Introduction
Crisis management has increasingly shifted toward wide-scale cooperation among state first responders, military agencies, and diverse Civil Society Organizations (CSOs). Yet cooperation remains challenging due to differing organizational cultures and communication systems. Much effort has focused on developing the “hardware” of governance (Boin & ’t Hart, Reference Boin and ’t Hart2010), formal structures, standardized procedures, and interoperable technologies, rooted in a command-and-control approach increasingly viewed as insufficient for dynamic crises (Drabek & McIntyre, Reference Drabek and McIntyre2003). As Capano and Toth (Reference Capano and Toth2025) argue, moving beyond rigid command-and-control requires systems characterized by polycentricity.
This article shifts focus to the relational “software” (Boin & ’t Hart, Reference Boin and ’t Hart2010), investigating networks, power dynamics, and interorganizational ties that constitute polycentric crisis systems (Koinova et al., Reference Koinova, Deloffre, Gadinger, Mencutek, Scholte and Steffek2021; Ostrom, Reference Ostrom1990). It asks: what are formal CSO attitudes toward informal actors, and what is the substance of relationships binding actors in polycentric systems? The study focuses on CSOs, broadly defined to include local branches of international aid organizations, first responder and emergency services, emergent self-organized volunteer groups, and grassroots community organizations, united by their participation in crisis response and their operational independence from government and military structures. Sampling across this spectrum captures how attitudes toward informal actors vary within civil society itself. While all respondents are from CSOs, their reflections on cooperation with state agencies, the military, and other formal actors provide rich data on how different governance architectures are negotiated in crisis management.
Drawing on interviews, focus groups, and case studies from a Horizon 2020 project, the European Union’s framework program for research and innovation, the study employs a modified grounded theory approach (Bryman, Reference Bryman2012; Charmaz, Reference Charmaz2006; Creswell, Reference Creswell2013) across three iterative steps. Framed through polycentricity (Ostrom, Reference Ostrom1990), the analysis examines how polycentric arrangements can include or marginalize informal actors.
The study reveals how four core elements of a functioning polycentric system, namely trust, flexibility, equality, and temporality, are built, sustained, or undermined. It contributes to understanding the relational dynamics within polycentric crisis governance by specifying how these elements are enacted in practice: trust emerges through recognition rather than hierarchy, flexibility through distinguishing specialized from transferable tasks, equality through co-production, and temporality as a resource for adaptive, short-term cooperation. In doing so, the study helps operationalize polycentricity for crisis management, a contribution explicitly called for in the literature (Boin & ’t Hart, Reference Boin and ’t Hart2010; Capano & Toth, Reference Capano and Toth2025; Drabek & McIntyre, Reference Drabek and McIntyre2003).
Literature review
The limits of command-and-control in crisis management
The organization of crisis response in many countries has been historically shaped by a command-and-control model, rooted in the paramilitary traditions of civil protection (Neal & Phillips, Reference Neal and Phillips1995; Quarantelli, Reference Quarantelli1987). This approach seeks to “order chaos” through strict hierarchies, predefined plans, and formalized procedures (Drabek & McIntyre, Reference Drabek and McIntyre2003). While the bureaucratic form provides clarity and breaks down complex tasks, it is also criticized for its inflexibility and slowness (Capano & Toth, Reference Capano and Toth2025; Weber, Reference Weber1978), traits ill-suited to the dynamic and unpredictable nature of crises (Drabek & McIntyre, Reference Drabek and McIntyre2003; Whittaker et al., Reference Whittaker, McLennan and Handmer2015). Boin & ’t Hart (Reference Boin and ’t Hart2010) question this model as a “myth,” arguing that an overemphasis on the formal “hardware” of institutional structures comes at the expense of the relational “software”—the communication, trust, and cooperative relationships between diverse actors that determine operational effectiveness. Bridging crisis management and public policy, Capano and Toth (Reference Capano and Toth2025) argue that moving beyond rigid command-and-control requires systems characterized by “polycentricity, and coordinated autonomy” (p. 14), insights that speak directly to the limitations of purely hierarchical approaches.
Command-and-control logic is not confined to state agencies but has also been adopted by some established CSOs. Through close cooperation with state and military actors, some CSOs have experienced a convergence of norms, leading to bureaucratization that mirrors top-down control (Chaulia, Reference Chaulia2011). This approach can shape perceptions of legitimacy, sidelining emergent groups and local communities who lack formal structure but are active in disasters (Drabek & McIntyre, Reference Drabek and McIntyre2003; Łapniewska, Reference Łapniewska2022). This lack of recognition inhibits the openness and flexibility required for an inclusive response. As Capano and Toth (Reference Capano and Toth2025) note, polycentric governance implies coordinated cooperation of many actors and does not necessarily mean the absence of hierarchy, but a recognition that effective crisis management requires moving beyond exclusive reliance on formal, hierarchical structures. While the organizations might still internally operate as hierarchies, they would not attempt to exercise hierarchical control over other actors in polycentric system.
The reality and potential of informal actors
In contrast to the rigidities of formal systems, the emergence of self-organized groups is a persistent feature of crises. Driven by solidarity and humanitarian impulses, people naturally come together to help, forming bottom-up, temporary organizations (Quarantelli, Reference Quarantelli1986; Simsa, Reference Simsa2017; Simsa et al., Reference Simsa, Rameder, Aghamanoukjan and Totter2019). Classified as “emergent organizations” (Dynes, Reference Dynes1970), they have ambiguous mandates and unstructured processes (Boin & ’t Hart, Reference Boin and ’t Hart2010), yet are not necessarily uncoordinated. During the refugee crisis and COVID-19 pandemic, emergent groups used digital tools to self-structure with remarkable speed and creativity (Łapniewska, Reference Łapniewska2022; Meyer & Simsa, Reference Meyer and Simsa2018). Simsa et al. (Reference Simsa, Rameder, Aghamanoukjan and Totter2019) note a shift toward episodic volunteering, where spontaneous volunteers emerge especially during social crises, often substituting for an absent response system. The temporality of these groups is a key feature, allowing for rapid mobilization and dissolution post-crisis, yet their contributions are often dismissed by formal actors constrained by a command-and-control worldview (Stallings & Quarantelli, Reference Stallings and Quarantelli1985). As Paciarotti et al. (Reference Paciarotti, Cesaroni and Bevilacqua2018) demonstrate, such integration is possible: during a flood in Italy, a local Caritas organization successfully managed spontaneous volunteers by improvising based on need rather than predrafted plans, contrasting with official responders who saw no role for spontaneous volunteers.
Local communities represent a vital yet underutilized informal resource, possessing invaluable local knowledge. However, the relationship between formal actors and communities often remains hierarchical, with communities treated as subjects of expertize rather than active agents (Li, Reference Li2005). This dynamic reflects Scott’s (Reference Scott1998) critique of high-modernist policies that privilege expert knowledge while systematically disregarding local knowledge, a pattern that reproduces itself in crisis management. Consequently, meaningful engagement that genuinely values local knowledge remains scarce, preventing relationships built on equality and mutual respect (Sanderson et al., Reference Sanderson, Heffernan, DeSisto and Shearing2025). Simsa et al. (Reference Simsa, Rameder, Aghamanoukjan and Totter2019) propose that CSOs can address this by creating space for new organizational structures, establishing framework conditions for self-organization, and providing resources for care and coordination.
Polycentricity as an alternative framework
To overcome the limitations of the command-and-control model, where some actors either exercise hierarchical control over other actors or exclude non-typical informal actors, this study adopts polycentricity as its core analytical framework. This concept describes governance systems where multiple, independent actors self-organize (Ostrom, Reference Ostrom1990), shifting the focus from static institutions to dynamic, relational networks (Koinova et al., Reference Koinova, Deloffre, Gadinger, Mencutek, Scholte and Steffek2021). This lens is particularly suited to crisis management, as it facilitates a deliberate shift from the “hardware” to the “software” of response, broadening the range of actors to include the full spectrum of crisis responders (Dynes, Reference Dynes1994). At its core, polycentricity examines the inter-organizational ties that form the “binding glue” of a network (Koinova et al., Reference Koinova, Deloffre, Gadinger, Mencutek, Scholte and Steffek2021, p. 1992). The analytical focus is on how these ties are built and maintained. A critical, though undertheorized, element of this relational glue is trust, which largely determines the capacity of these relationships.
Polycentricity as a concept is deeply indebted to Elinor Ostrom, whose empirical approach began with existing forms of governance as constructed by multiple actors (1990). Drawing on bottom-up observation, she outlined principles for effective polycentric systems that facilitate restoration of local ecosystems through co-produced multi-stakeholder arrangements. Although her work centered on environmental commons, the framework has proven highly transferable to other domains, including crisis management. Nevertheless, Jordan et al. (Reference Jordan, Huitema, Hildén, Van Asselt, Rayner, Schoenefeld, Tosun, Forster and Boasson2015) caution that enthusiasm for polycentricity must be tempered by a harder look at the evidence. Reliable metrics on whether polycentric governance is working as hoped remain scarce.
The current study acknowledges these limitations. Rather than focusing on quantitative performance data, it examines the power dynamics inherent in polycentric governance systems. Koinova (Reference Koinova2022) proposed a relational approach to polycentricity, yet argued that it requires further empirical investigation, a gap the current study seeks to address through qualitative inquiry focused on relationships within such systems. The analytical framework draws on Schobert et al.’s (Reference Schobert, Orru, Gabel, Nero, Windsheimer, Klaos and Nævestad2023) conceptualization of social capital as a situational resource (availability, accessibility, and activatability). To interpret the findings, the study also engages with Scott’s (Reference Scott1998) critique of high-modernist policies, both of which illuminate how expert privilege can marginalize local knowledge within polycentric arrangements.
Beyond its theoretical foundations, polycentricity enables the mapping of relationships among formal agencies, emergent organizations, and local communities. Empirical studies show that polycentric systems foster adaptability (Ansell et al., Reference Ansell, Sørensen and Torfing2021), though fragmentation can lead to coordination failures, highlighting the importance of bridging organizations and network leadership (Nowell et al., Reference Nowell, Steelman, Velez and Yang2018). Storr et al. (Reference Storr, Grube and Haeffele-Balch2017) argue that polycentric systems using dispersed local knowledge are better suited for local recovery than monocentric systems, offering two key advantages, namely the ability to draw on context-specific information and the flexibility to provide assistance at the appropriate scale. Crucially, they find that strong social networks, shared values, and prior disaster experience support collective action, a finding that resonates with this study’s emphasis on shared values and existing networks as drivers of self-organization.
The relational foundations: Social capital
Understanding how order and cooperation emerge in a diffuse polycentric system requires examining the substance of the relationships that constitute it. This involves a theoretical lens of social capital embedded within networks. Social capital, the “trust, norms, and networks” that enable collective action (Putnam, Reference Putnam1995), is the essential resource within the relational “software,” an aggregate of resources available through networks of relationships, functioning as a reservoir of trust and reciprocity that facilitates coordinated action. However, much of the social capital literature focuses on individuals, whereas the present study examines social capital mainly among organizations as a feature of a polycentric system.
Crises can dynamically reshape social capital, giving rise to spontaneous networks and bridging capital through mutual help based on shared fate, a phenomenon termed disaster social capital. Yet Schobert et al. (Reference Schobert, Orru, Gabel, Nero, Windsheimer, Klaos and Nævestad2023) caution that social capital is not automatically an asset as it can be inaccessible, disrupted, or diminished by social norms. They introduce situational social capital, distinguished by the three dimensions of availability, accessibility, and activatability, with activatability being the most potent determinant of whether social capital actually increases resilience. The present study contributes by examining how institutional support for ad-hoc groups can serve as a mechanism for increasing activatable social capital, while also fostering the openness, equality, and temporality necessary for meaningful cooperation with diverse actors.
In synthesis, this framework posits that effective crisis response is best understood through a polycentric model. The viability of this model depends on the activatability of social capital residing in its relational networks. Together, these concepts provide the tools to analyze how the “binding glue” of a resilient crisis response system, characterized by trust, flexibility, equality, and temporality, is actively formed, even within the constraints of temporary engagements.
Methodology
Research design and overview
This study was part of a three-year Horizon 2020 project, leveraging a multi-stakeholder consortium of emergency practitioners, academics, and industry partners. The component of the research into civil society actors aimed at investigating the “software” of cooperation in crisis management to complement its “hardware” focus (Boin & ’t Hart, Reference Boin and ’t Hart2010). The research employed a modified grounded theory approach (Bryman, Reference Bryman2012; Charmaz, Reference Charmaz2006; Creswell, Reference Creswell2013) and proceeded through three iterative steps, with findings from earlier steps informing the design of subsequent ones. By design, research conducted in Step 1, although with its own objectives and research questions, was intended to lay the foundation for research in Step 2. Step 3 did not involve new data collection. Figure 1 illustrates the application of the grounded theory approach, presenting examples of codes, concepts, categories, hypotheses, and the substantive theory at each step. Adopting a qualitative interpretivist epistemology, the study moves beyond description toward analytical explanation of relational dynamics in polycentric governance systems.
Modified grounded theory.
Fig. 1. Long description
The table is organized under five main headers: Stages of Grounded Theory Approach, Codes (Examples), Concepts / Themes (Examples), Hypothesis as Premises for Further Analysis, and Literature.
Step 1: Community resilience in crisis management.
Research questions focus on how community and community resilience are understood.
Methods: Literature review and case study collection from June 2020 to June 2021 involving 8 C S O s.
Main empirical findings: Community is understood in four ways: community of practice, interaction-based community, place-based community, and interactional field. Resilience is defined as an adaptive, transformative, and learning-based process.
Data analysis 1: Codes include Communication, Coordination, and Local knowledge. Themes include Crisis management community as network and Recognition of local volunteers. Hypothesis 1: Integration of volunteers with professional teams leads to greater operational effectiveness.
Step 2: The role of civil society in crisis management.
Research questions focus on the current role, key actors, and challenges of civil society.
Methods: Interviews and focus groups from September 2020 to December 2022 involving 40 organizations.
Main empirical findings: Formal C S O s are recognized but informal actors remain under-recognized. Key actors include established N G O s and spontaneous volunteers. Challenges include lack of feedback platforms and volunteer turnover.
Data analysis 2: Codes include Informal Volunteers Advantages and Challenges, Empowerment of local communities, and Policies and protocols. Themes include Changing landscape of volunteering and Top-down emergency plans. Hypothesis 2: Crisis management that integrates formal and informal actors is more adaptive. Hypothesis 3: Self-organized volunteers face systematic non-recognition.
Step 3: Crisis management as a polycentric system.
Research questions focus on attitudes of formal actors toward informal actors and the substance of their relationships.
Method: Re-reading transcripts and focus codes in 2023.
Main empirical findings: Formal actors have two orientations: command-and-control (dismissive) or polycentric-orientated (recognizing). Relationships in a polycentric system are built on trust, equality, and temporality.
Step 1: Community resilience in crisis management (June 2020–June 2021)
To address the research questions concerning how community is understood in crisis management and how community resilience is best defined and approached, an initial purposive sample of eight CSOs was selected across Austria, France, Germany, Greece, Italy, Poland, and Spain. The sample comprised two local branches of an international aid organization, four Search and Rescue organizations, one medical emergency organization, and one civil protection organization. From these eight CSOs, an initial set of 32 case studies was collected using two tailored data collection templates. Main empirical findings informed the four conceptualizations of crisis management community and established an elements-based approach to community resilience, pointing to the hypothesis that integrating diverse volunteers with professional teams leads to greater operational effectiveness and community resilience than treating the community as only formal actors in silos (Figure 1).
Step 2: The role of civil society in crisis management (September 2020–December 2022)
Research questions:
-
• What is the current role of civil society in crisis management (NGOs, volunteers, communities, and individuals)?
-
• Who are the key civil society actors and networks in crisis management?
-
• What are the main needs, challenges, and best practices within civil society in crisis management?
Theoretical sampling: Based on findings from Step 1, theoretical sampling was employed to discover categories and their properties (Glaser & Strauss, Reference Glaser and Strauss1967). The sample consisted of 53 entities, mostly European, to ensure consistency in institutional and policy contexts. This included various types of emergency organizations to capture the full spectrum of crisis management practices while enabling cross-national comparison within each type as detailed below:
-
• Formal CSOs (40 organizations): 28 first responder/emergency organizations and 12 international organizations from European countries, plus Australia and Israel. Within these organizations, individuals at different management levels, as well as fully trained volunteers formally part of the structure, were selected to capture diverse perspectives on the role of CSOs in crisis management, including views on informal civil society actors.
-
• Informal and Emergent CSOs (11 groups): Transient groups were accessed through personal contacts during active crises: in Poland, groups responding to the refugee crises on the Polish-Belarusian border and following the war in Ukraine; in Greece, groups operating after the 2015 refugee crisis. Fewer emergent groups were included due to limited access (they are temporary, dissolve quickly, and are fully occupied when active), their supplementary role in crisis management, and because perspectives on their role were also gathered from formal CSOs, a crucial source when investigating formal–informal relationships.
-
• Grassroots Community CSOs (two groups): Established, but not formally registered, place-based community groups from Poland and Greece.
A total of 45 semi-structured interviews were conducted using two distinct schedules: one for professional CSOs (end-users and NGOs) and another for community-based CSOs. Three focus groups were held, each engaging a specific actor type from multiple countries: professional responder CSOs (Austria, France, Germany, and Greece); medical CSOs (Greece, Italy, and Spain); and emergent CSOs (Greece and Poland). These three types were selected because they interact frequently in crisis settings yet operate under different assumptions about authority, expertize, and decision-making—professional responders under hierarchical, protocol-driven frameworks; medical CSOs with specialized, triage-oriented approaches; and emergent CSOs exemplifying ad hoc, value-driven coordination. Main empirical findings revealed that formal CSOs are recognized and incorporated into crisis management, whereas informal actors remain under-recognized by both the state and formal CSOs despite contributing de facto. Key challenges included the absence of feedback platforms, volunteer turnover, and coordination difficulties, while best practices included combining resources, joint training, and providing institutional support to self-organized groups.
Step 3: Crisis management as a polycentric system
Research questions:
-
• What are the attitudes among formal civil society actors in crisis management toward informal actors?
-
• What is the substance of relationships binding various actors in a polycentric crisis management system?
Methods: No new data were collected in Step 3. The objective for data collection in Step 2 was wider in scope, but it included questions that specifically asked about cooperation, roles, and relationships. This provided ample material for further focused analysis. In 2025, transcripts were re-read and codes were refined further through constant comparison.
Main empirical findings identified two distinct attitudes among formal actors: command-and-control-oriented actors often dismiss informal actors as an additional risk requiring hierarchical incorporation, while polycentric-oriented actors tend to recognize and legitimize informal actors without attempting to control them. The substance of relationships in polycentric systems was understood through four elements: trust (built through recognition rather than hierarchy), flexibility and openness (distinguishing specialized from transferable tasks), equality (emerging through co-production), and temporality (leveraged as a resource for adaptive, short-term cooperation).
Data analysis
Across all steps, data were analyzed using a modified grounded theory approach (Charmaz, Reference Charmaz2006; Creswell, Reference Creswell2013; Glaser & Strauss, Reference Glaser and Strauss1967). Bryman (Reference Bryman2012) notes that most applications of grounded theory come down to following “core processes, such as coding, memos, and the very idea of allowing theoretical ideas to emerge out of one’s data” (p. 575), resulting in “grounded concepts rather than grounded theory” (p. 573), which was the case here.
The process begins with formulating initial research questions, which drive initial sampling, based on theoretical and purposive rather than statistical criteria to “discover categories and their properties” (Glaser & Strauss, Reference Glaser and Strauss1967, p. 62). Data collection is followed by coding and memo-writing, with codes developing into concepts and categories/themes through constant comparison. Once saturated, categories provide the basis for formulating hypotheses (Bryman, Reference Bryman2012). The iterative process can be repeated until categories are refined enough for a substantive theory grounded in the data. In the current study, this process was accompanied by an ongoing iterative literature review, as suspending awareness of existing conceptualizations is neither feasible nor desirable (Bulmer, Reference Bulmer1979).
Within this framework, thematic analysis (Braun & Clarke, Reference Braun and Clarke2006) served as a detailed elaboration of theme development, complementing constant comparison. The analytical process can be compared with principles outlined by Gioia et al. (Reference Gioia, Corley and Hamilton2013) for rigorous qualitative data structuring, particularly the distinction between first-order concepts (informant-centric codes), second-order themes (researcher-centric categories), and aggregate dimensions (overarching theoretical constructs). This ensured transparent progression from raw data to abstract categories while remaining grounded in participants’ accounts.
The analytical approach was applied iteratively across the three steps (Figure 1):
Step 1 analysis
The literature review informed initial coding and category development. Data from 32 case studies were coded, generating first-order codes including Communication, Community, Coordination, Feedback and debriefing, Networks, Local knowledge, and Volunteers. These were organized into second-order themes (“crisis management community as a network,” “recognition of local knowledge and volunteers’ contribution”). This produced H1 concerning the integration of diverse actors for operational effectiveness.
Step 2 analysis
Data from 45 semi-structured interviews and three focus groups were coded. First-order codes included Informal volunteers, Advantages and challenges, Expertize from various fields, Recognition, Self-organized volunteers, Ways of support, Policies and protocols, Training, and Organization type. These were organized into second-order themes (e.g., “changing landscape of volunteering,” “lack of recognition by formal actors,” “command-and-control versus network approaches”) and the aggregate dimension contrasting command-and-control with polycentric arrangements. This produced H2 (on polycentric systems) and H3 (on institutional facilitation for self-organized volunteers).
Step 3 analysis
No new data were collected. Interviews and focus group data from Step 2 were re-read, and codes were refined further through constant comparison. First-order codes were organized under two main architectures: command-and-control approach (contestation, barriers, oversight, assuming incompetence, and competition) and polycentric systems (institutional support and mapping expertize). Relational codes were grouped into second-order themes—Trust, Flexibility and openness, Equality, and Temporal aspects of relationships—which together formed the aggregate dimension of polycentric system building blocks. This produced the substantive theory presented in the Findings.
Ethical considerations
This research adhered to strict ethical standards. For all data collection involving human subjects (interviews and focus groups), verbal or written informed consent was obtained. To protect participants, all individual names have been anonymized.
Findings
Command-and-control approach: Contestation as a relational mechanism
In discussions on cooperation, many participants from formal CSOs only referred to other formal actors. When asked specifically about informal and emergent groups, half of our research participants from established CSOs dismissed their efforts, reporting either that they never witnessed such organizations or that they were unimportant or problematic. As one first responder stated:
I consider them as additional risk since they are not trained. Sooner or later, they will be in need of active assistance, and this will make the professionals and the trained staff including the volunteers from the unit, it will make their job heavier. (First Responder Organization, Bulgaria, Interview 18).
This quote reflects the perception that if volunteers do not undergo specialized training and formally belong to an emergency organization, they lack useful expertize during a crisis. In this view, self-organized spontaneous volunteers must become part of a command-and-control system to be effective. This is echoed by this participant, a member of an international organization based in Italy:
It was purely self-management, then we stepped in, we are a militarized organization, so we are very structured, and then we put the stakes back on the activities to be done in that situation. (International Organization, Italy, Interview 28).
A more accepting approach to self-managed groups was demonstrated by an international aid organization in Austria. As one participant explained:
We have seen the possibility to involve those persons in the chain of command … a system on how to manage what we call spontaneous volunteers. (International Aid Organization, Austria, Interview 6).
The program built capacity and allowed for the incorporation of spontaneous volunteers into the command-and-control chain more long-term by turning their efforts into a regular and recognized resource in crisis management. The relationship, however, was still hierarchical, with the CSO holding a certain level of authority over volunteers who, in this scenario, could not self-organize, but were instead organized by experts.
Polycentric approach: Cooperation through recognition and legitimization
Institutional support—Recognition and legitimization
A different approach by CSOs and state actors recognized and legitimized emergent groups for what they were, informal, creative, and ephemeral, without attempting to incorporate them into hierarchical command-and-control. Data showed how such institutional support facilitated the scope and effectiveness of emergent groups. This was discussed by a volunteer involved in organizing a large citizen support point for refugees at the Polish-Belarusian border in 2021. What started as a small community initiative turned into a complex logistical operation. The local Volunteer Fire Service (VFS) made its premises available at the mayor’s request. Volunteers improvised a kitchen and basic sleeping area for people who crossed the border and got lost in the forests. A call for donations exceeded expectations. When coordination became a challenge, volunteers created shared online rosters where they signed up for shifts, assigned a coordinator, and broke down tasks into collecting, sorting, and distributing donations, while communicating online which goods were still needed. The VFS assisted with their expertize. The emergent organization was one of many ad hoc, self-organized centers with limited yet autonomous authority, operating without central command but still part of a larger system that provided support from the local CSO (VFS), the local community, and the local government. This was appreciated by the organization members as expressed by this respondent:
Two days ago an organization was created in a town nearby … they told me that it’s really hard because their local government is against it … So when I talked to them, they were a bit scared, they were saying, “You know, your situation is different because your mayor and local government support you so you can operate” … But you know, we are not asking the mayor what we are supposed to do … The Fire Service or us, we do it on our own and he just accepts it. (Emergent Organization, Poland, Interview 44).
The support of the local government and the VFS legitimized the self-managed group’s activity as an actor in a polycentric system and facilitated its effective and rapid crisis response while maintaining its creativity and flexibility. Conversely, the group in the town nearby was not recognized as part of the polycentric system and could not function in such an informal way. For them, the choice was to either cease to exist or to formalize. The lack of support from the other actors in the system forced the group to create a formal organization to attempt to gain legitimization by creating a bureaucratic structure which changed the character of the group and slowed their response.
Locality, pre-existing ties to other actors, and trust (or lack of it) were the features of the above scenarios. However, institutional support of a formal CSO could be offered where no pre-existing ties exist if such an organization is open and recognizes the potential and capacity of a self-organized group. This was discussed by this firefighter who was a first responder during an earthquake with tsunami:
It was in Indonesia … It was local students who self-mobilized to come and help the victims of a disaster. They did not have any structures, they got together as university students and set up their organization to provide mainly food aid. (…) They had a lot of contacts, so they managed quite well, they were well perceived by the population; they did not lack much, only someone to guide them a little in their organization, but the rest was very good. (First Responder Organization, France, Interview 11).
When discussing relationships, emergent groups referred to both other CSOs and public authorities, referring implicitly to a polycentric system. Most of the members of self-managed groups who participated in the research emphasized the importance of having positive relationships with formal CSOs as well as with the state agencies. This medical doctor from the Medics on the Border group operating during the refugee crisis on the Polish-Belarusian border discussed how the absence of state recognition could hinder the activities of a self-managed group and replace the anticipated relationship of cooperation with that of competition:
NGOs have been great…and very collegial (…) But they [state first responders] sometimes look at us as something weird … I mean, I counted on being able to take care of things in a brotherly … collegial way but not everything could be done in a collegial way unfortunately … (Emergent Organization, Poland, Interview 43).
A strained relationship with the formal state actors, despite support and recognition received from formal CSOs in the system, added another constraint to the already difficult situation of providing medical assistance with limited resources during the crisis.
Mapping expertize: Locating knowledge outside formal centers
Along with institutional support and recognition, a key feature of polycentric systems is the ability to locate expertize outside formal crisis management networks. In Warsaw, an established refugee CSO (abbreviated below as ERC) turned to an art gallery for support in the first days of the Russian invasion of Ukraine in February 2022. With the advice and support of this professional organization, the gallery was able to reorganize within days to respond to a need to receive refugees from the war who were not Ukrainian and who fell outside the formal system of support for people fleeing Ukraine.
The gallery space was turned into temporary accommodation. Once a week, a bus arrived from Ukraine carrying around 40 people who would stay for 3–4 days. A member of the ERC was always present to support the gallery staff and volunteers. Volunteers were recruited from personal networks of the two organizations. The gallery team and first volunteers were trained by ERC, after which trained volunteers trained new recruits. A sense of belonging founded on shared values facilitated trust.
Despite its temporality and flexibility, the work quickly became structured. As one organizer explained:
I realized I’m getting overwhelmed … So we sat down with another team member and we came up with a list of tasks that we grouped under teams, and then we created spreadsheets where people could start entering their names … So this structuring was our way of dealing with stress. (Emergent Organization, Poland, Interview 45).
Ideas were run through the ERC for feedback. The work was organized into four 6-hour reception shifts, three 8-hour remote shifts, and shorter shifts for specific tasks. Shifts were administered in a shared online calendar where volunteers subscribed themselves. More details were available on Slack, organized into sections. The process was democratic and co-produced, and volunteers decided on their availability and entered details in an online tool.
The gallery’s coordination expertize proved transferable to crisis management. Recognizing that expertize can be located in unexpected places required openness, flexibility, and trust from the ERC. The other features of polycentric crisis management, institutional support, recognition, and legitimization, were also present.
The operation lasted about 3 months. By May 2022, the crisis subsided, but even if it had not, the gallery would have needed to focus on its annual exhibition. The arrangement was temporary by design, supportive of a larger refugee aid system. The emergent organization was one of many local centers where crisis response was local, co-produced, and independent of central command-and-control authority.
Formal CSOs and local communities: From co-optation toward co-production
Another informal actor in crisis management is local communities. Almost all respondents from established CSOs actively reached out to communities, often through gatekeepers who connected them to various groups. However, despite participants’ awareness of resources residing in local communities, their expertize in local conditions and needs, data demonstrated no methodologies for tapping into local knowledge. Engagement was wide but mainly unilateral and top-down, taking the form of training, courses, campaigns, and events as discussed by this participant:
We make sure to raise awareness and inform local communities about the dangers … and we create rescue plans. We mainly inform the local communities through meetings with community/municipal councils, and speeches. (First Responder Organization, Greece, Interview 3).
The above quote resonates with other responses from members of established CSOs and demonstrates the perception of local communities as recipients of expert knowledge rather than knowledge producers. This mirrors the command-and-control approach discussed in the previous section, further elaborated by this participant:
Often the communities need a professional approach, even if we are an NGO and our actions are based on humanitarianism and volunteering, we absolutely want to have a professional approach. (First Responder Organization, France, Interview 11).
Two organizations discussed a deeper and more needs-based engagement with communities. A German international aid CSO developed a tool to support newly formed groups:
We support such groups by helping them to establish a formal NGO so that they can fit in their local civil society organisation system, and connect them to the relevant authorities. (International Aid Organization, Germany, Interview 13).
This effort formalizes grassroots initiatives and incorporates them into command-and-control. While beneficial in some ways, it risks changing local groups’ character and hindering their creativity by imposing bureaucratic constraints (Scott, Reference Scott1998). However, it does create space for communities to contribute and build more long-lasting, potentially balanced relationships over time. Another relationship-building approach was discussed by this participant from a British CSO:
It’s something that we try and encourage the team to learn to do … we are also trying to build a human terrain picture. Who’s who? Who can we get to talk to? Who are natural leaders? (First Respondent Organization, UK, Interview 36).
This methodology maps local communities and identifies diverse groups. Like others, this respondent emphasized identifying gatekeepers to access and build relationships with communities. However, what sets this participant apart is that he indirectly recognized communities as knowledge producers, not just recipients, by acknowledging their leadership qualities. This approach could be developed further to create a genuine knowledge-exchange process where local communities become recognized actors in a polycentric system.
Discussion
This discussion addresses the two research questions in turn. First, the contrasting attitudes of formal actors toward informal actors (RQ1) are examined, distinguishing between command-and-control and polycentric approaches. Second, to explore the substance of relationships binding actors in a polycentric system (RQ2), the concept of social capital is applied. As Schobert et al. (Reference Schobert, Orru, Gabel, Nero, Windsheimer, Klaos and Nævestad2023) demonstrate in the crisis management context, social capital is understood not as a static asset but as a situational resource whose availability, accessibility, and activatability determine whether relationships function effectively during crises. The four elements that emerged from the analysis, namely trust, flexibility and openness, equality, and temporality, operationalize this concept, revealing what binds polycentric systems together in practice.
System architecture: Command-and-control versus polycentricity
This research identified two main approaches among CSOs and other actors in crisis response: command-and-control and what I refer to as polycentric. As presented in the Findings first section, the command-and-control approach focuses on the “hardware” of crisis management, namely rules, procedures, formal hierarchical structures, and technological solutions (Boin & ’t Hart, Reference Boin and ’t Hart2010). Only formal actors are recognized while emergent organizations are overlooked, perceived as unimportant, or a nuisance. This aligns with Scott’s (Reference Scott1998) critique of high-modernist policies, which privilege expert knowledge while systematically disregarding local and informal knowledge. The dominance of command-and-control has been critiqued as unsuitable for dynamic crises (Capano & Toth, Reference Capano and Toth2025; Drabek & McIntyre, Reference Drabek and McIntyre2003; Whittaker et al., Reference Whittaker, McLennan and Handmer2015). While expected among state first responders (Neal & Phillips, Reference Neal and Phillips1995; Quarantelli, Reference Quarantelli1987), this approach can be surprising among CSOs whose declared focus is humanitarianism (Chaulia, Reference Chaulia2011). Half of our respondents from established CSOs were either unaware of self-organized groups or criticized them for not operating according to command-and-control principles, often basing opinions on anecdotes rather than experience.
Although expressed by a significant number of respondents, this view was not shared by all of our participants from formal CSOs. The other half of the respondents demonstrated a degree of openness towards informal organizations and arrangements, that is, a polycentric approach. Such an attitude was acknowledged and appreciated by the participants from emergent organizations who participated in the research. Combining views of formal and informal actors allowed for identifying the main features of a polycentric governance system in emergency contexts, with a focus on relationships binding various actors.
Polycentric systems combine diffusion and order, meaning decision-making is distributed across multiple centers rather than concentrated in a single authority (Koinova et al., Reference Koinova, Deloffre, Gadinger, Mencutek, Scholte and Steffek2021). This does not imply the absence of hierarchy; rather, polycentric governance implies coordinated cooperation among many public and private actors (Capano & Toth, Reference Capano and Toth2025). While Jordan et al. (Reference Jordan, Huitema, Hildén, Van Asselt, Rayner, Schoenefeld, Tosun, Forster and Boasson2015) caution that evidence for polycentric effectiveness remains limited, they acknowledge its potential for experimentation and tailoring to local contexts. Paciarotti et al. (Reference Paciarotti, Cesaroni and Bevilacqua2018) offer a concrete example of a local Caritas that successfully managed spontaneous volunteers by improvising based on need with the mayor’s legitimization, which contrasted sharply with official responders who insisted on traditional, thoroughly trained volunteering. The examples discussed in the Findings illustrate a dynamic where decision-making was diffused and independent from central command-and-control, with arrangements that were local, co-produced by inclusive networks of formal and informal actors, and uniquely tailored to each crisis (Ostrom, Reference Ostrom1990).
Relationships: Social capital as the binding glue of a polycentric system
Identifying the architecture of each arrangement is important. However, to understand how order is established in a diffused system, the content of relationships must be analyzed. Data illustrated that each local center, referring to the emergent groups and informal arrangements described in the Findings, such as the citizen support point at the Polish-Belarusian border, the Warsaw art gallery, or the student group in Indonesia, had limited yet autonomous authority. Success depended not on central authority but on self-organization (Koinova, Reference Koinova2022). Actors created temporary rules appropriate to local conditions (Ostrom, Reference Ostrom1990). Where successful, relationships were based on trust, flexibility, openness, and equality—the four elements that emerged from focused coding in Step 3. These four elements operationalize the concept of social capital in the context of polycentric crisis response. Following Schobert et al. (Reference Schobert, Orru, Gabel, Nero, Windsheimer, Klaos and Nævestad2023), social capital is not a static asset but a situational resource whose availability, accessibility, and especially activatability determine whether it enhances crisis response. Each of the four elements below illustrates a different dimension of how social capital becomes, or fails to become, activated within polycentric relationships between formal and informal actors.
Trust
Establishing trust, that is partners knowing what to expect and assuming goodwill (Holohan, Reference Holohan2005), can be challenging in ad hoc arrangements. For formal actors, trust requires confidence in both good intentions and expertize of emergent organizations. For informal actors, trust requires confidence that established organizations will provide support rather than control. As Schobert et al. (Reference Schobert, Orru, Gabel, Nero, Windsheimer, Klaos and Nævestad2023) note, social capital is not automatically an asset as it can be inaccessible, disrupted, or diminished by social norms. Where trust and recognition are absent, relationships of competition and contestation emerge, as discussed in the second section of the Findings.
Trust can develop from pre-existing ties or on the spot. In the Polish-Belarusian border case, trust was based on previous cooperation. The mayor and VFS trusted volunteers’ intentions and expertize, recognizing their right to create their own rules (Ostrom, Reference Ostrom1990). The emergent organization trusted that there would be no attempt to control them, retaining independence and enabling a speedy response. This illustrates what Schobert et al. (Reference Schobert, Orru, Gabel, Nero, Windsheimer, Klaos and Nævestad2023) term activatable social capital, the most potent determinant of whether social capital actually increases resilience. By contrast, as shown in the Findings, a nearby emergent group was contested by formal actors and forced to formalize, significantly slowing their response.
In the case of the Warsaw gallery, there was no previous direct cooperation among partners, although previous ties existed. Trust in this case was based on the assumed shared overarching values among like-minded people who belonged to a loose network of friends and friends-of-friends. Cooperation was facilitated by the formal organizations’ ability to bridge and to locate expertize in an unexpected place. They provided institutional support by sharing expertize, feedback, and advice, without trying to control the emergent organization (Simsa, Reference Simsa2017). In the Indonesian example, there were no previous ties between the formal CSO and the local organizations. Trust developed on the spot and was at a sufficient level to facilitate temporary cooperation, based on shared values and goals of humanitarianism and providing relief to victims of a disaster. The formal CSO was able to recognize the importance of spontaneous volunteers’ ties to local communities and their ability to handle tasks of distributing food and clothing. Storr et al. (Reference Storr, Grube and Haeffele-Balch2017) similarly found that polycentric systems using dispersed knowledge can be better suited for local recovery, identifying two key advantages: “their ability to make use of local knowledge, and their flexibility” (p. 881).
The regular interactions between local communities and formal CSOs who provide training and institutional support are conducive to building trust. This trust could serve as a foundation for transforming the present limited and one-way communication into a richer and multilateral exchange where communities are recognized as knowledge producers and not just knowledge recipients and can contribute to the crisis management system in a meaningful way.
Flexibility and openness
Distinguishing among tasks of various levels of specialization within crisis management was a common feature among actors applying a polycentric approach. Acceptance that some of the skills required for managing a crisis are transferrable from different fields requires certain openness and flexibility on the part of formal actors in the system. There are some tasks that are highly specialized and require rigorous training such as SAR under dangerous conditions, SAR with canines or medical assistance. For safety reasons, such tasks are best handled by formal organizations since they require a high degree of institutionalization: established rules, protocols, and procedures. These kinds of tasks are often the responsibility of formal international organizations in emergency contexts (Chaulia, Reference Chaulia2011). Other tasks, such as general coordination, provision of food, clothing or shelter, involve transferable expertize and can be handled by informal and emergent organizations, as this research and other research demonstrated (Łapniewska, Reference Łapniewska2022; Paciarotti et al., Reference Paciarotti, Cesaroni and Bevilacqua2018). The ability to distinguish between these two categories of tasks, and that only one of them must be formalized, facilitates the development of trust, recognition, and cooperation between established and emergent organizations. The same applies to cooperation with local communities. This flexibility reflects Scott’s (Reference Scott1998) concern with who gets to define the problem and whose knowledge counts, a question that cuts to the heart of how formal actors recognize or dismiss informal actors.
Equality
One of the main features of polycentric relationships between emergent and formal organizations discussed in this research is that they are relatively power-balanced, in contrast to the command-and-control dynamics described in the Findings. The emergent organizations are treated as equals who might need some institutional support due to their temporary nature and limited resources, but who nevertheless possess expertize and are knowledge producers and not just knowledge recipients. This marks a shift from the command-and-control system toward a polycentric one and is in line with Ostrom’s (Reference Ostrom1990) principle of self-organized groups’ ability to create their own rules.
In most formal CSO-local community relationships, however, an asymmetrical expert–recipient dichotomy remains (Li, Reference Li2005), with expert knowledge prevailing and local knowledge dismissed (Scott, Reference Scott1998). In crisis management, formal actors typically position themselves as experts while community voices are sidelined. Evolving toward equality would require trust, openness, flexibility, and overcoming the command-and-control constraints that perpetuate this imbalance.
Temporal aspects of relationships
Relations between formal and emergent organizations are usually short-lived, presenting both challenges and opportunities. Such relations can be labor-intensive and perceived as having no future. On the other hand, temporary relations do not require the maintenance work of long-term relationships. Short-term cooperation still allows combining resources and building overall effectiveness, as this research demonstrated. Temporality can also limit competition and contestation (Koinova, Reference Koinova2022). As Storr et al. (Reference Storr, Grube and Haeffele-Balch2017) note, “the ability to organize and solve collective action problems is supported by strong social networks, shared values and past experiences with disaster” (p. 895), but these need not always be long-standing. As the Findings suggest, relational burdens can be avoided if formal partners focus on mapping resources among informal actors, establishing trust, and providing temporary institutional support. By contrast, recognizing local communities in a polycentric system requires long-term investment in building ties and reshaping existing relationships into more equal ones.
Conclusion
This study set out to examine attitudes among formal CSOs toward informal actors and the substance of relationships binding actors in polycentric crisis management systems. The findings reveal a clear divide: half of formal CSO respondents dismissed informal actors as untrained risks requiring hierarchical control, while the other half demonstrated openness, recognition, and institutional support without attempting to absorb emergent groups. Turning to the second research question on the substance of polycentric relationships, the analysis identified social capital as the binding glue of a polycentric system. Where polycentric attitudes prevailed, relationships were characterized by trust built through recognition, flexibility in distinguishing specialized from transferable tasks, equality through co-production, and temporality leveraged as a resource for adaptive cooperation.
The contributions of this article are empirical and theoretical. The study applies a polycentric lens to crisis management, an approach that moves beyond cataloguing actors to analyzing the relationships and power dynamics that bind them and provides qualitative evidence of how polycentric governance operates in practice. Drawing on Schobert et al. (Reference Schobert, Orru, Gabel, Nero, Windsheimer, Klaos and Nævestad2023), it conceptualizes social capital as the binding glue of polycentric systems, specifying trust, flexibility, equality, and temporality as the relational building blocks that underpin functioning polycentric governance. It also introduces concepts from critical governance studies (Scott, Reference Scott1998) to understand how recognition, expertize, and power dynamics shape inclusion or marginalization in crisis response.
These findings contribute to operationalizing polycentricity for crisis management by specifying how trust, flexibility, equality, and temporality are enacted in relational practice. Nevertheless, the study has limitations. The findings are drawn primarily from European contexts, and while qualitative research does not aim at statistical generalizability, future research in non-European settings would enrich understanding of how polycentric dynamics vary cross-culturally. The study’s qualitative design was appropriate for capturing actors’ perceptions, sense-making, and negotiation of power dynamics, which are essential for understanding relational processes. However, future quantitative or mixed-methods studies could usefully complement these findings by focusing on performance metrics, measuring crisis outcomes, and examining other quantifiable aspects of the issues raised here.
Data availability statement
Data will be made available on request.
Acknowledgements
The author kindly acknowledges the contributions of the consortium partners who participated in the data collection for this study, namely NTUA, PSCE, SAN, SYNYO, UCSC, UGL, and VUB. Special thanks go to Dr. Waleed Serhan who coordinated the data collection with me and co-authored the initial Project Deliverable reporting on the research results.
Funding statement
This project falls under the European Commission Horizon 2020 framework with Grant agreement ID: 882897.
Competing interests
The author declares that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.
Open access
