Moving targets: social protection as a link between humanitarianism, development and displacement

Despite the widespread concern with social protection in the field of development, it has had little impact on displacement until very recently. UNHCR has had a Social Protection Unit since 2009, but social protection is barely mentioned in either the Global Compact on Refugees or the Global Compact for Safe, Orderly and Regular Migration, both agreed in 2018. The period since 2018 has seen a growth of research interest and new policy development. This represents the latest appearance of the humanitarian-development nexus in the field of displacement. This nexus is a long-standing one and is inherent in the term ‘durable solution’ as the only effective end of displacement. Unfortunately, there are few recognised successes, and durable solutions are enjoyed by only a tiny proportion of the growing number of displaced people in the world today. This makes renewed attention to the humanitarian-development nexus all the more urgent. In this paper we review the relationship between a humanitarian response to initial displacement and longer-term development planning, as well as the recent range of research and policy responses in this field. These demonstrate significant potential of social protection. We go on to consider six areas of developing theorisation in order to inform what would constitute success in the expanding inclusion of displaced people in social protection programmes or systems. We conclude with four suggestions where further research in this area can help to determine how and if the potential for social protection to offer more sustainable responses to displacement is being realised.


Introduction
Social protection has created much enthusiasm for its contribution to supporting people affected by forced displacement. The connection between social protection and responses to displacement has been widely discussed in a range of global policy forums (e.g. UNICEF 2017) and has several research programmes devoted to it ODI 2020). Yet it is sufficiently new that a 2016 systematic review identified only five publications exploring the link between social protection and displacement (Doocy and Tappis 2016) and it barely received a mention in the 2018 Global Compact on Refugees (GCR) (UN 2018). Despite this novelty, there are antecedents to the current interest of social protection in relation to forced displacement. These go back to the 1951 Convention relating to the Status of Refugees (1951 Refugee Convention), which is startlingly progressive on economic rights. But there have been regular attempts since then to link the broadly humanitarian architecture of the regime governing international responses to forced displacement to longer-term concerns, usually without any great success (Crisp 2001). The meaning of social protection varies widely. In this paper we draw on definitions that go well beyond a minimal safety net to provide wider, longer-term support for more substantial social transformation (Devereux and Sabates-Wheeler 2004). This includes social assistance, either in cash or in kind, provided by public, private or community actors, as well as access to education, health care and work.
The current round of global policy initiatives to link humanitarianism and development in response to forced displacement began with the 2016 New York Declaration for Refugees and Migrants. Two subsequent compacts were signed in 2018 -The GCR and the Global Compact for Safe, Orderly and Regular Migration (GCM). This separates refugees and migrants into distinct policy regimes, failing to recognise the multiple interconnections between the two. Nevertheless, they are a serious attempt to tackle the growing duration of displacement with near universal support from United Nations (UN) member states and have been at least partially welcomed by the various stakeholders concerned. The GCR references the three recognised 'durable' solutions that mark the end of refugee status: voluntary return, local integration and resettlement. A very small proportion of refugees now achieve these durable solutions; they have never been directed at internally displaced people (IDPs), who now significantly outnumber refugees. We use the phrase 'forced displacement' throughout this paper to recognise the interconnections between these different groups of displaced people. Forced displacement covers all migrants who are obliged to move because of targeted political violence, massive civil unrest, environmental disasters or large infrastructure projects, whether they have crossed an international border or not. This encompasses international definitions of refugees, under both the 1951 Refugee Convention and the 1969 Organisation of African Unity Convention Governing the Specific Aspects of Refugee Problems in Africa supplementary definition, but also recognises the role which environmental disasters increasingly play in displacement. It is a category that relates to both compacts.
The failure to achieve durable solutions for most forcibly displaced people and associated rise in protracted displacement are quietly recognised by the compacts' introduction of other ways of addressing displacement, which fall short of durable solutions, such as to 'enhance refugee self-reliance' (GCR para. 7). This emphasis has been widely criticised as a justification for reducing international support for displaced people (Morris 2020) but this is the clearest link to subsequent work on social protection. Nevertheless, caution is required in interpreting these policy initiatives. If they are to be successful, global attempts to resolve displacement crises must respond to the very different concerns of wealthier states of the global North, which retain dominance in decision-making and finance, and those of the global South, where most displaced people are hosted. The substantial new initiatives to support displaced people in the GCR are balanced by a clear objective to support them where they are and dissuade them from moving to the global North. This reveals the less explicit but widely noted objective of containing displacement crises in the global South. This has been going on for some time; indeed, the increasing restrictions on border crossing by forcibly displaced people are one explanation for the rise in IDPs since the 1990s (Dubernet 2017).
The focus on containment results from the dominance of wealthier states in the drafting process of both compacts. Forced displacement is managed at international levels in an almost permanent atmosphere of crisis. This is partly fuelled by rising numbers of displaced people, but as in the case of the regional European 'migration crisis' of 2015, the language of crisis can be more easily explained by a failure of response. The charge levelled by some (e.g. Betts and Collier 2017) that these failures are largely because international protection infrastructure is stuck in the past is not entirely fair. Granted, as it has just turned 70, the 1951 Refugee Convention is straining to remain relevant, but the reason it has not been opened for renegotiation is not stubborn attachment, but fear that in the current political climate it would be replaced with something that falls even further short (McAdam 2019). The more troubling lack of evolution is not tied directly to the text of the 1951 convention but to the concentration of decision-making capacity in the same limited number of wealthy countries that were central to drafting it in the immediate post-war period (Chimni 2018). This means that the international system remains orientated around the interests of those states, overlooking any responsibilities they may have for initiating displacement through armed intervention or helping to legitimate the disproportionately small role they play in offering protection, for example.
A more direct explanation for the current sense of crisis is that displacement contexts are becoming more complex and the political causes more intractable. Since 2006, when the World Bank began compiling an annual list of fragile and conflict-affected situations (FCAS) (World Bank 2021), this list has overlapped very significantly with data from the United Nations High Commissioner for Refugees (UNHCR) on the main countries of origin for refugees and Internal Displacement Monitoring Centre (IDMC) analysis of conflictrelated internal displacement. Displacement related to disasters has a slightly different geography, according to IDMC (2021), though there is still significant overlap between conflict-and disaster-related displacement, particularly including significant refugee destination countries in the analysis, such as Pakistan. The bulk of FCAS countries are low income, but the spread of conflict to middle-income countries over the past decade and the growing numbers of refugees and displaced people in urban areas makes this picture more complex. The increasing differentiation and complexity of displacement crises makes any kind of standard response particularly challenging.
A third explanation for the ongoing sense of crisis in displacement policy is the peculiarly fixed nature of responses. Provision of support for displaced people, from shelter to schooling has always assumed forcibly displaced people stay in one place until they are ready to return to their country of origin (Bakewell 2008). For decades, UNHCR was wedded to the 'care and maintenance' model of service delivery in campssometimes camps which were officially closed, in that residents were forbidden from leaving. Although this remains in place for the substantial minority of the world's refugees, UNHCR and associated agencies have developed much more progressive alternatives to camps and more innovative ways of engaging with refugees in urban areas. Nevertheless, these retain the basic assumption that displaced people do not move once they have been displaced. Movement after initial displacement has been overlooked or viewed with suspicion, often labelled 'secondary movement' by those who seek to eliminate it, though it typically reflects a lack of protection in earlier potential destinations (Garlick 2016). Displaced people who have abandoned the hope of a future in camps, as years turn to decades, are increasingly rejecting the care and maintenance model. A more dynamic, mobile lifestyle has significant attractions, particularly for younger people, enabling them to find a location which meets their needs. A key attraction of social protection is the potential it offers to mirror this mobility. If social assistance were linked to the individual, rather than the location, individuals would be able to access support where they needed it.
The focus of this paper is on social protection as a mechanism for linking humanitarian and development responses to displacement; this has enormous potential, particularly when social protection means something more transformative than a minimal safety net and encompasses eligibility to education, health care and work. An important measure of the success of social protection will be the extent to which a degree of agency can be restored to displaced people. This goes beyond the 'self-reliance' model to support genuine self-determination, providing choice and helping to address inequalities based on gender, age, ethnicity and ideally legal status. It involves recognising that mobility is a resource that may be used to create further opportunities. The paper falls into three sections. The following section considers in more detail current policy initiatives to bring humanitarian and development responses to forced displacement together. Section two turns to the even more recent flourishing of research into social protection and displacement, significantly occurring beyond the GCR. The final section examines theoretical engagement with social protection and displacement to direct further analysis. The paper concludes with indications of a research agenda to be developed in this area.

Linking humanitarianism, development and displacement
In 2015, the UN General Assembly described the Sustainable Development Goals (SDGs) as 'comprehensive, far-reaching and people centred', providing a succinct summary of current attitudes to international development. The World Humanitarian Summit the following year illustrated the continuing divide between humanitarianism and development, but also coined the language for current attempts to close it, in the humanitarian-development nexus (HDN). The policies resulting from both these initiatives consider refugees in some way. Only one of the 169 SDG targets is dedicated to migration -the disappointing SDG 10.7, which makes controlling migration a development priority -and refugees were not even mentioned in the initial set of indicators. However, after lobbying from UNHCR an additional indicator was added to target SDG 16.3 (rule of law): the number of refugees from a given country, compared to the population of that country (Nahmias and Baal 2019). The Agenda for Humanity (AfH), which arose from the World Humanitarian Summit, gives rather more prominence to migration and displacement, with one of the 24 'strategic transformations' devoted to each of them. On displacement, the goal is to reduce displacement by 50 per cent by 2030. The Grand Bargain is the second agreement related to the World Humanitarian Summit, which provides important context for responses to forced displacement. Of the ten work streams, a number are of direct relevance to forced displacement and social protection. Most obviously, the aim to enhance links between humanitarian and development actors, which cuts across the remaining nine themes. Others include the localisation agenda, increased use of cash-based programming and enhanced participation. These provide the framework for ongoing discussions on both national and international responses to forced displacement.

Changes to displacement are making it less of a humanitarian issue
The broad aspirational goals in the SDGs and the AfH give a very similar impression of displacement. Neither allow a clear sense of the difference between responding to displacement from a development or humanitarian perspective. The only distinction is that the SDGs refer to refugee movement and the AfH to displacement, but this is more about the greater degree of agreement on refugee movement statistics. The architecture of the global displacement regime has always been primarily humanitarian. UNHCR defines itself as a 'humanitarian organisation'. Yet the basis for this humanitarian framing is becoming increasingly questionable. The humanitarian focus relies on a characterisation of forced displacement as an exceptional, unpredictable, one-off event requiring an urgent response from external actors. As protracted situations account for an ever-larger share of global displacement, the characterisation of displacement as exceptional and unforeseen is increasingly inaccurate. Development is becoming more relevant in responding to displacement as it is currently experienced by the large majority of forcibly displaced people.
The sheer numbers of people forcibly displaced undermines any idea that it is exceptional. One of the headlines of UNHCR statistical reports since 2019 has been that 1 per cent of the world's population is forcibly displaced. And this is not unprecedented; historically, various major displacement events have resulted in what were probably similar shares of the world's population being displaced, though poor statistical records make this difficult to verify. The incredibly uneven distribution means that in some countries forced displacement has been normalised. Climate events are widely projected to further increase forced displacement, though measuring climate-related displacement is uncertain and controversial. Nevertheless, well-respected organisations such as the IDMC are attempting to refine methods. A 2021 IDMC report details displacements affecting more than 30 million people in 2020 that were attributable to climate events (IDMC 2021). This includes populations in wealthier countries that, having hosted refugees for many decades, are now experiencing displacement themselves.
Unpredictability is a second reason justifying the urgency of a humanitarian response and the capacity to respond quickly is vital in humanitarian planning. Yet, unpredictability does not characterise most current displacement situations. More than 80 per cent of refugees are in situations of protracted displacement. Although populations affected by long-term crises vary, the number of people experiencing displacement is mostly highly predictable and does not require fast response times. Urgent responses are only required for the minority of newly displaced people. Even among situations of new displacement, many are both predicted and planned for by the international community. Some of the most significant international displacements in recent decades were the predicted result of Western military intervention and involved detailed modelling of displacement scenarios before armed interventions even began (Vine et al. 2020). In most current situations of forced displacement, the task of international response is not to respond urgently to unforeseen events, but to continue to support those who were displaced decades ago. Responding to these long-term events through humanitarian channels has long been recognised as inappropriate.
This relates to the third reason undermining the characterisation of forced displacement as exceptional. Individuals whose lives are disrupted by forced displacement need support from somewhere. Yet displacement often affects the same people multiple times and previous experience makes a difference. In northern Sri Lanka, humanitarian workers report experiences of arriving within 48 hours of a large-scale displacement to find teachers already taking classes in makeshift camps; displacements were frequent and the community had developed a degree of expertise in dealing with them. This is not to suggest that communities such as this do not need external support, but any intervention must start from an appreciation of what responses are already proving effective and try to support them as much as possible. Multiple displacements may occur sequentially, as conflict spreads; or they may occur in a circular pattern, as individuals return to their original homes when conditions permit, only to be displaced again. Circular displacement appears to be more characteristic of displacement associated with short-term weather events (as opposed to slow-onset environmental change), though it is not uncommon in political crises. This suggests it may become increasingly common as weather-related displacement becomes more frequent.
The characterisation of displacement as exceptional and unforeseen is no longer as accurate as it perhaps once was. Although new displacement remains just as difficult to predict, newly displaced people make up a small minority of total global displacement. Most displaced people were already displaced five years ago; if current trends continue, we can be fairly sure that most of them will remain displaced in five years' time. The normalisation and predictability of most global displacement means that a humanitarian response is required much less often. There are situations in which urgency is still required, but for most people a development response is much more appropriate. The turn to development concerns in relation to displacement reflects the changing nature of displacement; the only surprise is that it is happening so slowly.

The current global policy framework on displacement and the HDN
A range of international agreements make some reference to universal social protection or specifically to social protection for forcibly displaced people (Table 1). As the numbers of signatories suggest, support for these agreements is far from universal, but they highlight how well established the basic principles are.
What is missing is a broader global initiative to implement these policies. There is plenty of evidence of such initiatives since 2016. The GCR is the most significant current influence on international policy on forcibly displaced people, but it is not the only one. The GCM also offers provisions that could potentially support provision for displaced people, particularly those displaced by climate events (Kälin 2019). Both compacts have received significant support from states: the GCR has 181 signatories and the GCM has 164. They both occur in a broader context of shifting attitudes to development itself. The SDGs and AfH set only broad, aspirational goals in relation to displacement, as already discussed. The Grand Bargain provides a clearer framework that is already being used to support innovative steps towards social protection for displaced people. Several initiatives beyond the GCR also identify best practices and develop research on using social protection to respond to displacement.
The 2016 New York Declaration for Refugees and Migrants is framed very broadly (UNGA 2016). The first paragraph recognises that some people move for economic activities, though movement also occurs 'to escape armed conflict, poverty, food insecurity, persecution, terrorism, or human rights violations and abuses', going on to note 'the adverse effects of climate change, natural disasters… or other environmental factors', finally recognising that movement may occur for a combination of these factors. Later (para. 20) the declaration connects those crossing an international border with internal displacement, highlighting the need for protection for both groups. This all-encompassing range of motivations for migration goes well beyond a strict definition of refugees to cover what -echoing the International Association for the Study of Forced Migration -we are calling forced migration in this paper. The declaration and resulting compacts have been criticised for leaving development undefined (Zetter 2019), though there are some indications as to how they understand development. The declaration emphasises that 'our challenge is above all moral and humanitarian' (para. 10), but immediately goes on to express the need for 'long term and sustainable solutions', which we interpret as a reference to development priorities. The declaration references the SDGs and links their success to addressing root causes of refugee movement (para. 17). It then goes on to set out three groups of commitments for both refugees and migrants together and separately. There is only a single reference to social protection in the declaration, at paragraph 83: 'We will also develop national strategies for the protection of refugees within the framework of national social protection systems, as appropriate.' This brief reference is not picked up in the GCR, but it captures the direction of most current initiatives to integrate support for refugees into national systems.
The Comprehensive Refugee Response Framework (CRRF) is Annex 1 of the New York Declaration and was incorporated into the GCR unchanged. In 19 paragraphs, it sets out the basis for a burden-sharing response for 'large movements of refugees' (para. 4), though both the CRRF and the GCR have been criticised for leaving 'large' undefined (Hathaway 2018). The CRRF sets out a progressive framework for a broad multistakeholder approach 'including national and local authorities, international organisations, international financial institutions, regional organisations, regional coordination and partnership mechanisms, civil society partners, including faith-based organisations and academia, the private sector, media and the refugees themselves' (para. 2). Although the CRRF does not mention social protection directly, there is plenty of language to support initiatives in that direction. States, multilateral donors and the private sector are called upon to support 'cash-based delivery mechanisms' (para. 6f). In addition, host states, in cooperation with UNHCR should 'deliver assistance to the extent possible through appropriate national and local service providers' (para. 7b). The three traditional durable solutions -voluntary return, local integration and resettlement -are all referenced: voluntary return, which receives two paragraphs (11 and 12), is emphasised as the 'primary goal'; local integration is not mentioned as such, though paragraph 13 gestures towards measures for local integration, detailing responsibilities of host states; and self-reliance is mentioned twice. The language is quite definite: in addition to providing the legal right to stay, host states would 'take measures to foster selfreliance' (11b) and 'invest in building… self-reliance' (11d). The replacement of local integration with selfreliance is characteristic of the CRRF in general. Although potentially positive, it shifts much of the burden of integration from states to displaced people themselves. Resettlement is referenced more briefly and the language is much weaker; third countries are only requested to 'consider' expanding the criteria for resettlement (para. 14c). 'Complementary pathways for admission' is almost given the status of a fourth durable solution (para 10) to include 'medical evacuation and humanitarian admission programmes, family reunification and opportunities for skilled migration, labour mobility and education' (para. 14a). There is a contrast between the non-committal language ('consider') and the ambition of the objective to deliver resettlement on a scale that would meet UNHCR's identified resettlement needs (para. 16); even at the maximum extent of resettlement, it barely meets 10 per cent of resettlement needs.
This is developed further in the GCR, which comprises three parts. Part 1 is a nine-paragraph introduction; the CRRF (unchanged) forms part 2; (para. 10) and part 3 is a 97-paragraph programme of action, divided into two further parts: 'Arrangements for burden and responsibility sharing' and 'Areas in need of support'. Like the CRRF, the GCR does not focus on all refugees but on 'countries particularly affected by a large refugee movement, or a protracted refugee situation' (para. 11). Although the World Bank term 'FCAS' is not used, such situations are vastly overrepresented within countries experiencing large and protracted refugee movements. Like the CRRF, the GCR goes on to recognise the difficulty of exclusively focusing on refugees: While the CRRF relates specifically to large refugee situations, population movements are not necessarily homogenous, and may be of a composite character. Some may be large movements involving both refugees and others on the move; other situations may involve refugees and internally displaced persons; and, in certain situations, external forced displacement may result from suddenonset natural disasters and environmental degradation. (UN 2018: 6) This captures the difficulty of any attempt to draw clear divisions between the wide-ranging and interconnected experiences of forced displacement. The GCR sets out to provide a basis for burden-sharing (para. 3) guided by four objectives (para. 7), to: (1) ease pressure on host countries; (2) enhance refugee self-reliance; (3) expand access to third-country solutions; and (4) support conditions to allow return.
The GCR includes a single reference to social protection (para. 81) in the context of food security and nutrition. This does not advance the framework for using national and local service providers already set out in the CRRF; indeed, Nimeh et al. (2020: 2) argue that 'the Global Compact [on Refugees] fails to recognise the role for social protection to support self-reliance from a durable perspective'. The Programme of Action of the GCR does add more detail on the broader HDN. The GCR recognises the barriers to a smooth transition between humanitarian and development actors. For example, paragraph 32 considers humanitarian and development actors under separate bullet points, with different overviews of their activities. Paragraph 35 states simply that 'humanitarian and development actors will work together'; the fact that such a straightforward statement is even necessary implies a recognition that such cooperation is not guaranteed. Later, the GCR even suggests that the use of development to tackle refugee crises is really in its infancy: 'there is also increasing recognition of the development challenges posed by large refugee situations and the advantages of shared and inclusive economic growth in refugee-hosting areas' (para. 64). It seems odd that this should need repeating in such a high-level document since this recognition can be traced to the 1970s (Crisp 2001). Paragraph 64 makes brief reference to the SDGs. The following paragraph considers both development and humanitarian actors, emphasising that 'development actors will work in a complementary manner to humanitarian assistance' (para. 65) and the next paragraph opens: 'humanitarian assistance remains needs driven' (para. 66). The overall impression is of a document put together by humanitarian actors only grudgingly acknowledging the new role of development. Rather than resolving the barriers that have separated humanitarian and development interventions involving displaced people, the GCR provides new evidence of the nature of the barriers that are still to be overcome.
Whereas the GCR mostly draws together accepted approaches and some recognised best practices, the GCM is much more aspirational and has received significant praise. It expects more of states, reflected in the lower signatory rate (164 vs 181 for the GCR). Although it is designed to be clearly separate from the GCR, several provisions address climate-related displacement directly. Specifically, articles 18.H (share information to better map and predict migration based on climate change and environmental degradation), 18.I (develop adaptation and resilience strategies that prioritise the country of origin), 18.J (factor in human displacement in disaster-preparedness strategies) and 18.K (support climate-displaced persons at sub-regional and regional levels).
These pieces of legislation represent the highest level of engagement on these issues -the New York Declaration, CRRF, GCR and Migrant Compact are a powerful set of guidelines for the next decade at least. In this they are valuable, but they are only part of the picture. The compacts have received a great deal of analysis, including special issues of the Journal of Refugee Law, International Migration and Refuge, as well as many additional articles. This early analysis provides a mixed picture. This is not surprising, since the compacts provide only a framework with relatively few details. There is plenty to like about them, including much progressive language in both. But there is also a lot of caution, highlighting the containment focus that underlies the GCR; in particular, worries about lack of specific targets, lots of meetings, vague language and warnings that all this high-level policy focus may not be going anywhere. The documents do indicate a substantial degree of political will at the highest levels. This is not nothing, but without specifics it is not enough. The test will be how the compacts are applied. They have been drawn up by a particularly powerful group of actors: international organisations, multilateral donors and primarily wealthy states. They regularly call for progressive partnerships with other actors, including host states, states of origin, municipalities, private organisations, community groups and refugees themselves. Ultimately, the extent to which the policies set out in these documents can be carried through will depend on the success of those partnerships.
Beyond the HDN, this paper considers the prospects for a social protection response to displacement. There are plenty of openings for this, particularly in the CRRF, and much of the language around mainstreaming social protection for refugees in national mechanisms of host states highlights the direction of travel in this area. Still, social protection is not a major focus of any of the documents and presents plenty of ongoing challenges. The next section turns to these challenges, reviewing the analysis of current best practices and the beginnings of systematic research in this area.

How does displacement change access to social protection?
Having reviewed the current international policy framework for harmonising humanitarian and development work in response to displacement, this section focuses attention squarely on social protection as one of the key tools for allowing that connection. The widespread assumption that mobility weakens social protection (Banting and Koning 2017) relies on three further assumptions. First, that displaced people enjoyed a degree of social protection before they were displaced; second, that social protection itself is immobile; and third, they have no access to social protection in their post-displacement contexts. There is plenty of evidence to support each of these three assumptions, yet the exceptions to this general picture are increasingly common. Overall, there is growing evidence of 'opening access' to social protection for forcibly displaced people (Sabates-Wheeler 2019) and the Covid-19 pandemic has provided greater legitimacy to this trend (Hagen-Zanker and Both 2021). Also, a few policy initiatives are operating at a sufficiently large scale to start to reach some conclusions that indicate directions for further research. This section considers the implications of displacement for the provision of social protection, considering the assumptions behind this provision, the basis for categorisation, including location, and finally reviewing recent examples of best practice.

Mobility and social protection
The three assumptions underlying the idea that mobility weakens social protection each have implications for post-displacement interventions. First, the nature of pre-displacement social protection varies widely according to the context of displacement. Those displaced from low-income countries typically have very limited access to social protection, at least in the form of direct social assistance. Displaced people from lowincome countries are therefore used to coping and strategising on their own in a pre-displacement context; indeed, migration is well known as one form of household insurance against shocks for those with no other alternative forms of provision (Lindley 2009). As the number of forcibly displaced people in and from middle-income countries has grown over the past decade, displaced people are more likely to have some experience of state-provided protection and may more readily seek such protection after displacement. Any intervention should start from an appreciation of the efforts that displaced people are already making prior to the intervention, so as not to disrupt or misdirect what may be more sustainable strategies.
The second assumption, that social protection is immobile, depends on the extent to which access to social protection is tied to a physical location. Certain forms of state-provided social protection, such as subsidies on essential products are open to everyone within a country and can usually be accessed regardless of location. More individualised forms of social protection are typically tied to location. In the very few examples where such protection is portable, such as India's Public Distribution System, migrants can only shift location in certain circumstances and even then require additional registration (Srivastava 2020). Where a degree of planning is possible and where displaced people continue to enjoy the support of state institutions, this may still provide some support. This may be the case for those displaced for climate-related reasons but is unlikely to apply to those forced to move because of conflict or political violence.
For most displaced people, even those internally displaced, opportunities for continued provision of social protection by their state of origin are very limited indeed, though there are certain notable exceptions to this, such as Colombia, which has excellent provision for IDPs. There is more chance that unofficial forms of protection through networks of community or religious organisations or, in particular, family members can continue. Although remittances are often used to support major projects, such as housebuilding, there is growing evidence of their value as a source of social protection (Savage and Harvey 2007;Boccagni 2017). We do not pursue this further in this paper, but the circumstances under which such networked social protection operates may have implications for policy in this field.
We consider the third assumption in greater detail here, since the provision of social protection postdisplacement by state or non-state authorities has generated particular interest in recent years and is the source of various policy innovations. In the global policy context examined in the previous section of the CRRF, the global compacts and the Grand Bargain, social protection provides a potential way of linking humanitarian and development approaches in a sustainable manner. Post-displacement support for registered displaced people has historically been provided through the humanitarian focus of international organisations, particularly UNHCR and associated non-governmental organisations. Such support has had a strong 'sedentary bias' (Bakewell 2008;Brun 2016), assuming that following an initial, unitary displacement, forcibly displaced people stay in one place until they return. Displacement may be an ongoing process with repeated or sequential displacements. Movement other than return, such as circular returns or onward movement, has been viewed with suspicion, often seen to undermine the legitimacy of an initial claim to require protection, whereas further mobility represents a way in which displaced people try to resolve their situation for themselves. In this way, providing social protection to those who have been forcibly displaced is a different proposition to providing social protection to those who are not mobile, though the two are increasingly connected (e.g. Hillier et al. 2020). A more universal approach to social protection provides a way of recognising this.

Registration, categorisation and location
Displaced people's eligibility for social protection often depends on post-displacement registration and categorisation. The flexibility of this categorisation is one of the clearest determinants of how transformative social protection may be. Mobility is relentlessly categorised and those categories are inevitably hierarchical. Expats, émigrés, tourists, foreign students, labour migrants, refugees, asylum seekers and undocumented migrants are ostensibly distinguished by characteristics of their mobility; but these labels say more about their status and crucially their need for and entitlement to assistance. The substantial literature on categorisation and labelling is highly critical of any attempt to use categories as a kind of universal indicator: the categorisation process is contingent on a multitude of politically influenced and locally contextualised factors (Crawley and Skleparis 2018). Yet it is often the changing forms of categorisation of mobility that determine the nature of a migrant's relationship with state authorities and their access to different forms of social protection (Zetter 2007).
In humanitarian crises, this in turn influences attitudes of humanitarian institutions and so further determines migrants' access to social protection. From a migrant's perspective, this means that access to social protection appears to be unpredictable, even arbitrary. In situations where host governments are lukewarm or openly hostile to migrants' presence, arbitrariness and unpredictability may be a deliberate element of a logic of deterrence. This uncertainty underlies one of the major challenges faced in organised crisis response. What should be a dynamic relationship between social protection and mobility is too often rigidly determined by categories that are increasingly rejected by those who are subject to them (Coddington 2018).
Governments' desire to safeguard internal security and to monitor the bodily health, wellbeing and movement of displaced populations drives the imposition of measures that seek to make these groups legible to the state, meaning that they are identified and fit into a categorisation scheme that is recognised by state authorities. These include efforts to isolate and contain displaced populations in remote camps, where sedentariness, severely restricted mobility and highly constrained participation in labour markets (outside of the camp) precondition receipt of social protection. Arguably though, interconnected chains of policies and laws spanning border movement, legal status, economic participation, mobility and residential status increasingly differentiate who gains access to what kinds of social protection in urban areas.
For decades, camps have been the template for organising large-scale humanitarian protection of refugees. Over prolonged periods of time these camps may take on the physical features of the urban fabric ('the urbanisation of camps' -see : Sanyal 2017;Ramadan 2013). Yet, humanitarian programming for urban refugees has long been frowned upon. Within UNHCR, such programming was deemed frivolous, unnecessary and subject to much censure until at least the mid-1990s, with residual resistance common (Crisp 2010(Crisp , 2017. However, evidence on global urbanisation is now irrefutable (Fox and Goodfellow 2016). Broad consensus has emerged that protracted displacement is urbanising (UNHCR 2020) and that social protection approaches need to take the particularities of these contexts seriously.
In contrast to the late nineteenth century development of European social welfare programming, social protection programming in international development as it has gained prominence over the past few decades has firmly rural roots. However, more recently momentum has been growing to rethink social protection for urban areas (Gentilini 2015;Gentilini et al. 2021;Cuesta et al. 2021), accelerated by Covid-19 responses in many countries (Roelen, Archibald and Lowe 2021). Key challenges social protection faces include the global urbanisation of poverty, driven by rapid urban population growth exceeding economic growth in South Asia and sub-Saharan Africa; the stubborn persistence and indeed global increase of informal jobs gainsaying decades of predictions of a decline; the generation of jobless growth; and the growing casualisation and stripping bare of social security features in formal labour markets and informal 'gig' economies.
Unlike populations that cross national borders, IDPs typically carry -or are entitled to -citizenship status, and thus prospectively deemed the rightful recipients of social protection by their governments. In case of cross-border mobility, displaced populations' eligibility for social protection (as non-citizens) is significantly dependent on successful completion of status verification procedures (e.g. by UNHCR) and the recording and issuing of legal identity documents/cards by host states. Once obtained, maintaining valid legal identity is not a given and can require regular renewal. Moreover, successful status verification may be made subject to displaced people having valid passports and proof of their legal crossing of borders. Failure to provide such proof can lead to incarceration, refoulement (forcible return) and the withdrawal of protection.
Ongoing mobility, shifting categorisations and a dynamic protection context are particularly characteristic of forced displacement in FCAS. Where conflict is ongoing, control of territory changes and threats shift, individuals may move back and forth across international borders or develop circular patterns of mobility, all of which may intersect with challenging climatic conditions. International displacement frequently intersects with internal displacement and periods spent living alongside people who may not be displaced but face similar conditions and are likely to have had previous histories of displacement. In such situations it is not practical to categorise or group individuals together, allocating a certain level of assistance to some but not others. A more universal approach to social protection offers a route to greater justice, equality and sustainability. Research into these questions is still very much in its infancy, but various policy innovations warrant further attention.

Current research and policy on displacement and social protection
Following the World Humanitarian Summit in 2016, the Social Protection Inter-Agency Cooperation Board issued a joint statement calling for social protection to be systematically applied to situations of extreme fragility, protracted crises and forced displacement. Reference to displacement was very brief; it simply highlighted that: 'Social protection can be incorporated as a cornerstone of any strategy to address the displacement, including IDPs and refugees' (SPIAC-B 2016: p2). This initiated various studies that in some cases have been linked explicitly to high-level negotiations around the global compacts but have more often taken place on a separate track. These include national approaches to refugees, often now also referred to as refugee compacts, influenced by the Jordan Compact the first such nationally focused policy approach, which was signed in 2016 (Hagen-Zanker, Barbelet and Mansour-Ille 2018).
Provision of social assistance in forms that are not explicitly connected to the global compacts is now extremely common. UNHCR's cash operations, a key tool of social protection, have expanded from 33 countries in 2015 to over 100 in 2020 (UNHCR 2021), though it appears only a minority of these occur at scale and provide a basis for longer-term engagement with national social protection programmes. UNHCR has provided social protection to refugees for decades, but for most of this period this has been exclusively for refugees and typically in the form of direct food assistance. In 2018, UNHCR conducted a study of 'social safety nets' in 18 countries. 1 The introduction to the study quoted a UNHCR report from 2017, highlighting the challenges of engaging with governments in this provision of social safety nets: While recognising an increasing trend in inclusion in national systems, refugees, in particular, rarely enjoy the same rights as citizens. Only in every second UNHCR operation, refugees can choose their place of residence; in less than 50% they access national health care systems; in two third they access the national education systems; and in less than 40% they are allowed to work according to law and policy and in practice. In approximately 10% of UNHCR operations, refugees are included in the national or local development plans. Initiatives related to shock-responsive social protection rarely address conflict and include displaced people.
(UNHCR 2018: 1) While the language of 'social safety nets' appears to reject the more transformative potential of social protection (Devereux and Sabates-Wheeler 2004), the study found high levels of inclusion of displaced people in social safety nets in four countries: Ecuador, Niger, Pakistan and Turkey. It concluded that full inclusion of displaced people could happen in these four countries and was possible in a further 10, though in the final four significant barriers remained. Further information on this study is unfortunately limited, but the relatively positive implications are important.
Since then, further large-scale research has been undertaken, significantly under the auspices of the United Nations Children's Fund (UNICEF), which takes a wide sweep of social protection policies, including access to education, the right to work and access to cash transfers. UNICEF's research division sponsored a series of studies into social protection in contexts of fragility and forced displacement that appeared in a special issue of the Journal of Development Studies (Brück et al. 2019), many of them applying randomised control trial methods to this challenging context. UNICEF's Regional Office for Latin America and the Caribbean, in partnership with the World Food Programme (WFP) and the International Policy Centre for Inclusive Growth (IPC-IG) have highlighted the regional situation for social protection of Venezuelan migrants, particularly during the Covid-19 pandemic (IPC-IG et al. 2021). IPC-IG et al. 2021 emphasises the importance of flexibility in responding to displacement. Some form of support was offered to Venezuelans in most of the nine countries studied, but this varied widely and usually required regular migration status. The response of the Brazilian state was highlighted; in particular, integrating large numbers of Venezuelans into national cashbased social protection schemes, particularly the Bolsa Família programme. A joint World Bank/UNHCR study examined the situation of Venezuelan migrants in Brazil more closely (Shamsuddin et al. 2021), underlining the significant contribution made by Brazilian social protection policies but also highlighting significant regional variations in the fate of Venezuelans in Brazil, with those in areas with a high concentration of Venezuelans faring worse than those elsewhere.
In June 2020, the Regional UN Issue Based Coalition on Social Protection was established in the Middle East and North Africa/Arab States region, jointly led by UNICEF and the International Labour Organization (ILO). The group produced a 62-page analysis of all social protection policies in the region responding to Covid-19, including a chapter on measures for foreign workers, IDPs and refugees (IBC-SP 2020). Of 21 countries (and the Occupied Palestinian Territories) considered, nine had interventions to support foreign workers, all provided by governments, and 12 had measures to support refugees and IDPs, all provided by UN agencies. Measures covering refugees and IDPs included delivering food and hygiene products, distributing vouchers for food and paying cash transfers. Cash transfers were the most common measure, conducted by UNHCR in Algeria, Egypt, Iraq, Jordan, Kuwait, Lebanon, Tunisia and Yemen and by WFP in Egypt. However, the programmes serving the largest numbers of displaced people were in Jordan and Lebanon, which also had the greatest success in aligning support for displaced people in national social protection policies. The report made various recommendations, including at a minimum, to extend national social protection measures to IDPs and ensure that social protection is gender sensitive in terms of targeting, design and implementation. This flourishing of high-level reports on social protection and displacement continued with a global study, sponsored by UNICEF's Egypt office (Andrade, Sato and Hammad 2021). The report considers Egypt in detail, before turning to a more global range of examples of best practice.
A more recent ODI study considers social protection provided specifically to refugees during the Covid-19 pandemic (Hagen-Zanker and Both 2021). Based on a systematic investigation of low-and middle-income countries, the study identifies ten countries that include refugees in government-financed social assistance programmes established in response to Covid-19: Argentina, Brazil, Cameroon, Colombia, Chile, Republic of Congo, Djibouti, Panama, South Africa and Trinidad and Tobago. All of these were new programmes, established as a national social assistance response to the pandemic. In addition, the study identifies six countries in which social assistance provided by UNHCR, UNICEF or WFP was aligned to governmentprovided social assistance: Colombia, Jordan, Morocco, Pakistan, Peru and Turkey. The study examines four programmes in more detail, two government-provided programmes (Colombia and Republic of Congo) and two aligned programmes (Jordan and Pakistan). In all cases, social assistance was delayed and partial. These problems were more significant in cases of government provision than in international organisationfunded provision since international organisations had access to more complete registration details for refugees.
It is striking how the same best practice case studies appear regularly in this emerging literature, highlighting the relatively few cases in which these ideas have been implemented at any significant scale. Programmes in low-and middle-income countries to support forcibly displaced people are still relatively few. UNHCR's cash transfer programme for Syrian refugees in Jordan, established in 2012, was one of the earliest to operate at significant scale. ODI evaluated it positively (Hagen Zanker et al. 2017;, highlighting indications of positive wellbeing effects of the programme. Turkey's Emergency Social Safety Net is one of the largest, reaching approximately 1.75 million Syrian refugees, though since this is funded exclusively by the European Commission as part of the 2016 European Union (EU)-Turkey statement, it has received a more mixed assessment. The EU's own reviews highlight the effectiveness and value of the support received. It is more widely recognised as a leading policy example of where an external funder has financed the expansion of an existing national social protection system to include refugees, rather than creating a parallel one (Andrade et al. 2021). Nonetheless, the substantial EU funding is explicitly part of a policy to contain Syrian refugees in Turkey and, like the EU-Turkey statement as a whole, prevent their onward mobility to the EU (Cetinoglu and Yilmaz 2020). In contrast, Brazil's Bolsa Família has incorporated Venezuelan migrants and South Africa's Covid-19 relief grant was extended to refugees with no external financing, providing a powerful example of inclusive social protection programming (UNHCR 2020b). In Pakistan, the Ehsaas programme of emergency cash transfers was mirrored to support refugees with UNHCR support (UNHCR 2020a).
The main barrier to further expansion of these systems is undoubtedly financial. They are most needed in states hosting large populations of forcibly displaced people and these are disproportionately poor countries. This is where the GCR can help. Although it makes little mention of social protection, it sets out a framework for international solidarity in this field. Assuming the international community can fulfil these promises of additional resources, it will provide one way of supporting the generation of new programmes and expansion of existing ones to benefit displaced and non-displaced people alike. The opening up of the World Bank's IDA18 Regional Sub-Window for Refugees and Host Communities (World Bank 2020), which provides access to development financing for countries hosting more than 25,000 refugees -UNHCR's cut-off number for a situation to be defined as protracted -has highlighted the possibilities of extending social protection to refugees (Hagen-Zanker and Both 2021).
However, finance is not the only barrier. A recent ODI study highlights policy design and delivery capacity as key elements of the delivery of social protection to displaced people that can be improved without additional finance (ibid.). Registration and accurate data systems are also areas where progress can be made at a lower cost, particularly given concerns around privacy and data protection raised by the range of biometric systems already in operation (e.g. Kuner and Marelli 2020). The ethical issues raised by the potential exchange of privacy for basic social support are undoubtedly problematic. This is developing quickly and raises issues that are complex and need time to think through. The fundamental question is how to interpret this changing landscape of social protection, which requires a more theoretical analysis than has been obvious in recent, largely applied literature. The final section now turns to these questions.

Theorising social protection and displacement
The recent array of policy developments and associated research share a strongly applied focus. The underlying analysis in these approaches is usually expressed as a straightforward 'what works', with further justifications around efficiency and value for money. This is certainly appropriate and, particularly where large sums of public money are being invested, it is important to retain a focus on reducing waste. The broader theoretical context is not usually interrogated. This does not mean theory is absent, since assumptions of a theoretical nature underly any attempt to project present circumstances into the future, as well as most analysis of the past, but theoretical analysis is certainly not explicit. In considering the wider implications of these developments, it is worth turning to more theoretical explanations to examine consequences in more detail. We consider six areas where social protection may form part of a more transformative solution.

Gender, equality and intersectionality
Attention to gender equality is a well-established feature of global policy towards displacement; the two compacts, CRRF and wider current framework are no exception. This goes significantly beyond the grouping of 'women and children' as an almost undifferentiated vulnerable group that has often characterised refugee policy responses. Such an approach both infantilises women and fails to recognise the distinct vulnerabilities that displacement produces. Current policy, particularly UNHCR's (2019) Age, Gender and Diversity Policy, has been widely praised (e.g. Taha 2019). In its recognition of the significance of 'intersecting personal characteristics' this policy moves towards a recognition of intersectionality or the compound impacts of multiple disadvantages. Although most critical gender analysis would consider gender as a vector of power relations rather than a personal characteristic, this policy sets out a progressive framework for interventions and is at least partially reflected in the GCR's reference to 'age, gender and diversity' (e.g. para. 13). The commitment to gender equality; the specific recognition of forms of harm, such as sexual and gender-based violence (SGBV) and trafficking, which become more likely in situations of displacement; and the ways in which they intersect with age, nationality or ethnicity are an essential step towards addressing them.
The difficulty is how to translate these progressive policy frameworks into action on the ground. There is increasingly widespread commitment to the minimal standard of differentiating gender in any data collection. Initiatives such as the Gender-Based Violence Information Management System, 2 implemented by an interagency partnership have sought to standardise collection of data on SGBV so that much more is now known about the prevalence of such violence and the differential ways in which it affects people according to age, gender and diversity. Yet although such data collection is a further step towards addressing SGBV, for the moment it has only helped to clarify how widespread the problem is in displacement contexts. The recent turn to social protection has highlighted the potential contribution that well-designed forms of social assistance can provide in this area. A review of 22 research studies of cash-based transfers highlighted how 70 per cent of them resulted in a reduction in intimate partner violence, though this was not among the initial aims (Buller et al. 2018). A recent UNICEF blog argues that cash transfers in addition to other aligned interventions such as gender-transformative curricula in schools or skills training can result in increases in empowerment of women and girls and reductions in SGBV (Wijesekera 2020). The nature of these connections, beyond reducing poverty levels which social protection should also produce, is worth further investigation.

Burden-and responsibility-sharing
Both the CRRF and GCR recognise the 'urgent need for more equitable sharing of the burden and responsibility for hosting and supporting the world's refugees' (GCR para. 1). The current system of responsibility for displaced groups is startlingly unequal. UN Special Representative of the Secretary-General for International Migration Peter Sutherland referred to 'responsibility by proximity' (Doyle 2018). There has been a succession of attempts, both global and regional, to identify a more egalitarian way of allocating responsibility for providing protection, 'yet none has had a significant or enduring legacy' (Betts, Costello and Zaun 2017b). Perhaps mindful of this, the CRRF and the GCR do not propose any clear formula but simply a regular pledging conference, the Global Refugee Forum (GRF), to be held every four years. These conferences should involve development actors much earlier in response planning; indeed, development actors sent large delegations to the 2019 GRF in Geneva. The underlying assumption in much of this is that the focus is on states; 'ease pressures on host countries' is the first objective of the GCR (para. 7). A widespread critical response has been to question the reactive 'solidarity with states' approach of wating until states themselves are burdened, in contrast to a more proactive 'solidarity with refugees' approach to try to avoid that situation (Wagner 2017). Betts et al. (2017b) draw on public goods theory to explain the failure of collective action in allocating responsibility for refugees between states. One of the key questions to be resolved here is identifying the costs and benefits of hosting refugees. If the aim of policy is simply to provide refugees -and eventually other forcibly displaced people -with the means necessary to sustain life, this will inevitably be perceived as a continual cost, a drain on resources that will only end when refugees leave. Willingness to fund such openended commitments declines over time (Whitaker 2008). In such cases, the 'burden' to be shared is not the refugees themselves but the policy choices taken to provide humanitarian relief over the long term. A more explicit shift to a development focus could potentially change this since the goal of development-focused policies is to support refugees to make a meaningful contribution to their host state. If this is successful, support for refugees should more properly be seen as an investment. With evidence that the wider application of social protection can achieve this, the debate around burden-and responsibility-sharing should shift in important ways and may allow a more proactive refugee-focused response.

Localisation and urban responses
In a world where annual international humanitarian funding falls short of global humanitarian need by more than US$15 billion and where global appeals consistently only ever manage to secure around 60 per cent of their required funding (Development Initiatives 2015: 22), the UN Secretary-General's High-Level Panel on Humanitarian Financing (UN 2016) belatedly arrived at the conclusion that greater investment in local responders is now indispensable. Yet, understanding who local responders are, and the economic value of their contributions to social and humanitarian assistance in FCAS, is not entirely clear. The salience of locally embedded responses to mass displacement takes on greater significance in contexts such as Lebanon where a 'policy of no policy' with respect to Syrian refugees manufactures deliberate institutional ambiguity to evade responsibility towards refugee populations (Nassar and Stel 2019). A clear theoretical account is required of how and why locally embedded responses to displacement choose to interact or disengage with social and humanitarian assistance mandated by the global North.
With situations of mass displacement increasingly taking on an urban character, the focus of researchers in recent years has turned to the multiple (in)formal sovereignties (Carpi and Fiddian-Qasmiyeh 2020b) that contest the right to intervene in the humanitarian field. Scholarship has increasingly become attuned to local actors embedded in displacement responses, identifying the integral role of municipalities (Boustani et al. 2016;Mourad 2017;Betts et al. 2017a;Zapater 2018;Easton-Calabria 2020) and mandates outside the reasoning, structures and networks of so-called formal humanitarian responses. This requires an analysis of multi-level governance; specifically, how different forms and spheres interact. These different spheres may include faith-based and faith-inspired humanitarian action (

Post-colonialism and South-South responses
With mass displacement largely prevalent in the global South, it is unsurprising that Southern states and regional actors have historically adopted dynamic approaches to responding to displacement. This had been in part driven by the recognition that international frameworks for responding to displacement had hitherto spoken to European experiences of displacement. The 1969 Organisation of African Unity Convention Governing the Specific Aspects of Refugee Problems in Africa and the 1984 Cartagena Declaration on Refugees situate displacement in a world shaped by colonialism and its persistent after-effects. States and regional actors in the global South have also led the way in formalising protection and assistance for IDPs: the 2009 African Union Convention for the Protection and Assistance of Internally Displaced Persons in Africa (Kampala Convention).
Until very recently, policymakers in the global North viewed responses to mass displacement in the global South that were located outside of those recognised and/or mandated by Northern actors through a securitisation lens, which blurred humanitarian endeavour with perceived threats to national and international security (Barakat and Zyck 2010;Binder, Maier and Steets 2010;Fiddian-Qasmiyeh 2018). Moreover, analysts, policymakers and practitioners embedded in Northern-mandated responses to displacement delegitimised Southern responses as not properly humanitarian by claiming that ideological and faith-based grammars and vernaculars used by humanitarian actors of the global South were fundamentally at odds with the core 'international' humanitarian principles of impartiality, neutrality and independence (Ferris 2011;Fiddian-Qasmiyeh 2011;Ager et al. 2015;Zaman 2016). In turn, locally embedded actors in situations of mass displacement equally viewed Northern-mandated responses to crises in the global South as motivated by political and ideological concerns rather than strict adherence to international humanitarian principles (HPG 2018).
The global financial crisis of 2008 precipitated interest among Northern donors in the potentialities of Southern-led responses to crises. These Southern-led responses have been well established for decades. Following the World Humanitarian Summit in 2016, there was considerable debate and discussion on the need to localise aid and to incorporate specific Southern actors into the 'international humanitarian system'. Very little has translated into action. Despite signatories to the Grand Bargain committing to channelling at least 25 per cent of international humanitarian assistance to local and national actors as directly as possible by 2020, trends in direct funding portray a dismal picture, with direct funding to local and national actors accounting for a mere 2.1 per cent of total humanitarian assistance in 2019 (Development Initiatives 2020).
Rather than thinking which actors best fit the Northern-mandated system of social and humanitarian assistance, Fiddian-Qasmiyeh (2018) argues for a closer reading of the diverse modalities, spatialities, directionalities, relationalities and conceptualisations that underpin South-South cooperation in FCAS. Space needs to be ceded to alternative histories and understandings of responding to crises that do not a priori begin with Western conceptions of humanitarianism. In doing so, how we understand Southern-led responses to displacement are not fixed at national or regional scales but allow us to consider, as legitimate, everyday neighbourhood-level responses.

Sustainability and self-sufficiency
Like other aspects of the GCR, the central goal to 'enhance refugee self-reliance' (para. 7) has a long history (Jacobsen and Fratzke 2016). The idea of supporting livelihoods is a significant move away from the care and maintenance model. This more sustainable response is central to the longer-term objectives of development planning. UNHCR published its first livelihood policy in 2005 and created a dedicated Livelihoods Unit in 2008. This responded to a vibrant research agenda around refugee livelihoods that Karen Jacobsen brought together in the Economic Life of Refugees (2005) which set out a powerful argument that, particularly in protracted situations, support programmes should go beyond the humanitarian imperative of simply saving lives and focus on building livelihoods. Restrictions on work are a control on livelihoods that are more frequently experienced by refugees and other forcibly displaced migrants who have crossed international borders but may also affect IDPs. Implementing a right to work, even to the extent of recognising labour migration as a further durable solution, has formed a key part of the livelihoods debate (Long 2015).
UNHCR's Global Strategy for Livelihoods (UNHCR 2014), set out a comprehensive policy response to these initiatives covering the period 2014-18. This is addressed more widely than to refugees only, but rather to 'all persons of concern', the broadest category of forcibly displaced people UNHCR uses. The right to work is central to this strategy, as is a preservation of productive assets. The overall focus is self-reliance, defined as 'the ability of an individual, household or community to meet essential needs and to enjoy social and economic rights in a sustainable manner and with dignity' (ibid.: 7). Social protection is widely referenced in the Global Strategy for Livelihoods though not the more transformative vision of social protection we draw on in this paper. Social protection is seen in terms of a social safety net, the initial stage in a 'graduation approach' that further livelihood interventions can build on to reach self-reliance. More recently, approaches to self-reliance have gradually shifted from livelihoods to a broader context of economic inclusion, which 'entails access to labour markets, finance, entrepreneurship and economic opportunities for all, including non-citizens in addition to vulnerable and underserved groups' (UNHCR 2019: 3). This reflects a more market-led approach to self-reliance, which is an important policy trend in the GCR and related recent initiatives.
Critiques of self-reliance are now widespread and it is clear that critical theoretical analysis must find an alternative language to discuss these developments (Morris 2020). The challenge is that the obvious interpretation of self-reliance is a progressive one, a way of interpreting the autonomy of refugees; yet this is not how it is being used. Indeed, a language of self-reliance is common in policy that seeks to prevent further mobility of displaced people. Hyndman and Reynolds (2020) suggest that self-reliance often lacks selfdetermination and this at least provides a way of discussing autonomy without wider policy interpretation. Social protection, in a broad transformative interpretation, provides a way of doing this, particularly where it involves regular provision of cash support, an ideal basis for supporting self-determination for displaced people.

Market-led solutions to social protection
Growing engagement with the private sector in refugee responses is one of the more controversial developments advanced by the GCR. Recent approaches to this by Betts et al. (2017a) are among the best known in this area, setting out proposals for state-or UN-led collaborations with private companies wishing to employ refugees. Betts and Collier (2017)'s book Refuge: Transforming a Broken Refugee System received very wide coverage for the suggestion that refugees in Jordan's Zaatari camp could find work in nearby special economic zones (SEZs). Although the Jordan Compact (OECD 2016) granted Syrian refugees the right to work and join contributory social protection schemes in Jordan, the number of refugees eventually involved was relatively small (Crawley 2017). There were a range of reasons for this, including the demeaning nature of much SEZ work, but it highlights the wider practical problems of relying on work as a potential durable solution without assurances that the work is 'decent' in ILO terms.
In addition to practical obstacles, there are broader theoretical objections to increasing the role of the private sector. Betts et al. (2017a) laid out the argument for market-based solutions more clearly in Refugee Economies: Forced Displacement and Development, which Skran and Easton-Calabria (2021) were concerned 'risks uncritically overemphasising the positive addition of markets in refugee assistance'. Zetter (2019) has set out one of the most comprehensive critiques of market access, and indeed the shift to development-led responses more generally, going beyond a neoliberal critique to argue that it maintains the dependency of the 'refugee impacted Global South' on the global North. This return to dependency theorybased analysis is despondent about the prospects for more progressive solutions to protracted displacement crises: Converting refugees from welfare recipients into market actors as consumers and producers through employment promotion and cash-based transfers (CBTs), and the increasing privatization of humanitarian space through entrepreneurial activity, align well with the micro-economic aspects of the neo-liberal agenda of economic globalization. (Zetter 2019: 8) This is one of the most sustained critical analyses of the GCR and the wider context of the shift from humanitarian to development approaches, but Zetter (2019) suggests few alternatives other than the wholesale abandonment of this current direction.
In terms of social protection, there is a contradiction between the approving reference to 'welfare recipients' in a more humanitarian model and the concerns that Zetter (2019) expresses in relation to cash-based transfers. This is perhaps a further example of the self-reliance approach discussed above: a potentially progressive idea that is now widely interpreted as a form of words used by those more interested in containment policies. Self-determination would surely be enhanced by cash-based transfers. Indeed, from a neoliberal perspective, self-reliance is often seen as opposed to forms of social protection (Easton Calabria and Omata 2018) since true self-reliance means managing without support. Neoliberal policies are always associated with a reduction of welfare.
The wider distribution of cash-based support, which has become more universally accepted during the pandemic, is associated with a more Keynesian form of market stimulation but one which does at least open opportunities for market integration or 'economic inclusion' for displaced people, so this is not entirely clearcut. The ultimate implications of this will take many years to unfold, by which time policy will doubtless have changed again. In the short term, it appears that the connection between more social protection and more containment is not theoretical but coincidental, in that they occur together in the same policy proposals. The argument of Zetter (2019) and others goes further than this to suggest that the underlying objectives of wealthy, powerful states are unlikely to change, though the language used to frame them may do (Adesina 2020). This unfortunately leaves no room for a progressive response (short of revolution), whereas we remain convinced that greater social equality is achievable through more incremental change. This review has raised more questions about what is needed to achieve a more transformative form of social assistance, which we turn to in the conclusion.

Conclusion: an agenda for further research
A shift towards a more development-focused response to displacement has been the focus of policy initiatives since the 1970s but is increasingly urgently needed. We began this paper arguing that most displacement is no longer exceptional or unexpected, which are the main justifications for a humanitarian response. This turn to development is now widely recognised at high level on a global scale, so there is reason to believe it may have more impact than in the past. Although development is not clearly defined in any of the policy documents we have considered, it is consistently framed in relation to the SDGs as long term, far-reaching and crucially sustainable. In addition to the SDGs and the World Humanitarian Summit, displacement has been the focus of a succession of high-level global policy commitments in recent years, encompassing the New York Declaration, CRRF and both global compacts. These documents all emphasise the need for more sustainable responses, though the discussions of humanitarianism and development highlight an ongoing tension in the way this need is presented, with the GCR, in particular, giving much more prominence to a humanitarian rather than development objectives, despite recognising the significance of protracted situations.
Social protection has generated a growing interest in recent years as a way of linking humanitarianism and development to support displaced people. It has great potential to enhance the agency of displaced people by providing greater choice, including reflecting differential perspectives in relation to gender, age, ethnicity and legal status. It also offers a way in which support to displaced people recognises the significance of ongoing mobility. The extent to which this potential can be fulfilled depends significantly on the modalities of registration. As social protection is provided to displaced people through existing national systems, whether or not it is funded externally, the nature of registration is key. Where this allows flexibility of location, further mobility within and even beyond national borders may be supported, where displaced people can make that choice. This will reinforce using onward mobility as an established household strategy to spread risk in crisis situations.
The provision of support by actors whose role in responding to displacement has not been widely recognised at the international level plays an important part in the new policy framework. The Grand Bargain's localisation agenda sets the context for this wider group of stakeholders, but this received particular attention in the displacement-focused policies, particularly the CRRF. The extent to which these new actors are involved, how they are involved and the impacts of recognising -and in some cases supporting -these roles, provides directions for future research. These will crucially involve urban contexts, where social protection provides the dominant mode of engagement with displaced people. It also covers the modalities of access to national systems, how this access can be financed and how access varies with different forms of registration, including refugees, other international migrants and IDPs. The developing role of South-South humanitarianism is an important development research focus that deserves continued attention. The CRRF particularly references refugees' own involvement in framing policy initiatives, though we have not been able to identify any examples of where this has occurred. Identifying the conditions that would allow refugees' agency would be an important contribution to the selection of examples of best practice that circulate widely in this literature.
Self-reliance is undoubtedly a vital focus of future research in this area. The prominence of self-reliance as an objective has only increased with the central role it plays in the GCR. Social protection is widely justified in terms of enhancing self-reliance and much of this discussion is progressive. Yet, the shift in language from local integration to self-reliance that we noted in the discussion of durable solutions in the GCR reflects an emphasis on individualised responsibility that has been associated with a more neoliberal policy direction. These concerns must be reflected in any further investigation of the more positive interpretation of this objective that social protection may provide. The language to discuss the more positive elements of greater individual responsibility must be chosen carefully; the sense of autonomy or self-direction provides one way of separating the more positive characteristics (freedom, choice) from the potentially more damaging (reductions in funding, withdrawal of international support) but these alternatives have not been widely investigated.
To summarise, drawing on the more theoretical review of policy areas in section three, we finish with four areas that merit greater research attention in this emerging field and which we hope to develop further soon: • The extent to which social protection supports greater levels of equality among displaced people based on an intersectional understanding of gender, age and diversity. • How registration allows flexibility of access to social assistance, supporting further mobility and transforming the sedentary nature of support for displaced people. • How different levels of government provide social assistance and how this differs across levels of government, considering the impact, role and financing of new actors at national, metropolitan and community scales, including displaced people themselves through organised collective action and individualised use of remittances. • The nature of the focus on measures to support self-reliance, particularly from the perspective of displaced people themselves, the connections between self-reliance and more neoliberal reductions in social assistance, as well as the ways in which a shift to 'self-determination' highlights the more genuine autonomous content of this aim.