementation of the project, including (a) collaboration with UNOPS as a project implementing partner to reach out to the six remote counties, two of which were in one of the most conflict-affected state; (b) contracting of Alpha Commercial Bank to handle the cash transfer to beneficiaries in all of the project locations; (c) development of biometric and GIS enabled verification system for the cash transfer; (d) rigorous supervision and handholding support by the World Bank team to government and implementing partners (AAH-I and UNOPS); and (e) reallocation of funds to the public work component to increase the number of beneficiaries. 23. Project implementation benefited from a collaborative effort between the World Bank, AAH-I, UNOPS, Alpha Commercial Bank and the Government, involving not only the PIU but also representatives of the Ministries of Finance, Gender, Labor and Agriculture. The efforts form the multilayered structures – PIU, core teams at the county (if rural) or city council (if urban) levels, development committees at the lower levels of government, and the community oversight and coordination structures, including the GRM – ensured a participatory, transparent and accountable project implementation with checks and balances to mitigate against elite capture and manipulation and other predatory behaviors. 14 The 2016 Juba conflicts were a series of clashes in Juba between rival factions of the Sudan People's Liberation Army (SPLA) loyal to President Salva Kiir and Vice-President Riek Machar respectively. Overall 300 people were killed, approximately 36,000 civilians fled parts of the city and there was massive destruction of properties 11 24. The collaboration with UNOPS, Alpha Commercial Bank and AAH-I has benefited project implementation in a number of ways including: a. UNOPS and AAH-I filled the implementation capacity gap of the Government and facilitated the scale up and smooth implementation of the project; b. They also enhanced local capacity (local government and community structures) by training and involving them in the implementation of the project to identifying and resolve implementation issues and challenges. Enhanced local capacity is perhaps the most important outcome of the project that will remain after the completion of the project; c. UNOPS’s nation-wide presence and economies of scale, including in the most conflict- affected and opposition held areas, allowed the project to scale up quickly in remote target areas which otherwise would have difficult within the short timeframe for NGOs; d. The use of Alpha Bank and biometric payment verification system in the cash transfer to beneficiaries ensured transparency and accountability, serving as an effective fiduciary risk mitigation strategy; e. Engagement of UNOPS and AAH-I created an opportunity to strengthen the linkages between the humanitarian-development nexus by meeting immediate needs while strengthening community and local institutional capacity for medium term recovery; and f. UNOPS and AAH-I benefited in deepening their understanding and strengthened implementation capacity of safety net interventions, which will help them to adapt to their mainstream humanitarian interventions in the future. 25. The Public Works component intended to benefit 34,000 poor and vulnerable households working for 90 days at US$3 a day. The project was carried out in different phases. Beginning in March 2016, PW was initially implemented in Juba City by AAH-I. The intervention initially targeted 6,000 poor and vulnerable poor households in 15 communities. Thereafter, it was extended to include 15 new communities targeting an additional 6,140 beneficiary households, bringing the total number of supported beneficiaries in Juba to 12,140 households at the end of the project in August 2018. At this time, SNV also supported some 4,000 households in Torit. Building on the experience of AAH-I in Juba, PW activities was further scaled up in six other counties 15 by UNOPS, including in Torit once SNV left after 15 The six counties are Kapoeta East, Gogorial West, Tonj South, Torit, Pibor, and Bor. 12 the 2016 conflict, with the purpose of targeting 35,000 beneficiaries, from December 2017 to December 2018. 26. The project used a combination of targeting mechanisms including geographic, administrative and community targeting. Geographical targeting was used to identify selected states and counties (both rural and urban) with high level of poverty and other development indicators as proxy for vulnerabilities, with relatively better accessibility and security to ensure year-round project cycle activities, with the assumption that gradually and with improved capacity and additional financing, the project would be able to roll-out to other areas. Then, community-based targeting was used to identify beneficiary households within the selected communities, with quotas given per community based on available funding. 27. To realize its primary development objective, the project was rolled out through nine series of activities: community mobilization; beneficiary selection and registration; PW subproject selection; environmental and social impact screening and assessment; establishment of beneficiary management committees; provision of working tools; monitoring of activities; GRM; and payment of wages. 28. Community mobilization was done through meetings and other exercises to create awareness and sensitization among selected communities about the project, including regarding the targeting criteria used for beneficiary selection. Community mobilization was a very important activity to ensure the community understands the project objectives, implementation guidelines, roles and responsibilities of each stakeholder (including beneficiaries, community committees and community members at large), and minimize grievances. It also supported capacity building of the beneficiaries and communities. In project locations where community mobilization and sensitization were not adequately conducted, the project experienced higher levels of implementation challenges, in rare cases culminating in escalated tension and violence between project staff and community members. 29. Households were targeted through a community-based approach using transparent and neutral targeting criteria established through wide-ranging consultations with key stakeholders. 16 The criteria included those with no means or limited income; high dependency ratio; and households headed by women, young men, or a child, with severely disabled person, whose livelihoods were disrupted, with a widow, and with no literate adult. Another intended goal in the selection process was to have a minimum of 50 percent women’s participation, although the Results Framework mandated 30 percent. All the beneficiaries were registered manually prior to biometric registration, and then biometrically once established (i.e. second phase in Juba and the six other counties). 16 These included relevant Government authorities, development and humanitarian actors, NGOs, academics etc. 13 30. Beneficiaries were trained on PW activities, and were given personal protective equipment (i.e. gumboots, reflective vests etc.) and first aid kits for treating injuries to improve and promote occupational health and safety. Cash was transferred through a commercial bank using GPS enabled biometric verification. 31. The project established management committees at all level of governance and trained them to support PW implementation. There were deliberate efforts to include all stakeholders in planning, implementing and monitoring and evaluation (M&E) of the project activities. The project management committees (City Council Core teams, Block Development Committees and Quarter Council Development Committees in urban areas, and County Core teams, Payam Development Committees, and Boma Development Committees in rural areas) played key roles in the implementation. Reports detailing the performance of the various management committees were produced by AAH-I and UNOPS that were shared with PIU and the World Bank. 32. Facilitated by the implementing partners (IPs), the communities selected various subprojects which they felt were priorities for their community and in line with the project guidelines. For instance, projects that needed an investment beyond the provisions of the labor intensive PW were discounted, such as constructions of culverts. Environmental and Social Safeguards 33. Environmental and social impact assessments were conducted before implementation of subprojects. Accordingly, mitigation measures developed and implemented to manage risks, as part of Environment and Social Management Plans (ESMP). Examples of the mitigation measures put in place to address environment issues were: sensitizing communities on solid waste management; opening of drains along the roads to facilitate flow of runoff water to avoid water stagnation; and installation of scour checks to reduce soil erosion, among others. 34. Another interesting feature of the project is the GRM established at all levels as part of social safeguards, with an Appeals Committee at each target community. The Appeals Committee’s role was to receive, record, and handle complaints from beneficiaries and community members. If any complaint could not be dealt with at this level, it was escalated to higher authorities (i.e. the development committees) as necessary, enhancing accountability between the community and local authorities. More details are in Box 2. Box 2: The Grievance Redress Mechanism 14 The GRM is a means to allow the beneficiaries and community members to voice their complaints and rights regarding any perceived unfairness in their work environment, without any restriction, so that corrective measures could be taken to address the issue, as appropriate. Given that PWs has multiple stakeholders and adopts a community-based approach to beneficiary identification, the implementation of PWs could lead to conflicts, disputes, and/or differences with respect to fairness of targeting and beneficiary selection, identification and implementation of activities, payments etc. To address these and other grievances, responsive and accessible GRMs were established in the form of Appeal Committees at the community level. The Appeal Committee was composed of 5 to 7 well respected people from the community who were not beneficiaries. After the formation of the Appeal Committees, awareness/sensitization trainings were conducted to explain their roles and responsibilities, including on how to receive, redress, and report grievances. Each Appeal Committee elected a secretary who received grievances, complaints, or feedback, and recorded the actions taken by the Appeals Committees in response. In Juba, a GBV focal point was introduced in the Appeals Committee on a pilot basis to sensitize community members, leaders, and beneficiaries regarding GBV issues. 35. In light of the prevailing violence against women and the magnitude of GBV incidents in South Sudan, the SNSDP also piloted the establishment of GBV focal points in the GRMs in Juba to sensitize communities, beneficiaries, Appeals Committee members, Chiefs and local leaders on GBV and gender issues. However, social norms around GBV made it difficult to openly discuss these issues within a community and resulted in under-reporting. Conversely, the lack of referral mechanisms and complementary support made it difficult to address the few GBV cases that were reported. 36. It thus became clear that changing mindsets and ensuring that GBV risks are mitigated within the project would require efforts to be mainstreamed more broadly throughout the project cycle. Specifically, it would require awareness raising campaigns prior and throughout implementation to ensure buy-in from local leaders and communities, and to communicate potential project-related risks and options for reporting or response should allegations arise. Consultations with relevant local organizations and communities, including separate focus group discussions with women and men, and specific outreach aimed at men would also be necessary to change mindsets. Public works would need to be designed to be gender sensitive (i.e. gender sensitive Codes of Conduct prohibiting sexual harassment, exploitation and abuse; work sites situated within communities; security at payment sites etc. Further, it would be useful for the GBV focal points to have information on response services and referral systems within/near their communities for potential survivors GBV-sensitive reporting channels should also be developed to be integrated into the GRM to enable use by survivors and to refer survivors to appropriate care. 15 Fiduciary and Safeguards 37. Implementing a cash-based program in an FCV environment such as South Sudan was not without high operational risk. Given the very poor governance, violent conflict, rampant corruption and mismanagement of public resources, the SNSDP was exposed to many risks. Key among them include the risk of (a) inadvertently warring parties accessing project resources, and thus fueling the conflict; (b) potential manipulation of the targeting process, and (c) elite capture of project benefits. 38. To mitigate these risks, the project developed robust and transparent beneficiary registration and payment mechanism. To this end, biometric solution was developed to assure registration of the right beneficiaries under the project and validation of those beneficiaries during payment. This was especially important in the context of the high-risk and complex South Sudan environment where most beneficiaries do not have national identity cards. Beneficiary registration was conducted in the field using portable biometric kits and the data was then uploaded to a secure server, which was linked to the MIS managed and hosted by a consultancy firm contracted by the MAFS. Due to challenges of unavailability or unstable internet connectivity, especially in remote areas where the program was implemented, the biometric system included offline functionality, which allowed automatic data transmission to the server once internet connectivity becomes available. 39. With biometric registration, payment of beneficiaries was highly secure. Once a beneficiary was initially identified using program ID and photograph, he/she then captured into the biometric kit using finger print identification, at which point the personal details were displayed and validated, only after which a payment was made. In this way, the system provided multi-factor authentication using beneficiary finger print, program ID, and photograph to ensure positive dentification of the beneficiary during registration and payment. Moreover, the biometric validation using finger print automatically generated a unique electronic signature for the beneficiary which is then used to secure audit trail for the payment. 40. In exceptional cases where a beneficiary could be registered biometrically due to disability, for instance, an alternate from the household was registered instead, which allowed him/her to access services just like the other beneficiaries. In case the principal beneficiary is unable to go for payment (due to sickness, old age or other reasons), the same registered alternate was allowed to receive payment on behalf of the principal beneficiary. 41. In order to address the risk of potential double dipping where a beneficiary could receive payments more than once, the following functionalities were built into the biometric system: (a) Geo- fencing of biometric kits – GPS coordinates captured at specific registration/payment centers were used to lock the biometric devices such that a beneficiary could not access payments in another center where 16 he/she was not registered. Under the SNSDP, beneficiaries are registered and paid in groups within Quarter Councils/locations on specific days; and (b) Graduation of household status – Once payment was made to a beneficiary household, its payment status automatically changed from Unpaid to Paid, and was removed from the pending payment list. This made it impossible to pay both principal as well as alternate beneficiaries at the same time in the same payment center. 42. Through these measures, the system provided additional assurance that targeted beneficiaries were actually enrolled and registered into the program. Similarly, the biometric system also resulted in enhanced fiduciary assurance by providing additional confirmation that cash transfer payments were made to the intended beneficiaries. The system also enabled payment to a large number of beneficiaries and faster reimbursements to the payment service provider. 43. However, there were also challenges in using the biometric system, including resistance from some community members to accept the introduction of collection of beneficiary’s finger prints and taking pictures. In some of the project areas, there were rumors that the biometric machines would suck blood and beneficiary pictures would be sold by project staff. It was also difficult to capture good quality finger print due to worn-out hands as a result of hard manual work and/or because of old age. Lastly, risks associated with privacy and personal biometric data security was a particular concern. South Sudan similar to most African countries, does not have an adequate legal framework for data handling and protection and the risk of beneficiary personal data falling in to the wrong hands was one of the main challenges faced by the project. Monitoring and Evaluation 44. The M&E system for the project included a MIS developed and hosted by an IT firm; a payment mechanism supported by biometric and GIS based verification; regular reporting by implementing partners (UNOPS and AAH-I); and participatory community engagement through Community Score Card exercises and a beneficiary feedback survey. Communities were facilitated to engage throughout the project implementation, including in the monitoring of implementation of project activities. 45. As part of the M&E framework, the implementing agencies gathered data on beneficiary feedback, documented case studies, generated periodic progress reports, and prepared end-of-project reports. An independent evaluation was done at the end of the project in Juba. However, this was not possibly for the other six locations due to the truncated implementation period resulting from the loss in implementation time after the July 2016 conflict. In addition, the beneficiary satisfaction survey was on- going during the writing of the report, and therefore was not available. 17 Capacity Building 46. Strengthening capacity of local government and community-based structures were among the main activities of the project. Training was provided to committee members at the community level on various aspects of the project cycle, including roles and responsibilities of the committee members in addressing implementation challenges; community mobilization; targeting criteria for the selection of the neediest households in the community; resolution of disagreements within the community as a result of project activities etc. Through these efforts, the project built implementation capacity that ensured successful completion of PW activities despite all the challenges faced during its implementation. At the end, strengthening of community structures empowered its members to actively engage in the project implementation and M&E process, which will remain an asset for future engagement in local development initiatives in these communities. 47. The project also provided training on basic financial literacy and water and sanitation practices that helped strengthen household capacity to better utilize the cash transfers and engage in livelihoods activities, as well as improve basic sanitation within households and communities for positive health outcomes. The project also encouraged group farming as an innovative public work activity which contribute to improved food availability and livelihoods. 48. The next section details the results of the PW design and implementation. Results 49. Based on the nine activities of the PW (i.e. community mobilization, beneficiary selection and registration, subproject selection, environmental and social impact assessment, establishment of beneficiary management committees, provision of working tools, monitoring of activities, GRM, and payment of wages), the following achievements can be considered. 50. Community mobilization appears to be one of the most important factors for the success of PW. The project adopted a community-based participatory approach in the selection of beneficiaries and the subprojects, as illustrated by the establishment of the community oversight and coordination structures, such as the beneficiary management committees. Through these measures, the project was found to contribute to enhanced community leadership, involvement, and ownership, as indicated by beneficiaries in end-of-project and completion reports. 51. As a result, in some communities, plans have been put in place to continue using the tools provided even after activities closed at least once per month in maintaining roads among other community works. In contrast, in areas where community mobilization was weak, the project faced 18 challenges due to misunderstanding and confusion regarding project objectives, approach, and benefits. In some extreme cases, there were violence against project staff by some community members (i.e. in three Bomas in Kapoeta East). 52. Moreover, project satisfaction among beneficiaries in Juba in particular was high according to the independent end of project evaluation. Beneficiaries stated that the project has improved their livelihood and their access to food, education, and health. They especially appreciated the additional trainings on WASH and financial literacy. With basic financial literacy training, beneficiaries were able to manage better the cash from the public work and some of them were able to start small income-generating activities from their savings, such as setting up tea kiosks and selling vegetables. WASH trainings contributed to improved household and community hygiene and sanitation practices, leading to reports by target communities of decreased incidents of communicable diseases such as cholera. 53. Beneficiaries pointed out that the project has improved social cohesion and unity, given that they had to work together regardless of difference in their ethnicity. The project provided a platform where community members came together, putting aside their differences, to work towards a common goal of improving community welfare. Figure 1: Before and after public works Before rehabilitation After rehabilitation * Photos from Project file 54. The project benefitted 53,163 households (355,871 individuals) of which women represented 74 per cent of the direct beneficiaries, indicating that they were the most vulnerable population based on 19 the project target criteria. Their participation was facilitated by ensuring that their gendered needs were met, such as ensuring that child care, breastfeeding, and provision of sanitation facilities, were taken into consideration in the project implementation. For instance, temporary shelter was provided where elderly women cared for the children as their mothers worked. 55. Public work activities supported by the project included road rehabilitation and maintenance, cleaning of drainage channels, cleaning of public places such as schools, collection and disposal of waste and group agricultural activities with majority of the priority given to roads, especially in urban areas such as Juba. This demonstrated the value addition of a good road network in terms of improved service delivery and mobility that is facilitated by passable roads. For instance, before the roads were repaired, delivery of vital services was hampered. However, after rehabilitation, water tankers supplying water to all households as well as ambulances, security patrols, and so on could access communities and households and provide services, accordingly. In addition, innovations were tried in rural areas with agricultural activities as public works, whereby poor households had divided up communal land for farming. Seeds were provided by the communities, and the yield was shared equally by the beneficiaries. While the cash from the public works supported families with immediate consumption gap, the harvest not only provided food for the family but has also improved economic welfare as it had allowed families to sell the surplus yield in the market, thereby enhancing food security and contributing to sustainability. 56. Concerning payment, the objective was to ensure timely and accurate payments to the intended beneficiaries with risks of error, fraud, and corruption and leakages mitigated, to the extent possible in high fiduciary risk environment such as South Sudan. Payments to beneficiaries were done by a Commercial Bank through a multifactor authentication biometric tool using finger prints and GPS tagging as verification methods. This was helpful in assuring accountability and minimizing risks to the project compared to a system based on manual payment and paper trail. Even in remote areas, this robust payment system enabled monitoring of payments – in terms of whether it was received by the intended beneficiaries and where, ensuring effective tracking of funds. The final report evaluation for Juba indicated that most of beneficiaries were paid on time while there were complaints about delayed payment in remote areas. 57. The GRM was also effective in addressing complaints from the beneficiaries, with a satisfaction rate of 77 percent in Juba, according to the independent end of project evaluation. The GRM dealt with complaints regarding targeting, payment delays,17 resistance to public works in some residential areas by 17Delay in payments was encountered, especially in the remote areas. Weaknesses in the MIS and biometric systems, coupled with lack of reliable internet connectivity, resulted in errors in the payroll and delays in payment. 20 non-beneficiaries, and some incidents of violence towards UNOPS staff in three of the remote Bomas in one of the other six counties. More details about these issues and how they had been managed by the GRM will be given in the next section that deals with challenges. 58. The project had tried to mitigate risks of gender-based violence (GBV). The project was implemented in an environment which is prone to high level of GBV incidents and the awareness on GBV and reporting of cases by beneficiaries is not easy. In the Juba communities, 405 focal points were selected and trained as a pilot on GRM and GBV issues to follow up GBV cases. However, challenges in underreporting continued to be perennial, and only one case by a husband refusing to allow a female beneficiary to work was reported, which was successfully mediated by the Appeals Committees. Lack of referral mechanisms also made follow up support difficult. The absence of reporting on GBV should not be interpreted as absence of GBV practice. It is rather lack of appropriate mechanism to capture and effectively solve the problem. Therefore, in future design, the issues of GBV and sexual harassment should be addressed by strengthening the GRM mechanism and devising a follow up support mechanism for survivors. Challenges and Lessons Learned 59. Despite these positive results, the project experienced many challenges, and addressing them allowed for lessons to be drawn that will improve PW project design in the future. 60. Well planned and coordinated community mobilization and sensitization at commencement and during implementation are critical to ensure buy-in and support from the community, as it will minimize tension and conflict between communities. Despite efforts to ensure effective community mobilization and sensitization, there were communication gaps, especially in the remote areas due to challenges and delays in implementation, which resulted in some communities and beneficiaries not receiving adequate information about the project. This in turn contributed to: (a) escalated tension and violence against project staff by community members in some locations; (b) a number of cases of conflict of interest among members of the QC development committee in targeting beneficiaries (i.e. selecting family members who are not poor or vulnerable as beneficiaries); and (c) some complaints about elite capture (i.e. influential community members/leaders dictating the selection processes) in the selection of beneficiaries.. 61. Regarding violence against project staff, there were three reported incidents of violence towards UNOPS staff in three Bomas of Kapoeta East. The three reported incidences were: a) detainment of the UNOPS Team Leader by non-beneficiary youths alleging that they were excluded unfairly; b) slapping of a 21 UNOPS Community Mobilizer by a beneficiary alleging that the biometric machine sucks his blood; and c) detainment and beating of the UNOPS Team Leader by local police because the beneficiaries complained that the project did not pay them on time. The incidents were addressed at the time through the GRM following established procedures. Following up, project activities in two Bomas were stopped as they had histories of recurrent violence against aid workers. 62. Besides the above extreme incidents, most of the complaints involved targeting of ineligible beneficiaries. These were deemed to have been addressed effectively through the GRM – the Appeals Committees reviewed the complaints about ineligible beneficiaries and removed those found to be included in error from the beneficiary lists. The process of review of such complaints took place during the targeting of beneficiaries and involved consultation with wider community members, including both beneficiary and non-beneficiary members of the community. Nonetheless, responses to complaints and grievances through GRM had been observed in some cases to be slow due to the time it took for the consultation process. Observed delays in addressing grievances by the Appeals Committees could also partly be attributed to lack of incentives and wages for committee members. This resulted in a general feeling of low motivation from some of the members, with some even deserting the committees. As such, it would be critical to further strengthen the GRM, particularly the Appeals Committees, and facilitate their work through alternate means of incentives (i.e. project paraphernalia), as the proper functions of the GRM is critical to minimizing social risks, improving relations between community leaders and members and beneficiaries, and ensuring that effective checks and balances to project governance is in place. 63. One of the complaints which was not successfully resolved during the life of the project was the exclusion of poor and vulnerable households without able-bodied adult to participate in PWs, as the project design only allowed cash transfers to poor and vulnerable households participating on public work activities. This should be adequately integrated in the design of future PW projects, so that more of the neediest and poor households can also benefit. 64. Concerning beneficiary registration and payment, there were numerous challenges, including reports of missing beneficiary names in payroll lists; multiple alternates18 per beneficiary in the biometric system; and inability of finger prints of beneficiaries to be read, among others. These challenges resulted in weaknesses in the MIS and biometric systems, exacerbated by unreliable internet connection (as the systems are web-based), and were most pronounced in the six counties under UNOPS. Weaknesses 18Beneficiaries are asked to nominate one family member to serve as their alternate to receive payments on their behalf if they are not able to come for payment or if their finger print can’t be recognized by the biometric reading machine. 22 included poor data quality, ineffective internal system checks to spot duplicates, generation of multiple payrolls etc. Missing beneficiary names in payroll list delayed payments and increased risk of violence against UNOPS staff. Multiple alternates made it difficult to track payments and ensure that payments reached the intended beneficiaries. 65. To address these issues, UNOPS had to buy a new VSAT19 equipment for Internet. In consultation with the PIU, UNOPS also devised a plan to accelerate both the public works and the payment process. For example, they changed work and payment cycles from 15 days to 20 days per month 20. This flexibility in the management of the project ensured the completion of all public works as well as the development of improved payment plan. Furthermore, the MIS consulting firm cleaned up the existing data to remove all duplicates, resolved data mismatch between the biometric database and the MIS, and strengthened internal checks to ensure that the system will not allow duplicate beneficiary information in the system. 66. Finally, the risks associated with beneficiary privacy and personal biometric data security was a concern. The lesson from the SNSDP indicate that while introducing a biometric system is good for transparency, countability, effectiveness and efficiency of a program it is mandatory that there are legal and technical safeguards in place to ensure the security of biometric data of beneficiaries. 67. Apart from issues related to project design, there was also challenges related to the general country context in South Sudan. 68. Due to hyperinflation in the country, the real purchasing power of the transfers continued to decline throughout the life of the project, risking welfare impact of the cash transfers. Moreover, due to the fluctuating exchange rate and the wide gap between the official and market forex rates, beneficiaries were confused about the wage amounts received every month, and constantly complained about the low official rate used for exchanging their wages. 69. This was addressed by pegging the transfer value to South Sudanese Pound (SSP) equivalent of US$ 3 per day, exchanged at a preferential rate negotiated with Alpha Bank. Moreover, Bank commission charges were also negotiated at a minimal level of 1 percent. These efforts ensured that beneficiaries received as much of the transfers in SSP as possible. It was also negotiated that Alpha Bank would provide the cash transfers using their own funds, which would be reimbursed by the World Bank through direct payment based on extensive verification of the payment evidence generated through the MIS and biometric payment reports. Despite all the effort to minimize the gap between the official and parallel 19 VSAT—Very Small Aperture Terminal—is a two-way satellite ground station with a small dish antenna used for the provision of satellite Internet access to remote locations, VoIP, or video. 20 Beneficiaries were paid after 15 days of work and to ease the delay of payment the payment and work cycle was changed, and beneficiaries were paid after completing 20 days of work. 23 exchange rate beneficiaries were not happy because they wanted to receive their wages in US dollars and exchange it in the parallel market. In future projects, it would be useful to ensure that the beneficiaries receive sufficient funds vis a vis real purchasing power to make the transfers impactful. It would also be good to ensure that favorable agreements with commercial banks are maintained such that beneficiaries can receive much of the transfers. 70. Insecurity and lack of infrastructure also led to high implementation costs and delays. Insecurity due to civil war was a major constraint as implementation was interrupted, and activities had to be paused for nearly a year. Moreover, the implementation budget did not foresee several critical expenditures, including increased field visits for close monitoring and supervision by the PIU, need for additional staff for biometric registration; scale up of beneficiaries given escalating needs; financing for environmental and social safeguards measures; secure storage facilities for PW tools to guard against theft; and allowances to smoothly facilitate the work of local level oversight and coordination structures and Appeals Committees, among others. For example, there was no budget allocated for air transportation to the field and the start of the rainy season made road delivery impossible. Hence, the project had to resort to using cargo planes to send the materials to the field, which was very expensive. Moreover, fuel cost and scarcity also affected the project activities. This was addressed through contracting with the United Nations Mission in South Sudan (UNMISS) to provide fuel in all the field locations, thereby providing more stability and certainty. 71. Lastly, the program could not support a majority of the neediest and most poor people due to limited resources, resulting in a lot of grievances and complaints from non-beneficiaries and non-targeted communities.21 This is a major issue in South Sudan, where over 6 million people face severe food insecurity and urban poverty deteriorated to 70 percent in 2016 from 49 percent in 201522. The number of households demanding to participate on the safety net project was way over the available fund and complaint about exclusion in the targeting was one of the major issues it had to deal with. In addition, the poor and vulnerable households who had no able-bodied members to participate in the PW did not receive benefits. 72. In order to effectively deal with the myriad challenges detailed above and given the rapidly changing context in South Sudan, the project had to be flexible in its design to handle a wide variety of risks related to conflict, insecurity, elite capture, and potential tensions between targeted and non- targeted communities, local authorities and IPs. Principally, strengthening the delivery tools, particularly 21 This a common feature with Social Safety Nets programs in Africa, as underlined by Beegle, Coudouel, and Monsalve (2018). 22 World Bank 2017. 24 the targeting mechanism, GRM, MIS, biometric and payments mechanisms has emerged as fundamental to ensuring successful management and implementation of PW programs, as well as to mitigating fiduciary and social safeguards risks. 73. The next section reviews the evidence in other countries. It will focus on core design features, including targeting, benefit structure (e.g., transfer level), assets, institutional arrangements (e.g., partnerships with UN and other actors), and linkages to other programs. 2. Evidence from Other Countries 2.1. Strategic Objectives / models for PWPs 74. In this section we review experience with public works in challenging environments by assessing the prevalence of five different ‘models’ of public works and making brief comments on the extent to which the outcomes of the respective models are achieved in our case study examples. Box 3 Models of public work programs applicable to South Sudan Model 1: Safety-nets – i.e. public works provide income support and consumption smoothing Model 2: Employment Generation – i.e. the provision of counter-cyclical and more permanent labor-intensive jobs in contexts of chronic poverty Model 3: Asset provision – i.e. infrastructure created and services delivered through public works create assets with a wider value in the local community and economy Model 4: Employability enhancement – i.e. where participation in public works generates human capital assets (skills, competencies, knowledge) that enhancements the likelihood of job market entry Model 5: Social cohesion and peacebuilding – i.e. governments gain legitimacy through the delivery of public works, including by providing required goods and services, by providing support to the demobilized and disarmed, and by strengthening the social contract between citizens and the state 75. Across the case studies projects, most public works projects have a combination of strategic objectives. In practice, this means that they align with more than one of the models listed. At a Project Development Objective (PDO) level, all the PWs align with the more modest aims of the safety nets model, but each has an additional angle or dimension or has followed a different transitional pathway between models. 76. Uganda. NUSAF’s model has changed substantially over time. Initially, NUSAF was a community driven development (CDD) project very heavily focused on conflict-resolution and recovery, and aimed to ‘empower communities in 18 districts in northern Uganda by enhancing their capacity to systematically identify, prioritize, and plan for their needs and implement sustainable development initiatives that improve socioeconomic services and opportunities – thereby contributing to improved livelihoods by 25 placing money in the hands of communities’. Under NUSAF 1, the ‘wage’ for public works was provided by a notional ‘community contribution’ to the project – i.e. program ‘beneficiaries’ worked for free with the logic being that they would benefit from skills development and the assets created (Models 4 and 3, respectively). Under NUSAF2, there was a more modest PDO to provide income support to and build the resilience of poor and vulnerable households in northern Uganda. There was a move away from a very direct engagement with conflict recovery and of the four elements of the Peace, Recovery and Development Plan (PRDP), which included consolidation of state authority; rebuilding and empowering communities; revitalization of the economy; and peace building and reconciliation. Rather, the focus was on rebuilding and empowering communities and the revitalization of the economy. In NUSAF’s third iteration it now focuses much more on resilience building, in relation to climate change and climate- related disasters rather than resilience in the face of conflict and violence, and thus has become more aligned with an adaptive social protection model. 77. Central African Republic. For LONDO23 project in CAR, the PDO ‘to provide temporary employment to vulnerable people throughout the entire territory of the country’ [author’s emphasis], is focused on basic tasks (vegetation clearing, debris removal, side drain cleaning and minor surface repairs) that are labor intensive and require little capital and so sits squarely within Model 1. In the PDO and in the underpinning rationale and more strategic objectives in the Project Appraisal Document (PAD), there is no attempt to be ambitious about the public works taking place (i.e. they are not expected to lead to substantial asset creation – so not Model 3; their limitations in terms of enhancing employability are recognized – so not Model 4 either). However, the PDO also stresses the importance attached to expanding program support outside the capital, Bangui, and extending it across all parts of the country in a way that is viewed as equitable and fair. So, in addition to the modest but highly focused aim under Model 1, the LONDO program also incorporates Model 5. Indeed, the PAD states ‘LONDO is a social cohesion project. However, social cohesion cannot be engineered from the outside. Instead, the project focuses on tangible outputs – payments of stipends and improved rural roads – to lead to the project outcome – social cohesion. The ways in which LONDO works – going beyond infrastructure building, promoting collective decision-making, collaborative behaviors and community ownership – underpins the social cohesion model. It aligns very clearly with what Andrews and Kryeziu (2013, p. 3) in their background paper for the 2013 World Development Report on Jobs have describes as a key outcome of 23 LONDO means “stand up� in Sango, the official language of the Central African Republic 26 public works – ‘a channel for voice of excluded groups; as well as an opportunity to interact with local government establishments and officials’ 78. Democratic Republic of Congo. In DRC, under the US$79.1 million Eastern Recovery Project (ERP) the objective is to improve access to livelihoods and socio-economic infrastructure in vulnerable communities in the eastern provinces of DRC. Conflict prevention and post-conflict reconstruction are listed as the major themes of the project with social protection as a supporting theme. The program is underpinned by a recognition that while labor-intensive public works (LIPW) have been an effective means to provide short-term employment to unskilled (and sometimes semi-skilled) workers (Model 1), they can currently only supplement meagre incomes from agriculture (Model 2 – the project works are aligned seasonally to avoid interfering with busy agricultural seasons), and not provide access to more remunerative alternatives nor offer a limited alternative to youth at risk of recruitment from armed groups. The PAD notes that ‘the bad state of the road network lends itself well to this type of initiative, though the quality/durability of some of the works carried out with labor-intensive techniques has been questioned’, suggesting that, at present, assets of only limited value are being created (so not Model 3). Thus, the ERP proposes an ambitious combination of activities contributing to all 5 models– the activities are: building on the shorter term employment programs that provide immediate support to households; avoiding overlap with the peak agricultural labor season; seeking to construct durable infrastructure that supports livelihoods; and supplementing public works with activities aimed at enhancing participants’ employability at the end of the works; while being underpinned by an approach that seeks to enhance community resilience and stability. 79. Sierra Leone/Liberia. In Sierra Leone and Liberia, the nature of the projects assessed (i.e. those providing a crisis response following the Food, Fuel and Financial crisis in 2008) means that they inevitably follow Models 1 and 2 – providing temporary support to households for a fixed duration. Elsewhere, the emphasis in PADs on delivering the program in the right season suggests a limited counter-cyclical element to the approach, though it is articulated more in terms of the practicality of doing work than the counter- cyclical nature of economic opportunity and vulnerability (Model 2). Implications for public works objectives in South Sudan 80. The lessons from other countries about appropriate and effective models for PWs suggest that a more modest model – one that supports consumption (Model 1), including through the lean months before the rainy season (Model 2) – may be more achievable in the South Sudan context than something that tries to face the drivers of conflict head on. In terms of Models 3 and 4, the elements of the SNSDP that focused on skills training were removed in 2017 after South Sudan slid back into violent conflict the 27 previous year. As such, it would seem that in the absence of institutional systems that have adequate capacity and are resilient in the face of political, societal and institutional instability, Model 4 should not be the immediate priority at present in South Sudan. There are no convincing examples of Model 4 being achieved in other countries experiencing violent conflict and political instability. Given that the simplistic formula of jobs for youth = peace has been widely challenged, and that where there are limited resources to provide sustained employment for people, there is a substantial risk of creating grievances when expectations are not met. Thus, a highly modest approach to Model 5 is more likely to be appropriate in South Sudan, i.e. an approach which focuses less on delivering jobs and more on delivering public works in ways that allow greater participation and engagement of local communities in decision-making (about targeting, about the work that takes place, about periods of work, etc.) and give a voice to local grievances may provide social cohesion dividends over time. 2.2. Design and Implementation Features 81. In this section, brief key findings are presented related to different elements of program design and implementation features from the selected case study programs. It is not the intention to be comprehensive in this analysis, but rather to highlight points of agreement, and of contention, that are most useful to the South Sudan context. Targeting 82. In any context, among beneficiaries, donors, government officials or civil society actors, targeting is frequently the most contested and divisive element of all social protection programs. It is unsurprising, therefore, that extra care needs to be taken over targeting of social protection, including public works programs, in fragile and conflict-affected situations. As Andrews et al (2012) note ‘… the challenge of achieving good targeting should not be underestimated. If inclusion errors cannot be limited and if positive externalities are not evident from the assets created, then the cost effectiveness and value of public works vis-à-vis other safety net programming options must be reconsidered.’ 83. In the case of eligibility and targeting, the issue of self-targeting is dealt with below in relation to wage-setting, but key findings are as follows. First, most programs use a combination of targeting mechanisms. For example, the Cash for Work (CFW) program in Sierra Leone had three layers: first geographical targeting was used to identify the beneficiary communities; then requests were submitted by communities to receive program funds for subprojects; and third, community-based targeting was used to identify beneficiary households within the selected communities. A number of lessons follow from this experience. 28 84. Where data sources such as food security, poverty or conflict-affectedness indicators are used for geographical targeting, they are less contested if they come from non-partisan actors. In Sierra Leone, for example, cash-for-work program resources were allocated to seven districts based on poverty and food insecurity rates and population ranking using the Comprehensive Food Security and Vulnerability Assessment (CFSVA) by World Food Programme (WFP) and Food and Agriculture Organization (FAO). Experiences in Yemen highlight the importance of relying on a multi-dimensional index (in the Yemen case to rank ‘distress’). So, in the case of ECRP, it was important not to focus only on IDPs as part of the targeting formula, but to also consider numbers of returnees, levels of conflict-affectedness and high malnutrition rates (Al-Ahmadi and de Silva, 2017). 85. Where data sources to support targeting are limited, the lessons from the countries are that lotteries might be prove effective. For example, in CAR a lottery system is used with beneficiaries distributed across all parts of the country (Figure 1). 29 Figure 2 Cartoon explaining the lottery system for LONDO in CAR At community level, the challenges associated with ranking of potential beneficiary households, the possible lack of trust between groups in the community, and the potential for elite capture also raises challenges for the approach taken in Sierra Leone in 2010. In Yemen, the importance of having a local implementing partner that is viewed as objective, neutral and impartial and with a good knowledge of the local community is viewed as paramount (Al-Iryani et al 2015). In areas where demand for participation in public works is high – in both fragile/conflict- affected and developmental context, the engagement of local committees or communities in decision-making often leads to rotation systems (where more households are included in the program but receive less work, and by extension, lower overall wages) – a good outcome in terms of community ownership of programming decisions but a poor one in terms of impact. The extent to which these systems can function effectively in fragile situations is less clear. In CAR, the LONDO project finds local selection and differentiating between different geographical regions is both logistically too challenging and likely to fan the flames of perceived unfairness. Instead, it distributes to all sous-prefectures (districts) and in each runs a lottery to identify beneficiaries. 86. Some of these approaches run counter to the orthodox operations of donor agencies. Furthermore, they raise questions about whether, in fragile and conflict-affected situations, implementers should be most preoccupied with minimizing errors of inclusion or exclusion. It remains unclear which are more likely to result in grievances, a break-down between households in communities under stress and a lack of trust and legitimacy of the program (and its officials) at local level. Where the principles of 30 (perceived) fairness are critical for program design, there is a need to better understand whether it is tackling inclusion errors or exclusion errors that is likely to create greater acceptance of program targeting choices. For example, it is the inclusion of people who shouldn’t benefit from the program that creates the greatest grievance or the exclusion of those who should be included? 87. Many of the programs drawn on in the case studies here emphasize the importance of youth employment as part of a stabilization or conflict-recovery strategy. Indeed, in making recommendations for future policy and programming, the World Development Report 2011 has the same emphasis: “National community-based public works programs should receive greater and longer-term support in fragile situations, in recognition of the time required for the private sector to absorb youth unemployment� (World Bank, 2011, p. 30). From the projects assessed here, however, there is limited evidence found in the ICRs of public works programs or evaluations of the case study projects to support this emphasis or demonstrate its effectiveness. 88. So why is the emphasis on generating work for youth so pervasive? The evidence trail (described in Chris Cramer’s background paper for the 2011 World Development Report (World Bank 2011)), tends to lead back to modelling by Collier and Hoeffler (1998) that has been highly influential but widely critiqued or contradicted by compelling empirical case studies and experimental work (Berman et al 2009, Blattman and Annan 2014). Slater and Mallett (Secure Livelihoods Research Consortium (SLRC), 2017) argue that in conflict and post-conflict settings, employment opportunities are often concentrated among the few� while an outcast majority finds itself in poorly paid, exploitative and precarious forms of work. As a result, labor markets become sites not where peace dividends are manifested, but where various forms of poverty and violence are maintained and, subsequently, where grievances emerge� (p. 7). Whilst the concern that unemployed youth may be drawn (back) into violent conflict drives much of the urgency of both economic development and public work programs, the SLRC suggests that �it is not necessarily a lack of employment that is most likely to lead to the expression of grievances, but rather the way in which people are incorporated into markets on adverse terms� (p. 7). 89. What are then the implications for public works programs in South Sudan? First, as Cramer (2015, p. 3) notes, the empirical evidence that exists “points us to labor relations and institutions rather than only to (often badly measured) levels of employment/unemployment.� Second, it suggests that PWs that are widely viewed as fair are required. In turn, this suggests that using lotteries for targeting beneficiaries (as seen in CAR with LONDO – Error! Reference source not found.) or providing support on rotation (as preferred by communities in Sierra Leone under the 2008 crisis response cash for work project) might work better than targeting systems that are subject to elite capture or patronage systems. 31 90. The SLRC goes on to flag the risks of raising expectations in conflict and post-conflict settings, and suggests that raising, and then failing to meet community and individual expectations may be the greatest threat of all to peace and stability. Indeed, as Andrews et al (2012) note in relation to the World Bank supported CFWs in Sierra Leone: “On the negative side, the program raised expectations among youths for long term employment opportunities which it was not meant to deliver, and these expectations will need to be managed.� Whilst in Sierra Leone there has been a relatively stable peace, in South Sudan this risk should be taken very seriously in the design of the next phase of public works programming. 91. The narrative about the importance of addressing unemployment among aggrieved youth is strong in South Sudan and influences the model and targeting for public works. In the SNSDP PAD, there is a substantial stress on this: South Sudan is a young country. Half of the population (51 percent) is under the age of eighteen and 72 percent under the age of thirty. Substantial numbers of young South Sudanese declare themselves as not working and looking for a job. Unemployment is highest among the youth, with around one fifth of those under 24 recorded as unemployed. Aside from being a cost to society and an individual welfare loss, youth unemployment, particularly among men, is recognized as a potential trigger for social instability. …. An inclusive and transparent community-based process with integration of Youth in the identification of beneficiaries, and prioritization and implementation of public work and skills development activities will contribute to strengthening the social fabric, promoting social cohesion and contributing to overall societal stability. This is of particular importance in South Sudan’s fragile and post conflict setting where “disengaged youth bear high short-term risks of social unrest and longer-term challenges to a nation’s economic development� (Andrews and Kryeziu, 2013). ( World Bank 2013, p. 2 and p. 3-4) Implications for targeting in South Sudan 92. The current context in South Sudan has most in common with the DRC and CAR, and the experiences of targeting there are likely the most transferrable to the South Sudan situation. At the local level, this means choosing between community-based targeting (where there is capacity in the implementing partner) or lotteries where there is not. The risk with community-based targeting is that, even where there are externally-imposed quotas on the numbers of beneficiaries, it often results in dilution of benefits as community members choose to spread working days among more households (the ‘we are all poor here’ effect) and in turn reduces the effectiveness of the support to households. Lotteries, on the other hand, whilst generally seen as giving all a fair chance, require good orientation and 32 communication systems at the local level to ensure that everyone understands them. They can also be weak at incorporating women if there are barriers to women entering the lottery in the first place and may require quotas for women’s engagement. 93. On paper, there are strong steps towards inclusiveness of those less able to work in the SSNDP. For example, the elderly is included as child care providers for the children of other beneficiaries working on public works tasks or are given light work such as fetching tools or water. Pregnant women work for up to four months into their pregnancy and then, if they have days outstanding, receive wages for the remaining days without participation on public work activities. Nursing mothers return to work four months after giving birth. 94. It is not altogether certain that these design features have translated into practice fully. In the small number of interviews carried out in beneficiary communities it was stated that those viewed as less able to do physically demanding work – for example the elderly and pregnant women – were not sufficiently included. Adjustments made over time were recognized, such as elderly women taking care of children in a shed provided by the AAH-I, and pregnant women doing lighter work. However, poor and vulnerable households with no able-bodied adults to participate on any form of public work were not included to benefit from the program. Beyond these examples, beneficiary interviews highlighted some incidents where women’s participation in the program resulted in gender-based disputes. For example, some men demanded for money received by women, or wanted to do the work themselves instead of the women or felt undermined. There were reports that this had triggered domestic violence, although few cases were reported to or solved by the Appeals Committee. And while these groups receive some complaints and seek to intervene by sensitizing and explaining that the income is meant for the whole family, it is probable that many gender-based conflicts go unreported. 95. The evidence reported here is based on a very limited number of interviews, so it is not clear whether it indicates widespread problems nor whether recent adjustments to the program have solved them. However, it is worthwhile considering the efforts in other countries to provide sensitization at community level to reduce the incidence of gender-based violence triggered by public works targeting. Wages 96. In many of the programs assessed here, wage rates were set to be aligned with the cash value transferred by WFP as the biggest cash transfer organization and pegged as the value of half of the food basket. In contexts where there is limited capacity and security is too volatile to allow an extended presence on the ground that is required for a more robust targeting process, self-targeting is an attractive 33 option. The principle being that, if wage rates are set below other labor opportunities, only the neediest will seek public works employment. It has its disadvantages too though. In many of the countries included here as case studies, poverty, malnutrition and vulnerability are generalized across the population. For example, in Sierra Leone, four out of five income/consumption quintiles were classified as poor at the time of the CFW response in 2009. There, lowering wage rates to self-target was not effective, the program was still oversubscribed, even though in some areas the wage rates were significantly under the prevailing wage rate (in Freetown Western Area, it was 50% less than the market rate) (Andrews et al 2012, p. 8). 97. In 2009, Freetown and much of the rest of Sierra Leone were very stable for a post-conflict country. But what about more volatile situations? There, self-targeting could be risky and counter- productive for a number of reasons. First, if food in local markets is not readily available and (as is often the cases in markets in fragile and conflict-affected settings) is priced high, it is likely that the calorific value of a low wage rate will undermine the safety net/consumption smoothing function of the public works program. Second, paying low wage rates can potentially lead to what Mallett et al (2016) have called ‘precarious work’ – that which is characterized by insecurity, bad pay and forms of exploitation. The SLRC has argued that it is bad jobs, rather than jobs per se, that can trigger grievances and violence against the state and other social groups. It therefore makes little sense for public works programs to join a ‘race to the bottom’. 98. Second, in terms of benefit levels, it is not clear that self-targeting based on low wages for public works would address oversubscription for SNSDP given the high levels of demand even in Juba where one has more chance of finding alternative work than any other place in the country. The poor performance of self-targeting in Sierra Leone suggests that where poverty and malnutrition levels are high, it may be ineffective. Implications for wage rates in South Sudan 99. At present, beneficiaries in the SSNDP are paid the equivalent of US$3 per day; work days are from Monday to Friday and Saturday in some cases; work is for 4-5 hours a day for 15 days a month, and so in a month beneficiary get equivalent of US$45. The extent to which this transfer value can provide an adequate safety net is debatable. Given rampant inflation in South Sudan, particularly Juba, it is worth considering an assessment of whether the calorific value of the wage is equal to the calories required by beneficiary households, and the calories burned in the work activities. In practice, a program that provides fewer calories that those expended in carrying out the work will not just be mildly counterproductive but damaging for participating beneficiaries. 34 100. The evidence from other countries also suggests that the economic rationale for setting wage rates (for example, offering wage rates below the prevailing casual labor rates in order to self-target and avoiding distortion of the local wage labor markets) may, in some case, have to give way to rationales based on principles of (perceived) fairness and equity. This is particularly the case in South Sudan where the demand for work far exceeds supply and labor markets clear at very low wage rates. 101. Finally, it is important to recognize in the design that there are risks associated with pushing down wages – this could undermine any fragile détente between different groups in the community, and between beneficiaries and local providers of the program. Public works activities / subproject selection 102. In the South Sudan context, public works subproject identification, prioritization and selection is largely the mandate of beneficiary communities, however there are some useful lessons from elsewhere. 103. The first lesson is about finding an appropriate level of ambition in sub-projects, and an appropriate balance between capital and laborintensive activities. In relation to asset creation, the ambition in the approaches of CAR and DRC are rather different. CAR maintains a very simple approach, arguing that ‘the execution of labor intensive public works (LIPW) in fragile and conflict-affected situations calls for a simple design, i.e. one type of public works’. The PAD goes on to stress that temporary employment programs are not meant to create sustainable assets that support livelihoods but are to be used as a stabilization mechanism. Spillover effects – such as assets in support of future livelihoods activities are not the first priority and the PAD is modest about those effects. “Tools and equipment (signage, wheelbarrows, etc.) will be transferred to local authorities (mairies 24) which will also benefit from the technical assistance; protective equipment (gloves, boots, etc.) is kept by beneficiaries at the end of the assignment, including the bicycles provided to workers as a means of transportation to work sites.� In DRC, the ambition is greater – with a broader range of works activities and greater expectations over the extent to which short-term employment can provide beneficiaries with the technical and soft skills to make them more employable in the future, create assets that will support future livelihoods, and contribute to social cohesion and peacebuilding. The extent to which these ambitions are achieved remains to be seen. 104. The second lesson is about the level of autonomy of local communities in selecting sub-projects. Here there are numerous trade-offs: the ambition of local communities may need to be matched with the resources available, or with the level of skills required for the creation of specific assets, or the speed at 24 Municipality 35 which work must be delivered with the capacity and needs of communities. In some contexts, communities submit requests to receive program funds for subprojects. This depends on high levels of capacity at community level, and good communication about programs. 105. It is not coincidental that Sierra Leone took this approach only in 2010 – some eight years following the end of its civil conflict. The approach also risks selecting on the basis of how good the public works proposals are, rather than the needs of the community – “Given the emergency nature of the first phase of the program, this implied that about 70% of communities were selected based on an ability to quickly implement ‘off the shelf’ road rehabilitation / maintenance projects in localities. The fact that the selection of the communities was done in such a manner and not based primarily on poverty/vulnerability criteria undoubtedly weakened the ability of the program to reach out to the poorest� (Andrews et al 2012, p.10). 106. A further tradeoff is the balance between “community� decision-making and the ensuring that the needs and capabilities of all groups in the community are considered. Different groups within communities have varying needs for asset creation and varying capacities to work but all too often ‘community’ decision-making translates into able-bodied men choosing projects and the exclusion of the elderly, pregnant women and others less about to work. Revisions to the SNSDP to ensure less demanding labor opportunities are available but there is further work to do in this regard. Gender Issues 107. The aim of this section is to document lessons about ensuring equity in public works programs (not solely for women and men, but also for other groups that might be marginalized, excluded or exploited). Because the majority of the reviewed documents tell us about design and outputs, there is rather less to say about how effective some of the design features have been in ensuring equity. 108. Women’s inclusion in public works programs. Most programs recognize the specific needs of women seeking to participate in public works activities. In DRC, the ERP notes the importance of encouraging the active participation of women and youth “both in decision making and in benefit sharing, including through the use of quotas when feasible (e.g. in village development committees and in LIPW) as well as in the design of employment-related activities (e.g. schedules mindful of women’s household obligations).� 109. All countries operate quotas of one kind or another, though quotas on participation in works predominate, and key performance indicators (KPI) always include a target for women’s participation. For example, for LONDO in CAR, the first KPI is the ‘number of direct beneficiaries, percentage of which are women,’ and the project seeks to incrementally attract women and reach 50% participation by the third 36 year of the project. LONDO also aims to ensure that the type of works involved is ‘not discriminatory’ – in contrast to Sierra Leone where evaluations and assessments suggest that the dominant focus on road maintenance and construction tended to exclude women – partly because it was traditionally seen as men’s work and partly because it required hard physical labor. In Sierra Leone, the response was to attempt to switch some work to activities viewed as more appropriate for women, including agricultural labor (Andrews et al 2012). In Liberia, the implementing organization, the Liberia Agency for Community Empowerment (LACE), was already familiar with working to a women/men ratio so had experience of making provision for women’s involvement. 110. The prioritization of women’s participation in public works is clearly a common practice, even in challenging environments. However, some substantial knowledge gaps remain. There is little evidence from program documentation of the impacts of men’s and women’s participation in public works, especially the impact of intra-household relations. Apart from some communications materials that highlight that sexual assault or harassment of women working on public works is forbidden (see Error! Reference source not found. for an example from LONDO in CAR), there is little evidence on the extent to which women are kept safe from exploitation and violence at work sites. Furthermore, the evidence and documents reviewed provide rather little in the way of: i) analysis of whether the prevalence of insecurity, fragility and violent conflict change the ways in which gender equity in participation and benefit sharing might be achieved; and ii) assessment of how successful attempts to enhance participation have been. Indeed, it may be the case that there is more to learn from exploring what is known about changing gender relations in South Sudan rather than seeking answers from elsewhere because contexts prove so important in these sorts of analyses (for example see Bubenzer and Stern, 2011). Figure 2 Cartoon about acceptable behavior for LONDO participants 37 111. There is a limited literature on gender-based violence and public works programs from further afield. Amaral et al (2015) find evidence from India’s MGNREGA that an increase in labor participation by women increased gender-based violence – including kidnappings, sexual harassment and domestic violence. Surprisingly, they find a reduction in dowry deaths which, although South Sudan’s bride wealth traditions are very different to India’s, is of interest to South Sudan given the extent of violent cattle raiding in rural areas to support the payment of bride wealth. From the wider literature on gender-based violence in situations of violent conflict the message is that the received wisdom and the widely shared 38 narratives on sexual violence and gender-based violence tend to greater simplify the drivers, the prevalence and the impacts of violence. For examples from DRC, see Hilhorst and Douma (2018) and Heaton (2014). Overall, the lesson from the broader literature on gender, violence and donor programming requires a far more sophisticated problem and risk analysis and a more nuanced approach to program design. Implications for gender-sensitive public works in South Sudan 112. Getting gender right in public works programs in South Sudan clearly matters. As the SNSDP PAD notes: Gender disparities [in South Sudan] are dramatic. Female headed households (28.6 percent) are among the poorest, with 56.9 percent below the poverty line compared with 48.1 percent of male headed households. War-induced poverty, displacement and trauma have weakened kinship and community ties and negatively affected social support mechanisms and intra-communal collaboration. There is thus a strong imperative that community mobilization and decision-making processes engender principles of participation, inclusion, transparency and accountability; to prevent the triggering of conflict stresses with roots in diverse historical, cultural and ethnic characteristics (World Bank 2013, p. 2). 113. Interviews in Juba suggest that SNSDP activities may, inadvertently, be contributing to exclusion, gender-based disputes and violence. It is important to pay more attention to support elderly and pregnant women, or to nursing mothers to identify lighter work options or different activities that allow women to combine public works with their reproductive/domestic roles. It appears that more sensitization and orientation work is required to ensure that beneficiaries view income from public works as household, rather than individual, income and adjust the household wage system accordingly (Slater and Mphale 2008). While the transition of household wage systems is a long-term process, it is clear that a better understanding of the threats to women’s safety and security due to their participation in SNSDP public works requires more attention. Reaching remote areas 114. The expansion of public works from urban to rural areas presents a range of challenges – some of which are directly related to the extent of violent conflict and instability, and others which are not but are equally important. First, seasonality is a noted challenge for a number of case study projects, for example, in the DRC ERP where it is noted that “in rural areas, LIPWs would be primarily implemented during the 39 agricultural off-season, to avoid interfering with agricultural employment/livelihoods.� In South Sudan, there is also the challenge of what public works subprojects is possible to implement – particularly during the rainy season when flooding and inaccessibility limit what is achievable. Second, lessons from Uganda highlight that the expansion into rural areas creates an additional layer of challenges for programming in post-conflict and unstable situations. McCord et al (2013) note that “expansion will require a rethink of any focus on asset building and supporting livelihoods, and climate will present a further element on top of the challenges associated with conflict and fragility.� Third, these challenges become particularly acute in conflict-situations where permanence of presence is important and yet the reach of the state apparatus might be constrained is how to maintain a consistent program presence throughout the year. Evidence from the SLRC notes the importance of sustaining support to remote communities in conflict-affected areas and seasonal interruptions are unlikely to help keep the state or project present and visible. 115. In CAR, public works are heavily focused on the capital Bangui and yet three quarters of CAR’s population lives outside Bangui and beyond the capital needs are greatest with the extreme poverty rate averaging 70 percent. The PAD for LONDO notes that investment beyond Bangui is required for two key reasons. First, a number of other rural development projects closed at the time of inception of LONDO and World Bank staff and government officials alike were concerned about the risk of ‘not passing the baton’ from one project to another and losing the technical expertise, operational expertise, resource network and reputation that project partners had established in CAR’s hinterlands. Second, there is a need identified to treat different geographical areas (many of which have been neglected in terms of public spending for a number of years due to conflict) and different groups of people equally and fairly. The PAD notes the need to ‘show that all citizens, in any region, and equal and fairly treated’. There is much in common in this regard with South Sudan. 116. A core element in CAR that allows the expansion into rural areas is the presence of AGETIP-CAF – a quasi-parastatal agency that has a longstanding presence throughout the country and is working through other donors (for example the Agence Française de Développement (AFD) is using the AGETIP- CAF to expand its professional training project outside Bangui). In the case of Yemen, the importance of an implementing partner with a wide geographical reach is also noted. There, the ECRP works through the Yemen Social Fund for Development (SFD)’s operational apparatus allowing it to deliver in a broad range of geographical locations including those controlled by Houthi groups. 117. Lessons from DRC highlight the importance of allowing some flexibility in sub-project budget levels for remote locations. “Construction costs vary within different parts of a country based on a number of factors, including distance from main markets, security, road quality, cost of labor, and supply of skilled 40 labor. For this reason, budget envelopes may need to be adjusted by province to reflect these variations; otherwise, sub-projects may need to be scaled down to fit within the budget in remote, underserved areas that may have the greatest need� (World Bank 2014). The PAD also notes, however, the challenges associated with delivering funding to the local/rural level in DRC – despite the commitment to a decentralization agenda based on the 2005 Constitution that would transfer 40 percent of national revenues to the provinces, and World Bank engagement in supporting provincial level public financial management, there are insufficient checks and balances on provincial authorities. This undermines the required flexibility to deliver in remote, rural areas. 118. In CAR, the required flexibility is achieved, in part, through the use of a ‘force account’. This is a rather unusual delivery mechanism for IDA, but it appears to work well in contexts where the work and outputs are difficult to plan in advance (Box 4). Box 4 Using a Force Account to delivery public works in the Central African Republic A unique delivery framework has been developed and tested, under which AGETIP-CAF executes road works itself rather than procuring them to third parties (private companies or non-governmental organizations (NGOs)). Force account (régie) is justified by the following circumstances: (a) the objective of the project does not define quantities of works to be included in a contract, i.e. LONDO is a social–not an infrastructure–project; (b) the works are small and scattered, and in remote or unsecured locations, for which qualified third-parties are unlikely to bid at reasonable prices, if at all; and (c) the risks of unavoidable work interruption, for instance because of insecurity, are better borne by the Government than by a contractor. However, in specific areas, depending on local capacity, access, security, and other operating criteria, AGETIP-CAF will retain the possibility to procure parts of works or related services to third-parties. Source World Bank (2015) Implications for extending public works in South Sudan 119. Deciding on the balance of investments between urban and rural in South Sudan is difficult. On the one hand, urban poverty is rising quickly and becoming a bigger challenge because of the lack of alternative coping strategies in urban areas where people are less able to fall back on their land and livestock. On the other hand, the slide back into conflict at scale in South Sudan is underpinned by disputes about the resourcing of different geographical areas (that have different political alignment). This intensified the need for SNSDP to work in more areas than originally envisaged. As in CAR, the dynamics of conflict are underpinned by grievances that relate strongly to ethnicity and geography, and this requires public works to cover a broad geographical area. 41 120. Initially, SNSDP was proposed to operate in four states. Three states would cover one urban center and one rural county while Juba City would make up the final site, totaling in 7 locations. The locations still covered only a limited share of South Sudan’s ethnic groups and states: Bor and Ayod (later changed to Pibor due to accessibility issues) Counties in Jonglei state, Gogrial West, and Tonj South Counties in Warrap state, Torit and Kapoeta East Counties in Eastern Equatoria state, and Juba city in Central Equatoria state. The acceptability of the program in the future will depend on the distribution of resources being shared across a broad range of locations and groups. Conflict-sensitivity 121. There are a number of lessons, with varying levels of transferability to the South Sudan context, about how to sustain and maintain program delivery when countries (back)slide into conflict or when insecurity deepens. This section explores these lessons with a focus on three core questions:  How do we ensure our design and implementation systems are conflict-sensitive and can be sustained in contexts where the WB offices evacuate?  What sorts of risk management and contingency mechanisms will allow programs to continue running on the ground (planning, financing, fiduciary, etc.)?  What can we realistically aim to leave running on the ground? What needs to be in place? 122. As previously noted, a key element of conflict-sensitive program design in CAR relates to recognizing where geographical location and membership of specific ethnic, linguistic, religious or other groups are the source of grievances that may result in a return to violent conflict. The work of the SLRC notes that, in relation to the impact of service delivery on people’s perceptions and grievances, ‘it ain’t what you do, it’s the way that you do it’. For public works programs, especially where there are geography and ethnicity-based grievances, this means that the way in which public works are delivered (for example the use of sensitization about targeting eligibility, consultation about the works to be carried out, and participation in decisions about community rankings to identify beneficiaries) can make a big difference for the acceptability of the program – and for its sustainability should program staff have to evacuate. 123. The institutional arrangements also matter if programs are to be able to continue running on the ground in the absence of program staff. In CAR, a community maintenance system has been established to sustainability. “Local authorities … have already provided office and storage space for AGETIP-CAF in each sous-préfecture. This subcomponent will sponsor the development of a community maintenance system (cantonnage communal) to promote the sustainability of the investment, in terms of both jobs 42 and roads. Upon project completion in each sous-préfecture, the project’s tools will be transferred by AGETIPCAF to local authorities (mairies), which can use them for future maintenance activities and other public works projects. In addition, AGETIP-CAF will provide technical assistance.� However, as the CAR LONDO PAD notes, there is still much that we do not know about ensuring that program activities can be maintained: “A study of experiences in CAR and other countries will be financed, when the project starts, to design the most appropriate approach to community maintenance in the very difficult circumstances of today’s CAR, and test it.� 124. In circumstances where World Bank offices are forced to evacuate, there are often also concurrent challenges of working with governments – either because of capacity constraints, or because in conflict-situations donor agencies are no longer able to work with institutions that are party to violent conflict. 125. In Uganda, where the situation is far more stable than South Sudan, steps have been taken to mainstream NUSAF activities into existing structures at the national and local government levels. NUSAF3 is implemented by the Office of the Prime Minister (OPM) with a technical support team (TST) headed by a director and staff by relevant technical experts. The Uganda experience may be something for SNSDP to aim for in the long term – there are few contingencies in place should government capacity to operate, particularly in remote areas, be rapidly undermined by conflict. But its early experiences with NUSAF are insightful – in his assessment of the institutional arrangement for NUSAF1, Robinson (2005) confirms the importance of working at a local level, through local governments, NGOs and community organizations in response to problems of endemic poverty, weak capacity and the legacies of violent conflict. 126. In CAR and DRC (and previously in Sierra Leone and Liberia), public works programs have operated through agencies – often parastatals – with varying levels of autonomy from the state. This system allows a greater compliance with financial and fiduciary risk management procedures in the World Bank – particularly by routing finances directly through private bank accounts to implementing agencies rather than through government (Error! Reference source not found.). In CAR, there was greater confidence in AGETIP-CAF meeting the financial management requirements – including should the direct contact between AGETIP-CAF and the World Bank be remote and virtual. In Liberia in the years following the peace agreement, the World Bank worked with the Liberia Community Empowerment Agency (LACE), an agency described as “a not-for-profit that operates like a private sector entity’ and that is ‘autonomous but accountable to the President.� 43 Figure 3 Routing of World Bank funding for the LONDO Project in the Central African Republic 127. In all the PADs for operations in situations of violent conflict, substantial attention is paid to financial management and risk mitigation. Yemen provides a particularly informative example, given that since March 2015 the World Bank has been operating under OP 7.30 – the World Bank policy relating to Dealing with de facto Governments – and suspended all disbursements and missions to the Republic of Yemen and evacuated all World Bank staff. To maintain operations, the World Bank has worked through the United Nations Development Programme (UNDP) – which oversees project implementation on behalf of the World Bank on the ground and transfers funds received from the World Bank to the executors – the SFD and the public works program. Al-Iryani et al (2015) note that the SFD “has been performing effectively in spite of a particularly hard context due to its collaborative relationships with local communities, its perceived political neutrality, its flexible mode of project funding and operation, and the importance of its interventions to beneficiaries thus enlisting their support and protection� (p. 334). Key for the World Bank is that the SFD is a public institution, with some autonomy from government, with 44 which the World Bank has worked for more than two decades, primarily delivering social protection interventions. 128. In CAR, the implementation support system is also detailed and includes strategies for managing the suspension of in-country missions, including reverse supervision missions in Yaounde, Cameroon, third-party monitoring, and substantial capacity development in preparation for these eventualities. The DRC PAD also notes that engaging with government is key to ensuring sustainability. The importance of investing in relationships with government that, in turn, support capacity in the organizations that can be trusted to deliver programs should conflict worsen or re-emerge is highlighted. A final key element of the DRC project is the direct linkage with work on conflict mitigation: “NGOs specialized in conflict prevention and management will be contracted to train key stakeholders in conflict assessment and management. In areas where existing conflicts may impact ability to deliver the sub-projects proposed (for instance areas recently liberated from rebel control, or where interethnic tension is high), local capacity to prevent and manage conflicts will be strengthened through targeted activities such as in-depth conflict analysis and participatory conflict transformation plans.� Implications for conflict-sensitive public works in South Sudan 129. The current conditions in South Sudan have much in common with CAR, DRC, and Yemen. Notwithstanding the substantial investments in the PIU and progress made there, the capacity of institutional partners, especially government, is weak; the security situation remains highly volatile; and there is highly constrained access to beneficiary communities outside of Juba because of geography, a poor infrastructure network, and ongoing violence. A further challenge for South Sudan is that it does not have a functioning local government, so following a Uganda NUSAF model would be very difficult. 130. Considering the advantages and disadvantages of the preferred solutions to these challenges in other countries, suggests that in the near future, it will likely be necessary to work in ways that allow greater flexibility (for example by continue working with a range of government and non-government (UN and NGO) partners as coordinating and implementing partners); further strengthening innovative systems put in place to manage the flows of finance and outputs; and building capacity in advance to deal with contingencies. Some of these elements can be costly (for example the overheads charged by UN agencies), but there are examples of negotiating these costs down (from Yemen). In any case, in the absence of existing local agencies, there are few alternatives. 131. Another key lesson is that long term investments in the capacity of parastatal agencies that work with some autonomy from the government is critical. Whilst this does not exist at present, taking steps towards this will be a key part of a longer-term strategy to make programming more conflict-sensitive. 45 This requires investments in institutional strengthening as an outcome of the SNSDP but with a longer timeframe for achieving results than the SNSDP itself. 132. Finally, the lessons from other countries highlight the importance of recognizing that this is a longer-term activity that make take some time to pay off. In the meantime, it requires more modest expectations of short-term outcomes and impacts of a PW program. Summary and Conclusions 133. A review of official and publicly available documents is not able to capture the nuances of programmatic decisions made in specific contexts, but nonetheless can offer some lessons for public works programming in South Sudan. Many of the lessons have been explored in each section above, but overarching lessons are also important. 134. Contexts such as South Sudan require us to recognise and grapple with trade-offs. As the DFID Building Stability Framework and the WB/UN ‘Pathways to Peace’ Report show, in fragile and conflict- affected situations it is not the case that ‘all good things go together’. Rather contradictions in programming options will be prevalent and navigating trade-offs is inevitable. Ranking priorities – for example a focus for public works programs in South Sudan on Models 1 and 2 (consumption and seasonality) – can take implementers some, albeit not all, of the way to better management of these trade-offs. 135. Second, the contrasting ambitions of different programs suggest that a tier of objectives and priorities might be most appropriate for PW programs in South Sudan. A first priority might be relatively modest (but still difficult to achieve given capacity constraints, accessibility of remote rural areas and continuing violent conflict) and be focused on regular and adequate income generation from PW, whilst further activities might focus on building assets that contribute in meaningful ways to improving livelihoods in both rural and urban settings, for example through building capacity through skills training, and linking public works to wider investments in infrastructure, public goods, climate change response, and rural development. 136. A ranked set of priorities should help to prioritize between trade-offs – just as was done when the skills development element of SSNDP was removed in 2017. And to build a ‘no regrets’ design in which broader objectives become part of the design but are quickly paused during difficult moments in order to focus on the most pressing priorities – ensuring that households can meet their basic consumption needs. This modesty is seen in other countries – for example in the modest level of ambition of asset creation (e.g. programs that have workers clean and maintain roads rather than building them) and also meets 46 sustainability principles: instead of building new roads that rapidly fall into disrepair and become unusable; the focus is on the maintenance and sustainability of existing assets. 137. Although using PW for achieving peacebuilding and state building objectives – especially if run through a government-affiliated agency/PIU – shouldn’t be at the top of the list right now, it is important to still recognise the potential impacts of PW interventions on state-society relations, if designed and implemented well. Thus, even if tackling conflict and insecurity aren’t at top of the list, there is still a need to follow the principles of ‘do no harm’. 138. Finally, it is recommended that the following topics for future similar studies: (a) link between public works and basic service delivery such as education and health particularly the impacts of support through the public work on beneficiary’s investment on human capital development. While the cash transfer programs usually attach soft or strict conditionality such as sending children to school, immunization, attending training to improve behaviors or to aquire skills, the public work safety net programs do not have such conditionalities. It is important to investigate the impact of public work programs impact on beneficiary’s human capital development; and (c) in-depth study on the GBV risks associated with public work intervention and mitigation measures is. Public work programs are providing income earning opportunities to women which in turn empowers them in many ways. However, it also exposes women to risk of gender-based violence at home as well as at the public work sites. It is very important to conduct an in-depth study to understand the root causes of the vulnerability and risks and mitigation measures. 47 Bibliography Al-Ahmadi, A., and S. de Silva. 2018. “Delivering Social Protection in the Midst of Conflict and Crisis: The Yemen Emergency Crisis Response Project.� Beegle, K., A. Coudouel, and E. Monsalve. 2018. Realizing the Full Potential of Social Safety Nets in Africa. Africa Development Forum; Washington, DC: World Bank. https://openknowledge.worldbank.org/handle/10986/29789. Dihel, Nora Carina, and Utz Johann Pape. 2017. “South Sudan - Economic Update: Taming the Tides of High Inflation - Policy Options (English).� Washington, DC: World Bank Group. http://documents.worldbank.org/curated/en/806291508505062484/South-Sudan-economic- update-taming-the-tides-of-high-inflation-policy-options. Gisselquist, R. (2015) ‘Good Aid in Hard Places: Learning from ‘Successful’ Interventions in Fragile Situations’, International Peacekeeping, 22(4): 283-301, https://www.tandfonline.com/doi/pdf/10.1080/13533312.2015.1059732 World Bank. 2002. Project Appraisal Document on a Proposed Credit in the Amount of SDR 80.1 Million (US$100 million equivalent) to the Republic of Uganda for a Northern Uganda Social Action Fund . Report No 23885-UG, World Bank, Washington DC. ———. 2009. Implementation Completion and Results Report (IDA-36970 IDA-3697A IDA-46260 TF-20972 TF-50522) Northern Uganda Social Action Fund. Report No: ICR00001211, World Bank, Washington DC. ———. 2010. Project Appraisal Document on a Proposed Grant to the Republic of Yemen for a Social Fund for Development Phase IV Project. Report No. 52365 YE, World Bank, Washington DC. ———. 2011. World Development Report 2011: Conflict, Security, and Development. World Bank, Washington, DC. https://openknowledge.worldbank.org/handle/10986/4389. ———. 2013a. Project Appraisal Document (PAD) - Safety Net and Skills Development Project, South Sudan. Report No: 75941-SS, World Bank, Washington, DC. ———. 2013b. Project Information Document (PID) Concept Stage - South Sudan Social Safety Net and Skills Development (P143915). Report No.: PIDC710, World Bank, Washington DC. ———. 2014. Project Appraisal Document - Democratic Republic of Congo Eastern Recovery Project (P145196). Report No 83571-ZR, World Bank, Washington, DC. ———. 2015a. Project Appraisal Document - LONDO (“Stand-up�) Project, Central African Republic . Report No: PAD1469, World Bank, Washington, DC. ———. 2015b. Project Appraisal Document - Third Northern Uganda Social Action Fund Project (NUSAF 3). Report No: PAD1188, World Bank, Washington, DC. 48 ———. 2015c. South Sudan - Systematic Country Diagnostic (English). Washington, DC: World Bank Group. http://documents.worldbank.org/curated/en/952191472743836373/South-Sudan- Systematic-country-diagnostic. ———. 2017. South Sudan - Country Engagement Note (English). Washington, DC: World Bank Group. http://documents.worldbank.org/curated/en/112071516734556800/South-Sudan-Country- Engagement-Note. ———. 2018a. “Feature Story: Once a Temporary Stop Gap, Public Works Project Now Provides People Income Year-long.� http://www.worldbank.org/en/news/feature/2018/02/06/once-a- temporary-stop-gap-public-works-project-now-provides-people-income-year-long. ———. 2018b. The State of Social Safety Nets 2018. World Bank, Washington DC. ———. 2018c. Project Appraisal Document on a Proposed Grant in the Amount of (Us$ 40 Million Equivalent) to The United Nations Office for Project Services (UNOPS) for the Benefit of The Republic of South Sudan for a South Sudan Safety Net Project (SSSNP). Report No PAD3194, World Bank, Washington DC. 49 Annexes Table A.1: Countries, Public Works Case Studies, and Sources of Information Country Projects Sources Central African LONDO (‘Stand Up!’) World Bank official project Republic documentation (Project Information Document [PIDs], PAD, mission Aide Memoire). Democratic Eastern Recovery Project (ERP) (and by World Bank official project Republic of extension the FSRDC) documentation (PIDs, PAD, Congo Implementation Status and Results Report [ISRRs]); PIU correspondence, mission Aide Memoire. Uganda Northern Uganda Social Action Fund (Phases 1, World Bank official project 2, and 3), documentation (PIDs, PAD, ISRRs, Implementation Completion and Results Reports [ICRs]), mission Aide Memoire. Yemen Emergency Crisis Response Project (ECRP) (and World Bank official project by extension the implementing partner for SFD) documentation (PIDs, PAD, ISRRs, ICRs) - Al-Ahmadi and de Silva (2018). 50 Table A.2: Project Development Objective-level Results Indicators by Project and Country Country Indicators Uganda -NUSAF  Number of household beneficiaries of the project  Number of female beneficiaries of the project  Number of households benefiting from post-disaster activities  Percentage of project beneficiaries satisfied with project intervention Yemen -  The number of direct beneficiaries of wage ECRP employment (disaggregated by gender, youth, and IDPs)—to measure coverage of income support  The number of people provided with access to key services (disaggregated by gender)—to measure potential coverage of community and social services  The percentage of core staff positions of the SFD and PWP retained—to measure preservation of the capacity of these critical national programs South Sudan -  Number of direct project beneficiaries, of which SNSDP female (%) (core indicators)  Social Protection Policy developed and submitted for discussion  Beneficiaries of public works complying with eligibility criteria  Targeting mechanism to select poor youth for skills development training designed and tested Central African Republic -  Number of direct beneficiaries, percentage of which LONDO are women  Number of worked days  Number of benefitting districts ( sous-préfectures) Democratic Republic of Congo -  Direct project beneficiaries (number), of which ERP female (percentage)  Average percentage increase in access to improved community social and economic infrastructure  Average percentage increase in annual revenues among beneficiaries of the livelihood support subcomponent (by sex)  Percentage of beneficiaries of the livelihood support subcomponent whose revenue increased (by sex) Source: Project Appraisal Documents of the different projects 51