Introduction
Global crises put international organisations (IOs) under pressure. Many IOs – including intergovernmental and non-governmental organisations as well as regional agencies – are expected to be competent crisis managers, yet they often lack the necessary resources to meet urgent and unexpected challenges. During the COVID-19 pandemic, for instance, the world turned to the ‘Public Health Emergency Operations Centre Network’ of the World Health Organization (WHO) to contain the spread of the virus, but member states did not provide adequate funding for the organisation.Footnote 1 Overlapping shocks, such as violent conflicts, climate extremes, and the disruption of global supply chains in the wake of the Russian invasion of Ukraine, fuel hunger crises. At the same time, humanitarian funding to IOs that are to strengthen food security, mainly the World Food Program (WFP) and the Food and Agriculture Organization (FAO), is declining drastically.Footnote 2 In the European migration and refugee crisis, actors like the European Border and Coast Guard Agency (Frontex), the Norwegian Refugee Council (NRC), and the United Nations High Commissioners for Refugees (UNHCR) were expected to organise accommodation, relocation, and returns quickly. Still, they were not well-equipped in terms of staff and mandates.
Notwithstanding these contradictions, IOs often exploit moments of crisis to change how transnational challenges are governed. In the course of the migration and refugee crisis, the European Asylum Support Office (EASO), for example, changed the definition of ‘vulnerability’. The term ‘vulnerability’ functions as a tool to categorise asylum-seekers; persons who are found ‘vulnerable’ are excluded from accelerated asylum procedures, which increases their chances of international protection.Footnote 3 EASO has thus effectively transformed how decision-making on asylum applications is organised at the joint European border – that is, how irregular migration movements are governed in Europe. As part of its crisis management efforts, it developed new routines and policy scripts. How did EASO manage to do this despite the conflicting pressures described above? How can IOs make and change policies during crises?
In this article, I argue that the creativity of frontline crisis managers is a vital yet overlooked resource for adapting and creating policies. IO staff members stationed on the ground use their professional knowledge to improvise provisional solutions for the public problems produced or highlighted by crises. Bureaucrats in headquarters stabilise these improvised solutions by devising training tools and operational guides. Through this dynamic interplay, they informally alter or produce policies, broadly understood as ‘courses of action over issues of common concern’.Footnote 4 Whether they actually achieve change through improvisation and stabilisation and whether the changed policy solutions are beneficial or detrimental to the governance of specific issues is difficult to establish before empirical analysis: not all crises will lead to change and not every change will be productive. My key argument is that we need to study systematically how frontline staff respond to crises to understand their actions’ effects.
Despite the ubiquity of crises in world politics, we still know little about how frontline crisis management by IO staff affects the governance of transnational challenges. International Relations (IR) scholars have studied how political elites handle crisesFootnote 5 as well as IOs’ operational capacities,Footnote 6 identifying institutional factors that contribute to organisational resilience and effectiveness. They have also studied crisis consequences in terms of changes in IOs’ institutional design and authority structureFootnote 7 and they have shown that crises cause a shift toward informal cooperation.Footnote 8 However, how crises lead to informal change in how issues are governed remains unclear. Similarly, there is a rich literature on policy change in IOs in reaction to crises, primarily concentrating on top-down dynamics.Footnote 9 Several studies have highlighted that policy entrepreneurs transform existing regulatory approaches in and between crises, which, in turn, informs how future emergencies are addressed, in the field as well. However, the question of how field dynamics during crisis episodes initiate policy change remains.
I address these omissions by developing a micro-level approach to how IOs act during crises, generally defined as episodes of uncertainty, urgency, and threat.Footnote 10 I draw on two established sets of IR research – the literature on organisational learning in peacekeepingFootnote 11 and studies of informal practices in IOsFootnote 12 – and complement both with insights from crisis management studies as well as the sociology of the professions and organisations to develop a framework for the study of the mechanisms that lead to bottom-up policy change. Broadly situated in practice-oriented IR scholarshipFootnote 13 and social-constructivist approaches in IO studies, this framework focuses on the agency of IO personnel. It draws attention to how they (1) make sense of crises, (2) test how to respond, (3) negotiate the appropriateness of improvised responses, and finally (4) routinise initial solutions which they then (5) consolidate in IO training programmes and best practice guidelines so that they become consequential beyond a concrete crisis setting. This inherently dynamic and open-ended process includes people working for different organisations – albeit in close physical proximity – and is likely to repeat itself, given the preliminary nature of frontline action and informal policy-making.
My framework relates in particular to operational IOs – IOs that implement and monitor projects and programmes in crises regions and countries that receive aid and assistance – as well as to fast-burning crises or humanitarian emergencies. I therefore examine the impact of frontline action on crisis policy-making through a case study of how IOs with in-built emergency management capacities responded to the 2015–16 European migration and refugee crisis in Greece. The case study draws on ethnographic fieldwork at the European external border during the crisis, interviews at the headquarters of seven organisations – including the International Organisation for Migration (IOM), UNHCR, Frontex, EASO, and three humanitarian NGOs – and an extensive document analysis of IO training material. It demonstrates that to reconcile the contradictions of having to cope with the emergency without being comprehensively equipped to do so, IO staff members located on the ground improvised to a great extent, using their professional knowledge. Their everyday, ad hoc solutions were, to some degree, stabilised by training and operations officers, which shows that migration and asylum policies did change from the field up.
With this, my contribution to the literature on crisis politics and change in IOs is twofold. First, by outlining the mechanisms at play during improvisation in the field and stabilisation at headquarters, I provide a framework which complements approaches that focus on political elites and top-down dynamics. Second, by identifying the institutional factors that are likely to shape the degree of improvisation and stabilisation, I pave the way towards a comparative research agenda on the effects of frontline action on global crisis governance.
The article proceeds as follows. Section one discusses the literature on crisis and change in global governance, highlighting the curious absence of analytical interest in frontline crisis managers and informal policy change from the ground up. Section two develops the iterative, five-step framework that conceptualises how IO staff members respond to crises across organisational boundaries and levels. It also sketches the organisational characteristics that enable and constrain the occurrence of these steps. To demonstrate the plausibility of my conceptual argument, Section three presents evidence from the case study on the 2015–2016 European migration and refugee crisis. The final section concludes by discussing how to study frontline action in non-operational IOs and outlining how to investigate the institutional features that are likely to condition the success of bottom-up change.
Crises and change in global governance
Crises are the rule rather than the exception in international relations. Although theorists of global governance have long been concerned with resolving the puzzle of global order under anarchy, this order is frequently disrupted by unforeseen events destabilising hard-won rules and agreements. Crises affect all world regions, policy fields, and issue areas. From natural disasters, violent conflicts, and terrorist attacks to the outbreak of diseases, energy price shocks, and the collapse of capital markets, crises are recurring features of international politics and many IOs are expected to respond.
IOs as crisis managers: Institutional resources and effects
To understand how IOs govern in crises, scholars of global politics have focused on two main aspects. First, at the perceptional level, they have analysed how political elites make sense of a crisis and react to it. Second, at the operational level, they have studied IO capacities for effective crisis management.Footnote 14 What is striking in both lines of research is that the focus lies on questions of institutional design and choice. As I show in the remainder of this sub-section, existing studies have largely overlooked how regular IO staffers interpret crises and how their reactions shape crisis outcomes, especially with regard to policy content.
When it comes to understanding how leaders cope with crises, most authors focus on heads of government and analyse how their perceptions and preferences shape international organisation.Footnote 15 They typically ask the important question of whether hard times open up opportunities for enhanced multilateral cooperation or whether a ‘nationalist reflex’ kicks in. Three findings stand out in this research agenda. First, in the long run, elite crisis management often leads to deeper integration, especially at the regional level.Footnote 16 Change is frequently ad hoc and incremental, producing new gaps in the institutional frameworks, but through the actions of heads of state, IOs are ‘failing forward’.Footnote 17 A second common theme is that emergencies produce a shift toward informality. Because less formalised settings allow for faster coordination, it is the preferred tool of cooperation in periods of discontinuity and contention.Footnote 18 Finally, the literature that focuses on political leadership in IOs has shown that authoritative individuals, such as executive directors and senior officials, can successfully navigate their IOs through crises if they rely on their personal skills and networks.Footnote 19 We thus know that the perceptions and actions of political leaders – on the part of the IO secretariat and member states – matter for IO crisis management and that this can induce change in key institutional features. The sense-making processes and strategies of regular IO personnel, especially operational staff, remain underexplored.
This is not to say that IO operations have been ignored in the literature. IR scholars have thoroughly investigated crisis management capacities. These authors focus on IOs’ official competencies and functional capabilities to explain variance in organisational effectiveness and resilience. They have shown that to handle crises well, IOs need to possess sufficient centralised decision-making powers, financial resources, and emergency personnel to intervene early and decisively in crisis situations.Footnote 20 Thanks to Hardt’s research, we also know that next to institutional capacities, organisational cultures and the social networks of IO bureaucrats are crucial features for explaining differences in the speed of IO responses.Footnote 21
Several authors have further studied whether IO actions during emergencies lead to (self-) empowerment and increased autonomy.Footnote 22 Kreuder-Sonnen, for example, has shown that the ‘state of exception’ that often comes with crises entails ‘authority leaps’, sudden and drastic increases in official competence.Footnote 23 He and colleagues have also demonstrated that IOs expand their functional scope in crises.Footnote 24 This is something that general-purpose IOs with large bureaucracies are best equipped to do, as Debré and Dijkstra have recently demonstrated.Footnote 25 In brief, we know that institutional design matters for IO crisis management capacity and resilience, and that emergencies provide windows of opportunities for institutional development. What the highly productive research agenda on IOs as crisis managers has yet to account for is how regular IO staff in secretariats and field missions cope with crises and how this reorders how global issues are governed.
Crises as change incubators in IOs
A second body of literature to address the question of how IOs change (policies) in crises puts a stronger emphasis on the change dynamics. Authors like Moschella and Tsingou start from the widely shared assumption that regulatory failures generate crises; redressing these failures, therefore, becomes a ‘key policy priority’.Footnote 26 These authors focus on the IO bureaucrats acting as policy entrepreneurs when things are in flux. For example, Braun, borrowing from Lévi-Strauss’ famous notion, has argued that emergency crisis managers are ‘bricoleurs’, people who use ‘whatever they have at hand’ to handle acute crisis situations.Footnote 27 Drawing on the euro crisis as a case study, he posits that the spontaneous actions by emergency crisis managers often leave a lasting imprint on the ideational and institutional landscape in which they are embedded.Footnote 28 There is, hence, a growing body of literature that reconstructs how the practices of IO bureaucrats in crises shape organisational perseverance and transformation,Footnote 29 and how the organisations actually address transnational challenges. However, studies in this line of research, too, pay little attention to how those stationed in the field shape these processes, as their analytical focus lies on the dynamics at headquarters.
In short, the question of whether and how frontline workers affect IO crisis policy-making and foster change remains unaddressed. This is all the more surprising given that it is precisely their job to act as first responders and adapt to unforeseen, local circumstances.
How frontline crisis management leads to informal policy change in IOs
To understand how decisions by frontline crisis managers can initiate change, I develop in this section a five-step model that draws attention to improvisation by ground-level staff and the consolidation practices of headquarters personnel. For analytical clarity, I divide the model into five ideal-typical steps, although I expect them to overlap in empirical reality: 1) sense-making, (2) testing provisional solutions, (3) negotiating the feasibility and appropriateness of these initial solutions, (4) routinising them on the ground, and (5) consolidating them at headquarters in IO training programs and best practice guidelines (see Table 1).
Table 1. Practices Driving Improvisation and Stabilisation.
This five-step process is not a goal-oriented, conclusive mechanism that linearly links crisis to change. Instead, it needs to be understood as a cyclical, iterative dynamic. The inherently improvisational nature of frontline crisis management means that even once codified in organisational practice, policy solutions developed in crisis moments will again be destabilised – adapted, stabilised, and destabilised – when confronted with new realities.
Further, the processes I outline in more detail in the following comprise staff members from different (types of) IOs. To be able to conceptualise how the practice of frontline crisis managers triggers informal policy change, I treat the people who work in organisations as the main unit of analysis. This means that instead of starting from the assumption that IOs have fixed, impermeable boundaries, and defining organisations as collective actors that can be differentiated by formal characteristics, transnational professionals take analytical primacy.Footnote 30 Accordingly, I can include personnel from IGOs, NGOs, and regional agencies in the analysis without making a conceptual distinction between public and private organisations.
Improvising in the field
Bottom-up policy change begins with frontline crisis management, which, at its core, amounts to improvisation. Improvisation, generally speaking, is the act of making or doing something with whatever is available at the time.Footnote 31 It is based on spontaneous decision-making, not deliberate or considered action. Frontline crisis managers ‘do not have a great deal of time to be reflective and, instead, must often improvise to find good interim solutions to problems that were unanticipated, or to problems that cannot be ameliorated through standard operating procedures in routine times’.Footnote 32 In the remainder of this subsection, I break improvisation into three practices: sense-making, testing, and negotiating.
The first step of improvisation is sense-making. Crises disrupt established patterns of action and known policy scripts: ‘The core of crises is thus the uncertainty created by exposed inconsistencies, contradictions and conflicts’.Footnote 33 In addition to uncertainty, crises generate urgency, stemming from an immediate threat to a given community. The intensity of the three defining crisis features is especially pronounced during a crisis' emergency phase and ‘fast-burning’ crises where shocks are instant and abrupt.Footnote 34 Notably, however, the degree and reference point of uncertainty, urgency, and threat cannot be objectively measured but, following a constructivist understanding of crisis politics, depend on the perceptions of those affected. ‘The prominence of perception, in turn, underlines the potential for framing contests and discursive struggles over the interpretation of a given social reality’.Footnote 35 It follows that personal and inter-subjective sense-making is the first step of crisis management, at any political level. Such sense-making dynamics not only define what constitutes a crisis but give way to more general negotiations over how to organise collective action on public problems: ‘crises are moments where […] agents attempt to persuade each other over “who they are” and “what they want”’.Footnote 36
Operational staff, such as peacekeepers, humanitarian workers, border guards, medical professionals, consular staff, and election observers are often the first to be confronted with crisis events.Footnote 37 In global governance institutions, they are the frontline responders, who ‘[i]n an early stage of the crisis, […] know close to nothing, while the cues, the bits of information, that they do receive or register, will often be confusing’.Footnote 38 They realise that they cannot rely on existing action repertoires and organisational resources because these prove to be ineffective in the changed conditions. At the same time, because of their close physical proximity to crisis outbreaks, they also cannot wait until they have gathered enough information and instructions to reach sound conclusions. The direct exposure to moments of crisis leaves little time for proper learning.Footnote 39 IO staff intervening in constitutional crises in Mali and Burkina Faso, for example, ‘were left scrambling to understand and position themselves on the rapidly changing events’.Footnote 40 In short, working under time pressure and with incomplete information, IO frontline staff need to make sense of what is happening in the field by making quick, practical judgments and acting on them.
This last statement indicates how they give meaning to crisis events: rather than cognitively processing information, they pragmatically test suitable ways to respond. They try out how to alleviate crisis pressures, starting these trial-and-error processes with the familiar. As we know from the literature on crisis management, frontline responders ‘will begin with taking actions that they have been trained to implement or that come intuitively’.Footnote 41 This pre-reflexive process of spontaneously acting on a situation by relying on well-known and well-practiced modes of action simultaneously ascribes meaning to a specific problem and makes a proposition on how to handle it. Here lies the spark for change: innovation comes from combining the known with the unknown, as, for example, Beaumier and colleagues have recently shown with regard to institutional invention.Footnote 42 In that sense, frontline crisis managers are ‘bricoleurs’ who do not work with a blueprint or innovate out of thin air but – more or less skilfully – patch together what they know and have with the conditions they find themselves in.Footnote 43
Importantly, these twin processes of making sense of a situation and testing ways to handle it do not occur in social isolation.Footnote 44 Field staff with different professional backgrounds – and often also from various organisations – work in close physical proximity to each other. Because they necessarily have competing understandings of what a situation means and requires,Footnote 45 their practical propositions might be inconsistent or conflictive. Frontline staff thus engage in contests over the appropriate course of action: they negotiate which preliminary solutions are feasible and acceptable.Footnote 46
Put differently, because crises present situations that were not anticipated, frontline workers need to engage in ‘bad behavior’,Footnote 47 they need to bend the rules and deviate from existing organisational norms, protocols, and routines.Footnote 48 These deviations are grounded in their professional knowledge and intimate understanding of how to get their job done in ‘normal times’. As frontline crisis managers work in complex organisational environments, what feels familiar to one might look unusual to another, likely to prompt social confrontations. Whether an improvised action thus constitutes ‘bad behavior’ or a ‘good interim solution’ depends not on an individual, personal assessment but on intersubjective negotiations over how to address a specific crisis situation. Here lies the second spark for change: Organisation scholars have shown that the chances of a single organisation or profession taking control are limited in turbulent times; the resulting inter-professional and inter-organisational struggles over how to understand and deal with specific situational challenges produce innovations.Footnote 49
While improvisation is thus likely to bring about change, the degree to which improvisation occurs during emergencies may vary depending on the leeway for mission personnel. Institutional features, the local crisis context, and the policy field’s degree of formalisation and politicisation are likely to condition the discretionary powers of first responders. Because of their distance from tight organisational cultures and principal oversight in IO secretariats, fieldworkers generally enjoy wide latitude in their decision-making.Footnote 50 However, IOs differ in their institutional design. More limited mandates, narrower task descriptions, stricter lines of command, and more extensive reporting requirements, for instance, might keep field staff from going off script and experimenting with new practices.Footnote 51 Short-term contracts, on the other hand, might lower concerns over early dismissal, leading to less risk-averse behaviour and fostering initiative.Footnote 52 The scope for discretion will also be shaped by the local crisis context and domestic laws, especially when they criminalise certain actions, provide strict codes of conduct, or refuse access to specific territories.Footnote 53 Emerging or under-regulated issue areas, in turn, may enable improvisation: ‘Bricolage is arguably more apparent in new global problems than in highly legalized policy domains’.Footnote 54 The politicisation of an issue area or specific crisis, by contrast, may constrain it: the higher the (perceived) stakes, the stronger the salience and contestation of what fieldworkers do. Perceptions of political leaders and the broader public thus might also affect fieldworkers’ room for manoeuvre.Footnote 55
Stabilising improvised solutions at headquarters
Just like leaders, crises compel IO employees on the ground ‘to make rapid decisions by relying on ad hoc mechanisms’.Footnote 56 These improvised decisions might not only create local innovations. In this sub-section, I discuss how routinisation and consolidation can stabilise local solutions.
In so doing, I build on Oksamytna’s argument that ‘[i]nnovations have to be institutionalized’ to lead to sustainable organisational change.Footnote 57 She defines change as ‘full issue institutionalization’, which comprises numerous steps, such as creating new posts and accountability mechanisms.Footnote 58 To detect policy changes at low institutional thresholds, I focus on two of the features that she identifies: (1) the creation of policy or guidance – that is, ‘tools for implementing the new task’ – and (2) issue-specific training, which ‘ensures that IO personnel are aware of policies and guidance documents and that they are able to apply them in practice’.Footnote 59 In my understanding, change through improvisation in the field is stabilised through creating and teaching guidelines at headquarters.
That said, the stabilisation process begins in the field. Once a new, improvised way of doing has been accepted as feasible and acceptable by first responders, it will likely be repeated. Routinising preliminary practices reduces the burden of testing and negotiating how to handle things imposed on frontline workers because they are directly exposed to global crises. It follows that they are likely to establish informal protocols, standard operating procedures, briefings, and on-the-job training programmes to routinise their on-site practice.Footnote 60 These routines are then further stabilised by personnel based at the head office.
IOs with a field presence usually have training and operations units at their secretariats. Their staff members produce practical guidance and instruct field staff. Peacekeepers, naval officers, diplomats, border guards, asylum experts, humanitarian aid workers, agricultural specialists, and immunisation experts all undergo training before entering the field. Since it is now widely accepted that professional training in organisations works most effectively through the active learning of skills, know-how, and competencies rather than the transmission of abstract knowledge, training and operations officers draw on the experiential knowledge of field staff to create training tools. The IMF, for instance, teaches through modalities that encourage economic policy officers to ‘apply theoretical models and macroeconomic formulas to practical examples of economic problems in order to foster common diagnostic skills’.Footnote 61 Accordingly, we can expect IO training officers to make use of operational knowledge when they are updating or designing new learning programs. They consolidate improvised ‘courses of action over issues of common concern’ emerging from frontline crisis management in textbooks, manuals, and scripts for best practice.
The extent of consolidation is likely to be conditioned by the institutionalisation of training programs which differs across IOs.Footnote 62 Explicit training mandates, well-funded training departments, and a training philosophy that is centred on the active learning of competencies, as well as operations officers with direct access to field staff will be most conducive to effective consolidation. Similarly, consolidation might be shaped by the degree to which an organisation’s staff rotation system allows it to retain mission staff. If field personnel cannot access secretariat positions, they leave an IO, which has been shown to obstruct the transmission of knowledge in organisations.Footnote 63
No matter the degree, consolidation is a political process – an ‘internal battle of ideas’ – during which ‘individual staff members [take deliberate actions to] embed their inferences from experience into organizational […] routines’.Footnote 64 Headquarters staff fight over which preliminary solutions count as valuable ‘lessons learned’ and suppress local knowledge deemed irrelevant.Footnote 65 Those practices from the field, which are codified in guidelines, form the basis for training future cohorts of professionals deployed to missions and projects worldwide,Footnote 66 which is how they are anchored in the organisations.
While bottom-up change thus starts with quick, spontaneous decisions, it becomes more reflexive as it cycles through its five steps, with IO personnel more consciously discussing and contesting which actions to accept, repeat, and consolidate. Of course, this does not mean that bottom-up change resolves the initial policy contradictions that led to a crisis in the first place. Stabilised frontline measures are informal, provisional solutions that only temporarily patch over gaps in policy frameworks, which is why I expect processes of improvising and stabilising to repeat themselves.Footnote 67 Still, they are consequential because they affect the life choices of the final addressees of global politics, and they are political because the field and headquarters staff fight over how to handle issues of public concern.
The European migration and asylum crisis: Changing how to govern mixed migration from the ground up
In the previous sections, I outlined the mechanisms at play during bottom-up policy change and identified potential scope conditions for their extent to develop the analytical tools that allow us to examine how frontline crisis management affects global governance, something that previous studies of IOs, crises, and change have neglected to do. In this section, I use the case of IOs handling the European migration and asylum crisis at the Schengen border in Greece to demonstrate the analytical value of my conceptual propositions. Although the case study does not allow me to systematically test the mechanisms and conditions across cases, it plausibilises the framework developed above, charting the road towards a comparative research agenda on the transformative power of frontline action in global politics.
The case study
In 2015 and 2016, over two million people irregularly crossed the land and sea borders into Europe. Mainly coming from Turkey and North Africa and traveling by foot along the Western Balkan route or in small rubber dinghies across the Eastern Mediterranean Sea, they created ‘spontaneous arrival asylum’Footnote 68 sites along the EU’s external border. The ‘international community was unprepared for such large numbers of people’.Footnote 69 As a result, a humanitarian emergency developed, and the Dublin system, according to which asylum-seekers need to claim international protection in the country of first entry, collapsed. Many transit and destination countries witnessed a xenophobic backlash as host communities felt overwhelmed with the task of registering, accommodating, and integrating the new arrivals, producing a crisis in governance. In response, the European Commission adopted the European Agenda for Migration, devised the ‘Hotspot Approach’ and deployed Migration Management Support Teams consisting of Frontex, EASO, and other EU agencies to Greece, Italy, and Malta. IOM, UNHCR, humanitarian NGOs, and transnational volunteer groups also started working there.
Before I show in what follows that the first line response by IO field personnel led to informal policy change, I offer a brief discussion of the case’s generalisability. The case is exceptional because the topic’s intense politicisation gridlocked international and European decision-making arenas. The resulting under-regulation meant that hardly any institutional solutions were in place to process the high number of people crossing the border in a very short time, resulting in an extremely rapidly burning crisis. Further, this crisis is exceptional with regard to the large and diverse number of organisations that responded on the ground. Situated in Europe, the crisis created high political salience. It attracted not only UN agencies, intergovernmental organisations, and a myriad of humanitarian NGOs, the specialised European agencies also started to upscale their operations. These exceptional features make it a case study that is particularly suitable for theory-building: Extreme or paradigmatic cases promise to be the ‘richest in information’,Footnote 70 allowing the reconstruction of the mechanisms and conditions of interest. I have thus relied on this case study to generate the propositions outlined above, going back and forth between the literature and insights from the empirical analysis.
At the same time, the case’s exceptionality limits the scope of inference regarding the conditions that increase the likelihood of bottom-up change emerging. Yet, the case study also exhibits typical features of operational IOs intervening in humanitarian emergencies. Previous studies have shown that in global health emergencies,Footnote 71 in violent conflicts,Footnote 72 in maritime security crises,Footnote 73 and after environmental disasters,Footnote 74 organisations such as the International Committee of the Red Cross (ICRC), the WHO, the UN Department of Peacekeeping, Médecins Sans Frontières, the UN Office for the Coordination of Humanitarian Affairs (OCHA), the WFP, and NATO send first responders who improvise in the field, and their actions shape informal policies beyond the immediate crisis context.Footnote 75 NATO, for instance, scaled up its on-the-ground cooperation practices with civil society actors on the protection of civilians. What started as a local, informal coordination policy in Afghanistan became an organisation-wide standard-setting mode.Footnote 76 Field-based ways of doing things ‘trickle up’Footnote 77 in many crises and IOs; also beyond migration and asylum.
These examples and the case study allow us to identify potential conditions likely to trigger improvisation and stabilisation across cases. For change to be initiated by frontline action, IOs need to have field staff directly exposed to crisis events. The immediate confrontation with an urgent, threatening problem and the uncertainty of how to deal with it will likely trigger improvisation. My propositions thus apply to organisations with an operational component rather than normative or programme IOs. Further, they apply to fast-burning crises where confusion and time pressure are intense rather than slow-burning, ‘creeping crises’ where systemic flaws are incubated and erupt eventually.Footnote 78 In humanitarian emergencies, this pressure is amplified by the fact that humans actively demand quick solutions. Finally, IO personnel must be exposed to contesting viewpoints. If they worked only with colleagues socialised into similar organisational frames and professional perspectives, they could more easily stick to ‘business as usual’ action patterns. Contestation can, of course, take place within organisational boundaries. Yet, the more organisations and professions operate within a specific place, the more likely disruption and negotiation will occur.
Two crises features and two organisational features are thus likely to be the main conditions for the occurrence of bottom-up policy change in reaction to crises: fast-burning crises and IO field presence (confrontation with a problem), and humanitarian emergencies and organisational complexity (confrontation with others). We can expect to find these features in policy fields ranging from food and maritime security to state-building and peacekeeping, environmental disaster management, and public health.
Methods and data
The two-level case study builds on fieldwork at the European external border in Greece and at the headquarters of the IOs operating on the ground. It includes 11 organisations: IOM, UNHCR, Frontex, EASO, and seven international humanitarian NGOs. I relied on ethnographic methods to gain first-hand insights into the frontline response. These are particularly useful for collecting data on experience-near phenomena such as negotiations, speech writing, and other practices, including those that drive improvisation and stabilisation.Footnote 79 The primary evidence for the first part of the case study consists of 530 hours of participant observation I made while volunteering on the Greek island Chios in March, April, and May 2016. Chios is one of the central border-crossing points along the Eastern Mediterranean migration route and one of the EU’s ‘hotspot’ sites, with over 700 people traveling through the island daily in 2015 and 2016. My observations were recorded in field notes and included over 60 conversational interviews. Another 16 recorded interviews were conducted with practitioners from different professional groups and all the IOs present.
As a member of a transnational volunteer team, I participated in shore patrol, helped with beach landings and port disembarkations, entered the hotspot called ‘Vial’, worked in the two open campsites on the island, attended weekly coordination meetings at the town hall, participated in local working groups, took part in a search and rescue training exercise, drank coffee on a Frontex ship after a night shift, lunched with EASO staff, and regularly went to a local bar with UNHCR and NGO personnel. I cross-checked and contextualised my fieldwork findings with formal and informal organisational documents, such as IO reports, meeting minutes, WhatsApp chat groups, and reports by other volunteers and researchers.
Interviews and training documents served as primary sources for the second part of the case study. Aiming for organisational balance, I spoke with 27 training and operations experts from UNHCR, IOM, Frontex, EASO, and several NGOs in February, March, and May 2017 as well as in January 2025 to trace crisis effects over time. As is typical for elite interviews, interviewee selection was based on purposive and snowball sampling. The semi-structured interviews combined factual questions with in-depth discussions of organisational processes. In addition, I coded 119 training documents to analyse whether and which routines were stabilised. For research ethics reasons, I only included transnational professionals in the study, disclosed my role as a researcher as openly as possible to everyone I talked to, and pseudonymised interview transcripts.
Improvising at the border: Developing spontaneous policy solutions on the ground
In what follows, I show that the analytical distinction between sense-making, testing, and negotiating is useful if we are to understand how IO first responders foster change in crises. At the European external border, fieldworkers were directly confronted with the emergency. People were dying at sea as their boats capsized; pregnant women, small children, and older people were sleeping on small patches of grass in public squares due to the lack of accommodation; asylum-seekers could not get registered because the number of new arrivals surpassed the capacity of officials in the reception facility, thus people crossed the border without any recording of their personal information. Because of their direct exposure to this fast-burning crisis, IO field personnel felt compelled to act: ‘Then, once you are here, you […] cannot avoid the responsibility to attend to these people’.Footnote 80 At the same time, they described the situation at the border as messy and chaotic. Because many organisations operated in the field, uncertainty about organisational responsibilities prevailed. In the words of a humanitarian professional: ‘Everything is just very scattered and confusing’.Footnote 81
Sense-making. To cope with the urgency and uncertainty at the border, frontline staff relied on their knowledge as border guards, lifesavers, camp managers, doctors, psychologists, social workers, humanitarian professionals, migration lawyers, asylum caseworkers, and return escorts. An EASO asylum expert stressed that when he first started working in the EU hotspot on Chios, where new arrivals were screened, debriefed, registered, and interviewed, he drew on his experiential knowledge of having worked at an airport reception centre.Footnote 82 A camp manager underlined the importance of her experience in refugee camps at the Horn of Africa before being deployed to Greece. It helped her give meaning to the events she encountered and understand what was required of her.Footnote 83 Volunteer lifesavers and firefighters, usually working for the Spanish Coast Guard, said they relied on their nautical skills and experience working under pressure to improvise search and rescue procedures in Greece. By applying their previous understanding of how to deal with particular issues, for example, responding to swimming accidents at Spain’s beaches, to a new situation, for example, migrants irregularly crossing sea borders, frontline crisis managers tried to make sense of what happened at the border.
Interestingly, these sense-making processes took place with very little interference from headquarters: ‘So, it’s been everybody, but here on the ground in Chios. It’s not been something that has been directed from a central level’.Footnote 84 The observational and interview data reveals that whenever mission personnel asked for instructions, they were put off, either being told that ‘Nobody thought of that!’ or ‘We’ll see!’.Footnote 85 Fieldworkers of all IOs agreed, ‘It’s very difficult to get your own organisation to commit to anything really because they know that things will change very quickly’.Footnote 86 As the mandates of their deploying organisations also provided field staff with only a limited base for responding to the unpredicted situational challenges they encountered at the border, they relied on their professional skills instead, trying ‘to adjust as much as possible and be creative’.Footnote 87
Testing provisional solutions. By relying on familiar action repertoires in unknown situations, IO field staff slowly started to create new procedures. This is shown in how ‘admissibility interviews’ were organised under the so-called ‘fast-track border procedure’. This procedure, initially introduced as a temporary measure by the EU-Turkey Statement, stipulated that an additional interview had to be conducted before the actual asylum interview. Its purpose was to establish if an asylum claim was admissible in Europe or if a person had travelled through a ‘safe third country’ or a ‘first country of asylum’ so that the asylum claim had to be lodged there.Footnote 88 For people arriving on Chios, that country is usually Turkey. EASO asylum experts were tasked with conducting the admissibility interviews and drafting opinions on the outcomes. Unaccompanied minors, other vulnerable people, and ‘Dublin’ cases (people eligible for family reunification under the Dublin III Regulation) were formally exempt from these interviews and allowed to travel to the mainland. Despite the seemingly straightforward legal basis for the interview, the question of who would qualify as ‘vulnerable’ remained. Through trial and error, asylum experts on the ground worked out how to determine ‘vulnerability:’
They set up the standard frame for the interview, and in the beginning, in my opinion, they thought of it as a little bit too easy. ‘Just ask some questions: “How did you get here? Do you have any family in Europe?” And, “Why can you not go back to Turkey?”’ That [was] basically the outset for this. But then in the course of time, after the first interviews were done and from the experience and also the comments from the Asylum Service, it became clear that we had to ask many more things during the interview. [And what was especially tricky, was how to determine vulnerable cases]. So that’s why a lot of times we had to discuss among the interviewers, you know, ‘How far do we go? And do we have enough [information] to know whether or not they are vulnerable?’Footnote 89
EASO hotspot staff had to learn how to establish ‘vulnerability’ and whom to grant access to the asylum procedure in Greece. The quote shows that although some formal policy exists on which the admissibility interviews are based – mainly the EU Asylum Procedures Directive, the EASO Special Operating Plan to Greece, and the Greek law implementing the EU-Turkey Deal – ‘the legislative provisions are not directly usable to carry out decision-making’ at the street level.Footnote 90 EASO hotspot staff needed to interpret, bend, and expand them substantially to make them work in the crisis context. ‘There was a lot of testing and lots of first experiences on the ground’.Footnote 91
Importantly, EASO staff did not make spontaneous decisions and only reflected upon them among themselves. To define what ‘vulnerability’ means, they had to cooperate, compete, and argue with staff from other IOs with different professional backgrounds. Legal experts, for instance, were eager to make sure that those who potentially qualified as vulnerable could assert their rights and not only be exempted from being returned to Turkey but also travel onwards from the border. However, the fast-track border procedure did not make provisions for legal counselling before or during the admissibility interview. While EASO’s field coordinator stressed that ‘Things should be as legal as possible!’ during one of the protection working group’s meetings,Footnote 92 neither EASO nor any other European agency provided legal support during the procedures. Migration lawyers working for small NGOs, assisted by UNHCR staff, therefore started to organise impromptu preparations for the interview.Footnote 93 In addition, asylum seekers began to request the doctors working in small clinics in the open campsites to provide medical assessments certifying their vulnerability. During one of the General Coordination Meetings, one of the exhausted doctors complained, ‘Doctors should only do doctory things!’Footnote 94 He did not consider it his task to help with the asylum procedure, but he and his team nonetheless drew attention to the medical conditions that put people at risk. As a result of these disputes among asylum caseworkers, lawyers, and medical professionals over how to determine ‘vulnerability’, the initially limited catalogue of questions and indicators defining who could claim international protection in Europe was extended significantly, drastically slowing down returns from the border.Footnote 95
Negotiating feasibility and appropriateness. The example of the inter-organisational and inter-professional struggles over defining ‘vulnerability’ illustrates that fieldworkers often disagree strongly over what is at stake in crisis situations. It is never immediately obvious what constitutes a ‘threat’ during crisis events, and assessments regarding who was at risk in the emergency under study differed widely. For border guards, the distinctive threat stemmed from the large number of irregular crossings. In 2015, fingerprints were only taken from eight percent of the arriving migrants in Greece, meaning the majority of new arrivals were not identified and registered.Footnote 96 For humanitarian workers and search and rescue professionals, the absence of safe migration routes and inhumane accommodation facilities constituted the most significant threat. For migration lawyers, the primary concern was protecting procedural safeguards and ensuring people could access their rights under international law.
These different professional perspectives and priorities over borders, people, and the law clashed whenever local practitioners with various backgrounds had varying approaches to the same situation, particularly because framing things according to your professional logic also means being able to claim control. In one instance, Frontex border guards tried to prevent humanitarian professionals from providing basic provisions for new arrivals during the landing of a small boat on Chios’ eastern shore. The border guards were trying to arrest someone they had identified as the boat’s pilot because they thought the most crucial thing in that situation was to arrest smugglers. Humanitarian workers, lifesavers, and medical professionals considered getting everyone onshore safely a priority. They helped navigate the boat to the rocky beach and the passengers to land, conducted medical check-ups, and tried distributing water, sweet snacks, space blankets, and dry clothes.Footnote 97 Both sides felt that the others were getting in the way of doing what was most urgently required at that moment to prevent harm. However, because fieldworkers operated in the same physical locations, they could not avoid each other and had to negotiate acceptable trade-offs for all professions involved. They negotiated the meaning of large numbers of people crossing the maritime border without valid travel documents, in rubber dinghies, wearing defective life jackets, and how to respond adequately. In so doing, they established search and rescue, disembarkation, and registration procedures that included ensuring the safety of people and their official registration.
In sum, first responders organised a series of provisional policy solutions ranging from search and rescue, registration, protection, accommodation, asylum interviews, and legal aid through the three interrelated steps of improvisation – none of which would have materialised without their initiative. During these processes, not all fieldworkers were equally proactive and resourceful. In the words of a camp manager, ‘Some people, they just follow what is in the book as well, you know?’Footnote 98 Interviewees agreed, however, that in general, ‘everyone is fully committed and trying to be flexible […] [and interacting] regardless of different mandates’.Footnote 99 They were aware that improvisation could have adverse outcomes and whether and which pragmatic solutions were considered suitable and legitimate resulted from negotiations among professionals with different sets of skills.
Stabilising in secretariats: Learning from the field
Routinszing accepted provisional solutions. The tentative solutions fieldworkers could agree on became routinized on the ground. The interview partners stressed that creating the operational knowledge to handle migration and displacement at the border ‘was an uphill struggle all the way’.Footnote 100 This is why they were eager to convey that knowledge to new practitioners and recorded it in detailed protocols, checklists, referral mechanisms, and standards operating procedures (SOPs). These documents were frequently updated in light of new experiences. As the EASO field coordinator for Chios said, ‘We are all learning by doing. [On] Lesvos, the SOPs have been revised five times’.Footnote 101 They also organised shadowing with experienced frontline actors, face-to-face workshops, and local simulations to transmit routine procedures.Footnote 102 The coordinator of a search and rescue NGO stressed how important it was to preserve informal routines:
The problem that I see right now is that because there is no massive coming of refugees, maybe we are losing some kind of knowledge [about] handling the situation and this is why I try to make some exercises, just to retain […] that. Because in wintertime, we had arrivals – five, ten boats per day – and everybody knew perfectly what to do […] and because now there are new volunteers, all this knowledge [is getting lost] – and this is one of my main concerns: not to forget that!Footnote 103
He and his team organised regular rescue exercises to retain provisional procedures.
Consolidating provisional solutions in training tools. Many headquarters’ staffers shared their concern about anchoring lessons learned during the height of the crisis within their organisation. The interviews with people who work in the training and operations departments of the IOs that operated on Chios show that they tried to consolidate operational knowledge further ‘to make sure that next time it [a migration crisis] comes around, [our response is] a lot better planned and we have systems in place’.Footnote 104
To do that, they tried ‘to capitalise on as much information as possible that we can gather from our target audience’, that is, the field workers.Footnote 105 As this brief comment from a UNHCR learning officer shows, secretariat personnel are not only training frontline responders to be aware of specific policies and programs but, in reverse fashion, are seeking to learn from them. Through briefings and debriefings, pilot projects, workshops, on-the-job training, and similar activities,Footnote 106 they have developed detailed guidelines for conducting search and rescue operations, providing shelter and managing a camp, screening and debriefing migrants, interviewing asylum-seekers, protecting refugees and their rights under international law, preventing gender-based violence, providing psychosocial support as well as returning rejected asylum-seekers or migrants who want to return to their home countries voluntarily, thereby embedding operational knowledge developed on the ground in organisational resources.Footnote 107
Take the admissibility interviews again as a concrete example: how EASO asylum experts working in the hotspots decided to carry out the interviews had important institutional repercussions because their practice defined what ‘vulnerability’ actually means, especially in light of the absence of comprehensive formal definitions.Footnote 108 Their street-level decisions then became normalised in EASO’s learning tools and diffused through training across the EU.Footnote 109 The agency thus created guidance, which effectively substituted for official policy, as the former head of training clearly confirmed: ‘Normally, the training is about something [e.g., a policy guideline] which already exists. But indirectly and quite often, the training content becomes guidance [as such] because, in many situations, there is not really a pre-existing guideline’.Footnote 110 Through the creation of such policy scripts and training fieldworkers, secretariat staff consolidate local innovations.Footnote 111
This is a contested process. A UNHCR official stated that especially during inter-organisational consultations on learning materials, staff from different IOs compete for ‘Deutungshoheit’ (interpretational sovereignty).Footnote 112 They defend their jurisdiction and, simultaneously, are careful not to press too far into new issue areas. They become knowledge brokers who translate knowledge from the field and assert their competence vis-à-vis other organisations, also to signal control to their superiors in their home organisation or external principals such as donors.Footnote 113 This also means that some knowledge from the ground does not travel. For example, the Sphere Standards, negotiated by over 400 humanitarian NGOs, do not include some of the local experiences. One trainer who was closely involved in writing and updating them said, ‘A lot of things are simply not covered in Sphere. They are not covered because they couldn’t be agreed – not that they are not important’.Footnote 114 An IOM staff member confirms that when headquarters staff devise handbooks, they try to prevent other IOs from encroaching on what they view as their area of expertise:
With some topics, collaboration is more difficult. For example: ‘Do the guidelines only refer to migrants or to migrants and refugees?’ It has been quite difficult to agree with UNHCR on a definition – it is always quite difficult to find a common denominator on that [laughs].Footnote 115
These examples illustrate that some field knowledge gets lost in translation. They also show that while we observed boundary blurring during improvisation on the ground, organisational politics and boundary making become more important during stabilisation processes at headquarters.
Finally, it is important to underline that the improvised and stabilised ‘courses of action over common concern’ remain provisional, which many secretariat interview partners acknowledged.Footnote 116 This is also expressed in many training documents. One of IOM’s psychosocial support tools explains that handbooks collect pragmatic solutions:Footnote 117
The tools […] are tailored to suit the operational aim, emergency timing, and limited quantitative and qualitative capacity of IOM assessments teams in emergency. As such they are not the best tools in general, but proved to be the best possible ones in order to grasp the complexity of the situation in the given time and operational capacity.
The UNHCR Reference Guide of Good Practices for operational protection in camps also makes clear that these practices are ‘not perfect’:Footnote 118
More than forty ‘good practices’ are included in this Guide, drawn from some twenty-three different country operations. The good practices included are not exhaustive and certainly not perfect. […] They likely present the best achievable solution to an issue in a difficult working environment.
The preliminary nature of the informal, practice-based policies that result from frontline crisis management means that they will be adapted again in new processes of improvisation and stabilisation: ‘It’s a never-ending circle’.Footnote 119
The empirical analysis has shown that without the creativity of first responders, new ways of handling migration and asylum would not have emerged: ‘2016 was a game changer’.Footnote 120 Without proactive training and operations officers, people at the headquarters would not have known about the ad hoc solutions developed in the field. ‘Thanks to 2016, we have massively scaled up our operational capacity. We now have good routines, templates, and SOPs’ to govern migration and asylum.Footnote 121 In the case of the European crisis, the frontline response did lead to – albeit incremental and informal – policy change from the bottom up.
Conclusion
‘Crises are open decision-making situations’Footnote 122; they reshuffle how global problems and adequate solutions are defined. How the actions of political elites and institutional crisis management capacities shape crisis politics and results is well explored. What is under-researched, by contrast, is how the practical judgments by regular IO bureaucrats influence crisis responses and outcomes. To address that gap, I have developed a theoretical framework that foregrounds improvisation in the field and the ex-post stabilisation of field-based policy decisions at headquarters. This framework helps account for how the actions taken by everyday crisis managers in operational IOs lead to informal, bottom-up policy change, recentring attention from institutional effects to substantive governance outcomes. To illustrate the analytical value of my theoretical propositions, I have relied on a two-level case study of the IOs handling the 2015–16 European migration and refugee crisis at the Schengen border in Greece. The analysis has demonstrated that first responders and training and operations officers changed how their organisations govern migration and asylum, even if that change remained provisional.
My framework relates in particular to fast-burning, humanitarian crises and IOs with in-built emergency response mechanisms that can deploy operational capacities to protect civilians in conflict, assist victims of environmental disasters, and provide aid to displaced communities. Further research is required to establish whether frontline crisis management also causes policy change in IOs where the ‘frontlines’ look differently. In the 2008 global financial crisis, for instance, the world ‘turned to the IMF to be a financial firefighter’.Footnote 123 How did the practice of IMF personnel conducting bilateral surveillance and negotiating loans in countries such as Mexico, Pakistan, and Iceland, change the organisation’s overall policy approach to financial supervision?Footnote 124 In the 2010 Euro crisis, the European Commission and the European Central Bank developed adjustment programs to stabilise crisis-ridden member states, including Greece. How did the pro-growth strategy pursued by the task force of EU officials that was providing technical assistance to implement these programmes on the ground (i.e. in Greek ministries) change European economic governance?Footnote 125 Investigating the longer-term implications of decision-making by frontline bureaucrats in IOs with crisis management structures that differ from typical operational organisations will provide additional insight into how policies are changed from the ground up during turbulent times.
Similarly, further research is needed to understand how decisions by frontline workers influence the governance of slow-burning crises. While the agency of first responders might be more restricted in less acute crisis contexts, initial evidence suggests that ‘lagged learning’ can result from spontaneous decision-making in crises – leading to policy change in the long run.Footnote 126 Follow-up research will help us understand the scope for bottom-up improvisation and stabilisation in crises that are characterised by less tempo and intensity than emergencies.
Finally, future research should investigate the conditions under which bottom-up processes are productive: it is not difficult to think of cases where improvisation in the field and the subsequent normalisation of on-the-ground patterns of behaviour caused harm.Footnote 127 Follow-up studies should analyse variance in institutional characteristics – such as discretionary powers for mission personnel, staff rotating systems, and local accountability mechanisms – to determine when change from the field up creates pathologies and when it enables IOs to be innovative global governors that respond to transnational challenges effectively.
Interviews
Acknowledgments
I would like to thank Len Seabrooke, Jens Steffek, Lisbeth Zimmermann, Nicole Deitelhoff, Christian Kreuder-Sonnen, Stephanie Hofmann, Benjamin Faude, Nele-Marianne Ewers-Peters, Benjamin Dassler, Andrea Liese, Jutta Joachim, Tobias Wille, Alex Tokhi, Ben Christian, and Max Lesch as well as the participants at the IT workshop in Montréal 2023, at the EWIS workshop in Amsterdam 2023, and my colleagues in Florence, Frankfurt, and Darmstadt for their helpful comments on earlier versions of this article. I particularly thank my interview partners for their time and insights.
Open access
