Working paper  Linking humanitarian assistance and social protection in response to forced displacement  Developing an analytical framework Christina LoweID, Cécile CherrierID and Rebecca HolmesID November 2022 Readers are encouraged to reproduce material for their own publications, as long as they are not being sold commercially. ODI requests due acknowledgement and a copy of the publication. For online use, we ask readers to link to the original resource on the ODI website. The views presented in this paper are those of the author(s) and do not necessarily represent the views of ODI or our partners. This work is licensed under CC BY-NC-ND 4.0. How to cite: Lowe, C., Cherrier, C. and Holmes, R. (2022) Linking humanitarian assistance and social protection in response to forced displacement: developing an analytical framework. London: ODI. Acknowledgements We sincerely thank the entire project team (Jessica Hagen-Zanker, Simon Levine, Steve Commins, Nathalie Both, Heiner Salomon, Amanda Gray Meral, and our research partners at CASS-RT, Universidad de los Andes and EKKE) for their invaluable contributions to the research project on which this paper is based. We are also immensely grateful to Mattias Lundberg, Jo de Berry, Paolo Verme and Ugo Gentilini from the World Bank, Andrew Mitchell and Theresa Beltramo from UNHCR, Emily Henderson from the UK’s Foreign, Commonwealth and Development Office, Claudine Mensah Awute from CARE, Michael Köhler from the European Civil Protection and Humanitarian Aid Operations, Loren Landau from the University of the Witwatersrand and Francesca Bastagli from ODI, all of whom provided helpful comments on earlier versions of this framework. Thanks also to our project management and editorial colleagues who saw this paper through to publication, notably Roni Lee and Aaron Griffiths. Finally, we thank the many displaced people and host community members, as well as the representatives from government, local, national and international organisations, who gave up their time in challenging circumstances to share their reflections and experiences with us throughout this project. Without their generous participation, none of this research would have been possible. About this publication This paper is one of a series of outputs from a wider project exploring effective mechanisms for linking social protection programmes and humanitarian assistance in contexts of forced displacement. By providing guidance on when, how and why different linkages might be considered, the project aims to develop the theory, evidence base and operational guidance on how social protection systems and humanitarian systems can work together to meet the needs of those affected by displacement crises, including not only displaced people but vulnerable households in their host communities as well. The research is grounded in three country contexts that present different contexts of displacement and humanitarian response: Greece (Athens and Ioannina), Colombia (Bogotá and Cúcuta) and Cameroon (Far North and East). The project is led by ODI, who work in close collaboration with the Centre for Applied Social Sciences Research and Training (CASS- RT) in Cameroon, the School of Government at the University of Los Andes in Colombia and the National Centre for Social Research (EKKE) in Greece. This work is part of the programme ‘Building the Evidence on Protracted Forced Displacement: A Multi-Stakeholder Partnership’. The programme is funded by UK Aid from the United Kingdom’s Foreign, Commonwealth and Development Office (FCDO), it is managed by the World Bank Group (WBG) and was established in partnership with the United Nations High Commissioner for Refugees (UNHCR). The scope of the programme is to expand the global knowledge on forced displacement by funding quality research and disseminating results for the use of practitioners and policy-makers. This work does not necessarily reflect the views of FCDO, the WBG or UNHCR. About the authors Christy LoweID is a Senior Research Officer in ODI’s Equity and Social Policy programme. Cécile CherrierID is a Research Associate in ODI’s Equity and Social Policy programme. Rebecca HolmesID is a Research Associate in ODI’s Equity and Social Policy programme. Contents Acknowledgements / i Display items / iv Acronyms/Glossary / v 1 Introduction / 1 2 Initial analytical framework / 3 ‘Determining factors and processes’   /  5 ‘Typology of integration approaches’   /  9 ‘Benefits and drawbacks’   /  12 3 Refining the framework  / 14 Lessons on ‘determining factors and processes’ (now re-named ‘determining factors and actors’) / 15 Lessons on the ‘typology of integration approaches’ (now re-named the ‘typology of linkage approaches’) / 22 Lessons on the ‘benefits and drawbacks’ (now re-named the ‘contribution to benefits and drawbacks over time’)  /  29 4 The final framework / 36 References / 37 Display items Tables Table 1 Provisional list of determining factors and processes, and examples of how these might determine the adoption and impacts of different integration approaches in a displacement context / 6 Figures Figure 1 Initial analytical framework  /  4 Figure 2 Column 1 from initial analytical framework  /  5 Figure 3 Column 2 from initial analytical framework  /  9 Figure 4 Integration profile for a hypothetical programme  /  11 Figure 5 Column 3 from initial analytical framework  /  12 Figure 6 Initial and final column 1  /  15 Figure 7 Initial and final column 2  /  23 Figure 8 Initial and final column 3  /  29 Figure 9 Approaches for linking humanitarian assistance and social protection in displacement settings, at the programme level  /  36 Acronyms/Glossary EU European Union HELIOS Hellenic Integration Support for Beneficiaries of International Protection IDP internally displaced person NGO non-governmental organisation SISBÉN Identification System of Potential Beneficiaries of Social Programmes SP social protection SPACE Social Protection Approaches to COVID-19 UN United Nations 1 Introduction The number of people in the world who are forcibly displaced has more than doubled in the last decade, passing the 100 million mark in 2022 (UNHCR, 2022). This displacement is frequently long term, and rather than residing in designated camps, those affected typically live among host communities, who are often socio-economically vulnerable populations in low- and middle- income countries where rates of poverty, inequality and precarious employment tend to be high and growing. Both displaced and host populations are increasingly based in urban areas that appear to offer better prospects but are not always set up to accommodate large additional populations. These shifts in the nature of displacement have required those responding to revise their approach, and recent decades have seen a growing shift away from traditional ‘care and maintenance’ models of humanitarian assistance – based on the immediate relief of emergency needs – towards development-oriented solutions. In the search for new approaches, there has been growing interest in the role that national social protection systems can play in supporting those affected by forced displacement (OCHA, 2016; SPIAC-B, n.d.; UNHCR, 2019). This has created a need to better understand when and how humanitarian (NGO/internationally led) assistance can be linked with state-led social protection systems to support displacement-affected populations (Peterman et al., 2018). To start to fill this knowledge gap, the World Bank, United Nations High Commissioner on Refugees and UK Foreign, Commonwealth and Development Office’s ‘Building the Evidence on Protracted Forced Displacement’ programme commissioned an ODI-led consortium to undertake research on the optimal approaches for linking humanitarian assistance and social protection systems in different displacement contexts. One of the principal outputs of this project is the development of an analytical framework that can be used to explore the potential approaches for linking humanitarian assistance and social protection systems in a given displacement context. The initial version of this framework was developed in April 2020 during the inception phase of the project, based on a February–March 2020 review of the existing theories and frameworks for linking social protection and humanitarian assistance, particularly in contexts of forced displacement (Annex 1, available on ODI website). This initial framework was then tested and refined over the course of the project using new evidence generated through a desk-based review of linkage approaches, a global quantitative analysis of available data sets analysing displaced and host populations’ access to assistance across multiple countries, and primary mixed-methods research through in-depth case studies in three countries. The final framework was developed at the end of the project in June 2022. 2 ODI Working paper The purpose of this paper is three-fold: to present the initial analytical framework that guided the research project (Section 2); to present lessons learnt when applying the framework and explain changes that needed to be made (Section 3); and to present the refined analytical framework that can guide future research and practice (Section 4). The final framework is also available summarised as a standalone document (Lowe et al., 2022a). 3 ODI Working paper 2 Initial analytical framework Our March 2020 review of the literature (Annex 1, available on ODI’s website)1 identified various theories relating to humanitarian and social protection provision during crisis response but found that there was not yet a single framework for understanding when, how and with what outcomes humanitarian and social protection systems might link in response to the forced displacement of both refugees and IDPs. At the outset of our project in April 2020, we therefore developed an initial analytical framework which built on the components of existing theory and which was adapted for displacement contexts, considering both refugees and IDPs. The framework was then further refined based on learning throughout the project (Section 3), to produce the final framework (Section 4). In this initial version of the framework, our focus was on the integration of social protection and humanitarian assistance in contexts of forced displacement. During the project, it became evident that the term ‘integration’ was problematic. Integration is not necessarily the goal, and the term can only be used to describe actual integration of one system with another (see finding 1 under column 2 in Section 3). For this reason, we later adopted the more neutral terminology of ‘linking/ linkages’. However, in this section, we present the initial framework as it was developed early 2020, with the integration terminology. Figure 1 presents the initial analytical framework developed for the project. As represented in the diagram, our initial analytical framework comprised three main components: 1. the determining factors and processes that are likely to influence the adoption and impacts of an integration approach in a given displacement context (left-hand column/column 1) 2. the typology of approaches for integrating social protection and humanitarian assistance (middle column/column 2) 3. the benefits and drawbacks that different approaches may generate, and a structure for thinking about the trade-offs between different outcomes (right-hand column/column 3). Following Figure 1, we describe each of these three components in detail. 1 See https://odi.org/en/publications/linking-humanitarian-assistance-and-social-protection-in-response- to-forced-displacement-an-analytical-framework/ 4 ODI Working paper Figure 1 Initial analytical framework Determining factors and Approaches to integrating Benefits and social protection and humanitarian assistance drawbacks processes Options PARALLEL ALIGNMENT PIGGY BACKING NATIONAL SYSTEMS-LED FACTORS Effectiveness in Nature of Standalone Response uses Response Standalone addressing displacement response aligns elements of integrated in the humanitarian displaced Host’s structural with existing or the national national SP response population’s needs features, including future SP system system system state of SP & and risks humanitarian systems Host’s laws, No integration Full integration Effectiveness in institutions, norms meeting host governing rights of population’s needs displaced & response Policy Wider govt. and intl. Connection points: financing, legal & policy frameworks, governance humanitarian response and coordination Equity strategies Current situation of affected populations Programme design Cost and Connection points: programme objectives, targeting criteria, transfer design, links efficiency PROCESSES to other services Decision-making processes Accountability and acceptibility to Actors involved, and Administration stakeholders their motivation, Connection points: outreach, registration, enrolment, payment, case management, relationships, balance complaints and appeals, VAM/M&E, protection, information management of power, incentives, decision logics Sustainability 5 ODI Working paper ‘Determining factors and processes’ Figure 2 Column 1 from initial analytical framework Determining This column presents key factors and factors and processes that influence both the choice processes of integration approach in a given context (Column 2) and the benefits and drawbacks FACTORS of the approach for different stakeholders Nature of displacement (Column 3). Host’s structural features, including state of SP & Our review of the literature highlighted humanitarian systems Host’s laws, that there was not yet an established institutions, norms governing rights of framework outlining the set of displaced & response considerations that are most likely to Wider govt. and intl. humanitarian response be significant when considering social strategies Current situation of protection-humanitarian system integration affected populations in a given displacement context. Our PROCESSES main aim over the course of the project Decision-making processes was to fill this gap, by generating a more Actors involved, and robust understanding of the set of factors their motivation, relationships, balance and processes that appear to be most of power, incentives, decision logics instrumental in determining the adoption and impacts of a particular approach. The list proposed in Column 1 was an initial conceptualisation of relevant factors and processes. We developed this draft list by drawing on our review of all the considerations that were identified in earlier literature as potentially relevant (summarised in Table A1 in Annex 1). Through a number of workshops in the inception phase, our cross-disciplinary team then explored how these considerations might be best structured in relation to our specific area of research. We hypothesised that each of the components of our provisional list may influence the feasibility or desirability of different integration approaches, as well as the effects that an approach generates in a particular displacement context. Table 1 provides further explanation of our rationale, including examples of how each of the factors and processes might influence the adoption and potential impacts of different integration approaches. 6 ODI Working paper Table 1 Provisional list of determining factors and processes, and examples of how these might determine the adoption and impacts of different integration approaches in a displacement context FACTORS Factors relating to the Many of the factors that determine the optimality of a particular integration initial context approach are likely to relate to the initial context in which the displacement situation is developing. The nature of the For example: displacement: shock, • If IDPs are displaced due to a conflict in which the government is an active setting, profile of party, this may reduce the feasibility and desirability (for international/ affected populations humanitarian actors) of using a national systems-led approach. • If refugees are primarily based out-of-camp and are likely to be displaced on a protracted basis, this may reduce the feasibility and desirability of a traditional ‘parallel systems’ approach. • If the displaced populations are of the same ethnicity as the host population, speak the same language and are of similar socioeconomic backgrounds, this may increase the political feasibility of serving both populations together through a national systems-led approach. The host context (nationally and in the specific affected regions): Structural features, including: The political, economic, For example: social, cultural, • If the government is an authoritarian regime with a poor human rights record, environmental, security this may reduce the feasibility and desirability (for international/ humanitarian context actors) of national systems-led approaches. • If the country is a high-income country but with high levels of inequality leaving many citizens feeling left behind, this may mean there is less public and in turn political support for prioritising coverage for refugee populations. • If the cultural norms of the host region heavily emphasise shared ownership, this may increase the feasibility of integrating newcomers into existing social protection systems. • If the host region is regularly affected by other shocks (e.g., droughts) that necessitate external assistance, this may increase the feasibility and desirability of extending a shock-responsive social protection system to this region (i.e., to support the response both to drought- and displacement-related shocks). • If the specific region where IDPs are now residing is still affected by conflict and governed by non-state actors, this reduces the feasibility of using a national systems-led approach. • If the social protection system is only nascent but is in the process of being extensively developed, this may limit the viability of certain approaches (e.g., national systems-led) but increase the desirability of other approaches (e.g., alignment). 7 ODI Working paper The state of the • If the system has high capacity nationally, but capacity has been weakened in existing social the specific host regions (e.g., by a shock unrelated to the displacement), this protection system may bolster the case for using the humanitarian response in a way that builds (at the policy, capacity back up in the affected regions. programme design • If the system is notoriously ineffective (high leakage rates, large exclusion and administration and inclusion errors), this may reduce the desirability of relying on the current level), its capacity, national system. performance, shock- • If the system already has many shock-responsive features (e.g., a large social responsiveness and registry and payments mechanisms that can deliver to existing registrants the laws governing regardless of their current location), this may support a national systems-led the displaced approach in a context where IDPs have been temporarily displaced by a natural population’s hazard. access to it • If the social protection system has strict laws about non-citizens’ access to the system, and the process for changing such laws is slow and complex, this may reduce the feasibility of a primarily national systems-led approach in a refugee setting. The existing • If there is no existing or recent humanitarian presence in the country, this may humanitarian reduce the feasibility and desirability of a parallel systems approach. presence • Conversely, if there is already a functioning parallel system operating for a different displacement crisis in the specific host region, this might increase the likelihood (but not necessarily the desirability) of adopting a parallel systems approach. Laws, institutions • If the country has not ratified the 1951 Refugees Convention, does not and norms governing recognise the status of the refugees in the country, and does not accord the rights of the them rights to work, education, health systems etc, then this may reduce the displaced and the feasibility and desirability of a more national systems-led approach. national response to • Or, if the country has ratified all conventions and on paper accords refugees displacement many rights, but in practice these rights are rarely upheld and discrimination against refugees is widespread, this may also limit the consideration of a fully integrated approach. Laws, institutions • If there is an active conflict in the hosting area, humanitarian principles and norms governing influence the ways in which humanitarian agencies can engage with state and the international non-state actors in the delivery of assistance. humanitarian • If the main donors have restrictions on the direct transfer of funds to the response in this host government, this restricts the feasibility of a wholly national systems-led context approach. • If the country is in the process of implementing the Comprehensive Refugee Response Framework, this may improve humanitarian system perceptions of the feasibility and desirability of trialling a more integrated approach. (table continues over page) 8 ODI Working paper Factors relating to the For example: overall national and • If the political leadership has committed to a comprehensive response to IDP international response needs, has formed a strong cross-ministerial committee to deliver this, and has to the displacement explicitly earmarked funds for this initiative, it is likely to increase the feasibility, context: overarching desirability and potential impacts of working through this structure. response policies/ • If the wider international humanitarian response is providing other types of strategies, stakeholders, support to the displaced through national systems (e.g., through the national coordination education, national health system, etc), then this may increase the feasibility and mechanisms, funding desirability of adopting a similar approach for the social protection system. • If the wider international humanitarian response is large, well-funded and generally perceived to have been effective, this needs to be taken into consideration when exploring the perceived benefits and drawbacks of the specific integration approach. Factors relating to For example: the current situation • If host populations are poorer than displaced populations, this may make it of displaced and host less feasible and desirable to extend social protection to the displaced without populations: poverty and adequately supporting the host neighbours. wellbeing levels, social • If the displaced population have no access to livelihoods, this will increase the cohesion levels levels of assistance that they will require to meet their basic needs and limits the feasibility of mirroring the transfer values used in the national system. PROCESSES The nature of the The decision on a particular approach and the outcomes achieved through decision-making an approach relate to the process used to compare and evaluate options, the processes level of consultation undertaken and the speed with which decisions are made, among other factors. The actors involved The specific combination of actors involved in the decisions will directly in these decision- influence conclusions about what is feasible and what is desirable. Each making processes, actor has particular interests and concerns that may strongly affect the and their motivations, implementation of a particular approach down the road. For example, if the relationships/balance of government is not adequately involved throughout the decision-making process power, incentives and but has major concerns about being left with the bill of closer integration or decision logics losing public support ahead of an upcoming election, this is likely to hinder the implementation of aspects of the response involving national systems. 9 ODI Working paper ‘Typology of integration approaches’ Figure 3 Column 2 from initial analytical framework Approaches to integrating social protection and humanitarian assistance Options PARALLEL ALIGNMENT PIGGY BACKING NATIONAL SYSTEMS-LED Standalone Response uses Response Standalone response aligns elements of integrated in the humanitarian with existing or the national national SP response future SP system system system No linkages Full integration Policy Connection points: financing, legal & policy frameworks, governance and coordination Programme design Connection points: programme objectives, targeting criteria, transfer design, links to other services Administration Connection points: outreach, registration, enrolment, payment, case management, complaints and appeals, VAM/M&E, information management Source: Authors This column shows the ‘typology of integration approaches’ that can be used to connect social protection and humanitarian assistance in displacement contexts. The top half of the column shows the continuum of options (from an entirely parallel systems approach of zero integration to an entirely national systems-led approach of full integration). The lower half shows the ‘connection points’ (formerly known in the literature as ‘entry points’): the components of the social protection system to which a degree of integration can be applied. Prior to our project, there had already been considerable efforts to conceptualise the range of approaches for using social protection systems to respond to different types of shocks. Recent work by Seyfert et al. (2019) drew on existing theories to explore the typology of approaches for integrating social safety nets and humanitarian assistance specifically in refugee settings, considering both the continuum of integration options and the points in the social protection system where integration can occur. It also highlighted the various ‘entry points’ where humanitarian actors can look to connect with the social protection system, at the levels of policy, programme design and administration. 10 ODI Working paper This meant that our typology of integration approaches did not need to develop a conceptualisation from scratch but instead took Seyfert’s typology as a starting point, to refine over the course of the project, testing for validity for both refugee and IDP settings, and for integration with all aspects of the social protection system (beyond just social safety nets). The central column in Figure 1 therefore largely consists of the framework presented in Seyfert et al. (2019); however, we adjusted the name ‘entry points’ to ‘connection points’. This is because entry points are often associated with the first entrance; once you have used a particular combination of points to ‘enter’ the system, it may be confusing to still label the subsequent set of connections as ‘entry points’. In addition, the term ‘entry points’ sometimes has connotations of ‘progress’ or ‘going in the right direction’; while increased integration is often the desired goal, there are also cases where zero integration may be the best option, so we prefer ‘connection points’ as a more neutral term. As illustrated in the Column 2 diagram, a programme’s integration approach is defined by the different integration options taken at each connection point in the social protection system. A detailed description of an integration approach would look systematically at each connection point in turn, assessing which integration option was used (parallel systems, alignment, piggybacking, national systems-led) at each connection point. An example of a detailed description of an integration approach is shown in Figure 4, for a hypothetical programme. We refer to this detailed description as the ‘integration profile’ of the programme. The horizontal axis in the figure shows the continuum of integration options (from zero integration with entirely parallel systems, to full integration with an entirely national systems- led approach). The vertical axis shows the connection points in the social protection system to which different degrees of integration can be applied. The marks on the table illustrate the different degrees of integration that were applied to the different connection points. 11 ODI Working paper Figure 4 Integration profile for a hypothetical programme Integration options No integration Full integration PARALLEL ALIGNMENT PIGGY BACKING NATIONAL SYSTEMS-LED Standalone Response uses Response Standalone response aligns elements of integrated in the humanitarian with existing or the national national SP response future SP system system system CONNECTION POINTS Policy Financing Legal and policy framework Governance and coordination Programme design Eligibility criteria and qualifying condition Transfer type, level, frequency and duration Administration Outreach Registration Enrolment Payment Case management Complaints and appeals Protection VAM/Monitoring and evaluation Information management Source: Authors, adapted from Seyfert et al. (2019) In theory, the full ‘typology of integration approaches’ therefore consists of all the possible integration profiles – i.e., all possible combinations of integration options and connection points. However, for practical reasons, analysis of different programmes’ integration approaches may not always describe the integration options used at every connection point so systematically, and may instead focus only on the parts of the integration profile that are most distinctive in that particular case. 12 ODI Working paper ‘Benefits and drawbacks’ Figure 5 Column 3 from initial analytical framework Benefits and Column 3 presents a set of criteria that drawbacks stakeholders may consider and/or prioritise when assessing the benefits and drawbacks of a particular approach. The criteria presented in this Effectiveness in column drew on the existing conceptualisation addressing of evaluation criteria in the shock-responsive displaced population’s needs social protection literature (O’Brien et al., and risks 2018; European Commission, 2019; WFP, 2019; Effectiveness in UNICEF, 2019; Seyfert et al., 2019). However, the meeting host existing frameworks needed to be tailored for population’s needs displacement contexts and for linkages with wider Equity social protection systems because none were specifically designed for both refugee and IDP Cost and settings, and most were designed with only social efficiency safety nets in mind (rather than broader social Accountability and protection systems). There was also a need to acceptibility to think further about the substantive dimensions of stakeholders certain criteria and how these dimensions might be quantitatively compared, for example for the Sustainability ‘accountability and political economy’ criterion proposed in Seyfert et al. (2019). One of the aims of our project was therefore to develop a refined understanding of the criteria for evaluating integration approaches in displacement contexts. The draft of this revised criteria list is shown in the right-hand column of Figure 1 and Figure 5. We developed this draft list during the inception phase in consultation with our cross-disciplinary project team by going back to the literature’s conceptualisation of each category of criteria, scrutinising what was meant by some of the less quantifiable terms (e.g., political economy), checking for overlap between certain criteria (e.g., ‘coverage’ and ‘addressing needs’), and identifying gaps or revisions that might be needed for our displacement-focused research question. The list consisted of six dimensions, which collectively aimed to encapsulate the key criteria of importance to all stakeholders (including the displaced and host populations, host government, humanitarian agencies and donor organisations): 13 ODI Working paper 1. Effectiveness in addressing the needs and protection risks of the displaced population – this considers whether the key needs and protection risks of the displaced population are addressed in a timely manner, with a well-designed and well-implemented intervention (appropriate type and amount of support, properly targeted and reliably delivered), that has sufficient population coverage. 2. Effectiveness in meeting the needs of host population – this considers whether host populations continue to be served by existing support systems and receive additional support (through existing or new support mechanisms) to meet new displacement-related needs. As in the case of the displaced population, the support that is provided should be delivered in a timely manner, with a well-designed and well-implemented intervention that has sufficient population coverage. 3. Equity – this considers how fairly the results of the approach are distributed across different groups in the target population (including between the displaced and host populations, and across groups within these populations). This dimension pays particular attention to how well the approach reaches the most marginalised, and it accounts for the heightened needs and risks that they may face based on gender, age, disability, ethnicity, religion, socioeconomic status or other characteristics. It also considers whether the intervention may have unintended negative effects on or between different groups, for example by damaging displaced-host relations. 4. Cost and efficiency of systems – this considers the overall financial costs of the approach and the extent to which it reduces the duplication of programmes, delivery systems and processes in the short and long term. 5. Accountability and acceptability to all stakeholders – this considers whether the approach enables both top-down and bottom-up accountability, and is perceived to be acceptable by the displaced population, host population, and host government at national and local levels, as well as international and non-governmental actors and donor agencies in contexts where they are engaged. This includes considering whether the approach meets the requirements and preferences of each stakeholder, including ensuring that it respects humanitarian principles. It also includes a consideration of whether there are any restrictions on the provision of funding to particular entities, for instance in cases where donors may not be permitted to provide direct budget support to a particular government entity. 6. Sustainability – this considers sustainability from two key dimensions: (i) the viability of sustaining the approach for as long as needs remain, and (ii) the approach’s contribution to bringing about long-term solutions to current and future displacement situations. 14 ODI Working paper 3 Refining the framework We used our primary research to test our initial framework, exploring the assistance approaches in place in different displacement settings, their perceived benefits and drawbacks, and the factors and processes that determined the selection and outcomes of the approach. In this section, we draw out the key lessons from the research and its implications for the framework, and then propose specific adjustments to the framework. Alongside a literature review of global evidence (Gray Meral and Both, 2021), we conducted primary research through mixed methods data collection in six sites across three countries: 1. Cameroon: in East region, refugees from the Central African Republic (mostly since 2015 in our sample); and in Far North region, Nigerian refugees and IDPs, displaced by conflict with Boko Haram since 2015. 2. Colombia: in the capital city of Bogotá and border city of Cúcuta, IDPs displaced by decades of internal conflict and violence, and Venezuelans displaced since 2015. 3. Greece: in the capital city of Athens and the smaller north-western municipality of Ioannina, asylum seekers and refugees fleeing Syria and elsewhere, mostly arrived since 2014. A survey of approximately 1,500 displaced and host community members was conducted in each country from January to April 2021. In addition, focus group discussions and in-depth interviews were held with members of displaced and host populations, and key informant interviews with representatives of the government, UN, non-governmental and donor organisations. The full case studies are documented in Levine et al. (2022 – Cameroon), Ham et al. (2022 – Colombia) and Tramountanis et al. (2022 – Greece). In this section we use the case studies to illustrate how they informed our conceptual understanding – and the revisions we made to the framework as a result. A variety of assistance approaches were in place in the case study countries, from a largely parallel response in Cameroon (with some elements of alignment in one pilot humanitarian programme) to a largely nationally led approach in Colombia (particularly for IDPs, and to a slightly lesser extent for Venezuelans). Greece fell somewhere in between the two. The dominant cash assistance for asylum seekers in Greece was EU-financed and delivered by humanitarian agencies at the time of the research but aligned with a social protection programme in transfer value (on the requirements of the Greek government). Although refugees are theoretically eligible for certain social protection programmes, de facto access was negligible at the time of our study, meaning the only assistance available (specifically for newly recognised refugees) was a European Union-financed programme delivered by UN and non-governmental agencies at the time of the research, with the support of the Greek government. 15 ODI Working paper Lessons on ‘determining factors and processes’ (now re-named ‘determining factors and actors’) Column 1 in the initial framework showed a provisional list of ‘determining factors and processes’, hypothesising that the approach taken and outcomes emerging in a given displacement setting were likely to be determined both by the key contextual factors and by the decision-making process and actors involved in designing and implementing the approach. Our research tested this provisional list by examining the factors and processes that determined the choice and outcomes of the assistance models in place in our primary case studies, as well as in our broader Global Evidence Review (Gray Meral and Both, 2021). As a result, we revised column 1 as presented in Figure 6. This section then explains the different adjustments made. Figure 6 Initial and final column 1 Determining Determining factors and actors factors and processes FACTORS FACTORS ACTORS Nature of displace- ment Initial context Actors involved and their: Host’s structural features, including Nature of displacement state of SP & – motivation, humanitarian systems Host context: incentives and decision logics Host’s laws, – structural features, institutions, norms state of existing SP & – relationships, governing rights of humanitarian systems interactions and displaced & response balance of Wider govt. and intl. – laws, institutions, power humanitarian response norms governing strategies government and Current situation of international response affected populations PROCESSES Wider government and Decision-making and international processes humanitarian response strategies Actors involved, and their motivation, relationships, balance Current situation of of power, incentives, affected populations decision logics 16 ODI Working paper Finding 1: Actors matter more than factors Our major finding was that the approach taken – and the outcomes of that approach – depend far more on the political will and incentives of key actors than on any of the contextual factors (such as the country’s income status, maturity of social protection system, size of displacement crisis). In our research, this was most clearly evidenced by the Greek-Colombian contrast. Although Greece is a higher-income country, with a more advanced social protection system in terms of expenditure and coverage and a smaller-scale displacement crisis (120,000 refugees/migrants settled since 2015/16), its assistance approach was much less integrated and nationally led than Colombia’s approach to assisting its 8 million IDPs and 1.8 million Venezuelans. In both country studies, this was primarily explained by the very different levels of political will and incentives, particularly of the national government. In Greece, the absence of political will to take a nationally led, integrated approach to assisting displaced people is driven by a desire to ‘make the asylum system unattractive to third-country nationals’. As stated by the Minister of Migration and Asylum: ‘benefits and hospitality act as a pull factor to come to our country and take advantage of these benefits’ (Proto Thema, 2020). By contrast, Colombia has been praised for its exceptional policies and will to integrate Venezuelans into many aspects of society and services (UNHCR and IOM, 2021). Although other elements also played a role, the primary reason given for this in the government’s official policy was the economic gains that Colombia could expect to achieve through well-managed integration, based on economic modelling. At a political level, the incentives are also likely to have been influenced by the inter-governmental relations between Colombia and Venezuela; there may be some political gains for the government from demonstrating that the opportunities for Venezuelans in Colombia are superior to those they have back home. Besides the actors and incentives at the national level, the Colombia case study also illustrated the importance of local governments as key actors in this analysis. They often played a major part in determining whether or how displaced people are served through the social protection system at the local level, as well as the extent of coordination between governmental and non-governmental agencies. Proposed revision(s) in the final framework In relation to our framework, the above finding implies that the most critical component of Column 1 is the bottom half (previously-labelled the ‘processes’, referring to the nature of the decision- making processes and the actors involved) (Figure 2). It also implies that within that component, the emphasis needs to be placed on the actors involved and their incentives and agency. We therefore made several changes to Column 1 (Figure 6): 17 ODI Working paper 1. We replaced the term ‘processes’ with ‘actors’, to emphasise the agency of the actors in shaping the approach and its outcomes, and to move away from a term that may be associated with technical or bureaucratic procedures. 2. We re-arranged the column to give greater prominence to the ‘actors involved’, an aspect that was previously somewhat buried at the bottom of the list. We still show the contextual factors at the furthest-left part of the column, since these can be seen as the background in which the actors are operating. However, we have now placed the ‘actors’ component immediately next to this, on the same level, to give it greater prominence. 3. We re-ordered and re-labelled the bottom two components ‘nature of the decision- making processes’ and ‘actors involved’ into a single overarching category – ‘Actors’ – with two sub-categories: (i) their motivation, incentives and decision logics, and (ii) their relationships, balance of power and interactions. 4. Although it does not affect the overarching category, the prominence of intergovernmental and intragovernmental relations also emerged as an important consideration, with local and foreign governments also emerging as potential key ‘actors’ to include in the stakeholder analysis. Finding 2: Although secondary to the political will of the actors involved, the factors identified in our initial framework also played a role in determining the assistance approach and outcomes generated We found evidence of each of the factors in our provisional list helping to determine the assistance approach taken or the benefits/drawbacks it produced, to greater or lesser degrees. Nature of the displacement The type of displacement shock, setting and characteristics of displaced populations all emerged as potential determinants of the approach taken. For example, in Colombia, both the government and the international community endorsed a nationally led approach to the Venezuelan influx, in part because of the expectation that the population will be displaced for a long time, given the lack of any obvious resolution to the situation in Venezuela. The fact that around one-third of those arriving from Venezuela were Colombian citizens who had been living in the neighbouring country either for work, family or as refugees themselves – speaking the same language and often living in mixed-nationality households – created a strong impetus to integrate Venezuelans directly into host communities, and to serve them through the host state systems. In Greece, by contrast, the initial trajectory of the displacement shock created a precedent of asylum seekers and refugees continuing on to other countries in Europe, limiting the impetus from both the government and the displaced population to look to national integration as a primary solution. Unlike in the Colombian case, the displaced population in Greece was relatively distinct from the host population linguistically and culturally, and lacked familiarity with the Greek state system, having arrived from much further afield and generally having few existing ties to Greece. 18 ODI Working paper The host context – structural features – Political, economic, social, cultural, environmental and/or security context Across the case studies, there was strong evidence that wide-ranging structural elements in the host context play a role in shaping both the selection of the assistance approach and the benefits and drawbacks that emerged from the approach taken. Structural features of the Cameroonian context illustrate why a nationally led assistance approach has not been taken there. The low national income level and weak economy mean that the government is struggling to invest adequately even in the development of the general population, rendering a nationally led approach to assisting displaced people (especially displaced non-citizens) a distant prospect. Security concerns have clearly influenced the government’s approach, for example active conflict between the government and the Boko Haram insurgency in the Far North has meant that the government has tended to respond to the displacement situation primarily through a security lens, focusing on (re-)securing territorial control, rather than supporting IDPs to rebuild their lives in their temporary places of residence. From the perspective of international actors (and many displaced and host people themselves), there was also little desire to see a transition to a wholly nationally led assistance model given historically high levels of corruption and concerns about the transparency and accountability of state systems. Meanwhile, Colombia illustrates how structural features also affect the benefits and drawbacks of a given approach. In the context of high inequality, a fragile social contract, and extensive unrest about the government’s failure to address longstanding citizen concerns, social discontent was already high even before the Venezuelan influx. Amid widespread discontent, the government’s apparent prioritisation of non-citizens’ concerns through a strong, nationally led approach to assisting Venezuelans has exacerbated social tensions, with drawbacks for both horizonal cohesion (relations between host and displaced communities) and vertical cohesion (relations between communities and the state). – The existing social protection system For obvious reasons, the assistance approach and its outcomes are closely linked to the host’s existing social protection system – its formation, capacity, performance, and shock- responsiveness at the policy, programme design and administration levels, as well as the displaced population’s existing legal and de facto access. In Cameroon, for example, it is not possible to integrate assistance into the national social protection system when the ‘system’ itself has not yet been formed. Fragmented social protection initiatives exist, offering short-term support with low coverage, spread across multiple ministries. But social protection is not rights-based, access is not enshrined in law, and the national social protection policy has yet to be fully ratified. Although shock-responsiveness and displaced inclusion is being built into the social safety net from the outset, its very nascent state makes it difficult to rely on the national system as the dominant assistance channel for displaced people in the short to medium term. 19 ODI Working paper Meanwhile in Colombia, the right to social protection is enshrined in law for all residents, and there is a much more developed system of social assistance to identify and support vulnerable households. Although Venezuelans did not officially have access to major national cash transfer schemes until the pandemic, they had been supported through various other aspects of the social protection system (including Early Childhood Development services, the school feeding programme, and the subsidised health insurance scheme), encouraging government and international actors alike to consider ways to meet Venezuelans’ needs more comprehensively through that system. – The existing humanitarian presence Our research indicated that the existing humanitarian presence in the country can shape the approach taken to linking with the social protection system – but that this occurs in diverse ways, with no clear pattern. In Cameroon, humanitarian agencies were already operating through the traditional model of largely parallel food assistance programming in the East Region when the Boko Haram-related insecurity began in the Far North. This existing way of working in the country may have set a precedent for the response. However, such precedents are not always necessary for a largely parallel humanitarian operation to be newly established. In Greece, there was no international humanitarian presence prior to 2015, yet after the refugee/migrant influx, internationally implemented assistance programmes were established, with limited operational links to the state social protection system. Host context – laws, norms and institutions (‘rules of the game’) governing the response to the displacement Alongside the structural features of the host context (discussed above), the laws, norms and institutions governing the response to the displacement situation also sometimes shaped the assistance approach. In line with political science definitions, institutions here refer to ‘the rules of the game’, including both formal and informal rules that govern behaviour (Harris, 2013). Laws refer to formal laws and regulations, while norms can be thought of as informal or unofficial social/political/cultural rules. We noted various instances where official laws appeared to be applied to very different degrees across contexts (e.g., obligations under the 1951 Refugee Convention, or guarantees of the right to social protection under the Universal Declaration of Human Rights, which theoretically applied in all three case study countries). In many cases, norms (at the international and national level) seemed to have a clear impact on the linkage approach taken. In Cameroon, for example, the international humanitarian principles of neutrality and independence likely reduced the feasibility and desirability (for international/humanitarian actors) of integrating assistance with state systems in response to the Anglophone IDP crisis in the country’s North-West and the South-West regions, given the state’s active role in that conflict. 20 ODI Working paper However, nascent linkages with state systems were perhaps considered a more viable option in response to the Central African Republic refugee influx in the East region, given the state’s lack of direct involvement in the cause of that displacement. Cameroon also demonstrates the role of informal institutions or ‘rules of the game’ at the local level in influencing how a linkage approach was implemented. There, it seemed that local officials and community leaders sometimes had an unofficial policy to prevent ‘double-dipping’2 by excluding those who were already receiving humanitarian assistance from the nascent rollout of the state-led social protection system. This meant that even though IDPs technically had the same right to social protection as host populations, and even though there was no official eligibility criterion restricting their access, they were sometimes de facto excluded when hosts were seen to be more deserving of the limited social protection available given their lack of access to other assistance. The wider national and international displacement response International and national actors’ decisions about linking with the social protection system to assist displaced populations are in part shaped by the wider displacement response (the overarching policy, strategy, mechanism and financing). In Colombia, the state holds official responsibility for leading the response to internal displacement under the Victims’ Law, and has a strong overarching national policy, strategy and coordination mechanism (the government’s established Victims’ Unit) for the IDP response. This wider response encouraged a reliance on a nationally led approach to assisting IDPs among both governmental and international actors. Similarly, for the Venezuelan influx, the Colombian government’s overarching response has been characterised as exceptionally progressive and has been strongly nationally led. This created a greater expectation that any international assistance should be provided as a supplement to state systems and linked with them – rather than as a parallel response. By contrast, in Greece, the overall national response to the refugee influx was for many years characterised by an unofficial policy of non-integration. Asylum seekers remained geographically and socially isolated for protracted periods on the islands in camps and reception centres, and the thorny question of long-term integration was largely side-stepped. This resulted in an internationally led assistance approach for many years, financed and delivered by international agencies, with minimal linkages to the state social protection system. Meanwhile, Cameroon demonstrates the role that the financing mechanism for the overall displacement response can play in shaping the extent to and ways in which assistance for displaced populations links with social protection. Historically, assistance for refugees and IDPs has been financed and delivered through traditional international humanitarian mechanisms. However, the entry of the World Bank (traditionally a development partner rather than a 2 ‘Double-dipping’ refers to a recipient obtaining assistance from more than one source, typically illegitimately. 21 ODI Working paper humanitarian donor) in financing components of the displacement response has created more impetus to integrate displaced populations into state systems. Through the World Bank’s IDA18 Sub-Window for Refugees and Host Communities, the government of Cameroon was awarded funding for the inclusion of displaced populations in various state services and systems, including a component to include populations affected by displacement in the nascent World Bank-funded national Social Safety Nets project. The situation of affected populations One of the main factors that might be expected to determine the linkage approach is the situation of the affected populations – their needs, preferences and capabilities, including both narrow and broad dimensions of material and subjective individual and household wellbeing, as well as the social wellbeing of the communities as a whole (i.e., social cohesion). In our research, we found clear indications that the situation of the affected populations could influence the outcomes of the approach taken but found fewer examples of their situation actively determining the decision about which approach to take. Displaced populations generally have distinct, and often higher, needs than hosts, given the losses incurred before and during their displacement, and their more limited access to land, livelihoods and community networks. This means that linking with lower-value social assistance may leave displaced households unable to meet their needs. For example, in Cameroon, the main national cash transfer value amounts to only 12% of the national poverty line and is widely seen as inadequate for the host population. This has given rise to concerns that aligning or integrating with the social protection system more closely in terms of programme design (transfer amounts) will lead to a neglect of the displaced populations’ needs, as well as potential tensions between the communities if the lower-value assistance reduces displaced populations’ ability to contribute to the local economy and pushes them to resort to coping strategies with negative social impacts (e.g., begging or theft). In Greece, asylum seekers and refugees’ ability to fulfil their basic needs independently were extremely limited given the constraints on their access to decent work, due to initial restrictions on their ability to work when they first apply for asylum, combined with few formal employment opportunities, language barriers and discriminatory practices. Only 4% of our sample was employed. This created challenges for the ability of the current assistance model to meet the displaced populations’ needs, particularly for refugees. The main assistance programme for refugees is the HELIOS scheme, which lasts only for up to one year and which had only been accessed by around one in seven recognised refugees. After this, refugees can theoretically turn to mainstream social assistance programmes for support – but the long-term residence criteria of many of the programmes exclude them in practice, and only two refugees in our sample had accessed this system. 22 ODI Working paper Proposed revision(s) in the final framework We broadly found the factors listed in this section to be helpful categories for thinking about the selection and outcomes of the assistance approach. However, to streamline the analysis, we have grouped the ‘Laws, institutions and norms governing the national response to the displacement’ and the ‘Laws, institutions and norms governing the international response to the displacement’ into a single category (thereby mirroring the combined national and international dimensions of the ‘Factors relating to the overall national and international response to the displacement context’ category). Lessons on the ‘typology of integration approaches’ (now re-named the ‘typology of linkage approaches’) Finding 1: Integration can only be used to describe actual integration of one system with another – and is not necessarily the goal We started the project using the term ‘integration options’ to refer to all possible options on the ‘continuum of integration’, ranging from no integration at all (parallel systems) to full integration (national-system-led). We illustrated this integration continuum with a uni-directional arrow moving from zero to full integration on our framework diagram. However, our application of the framework indicated that this conceptualisation was sometimes problematic. This language and visual representation implied that the national-system-led approach is the ultimate goal or ‘end game’, which may not always be the case. Similarly, the use of the term ‘integration options’ or ‘approaches for integrating’ implied that all the options on the continuum could generally be thought of as a form of (greater or lesser) integration. In reality, ‘integration’ describes only the right-hand extreme on the continuum, where one system becomes part of the other. By contrast, ‘parallel’, ‘alignment’ and ‘piggy-backing’ are conceptually distinct categories, where humanitarian programming intentionally avoids integrating with the national social protection system. Proposed revision(s) in the final framework To avoid the implication that an integration (national-system-led) approach is the ‘end game’ or optimal approach across all connection points, and to avoid implying integration where no integration occurs, we have: • Removed the ‘integration continuum’ arrow and label from the diagram. • Replaced the term ‘approaches for integrating’ with ‘approaches to linking’ and replaced ‘integration options’ with ‘type of linkages’. ‘Linking’ is a more neutral term that 23 ODI Working paper unambiguously describes ‘alignment’ and ‘piggybacking’, while still being relevant when a ‘national-system-led’ (fully integrated) approach is taken. This change is in line with the wider linguistic shift taken by some of the actors working on related topics, such as the expert advice helpline on Social Protection Approaches to COVID-19 (SPACE). It thereby avoids unnecessary divergence and conceptual confusion among those working on related initiatives. Figure 7 Initial and final column 2 Approaches to integrating social protection and humanitarian assistance Options PARALLEL ALIGNMENT PIGGY BACKING NATIONAL SYSTEMS-LED Standalone Response uses Response Standalone response aligns elements of integrated in the humanitarian with existing or the national national SP response future SP system system system No integration Full integration Policy Connection points: financing, legal & policy frameworks, governance and coordination Programme design Connection points: programme objectives, targeting criteria, transfer design, links to other services Administration Connection points: outreach, registration, enrolment, payment, case management, complaints and appeals, VAM/M&E, information management Approaches to linking social protection and humanitarian assistance Linkage options and connection points Promoting collaboration PARALLEL ALIGNING LEVERAGING STATE-LED Humanitarian-led, Humanitarian-led, Humanitarian-led, National social not linked to state using state SP as a involving the state protection system SP reference point and using or used by SP Tailored for displaced people vs standard SP Policy Connection points: financing, legal and policy frameworks, governance and coordination, functional and technical capacities Programme design Connection points: vulnerability and risk profiling,  eligibility criteria, objectives/type/package of benefits/services, transfer amount Administration Connection points: outreach, registration, enrolment, provision of benefit/service, accountability, case management (incl. protection), M&E, information management 24 ODI Working paper Finding 2: This framework primarily describes how humanitarian actors can link with the state to assist displaced people. But even where linked with social protection or led by the state, the approach can (and typically should) still be tailored for displaced people. Throughout the project, we noted some confusion in the way different linkage options were being interpreted (within the project team and among wider stakeholders). There were two main and sometimes divergent ways in which linkage options were being conceptually categorised: A) In some cases, the classification was made based on whether and how the provision for the displaced was linked to the existing social protection system (at the policy, programme design, or administration level), i.e., whether it operated in parallel to, was aligned with, piggy- backed on, or was fully integrated into the existing social protection system, overall or at that connection point. B) In other cases, the classification was made based on who was responsible for provision at a given connection point: i.e., whether that aspect of policy, programme design or administration was state-led, or whether it was led by international/non-governmental agencies – and if the latter, whether those agencies operated in parallel to, aligned with, or piggy-backed on state systems at that connection point. For example, if an international donor funds a new programme for displaced people that is entirely distinct from any existing social protection programme but is run by the state with humanitarian agencies’ technical support, this would broadly be categorised as Parallel in Method A, but would be categorised as an Integrated/national-systems-led approach in Method B. Our initial framework – following others – took the programme as the unit of analysis, reflecting the dominant view that the issue was about exploring the possible integration of a given programme targeted at displaced populations into an existing flagship national social protection programme designed for non-displaced citizens – in other words, following the classification approach in Method A. However, a key insight from our research is that having unique programming for displacement-affected populations, distinct from existing social protection provisions, should not imply that it is not integrated into the state-led system. The key conceptual distinction determining whether an approach is parallel versus linked is how humanitarian actors work in relation to the state (i.e., Method B), not how closely the provision for displaced people relates to existing social protection provisions (i.e., Method A). 25 ODI Working paper Proposed revision(s) in the final framework We feel that Method B from our classification above provides more room for thinking about the steps towards full and effective inclusion of displaced populations in social protection systems because the ‘integrated’ approach would mean not only allowing displaced populations to access the existing social protection system, but would also require the state to adjust the existing system to tailor policies, programmes and administrative mechanisms to the displaced population’s unique needs. Method B also recognises that state-led social protection systems are themselves diverse, often with a wide range of programmes under the remit of multiple governmental agencies, some of which may be contracted out to non-governmental organisations in the implementation phase. The key concern is whether or not the state plays the leading role and considers the programme part of its social protection system (loosely defined). We therefore clarified that the linkage option categories refer principally to who is responsible for provision at that connection point –whether it is led by the state or by humanitarian actors (NGOs/international agencies), with or without links to the state’s social protection system. This means that we have adjusted the description of the linkage options as follows: a. Parallel approach – NGOs/international agencies use a standalone approach to humanitarian programming, with no link to the state social protection system. b. Aligning – the programming approach is humanitarian-led but uses the state’s existing or future social protection system as a reference point.3 This could mean that the humanitarian actors ‘mirror’ the state-led social protection approach, but it could also mean that humanitarian arrangements are informed by the state’s existing approach but tailored by the humanitarian agencies. c. Leveraging – the programming approach is humanitarian-led, but the state is involved and the approach directly uses or is used by the social protection system.4 ‘Leveraging’ means that the humanitarian approach is not just informed by state social protection, but that the state is engaged and social protection is in some way explicitly leveraged for the humanitarian programming approach, or vice versa. d. State-led – the programming approach is state-led (operating with or without humanitarian support). 3 The term ‘future social protection system’ acknowledges that the system in many displacement settings is not yet fully established, but humanitarian actors may nevertheless refer to the expected/future plan for the social protection system when developing their approach. 4 As explained in the next sub-section, we changed the term ‘Piggy-backing’ to ‘Leveraging’ because piggy-backing was not well-understood by non-native-English speakers. 26 ODI Working paper Furthermore we added a ‘tailored vs standard’ option under the three latter categories - this indicates that even if an approach is linked with the social protection system or led by the state, it can – and almost always should – be tailored to the specific needs of the displaced population, rather than adopting the ‘standard’ approach used for non-displaced populations. Finding 3: It is not just about humanitarian assistance linking with or building on a fixed and established social protection system; the social protection system can also link with and build on the humanitarian provision Our initial analytical framework was focused on the ways in which humanitarian assistance could operate in parallel to, align with, piggyback on, or integrate with the social protection system. This uni-directional line of engagement was designed to keep things as conceptually clear and simple as possible, since our primary focus was to understand how changes in the assistance model (from a parallel approach to something more closely linked with or integrated into the state system) might influence provision for the displaced population (or their relations with the host community), rather than being concerned with wider social protection system-strengthening as a primary objective or outcome. However, we found our initial framing too limiting. It portrayed the state social protection system as fixed, rather than recognising its potential to adjust based on learning/contributions from the humanitarian system. In Colombia, for example, key informants highlighted the benefits for the state social protection system when the government was able to make use of the data collected by humanitarian agencies to update Venezuelans’ entries in the SISBÉN, off the back of a joint humanitarian-government pilot (which was largely humanitarian-run but with government resources engaged for various aspects of policy, programme design and implementation). Examples of social protection actors leveraging humanitarian systems have been seen in other contexts. In Lebanon, for example, the EU Trust Fund was mobilised to assist both refugees and poor and vulnerable Lebanese populations. It supported two systems: a parallel response managed by WFP, and an expansion of a World Bank-supported government social assistance programme (NPTP) that had a very low caseload. Adjustments were made to allow the EU funding to also be used for people affected by Covid-19 restrictions. Proposed revision(s) We have: • Clarified that ‘leveraging’ can happen in both directions, by changing the sub-description from ‘assistance uses elements of the national system’ to ‘humanitarian programming involves the state and uses or is used by the social protection system’. • Changed the term ‘piggybacking’ to ‘leveraging’ to help clarify for non-native-English speakers the concept of an actor directly using another actor’s approach. 27 ODI Working paper Finding 4: Regardless of whether/how humanitarian actors link their programming with state systems, it is still (almost always) a good idea to have a collaborative approach Even if humanitarian actors choose not to link operationally in any way with the social protection system, it is still important to be aware of that system, to aim to complement it as much as possible, and to communicate effectively with the state regarding the provision of assistance. This is important to deliver an effective response, to identify and avoid unnecessary duplication in efforts, and to improve coherence from the perspective of the affected populations (e.g., so that humanitarian agencies can inform them of the national social protection programmes that are in place, even if they do not directly refer to or operate with them). Proposed revision(s) Based on this revision, we have added an overarching layer of activity – ‘promoting collaboration’, which can apply across all types of linkages and all connection points. We avoid the firm directive to ‘collaborate’ because of a recognition that in a few cases it may not be possible to engage at all with the state (e.g. because the state is non-complicit or because doing so would legitimise actions that violate international humanitarian or human rights law). Finding 5: The thinking on the potential connection points between social protection and humanitarian assistance have been further refined in the broader field/ literature, particularly as a result of the Covid-19 response Covid-19 accelerated both the theories and practice of delivering social protection in response to a crisis, including the ways in which humanitarian assistance and social protection systems can work together in the crisis response. The conceptualisation of the potential connection points between the two systems has therefore been developed further since we first started working on this project in January 2020. Given the need to promote conceptual coherence across actors, we have aligned with the latest framework from the SPACE team, which was developed building on available literature and insights from most recent experiences (SPACE, 2021). Proposed revision(s) We have updated the connection points at the three levels. This includes: • At the policy level: adding ‘functional and technical capacities’ • For programme design: adding ‘vulnerability and risk profiling’, regrouping objectives, transfer type(s) and links to other services into ‘benefits/services package’ and relabelling other aspects of transfer design as ‘amount of support’ • For administration: renaming some functions along the delivery chain, now considering protection issues under ‘case management’. 28 ODI Working paper Finding 6: The overall benefits and drawbacks that may result from linkages between humanitarian assistance and social protection relate to the way in which many programmes operate in combination – and can only be understood when looking across the systems Social protection systems use multiple programmes to meet the diverse needs of different groups in the target population; likewise, humanitarian responses use multiple forms of programming to address crisis-affected populations’ needs. While there is certainly a need and opportunity to reduce duplication where this exists, in most cases effectively addressing the basic needs and protection risks of displacement-affected populations, and promoting the broader realisation of their rights, will only occur when multiple programmes operate in combination. Beyond the effectiveness dimension, this is also true in relation to other outcome areas; a suite of humanitarian-run refugee programmes operating largely in parallel to the social protection system may be acceptable to host governments – and may produce net benefits for social cohesion – but only if accompanied by simultaneous support to expand social protection programming for host communities in the affected areas. For this reason, benefits and drawbacks from linking humanitarian and social protection responses can only be properly assessed when mapped across the humanitarian and social protection systems. But the linkage profiles cannot be meaningfully mapped for all programmes simultaneously; firstly, the type of linkage that can be or has been adopted at each connection point will be dependent on a particular programme (or multiple programmes if they are identical in their approach). Secondly, even if multiple linkage profiles for displacement-focused programmes are represented simultaneously on a linkage profile diagram, the benefits and drawbacks will not be evident unless considered in combination with all other (non-displacement- focused) social protection programmes – many of which will not have linked at all with humanitarian assistance. Proposed revision(s) in the final framework Our initial framework – following others – had taken the programme as the unit of analysis. But our field application highlighted the limitations of this programme-level focus when it comes to trying to assess the outcomes of an approach. To understand outcomes of linking, it is the system that matters – the whole suite of humanitarian and social protection measures in place and any linkages between them. Yet, linkage profiles wouldn’t be drawn in the same way for programmes and systems – firstly because different programmes in the system may all have different linkage profiles, and secondly because the diagram would also need to map the various (non- displacement-focused) social protection programmes serving the general population, and these would likely not be linked in any way to the displacement-focused humanitarian response. For this reason, we have: 29 ODI Working paper • Replaced the references to ‘(humanitarian) response’ with ‘(humanitarian) programming’ in the descriptions of the ‘type of linkage’ categories to clarify that the linkage only refers to a programme rather than the response as a whole. • Clarified that the framework diagram refers to ‘approaches to linking social protection and humanitarian assistance at the programme level’ and that the right-hand column refers to the programme’s contribution to benefits and drawbacks. This language aims to indicate that the actual benefits and drawbacks that emerge can only be understood when looking at the programme in combination with other programmes in the humanitarian and social protection systems. Lessons on the ‘benefits and drawbacks’ (now re-named the ‘contribution to benefits and drawbacks over time’) Column 3 in the initial framework showed a provisional list of ‘benefits and drawbacks’, hypothesising that the key outcomes of a given assistance approach related to both the effectiveness and equity of the approach for the affected populations (displaced and host), as well as the benefits and drawbacks for wider stakeholders (notably the accountability and acceptability of the approach, cost-efficiency, and sustainability). Our research tested this provisional list by examining the benefits and drawbacks that were found to emerge from the assistance models in place in our primary case studies and in our broader Global Evidence Review. The changes to column 3 are shown in Figure 8. Figure 8 Initial and final column 3 Benefits and Contribution to benefits and drawbacks drawbacks over time Effectiveness in addressing First order / direct Second order / indirect displaced popula- tion’s needs and Effectiveness and risks equity in Effectiveness in – meeting basic meeting host Wider economic needs and effects population’s needs addressing protection risks Wider social Equity – promoting effects broader realisation of rights/agency Wider political Cost and effects efficiency Cost and efficiency of Wider Accountability and institutional systems acceptibility to effects stakeholders Accountability and acceptibility Sustainability to all stakeholders 30 ODI Working paper Finding 1: It is important to recognise second-order/indirect impacts, alongside the first-order/direct benefits and drawbacks from the assistance provided Over the course of our project, it became increasingly clear that there is an important distinction to be made between the direct/first-order benefits and drawbacks of a particular approach – for the immediate stakeholders affected and in relation to the specific assistance provided – compared to the wider indirect/second-order benefits and drawbacks that might be generated. The first-order impacts relate to the effectiveness and equity of the assistance model for populations directly affected by displacement (displaced and host communities in the affected area), as well as for those directly financing/providing/ overseeing the assistance programme and displacement response. By contrast, the second-order impacts relate to the wider economic, social, political and institutional benefits and drawbacks that might be generated for the host economy, society and state more broadly. Effects on the local and national economy (wider economic impacts) The assistance itself may serve as an economic stimulus, increasing consumption of local goods and services. This was noted repeatedly in the Cameroon case studies, where the relatively high- value humanitarian assistance for displaced populations was felt to have stimulated local business. However, the decreasing value of humanitarian assistance over time reduced the potential economic gains, highlighting a likely disadvantage of aligning humanitarian transfers more closely with lower-value social protection systems. Across all three contexts, assistance was also associated with increasing access to financial accounts (which was evident both when the assistance was provided by humanitarian agencies and through the social protection system). Such increases may help stimulate access to financial products and services and subsequently support the growth of banking or mobile money markets. However, there was also evidence (particularly from Cameroon and Colombia) that recipients tend to immediately ‘cash out’ their transfers, suggesting relatively limited gains in terms of true financial inclusion. In Colombia it was also noted that the extent of Venezuelan participation in the social insurance system influenced the taxes and social security contributions that accrued to the host economy. However, these gains had not yet reached their full potential given the widespread informal barriers hindering Venezuelans’ (as well as many Colombians’) access to formal employment. 31 ODI Working paper Effects on relations between host and displaced populations (wider social impacts) Assistance provision is unlikely ever to be the primary determinant of relations between host and displaced populations. Nonetheless, the assistance approach was felt to play a role in influencing inter-community relations in all three case studies (as documented in greater depth in Lowe et al., 2022b). In Colombia, the adoption of a nationally led approach to assisting Venezuelans had heightened social tensions, since it was often perceived to come at the expense of vulnerable citizens (including IDPs). By contrast, in Cameroon, host populations’ perceptions of fairness did not seem to distinguish between assistance from humanitarian versus governmental sources, and there was little sense that governmental assistance – were it ever to be the dominant source of provision – should be restricted to citizens. Where assistance was being provided to host populations in Cameroon (either from humanitarian agencies or through the limited government social protection system), there was a perception that this had only arrived because of the displaced population’s presence and the international community’s involvement. Such assistance was greatly appreciated, although displaced and host respondents alike argued that more assistance should be provided to the host population given how little was available to them (offering strong evidence of the need for and likely benefits of an approach ‘stepping up’ provision for hosts alongside coverage of the displaced). Meanwhile in Greece, the camp-based humanitarian assistance model was to some extent felt to have contributed to the ‘ghettoisation’ of the displaced population in geographically isolated housing, thereby limiting both interaction and integration with host communities. The value of assistance given to displaced populations was at times felt to contribute to social tensions, since some host community respondents viewed it as unfair that asylum seekers received cash assistance in addition to subsidised accommodation when unemployed Greek citizens did not. Effects on social contract/state–society relations (wider political impacts) As well as influencing displaced–host relations, our research also indicated several potential effects of the assistance approach on state–society relations. This impact can be categorised as either political or social, since it refers to the relations between people and the state (also known as the social contract, particularly when referring to relations between the state and its citizens). As documented in Lowe et al. (2022b), our research found some evidence that state involvement in effective provision for displaced people can potentially lead to improved relations between displaced people and the state. Assistance may help strengthen trust in the government (as was shown in the Colombia and, to a lesser extent, Cameroon survey samples – even though assistance for the latter was primarily humanitarian). In some cases, transfers associated with government may also help to foster a greater sense of belonging and affiliation to the nation-state. Where 32 ODI Working paper enshrined in comprehensive legal and policy frameworks (as under the Victims’ Law, in Colombia), assistance may also help strengthen recipients’ voice, agency and consciousness of their rights, as well as their access to wider state services. However, these effects are only likely when people are aware of their entitlements and able to access them fairly and effectively. In the Greek case study, social assistance recipients in the host population had lower levels of trust in the government than non-recipients; the same was also true for refugees receiving assistance (primarily through the International Organisation for Migration-run, Greek government-supported HELIOS programme). This suggests that relations between recipients and the state are unlikely to improve, and may even worsen, where such awareness is marginal or provision felt to be flawed. Ultimately, recipients’ trust in government will be largely mediated by their prior and wider interactions with government officials. Moreover, possible improvements in state–society relations from assistance provision may also be undermined by the potential for improved assistance for one group to be regarded as evidence of neglect of other groups. This was seen in Colombia when state provision for Venezuelans was perceived to take government resources or attention directly away from vulnerable citizens. As for horizontal cohesion, this outcome is far more likely if there is a pre-existing sense that the state has overlooked the needs of vulnerable citizens. Finally, our research noted that recipients and non-recipients do not necessarily know who is financing or delivering a particular assistance programme. Greater state involvement in assistance provision may therefore not necessarily be ‘credited’ as such (preventing both the potential benefits and downsides of enhanced government provision from emerging). Effects on strength of state systems (capacity, shock-responsiveness, etc. – wider institutional impacts) When discussing the potential benefits and drawbacks of assistance, wider stakeholders from both governmental and non-governmental agencies also identified important institutional impacts from linking humanitarian assistance and social protection. Such linkages were generally felt to be important for strengthening both the capacity and shock-responsiveness of the social protection system. In Colombia, for example, closer alignment of international humanitarian assistance and the government social protection system was seen (by government and humanitarian agencies alike) to help strengthen the social protection system, including by using humanitarian agencies’ databases and capacity to help update and fill gaps in the national social registry (the SISBÉN). More generally, cross-system linkages were seen as a way to increase the government’s capacity to improve its overall response to the displacement influx, by identifying gaps in provision and better understanding and addressing the population’s needs. 33 ODI Working paper Proposed revision(s) in the final framework Based on the findings above, we have: • Added a second column within the final ‘benefits and drawbacks’ section, to highlight the potential second-order/indirect effects that can emerge from a particular assistance approach. While this column could be organised more generally under the headings of wider economic, social, political and institutional benefits, we have also opted to use the sub-headings above (effects on local and national economy, on inter-community relations, on state-society relations, and on the strength of state systems) to make the benefits and drawbacks more concrete. • Drawn an arrow between the second-order (wider economic, social, political and institutional) effects and the first-order effects, to demonstrate that the effects on the host economy, society, political settlement and state systems more broadly will also influence the accountability and acceptability for wider stakeholders, as well as the effectiveness and equity of the approach. At the system level, the wider economic and institutional/system- strengthening effects may also influence the cost and efficiency of systems over time. Finding 2: The effectiveness of assistance depends both on its ability to meet basic needs/address immediate protection risks, and its ability to promote the broader realisation of the affected populations’ agency and access to their rights When considering the effectiveness of the assistance approach, we repeatedly noted a need to distinguish – and acknowledge the tensions – between meeting urgent needs (e.g., meeting direct consumption needs and addressing immediate protection risks, such as exposure to exploitation and violence) and supporting the realisation of recipients’ broader rights (including their social inclusion and economic agency). In Colombia, recipients generally felt that assistance met the former objectives quite well but failed to contribute to the latter, leaving recipients unable to break the cycle of socio-economic disadvantage. In Greece too, assistance was felt to focus narrowly on ensuring the survival of displaced recipients, but not on providing access to the necessary resources to enable them to thrive independently in their new place of residence. There, it was noted that assistance should also be complemented with services to facilitate access to the labour market, education and training, and empowerment. Proposed revision(s) in the final framework We have therefore developed two distinct sub-criteria within our effectiveness dimension, namely: (1) effectiveness in meeting basic needs and addressing immediate protection risks; and (2) effectiveness in promoting broader realisation of rights and agency. 34 ODI Working paper Finding 3: Equity is best explored together with the effectiveness criterion, since it is another dimension with which to measure the approach’s ability to meet needs/promote broader realisation of rights We found it challenging to explore equity as its own criterion, since in many ways it relates to the same dimensions as effectiveness (ability to meet basic needs and address immediate protection risks; ability to promote broader realisation of rights and agency), but considers whether the approach has been effective in an equitable manner – for all groups, and particularly for marginalised sub-groups. We do not propose to consider perceptions about equity here from a social cohesion perspective (i.e., perceptions about whether the distribution of assistance between hosts and displaced populations is fair). These considerations are instead now covered by the indirect social effects considering inter-community relations/horizontal social cohesion. Instead, the Equity category focuses on actual equity in the distribution of assistance given the relative needs of different populations. Proposed revision(s) in the final framework Based on this, we have combined effectiveness and equity in the same category, recognising that equity/inclusion considerations are a key aspect of the overall effectiveness of a given approach. Finding 4: Effectiveness and equity should be considered in relation to displacement-affected populations as a whole (referring to both displaced and host populations for systems level analysis, and to whichever population a specific intervention aims to serve for programme level analysis) ‘Displacement-affected populations’ refers both to the displaced and to those in the host communities who are impacted by a displacement influx. When considering the effect of linking humanitarian assistance and social protection at the programme level, we felt that it made more sense to refer to ‘effectiveness for displacement-affected populations’ together, rather than looking at displaced and host populations separately. This is because some programmes will be aimed only at a specific group within the displacement-affected population, for example a specific population of refugees. While that programme can consider its indirect contribution to wider social impacts (such as relations between displaced and host communities), it should not have ‘effectiveness in meeting needs / promoting realisation of rights for host communities’ as a primary aim or be assessed against that as a primary outcome. Proposed revision(s) in the final framework: Based on the finding above, we have grouped the previously separate ‘effectiveness for displaced population’ and ‘effectiveness for host population’ into a single dimension 35 ODI Working paper looking at the effectiveness and equity for displacement-affected populations. When conducting the analysis at the programme level, this allows programmes to consider their contribution to meeting needs and to promoting rights and agency among whichever target population is most relevant (i.e. the displaced and/or host populations). When considering the effects of linkages at the system level, displacement-affected populations would by definition refer to both displaced and host populations, and would require an analysis of how effective the overall systems – and linkages between them – are in meeting the needs equitably across and within both population groups. Finding 5: Sustainability should not be considered as a single criterion – it instead relates to the ability to sustain the other benefits and mitigate the other drawbacks over time across all areas In our initial framework, we considered two aspects under the ‘sustainability’ dimension: a. The viability of sustaining the approach for as long as needs remain b. The approach’s contribution to bringing about long-term solutions to current and future displacement situations However, aspect (a) is difficult to gauge and depends on many different factors, including: • Political and social acceptability – which is captured in the ‘accountability and acceptability’ dimension, as well as the second-order wider economic, social and political effects • Nature of funding – which is better captured by assessing ‘cost and efficiency’ over time, alongside ‘accountability and acceptability’ over time to the stakeholders who are funding or are expected eventually to fund the response • Future capacity to respond – which is better captured in the second-order wider institutional impacts/system-strengthening dimension. Furthermore, aspect (b) is now captured under ‘promoting broader agency and realisation of rights’ and second-order dimensions. Proposed revision(s) in the final framework Based on this, we have eliminated the sustainability criterion from the ‘benefits and drawbacks’ list but have highlighted the need to consider how other dimensions (effectiveness, acceptability, etc.) may change in the short versus medium versus long term, through the addition of the phrase ‘over time’ to the ‘benefits and drawbacks’ heading. 36 ODI Working paper 4 The final framework The final framework is presented below in Figure 9, as well as in an accompanying paper without the initial framework or the details of how that framework was refined (Lowe et al., 2022a). Figure 9 Approaches for linking humanitarian assistance and social protection in displacement settings, at the programme level Determining factors and actors Approaches to linking social protection Contribution to benefits and drawbacks and humanitarian assistance over time Linkage options and connection points Promoting collaboration FACTORS ACTORS First order / direct Second order / indirect PARALLEL ALIGNING LEVERAGING STATE-LED Initial context Actors involved and their: Effectiveness and equity in Nature of displacement – motivation, incentives and – meeting basic Host context: Wider economic decision logics needs and effects addressing – structural features, – relationships, Humanitarian-led, Humanitarian-led, Humanitarian-led, National social protection risks state of existing SP & Wider social interactions and not linked to state using state SP as a involving the state protection system humanitarian systems effects balance of SP reference point and using or used – promoting power by SP broader realisation – laws, institutions, of rights/agency Wider political norms governing Tailored for displaced people vs standard SP effects government and international response Cost and Policy efficiency of Wider Connection points: financing, legal and policy frameworks, systems institutional Wider government and governance and coordination, functional and technical capacities effects and international Accountability humanitarian response and acceptibility strategies Programme design to all Connection points: vulnerability and risk profiling,  eligibility criteria, stakeholders objectives/type/package of benefits/services, transfer amount Current situation of affected populations Administration Connection points: outreach, registration, enrolment, provision of benefit/service, accountability, case management (incl. protection), M&E, information management References Gray Meral, A. and Both, N. (2021) Social protection and forcibly displaced people: a literature review. London: ODI (https://odi.org/en/publications/social-protection-and-forcibly-displaced- people-a-literature-review). Ham, A., García, S., Dedios, M.C. et al. (2022) Social protection responses to forced displacement in Colombia. London: ODI (https://odi.org/en/publications/social-protection- responses-to-forced-displacement-in-colombia). Harris, D. (2013) Applied political economy analysis: a problem-driven framework. London: ODI (www.odi.org/sites/odi.org.uk/files/odi-assets/publications-opinion-files/8334.pdf ). Levine, S., Socpa, A., Both, N. et al. (2022) Integrating assistance to the displaced into a social protection system in Cameroon: the ideal, but in whose interests? London: ODI (https://odi. org/en/publications/integrating-assistance-to-the-displaced-into-a-social-protection-system- in-cameroon-an-ideal-but-in-whose-interests). Longhurst, D. and Smith, G. (2020) Identifying Practical options for linking humanitarian assistance and social protection in the COVID-19 response. Social Protection Approaches to COVID-19 Expert Advice Service (SPACE). Lowe, C., Cherrier, C. and Holmes, R. (2022a) Linking humanitarian assistance and social protection in response to forced displacement: an analytical framework. London: ODI ( https://odi.org/en/publications/linking-humanitarian-assistance-and-social-protection-in- response-to-forced-displacement-an-analytical-framework/). Lowe, C., Salomon, H., Hagen-Zanker, J. and Gray Meral, A. (2022b) Humanitarian assistance and social protection in response to forced displacement: effects on social cohesion. London: ODI (https://odi.org/en/publications/humanitarian-assistance-and-social-protection-in- contexts-of-forced-displacement-effects-on-social-cohesion). O’Brien, C., Holmes, R. and Scott, Z. (2018) Shock-responsive social protection systems toolkit appraising the use of social protection in addressing large-scale shocks. Oxford: OPM (www.opml.co.uk/projects/shock-responsive-social-protection-systems). OCHA – United Nations Office for the Coordination of Humanitarian Affairs (2016) The grand bargain: A shared commitment to better service people in need. Istanbul: World Humanitarian Summit (www.agendaforhumanity.org/sites/default/files/resources/2018/Jan/ Grand_Bargain_final_22_May_FINAL-2.pdf ). Peterman, A., de Hoop, J., Cuesta, J. and Yuster, A. (2018) Evidence on social protection in contexts of fragility and forced displacement. Florence: UNICEF Innocenti (www.unicef-irc.org/ publications/pdf/IRB-2018-22.pdf ). Proto Thema (2020) ‘Πρόσφυγες: Τέλος τα επιδόματα και οι παροχές σε όσους έχουν άσυλο λέει ο Μηταράκης’ 7 March (www.protothema.gr/greece/article/982128/prosfuges-telos-ta- epidomata-kai-oi-parohes-se-osous-ehoun-asulo-leei-o-mitarakis). Seyfert, K., Barca, V., Gentilini, U. et al. (2019) Unbundled: a framework for connecting safety nets in refugee settings. Social protection and jobs discussion paper 1935. Washington DC: World Bank (http://documents.worldbank.org/curated/en/970701569569181651/Unbundled-A- Framework-for-Connecting-Safety-Nets-and-Humanitarian-Assistance-in-Refugee-Settings). Smith, G. (2021) Deciding when and how to link humanitarian assistance and social protection: guidance and tools for response analysis. Social Protection Approaches to COVID-19 Expert Advice Service (SPACE). SPACE (2021) Social protection in the COVID 19 response and beyond. Video (www.youtube.com/watch?v=EIGlWkAnCu4). SPIAC-B (n.d.) ‘SPIAC-B Working Group on Linking Humanitarian (Cash) Assistance and Social Protection’. Webpage (https://socialprotection.org/connect/communities/social- protection-crisis-contexts/spiac-b-working-group-linking-humanitarian). Tramountanis, A., Linardis, A., Mouriki, A., et al. (2022) Humanitarian assistance and social protection responses to the forcibly displaced in Greece. London: ODI (https://odi.org/en/ publications/humanitarian-assistance-and-social-protection-responses-to-the-forcibly- displaced-in-greece). UNHCR (2019) Aligning humanitarian cash assistance with national social safety nets in refugee settings: key considerations and learning. Geneva: UNHCR (www.unhcr.org/5cc011417). UNHCR (2022) ‘Ukraine situation: Flash Update 11’. 5 May (https://data2.unhcr.org/en/ documents/details/92542). UNHCR and IOM (2021) ‘UNHCR and IOM welcome Colombia’s decision to regularize Venezuelan refugees and migrants’. UNHCR–IOM Joint Press Release, 8 February (www.unhcr.org/uk/news/press/2021/2/60214cf74/unhcr-iom-welcome-colombias-decision- regularize-venezuelan-refugees-migrants.html). UNICEF (2019) Programme guidance: strengthening shock responsive social protection systems. New York: UNICEF (www.unicef.org/media/63846/file). WFP – World Food Programme (2019) ‘WFP basic country capacity assessment: shock- responsive social protection’. Online tool (https://socialprotection.org/discover/ multimedia/wfp-basic-country-capacity-assessment-shock-responsive-social-protection).