Interrogating the conceptualisation and use of ‘agency’ in environmental relocation

Environmental relocation, the organised movement of individuals or communities from at-risk areas to perceived safer locations, raises complex questions about people’s agency in shaping their futures amidst a rapidly changing climate risk landscape. In climate mobility research, policy, and advocacy, the notion of ‘agency’ is invoked alongside other value-laden terms such as empowerment and resilience. Yet, the term is often used inconsistently or without clear definition, creating an ambiguity that limits the capacity of both research and policy to operationalise agency substantively rather than simply rhetorically. This paper addresses this gap by undertaking a scoping review of the literature on environmental relocation, a field that repeatedly grapples with the tension between voluntary and forced movement. A total of 163 articles were reviewed. This paper examines the dominant framings of agency, alongside an exploration of its scalar and temporal dimension, and its orientation, the sphere or direction in which agency is being exercised. Identified literature gaps include the need to explore temporal dimensions of agency within environmental relocation, adopt intersectional approaches, to explore intra-household dynamics, and further theorise the role of coercion and possible operationalisation of enabling or enhancing agency to name a few.

Rising sea levels, floods, and other environmental hazards are increasingly leading individuals, households and communities to relocate, yet these movements are shaped as much by social and policy constraints, as well as individual choices and capabilities, as they are by environmental risk. In the context of environmental relocation, agency—a key concept for grasping individuals’ desires and actions regarding mobility—remains underexplored and undertheorised. The Intergovernmental Panel On Climate Change (IPCC) AR6 Technical Summary on Impacts, Adaptation, and Vulnerability defines agency in wider climate migration as the “degree of voluntarity and freedom of movement” (IPCC, 2023, p. 52). According to the IPCC,1 migration outcomes widely vary with agency: high-agency movements can benefit both migrants and communities, while low-agency or involuntary migration often produces vulnerability and socioeconomic hardship. However, in the case of governed relocation, these dynamics become more complex: there is an intricate web of constraints and choices shaped by policy, planning, and cultural, historical and socio-economic contexts.

The term ‘agency’ floats in and out of policy, guidance documents, advocacy and research. For example, in the guidance ‘A Toolbox: Planning Relocations to Protect People from Disasters and Environmental Change’., a clear call is made: “the agency, resilience, and empowerment of Relocated Persons should be recognised, promoted, and enhanced throughout a Planned Relocation” (IOM et al., 2018, p. 7). More recently, an examples include the A/HRC/56/47 report by Paula Gaviraria Betancur, the Special Rapporteur on the Human Rights of Internally Displaced Persons, which underlines: “The agency, resilience and empowerment of relocated persons should be acknowledged, promoted and strengthened, and the specific needs, circumstances and vulnerabilities of relocated and affected persons and communities should be considered and addressed throughout planned relocation” (Gaviraria Betancur, 2024, p. 24). The term “agency” also appears occasionally in national adaptation policy. For example, Vanuatu (2018) ‘National Policy on Climate Change and Disaster Induced Displacement’ underlines the need for: “fostering self-reliance—ensuring that displaced people are respected, empowered and viewed as economically productive members of society and agents of their own recovery, including promoting interventions which aim to strengthen people’s skills, assets, networks and agency” (p. 17). However, this is rarely accompanied by a definition. Overall, the term ‘agency’ appears in policy arenas, but its inclusion is often one in a list of multiple value-charged words: there is a need for further research on how this term is utilised and operationalised.

This ambiguity has been identified as a critical area for further study. Simpson et al. in their article ‘Research Priorities for Climate Mobility’ note that the “concept of agency could provide an important framing for future research on involuntary immobility and vulnerability. When households have high agency, mobility responses are more likely to have positive outcomes. However, interactions between the factors that promote agency of mobility (...) and immobility are less well understood” (2024, p. 595). Additionally, researchers such as Hoffmann et al. have stated that “a stronger focus on agency would help identify cultural and psychosocial factors underlying migration decisions” (2020, p. 910). It appears that while we normatively call upon terms such as agency, consent or voluntariness regularly within this complex and interdisciplinary field, we are left with a significant lack of clarity about what these terms actually require under conditions of environmental risk and how they may relate to local contexts.

With this context in mind, this paper asks: how is “agency” framed, conceptualised, and debated in literature on environmental relocation? By systematically engaging with the literature, this research aims to develop a more coherent understanding of how agency is conceptualised within a field that regularly engages with questions of voluntary vs. forced movement, as well as what it means to conduct relocations in a just manner. Through a scoping review, we analysed a total of 163 articles. We first outline a theoretical background and our methodology. The subsequent analysis is organised across understanding how the literature defines and categorises agency, and then how we may critically engage with it across scalar (agency for whom?), temporal (agency when?) and operational considerations (agency how?).

It is important to note that here we are choosing to engage specifically with environmental relocation rather than the broader field of climate mobility. It is pertinent to narrow the study to environmental relocations as these programmes propose an even more constrained window of action, as opposed to all possible mobility options. Environmental relocation here is an umbrella term, as defined by Yarina and Wescoat: “a subset of environmental mobilities that involve the movement of individuals or groups of people from one site to another because of environmental risk” (2023, p. 1). Under this umbrella, we are considering a range of governed processes from home buyouts, to planned relocation, to disaster resettlement, to name a few (Kaenzig and Piguet, 2021). The key determinant in their inclusion is whether the movement of individuals, households or communities was an organised, governed process rather than movement initiated independently. This colours the analysis of agency, grounded in the consideration that it relates to a governed context—as diverse as this may be—engenders an analysis informed by both public policy and wider social theory.

A background on ‘agency’

While the focus of the review is to understand how the literature to date has grappled with agency in the context of environmental relocations, it is necessary first to outline the wider legacy of the term ‘agency’ and consider a theoretical starting point. ‘Agency’ is not neutral yet can be often treated as such, rendering it heavy with assumptions and normative interpretations. It is a term used in everyday language making it deceptively difficult to define (Barnes, 2000). Furthermore, it has a long-standing presence in a variety of disciplines—from economics with rational actor paradigms to sociology with wider debates on the relationship between agency and structure, to philosophy with debates of free will—which complicates how we choose to discuss and employ it within interdisciplinary topics such as environmental relocation. As previously mentioned, the indiscriminate use of ‘agency’ in policy similarly obscures how agency should be applied or operationalised. This tension between its indiscriminate use in policy guidances and the epistemological diversity of the concept of agency in academia creates a balancing act between recognising the disciplinary diversity and possible issues attempting to aggregate this diversity, and the need to move forward a conceptualisation that is relevant to the policy and social problem of environmental relocation. Migration scholar Bakewell underscores “the meaning of the terms agency and structure lie at the heart of the problem (...) They are fundamental metaphors of the social sciences and widely used across the literature often without any attempt at definition” (Bakewell, 2010, p. 1694). In tandem, as Hein de Haas warns that the inclusion of agency and structure “risks becoming a mantra, which is repeated time and again, while it often remains vague how this should be done in practice” (De Haas, 2014, p. 22).

As a point of departure, this paper takes an approach rooted in social theory which recognises the social embeddedness of agency from the outset. Social theory conceptually underlines the role of structures, power and social inequality in shaping agency which will subsequently be explored. This is paired with an anchorage in public policy perspectives. A general definition stemming from social theory is provided by sociologist Barnes: “for an individual to possess agency is for her to possess internal powers and capacities, which, through their exercise, make her an active entity constantly intervening in the course of events ongoing around her” (2000, p. 25). This reasoning asserts the link between agency and the ability or capacity to act.

It is then necessary to underline the argument that agency is a relational property, emerging in relation to others, structures, and wider environments (Bakewell et al., 2012). As Borrelli et al. state: “agency is an enactment, rather than a property to be had, an effect of the discursive and material relations in particular environments” (Borrelli et al., 2022, p. 1157). This places an importance on its dynamic nature, pushing us to consider how agency can be shaped, or negotiated. It is therefore equally relevant to consider that agency is intimately tied to power (De Haas, 2014; Giddens, 1984) and that agency can be unequal due to varying access to structural resources (Whittington, 1988). Our focus now narrows to the policy dimension, where we also consider that agency operates within “arenas” or fields of action delimited by politics, policies or programmes (Ostrom and Ostrom, 2004, p. 115). Lastly, agency is often considered at the individual level but collective agency is a key and pertinent angle to consider, particularly for community-wide programmes such as planned relocation (Capano and Galanti, 2018; Cinner et al., 2018; Arnall et al., 2019). To the latter, there is also a need to recognise that the relative emphasis or importance of individualised or collective interpretations of agency will be sensitive to cultural contexts and differing epistemologies. This is incredibly important, with anthropologist Ahearn proposing that agency ought to be defined as the “socioculturally mediated capacity to act” (2001, p. 112).

A further angle to consider is the relationship between agency and choice, terms often intertwined (Barnes, 2000). Choice can refer to the range of options available (De Haas, 2014; Dowding and John, 2009; Ostrom and Ostrom, 2004) but we can also understand it as a potential action to be taken, and therefore an exercising of agency. Once more, this can be a complicated topic—Dowding and John reflect on this within public policy when stating “the very meaning of ‘choice’ as well as how increasing it might be measured are deeply problematic” (2009, p. 221). They stress that choice can be valued both instrumentally—for the breadth of options—and intrinsically—for simply providing choice (Dowding and John, 2009). We must further consider whether these choice are voluntary, which philosopher Straehle proposes is “the result of deliberation and the active, intentional decision to choose one option over another. Others suggest that voluntary decisions are those that are made between acceptable alternatives” (Straehle, 2023, p. 456). The latter is particularly difficult as acceptability is subjective. Straehle further asserts that the quality of this voluntary decision is for it to be “...reasonably free from coercion when choosing among viable options; and of being reasonably able to implement these choices” (2023, p. 455). Coercion erodes voluntarity, and by extension agency. As political philosopher Nine argues, “coercion undermines autonomous capacities by subverting or warping processes so that the agent responds to beliefs, values, or desires held by others” (Nine, 2016, p. 42). When coerced, individuals’ agency is warped; choices are constrained, pressured or altogether lost. In environmental relocation, this can also occur through economic pressure, purposeful dis- or misinformation or withholding of services or social pressure (Grannis, 2011). Overall, we recognise that agency and choice are closely linked: agency being required to engage with choices, choice providing a direction to agency; yet sensitive to coercion. This is significant in environmental relocation, where the recurrent call for voluntariness demands careful consideration of what constitutes a voluntary choice.

Furthermore, to examine choice and agency, many scholars argue for careful considerations of the factors, or capabilities, that permit them: as Ibrahim and Alkire state, “agency is influenced by people’s individual (material, human, social and psychological) and collective (voice, organisation, representation and identity) assets and capabilities” (2007, p. 385). This relates to the seminal work of Amartya Sen on the Capability Approach, where Sen proposes that wellbeing and justice should be understood and evaluated in terms of individuals’ real freedoms and opportunities to achieve their goals or valued ways of life. This departure from focusing on resources alone, to what various goods allow people to do—their capabilities and achieved ‘functionings’—further values the freedom people have to act as they see right (Sen, 1985). This approach has further been conceptualised in regards to international migration with the aspirations–capabilities framework (Carling and Schewel, 2018; De Haas, 2014) Furthermore, Amartya Sen’s approach to agency is key as it too bridges social theory with public policy by interrogating whether people have the real opportunities to pursue the goals they value, the latter often being expanded or restricted by policy decisions.

Lastly, we also should consider that there is a temporal element to agency (Straehle, 2023; Emirbayer and Mische, 1998). It is not a static attribute. Sociologists Emirbayer and Mische highlight this through their “Chordal Triad of Agency” framework, in which they consider the interplay between the following dimensions (1) past, iteration agency which includes personal values and capacity, (2) practical-evaluative agency of the present that is shaped by cultural, structural and material factors, and (3) the projective element of agency which concerns future aspirations. This temporal orientation allows for an analytical approach to agency which they summarise as “the temporally constructed engagement by actors of different structural environments—the temporal and relational contexts of action—which, through the interplay of habit, imagination, and judgement, both reproduces and transforms those structures in interactive response to the problems posed by changing historical situations” (Emirbayer and Mische, 1998, p. 970).

Applying these reflections specifically to environmental relocation, agency would be exercised not only in having a say over whether or how to relocate, but also in the ways individuals, households and communities negotiate, adapt to, and respond to the structures and processes shaping this movement. The primary way that these concepts have been engaged with in the field of climate mobility is through the theorising of voluntary versus forced movement, broadly referred to as the im-mobility spectrum (Ayeb-Karlsson et al., 2022; Blondin, 2021; Farbotko et al., 2020; Piggott-McKellar and McMichael, 2021; Schewel, 2020; Zickgraf, 2021). Yarina and Wescoat propose that agency should be understood as a spectrum: on one end “actors have the agency and freedom to choose, either to remain in a place and adapt to its risks or move to a preferable location of their choosing. A very high level of agency can lead to further risky location decisions as well as safe ones. At the other end of the spectrum structural constraints limit relocation to safer locations. Progressive structural forces, e.g., anti-discrimination housing policies and regulations, can enable as well as constrain relocation decisions” (2023, p. 3). We see that overall the focus within the field of climate mobility is on the voluntariness of relocation and we aim to expand beyond this (Sengupta, 2025).

To conclude, there is a rich and diverse theoretical background to the concept of agency. We propose a reading of the literature through a lens informed by both social theory and public policy, and the Capability Approach, which allows us to consider how these key angles—that agency is a relational property; that agency is shaped by a range of temporal considerations; that agency is shaped by power; that agency occurs at multiple levels from the individual, to households, to wider families and communities to the broader collective, that agency is shaped by what choices are available, as well as capabilities and wider socio-economic and cultural contexts, for example—appear, or not, within environmental relocation literature and how we may move forward from this point.

Conceptual complexity is rife in growing interdisciplinary fields. Literature reviews have frequently been used to elucidate these ambiguities. For example, Troncarelli et al. undertook a literature mapping to understand how the term “climate change perceptions” is used concerning small-scale rural populations (Troncarelli et al., 2023) and Pham and Saner undertook a systematic review to define the fickle term of “inclusivity” within climate change adaptation (Pham and Saner, 2021).

Here, a scoping review was deemed more appropriate than a systematic review because the aim of this study is to explore and clarify a concept rather than to aggregate findings from a well-defined body of evidence (Munn et al., 2022; Arksey and O’Malley, 2005). The scoping approach allows for a more exploratory identification of conceptual patterns and gaps across multiple disciplines and approaches within the climate mobility field. This scoping approach further allows for the inclusion of conceptual or discussion articles, rather than focusing on empirical research alone, contributing to the paper’s aim to consider the wider framings and context to the conceptualisation of agency in environmental relocation This paper also aims to identify limitations within this diverse literature and critically discuss the concept of agency in relation to the policy context of environmental relocations.

Search strategy and eligibility criteria

A Boolean search string was developed iteratively from the main concepts of the research question. Key terms and synonyms were identified through preliminary scoping searches and knowledge of commonly-used terminology within the field. Truncation was used to retrieve multiple word endings (e.g. decision, decisions, decision-making), thereby increasing comprehensiveness. Additional terms were included to limit the context to environmental drivers (“disaster” OR “climate change” OR “hazard”) resulting in the following search string:

(“relocation” OR “planned relocation” OR “managed retreat” OR “home buyout” OR “resettlement” OR “community relocation”)

AND

(“agency” OR “autonomy” OR “choice” OR “decision” OR “participation” OR “empowerment”)

AND

(“disaster” OR “climate change” OR “hazard”).

We chose to not only search for “ AND (“Agency”) ” because this proved overly restrictive. The literature frequently engages with dimensions of agency through related concepts such as autonomy, participation, empowerment, or decision-making. Furthermore, we see in policy references mentioned in the introduction that agency is rarely used on its own. Including synonyms and related terms was therefore necessary to capture the full conceptual range, as well as supporting the comprehensive aim of a scoping approach (Arksey and O’Malley, 2005). We used Scopus and Web of Science as the main interdisciplinary scientific databases with the timeframe of 2000–2025, so as to prioritise treating 21st century literature on the subject.

Papers were screened to meet the following inclusion criteria: (1) the study focuses on populations who have experienced or are at risk of relocation; (2) the form of relocation discussed is a governed process; (3) the relocation pertains to environmental hazards; (4) the study explicitly or implicitly discuss agency or related concepts; (5) was predominately concerned with environmental relocation as a process either through policy, stakeholders, conceptual engagement or lived experience of relocated communities.

Inversely, the following exclusion criteria were identified: (1) Wrong problem: If the article considered relocation in the terms of human resource management (employment relocation, firm relocation), tourism, classic migration studies, biological sciences (animal relocation), political science (government agency) or more; (2) Wrong focus: If the study considered wider, unmanaged climate displacement or cross-border migration; and if the study discussed environmental relocation in overly abstract terms such as through modelling exposure to relocation with no meaningful engagement with relocation as a process (3) Wrong exposure: If the study focused on a range of other hazards such as chemical, medical or technological hazards, or conflict settings without explicit link to environmental considerations; (4) Wrong format: this review did not accept conference papers, presentations or other formats that were not peer-reviewed; (5) Accessibility: excluded studies where the article was not fully accessible through institutional access or where a.pdf format was not available, and excluded articles not in English (Figure 1).

Study selection and analysis

Using the chosen databases and search strings, the first search compiled in December 2025 resulted in 1513 articles, leaving a total of 1,092 papers after removing duplicates. We then completed a screening process: after screening titles for relevance, we had 421 articles, and the subsequent abstract screening led to a total of 213 relevant articles. This filtering process was carried out between December 2025 and January 2026 using Excel. The last screening stage occurred while thoroughly reading the articles between January and April 2026, leading to 50 excluded papers. A final 163 articles moved towards analysis. These papers were systematically aggregated across an Excel spreadsheet and coded across categories describing the key characteristics of the papers such as type of hazard or the type of environmental relocation studied, the inclusion of agency whether explicit or implicit, whether a definition was provided, or how else it engaged with the diverse complementary concepts, notably: choice, decision-making and participation, terms which were determined from conceptual mapping within the literature (Anderson et al., 2008). The latter allowed for a directed content analysis (Hsieh and Shannon, 2005).

Limitations

Limitations of this paper include that we only analysed English-language sources, which introduces a language bias and may have led to the omission of relevant case studies from non-anglophone literature. This could possibly explain the lack of case studies found in Africa and Latin America which were heavily understudied regions compared to Asia and North America, for example. Additionally, this study’s linguistic limitation may influence the resulting conceptualisations of agency that appear to be heavily dominated by Western epistemologies, bar a few papers anchored in indigenous knowledge systems and approaches (see: Gini, 2025; Jarillo and Barnett, 2022). It is important to note that this is not only due to the choice of language but also about broader patterns concerning where research on environmental relocation is conducted. In the reviewed literature, 45 studies were about the United States alone which introduces a significant skew to this review. Sensitivity to this distribution is significant as agency is not a culturally neutral construct but is embedded within specific knowledge systems. For example, in certain cultures collective understandings of agency would have greater importance than individual frameworks; or furthermore the choice to focus on human agency alone reflects a specific knowledge system (Ahearn, 2001). Non-human agency, such as the agency of nature or water, is an important concept in a range of knowledge systems but this consideration is beyond the scope of this article (Watts, 2013). This reflection also requires a careful interrogation of what it means to attempt to engage with such a broad range of epistemologies and disciplinary legacies, noting the danger of flattening the diversity of conceptualisations of agency, or foregrounding certain epistemologies over others. To limit the extent of possible over-simplification, we draw on explicit definitions where possible, though they are limited, and contextualise the discussion within broader theory, aiming to recognise the plurality of agency while paying heed to the vague and possibly indiscriminate use of the term within policy arenas. Lastly, it is necessary to underline that agency is one term of many invoked in policy and in the literature: as mentioned in the introduction, the policy documents we refer to include agency within a list of other terms, such as ‘empowerment’ in Betancur’s 2024 report, or ‘self-reliance’ in Vanuatu’s 2018 national displacement policy. Exploring these terms would be as rich and pertinent, further reminding us that this analysis must be situated within a broader discourse.

Key characteristics

The analysed literature spanned from 2007 to 20252 across a diverse geography of empirical case studies that covered 160 localised case studies (see Figures 2, 3). Examining the geographic distribution of our selected papers helps contextualise the represented conceptualisations. Asia and North America together represented 30% of the empirical studies. The United States of America was the most represented country, accounting for 45 studies. South and Central America was the least represented, with only 2.5% of the studies, followed by Africa with 5%—making both of these regions heavily underrepresented within anglophone literature. Over the last 25 years, there has been a clear increase in considerations of agency, participation, choice, and empowerment. We observe a dip between 2023 to 2024 which can possibly be attributed to a post-Covid decline in empirical data collection and therefore in publishing.

Using UNDRR’s Hazard Information Profile 2025 categorisation, we found that the majority of the literature covered multiple forms of hazards (49.1%), approximately a quarter of which were mixed coastal hazards (Jacot Des Combes et al., 2025). This could reflect that most relocation programmes are responding to complex risks with multiple hazards. This was followed by meteorological and hydrological hazards (29.4%), of which 35.4% studied flooding, 22.9% hurricanes and 22.9% typhoons, followed by a small percentage looking at tsunamis, storms and cyclones, respectively. 11.7% of the total studies looked at environmental hazards, the large majority of these looking at sea level rise. Finally 9.8% of the studies treated geological hazards with 62.5% of those studying earthquakes, 25% studying land- and mud- slides, and finally 12.5% looking at volcanoes (see Figure 4). As to the form of environmental relocation that these papers covered, there were relatively equivalent numbers of papers concerning planned relocation, buyouts (or buybacks, depending on terminology employed by researchers) and post-disaster resettlement (all between 20 to 23%). Focusing on the terminology used, 15% referred to relocation as a general term—not always specifying the governance mechanism. Less than 10% used the term managed retreat (see Figure 5). Lastly, 39.3% of the studies looked at reactive relocation, vs. 19% looking at proactive relocation and the remainder either had both or considered hypothetical future scenarios (see Figure 6).

The inclusion of agency

There was an explicit inclusion of agency in 56 papers, either as a central theoretical framework or within broader background and discussion sections. We found that an additional 55 papers discussed agency through the use of accompanying terms or synonyms, such as through autonomy (Bailey-Winiata et al., 2024), self-organisation (Albert et al., 2018), self-determination (Aidoo, 2021), voice or will (Siders, 2022) or independence (Yang et al., 2023) to name a few. Not intending to conflate these terms, the underlying focus on an individual’s or group’s capacity for self-determination, choice, and autonomous action was clearly present. The remainder of the included papers explored elements derived from the wider search-string, deemed critical for a holistic understanding. We found that the three most-frequently discussed in the chosen literature were: place attachment, resilience and livelihoods. This was determined through a form of summative content analysis through the counting of recurring keywords and key analytical frameworks, visually presented in the word cloud below (Figure 7), mapping the scope of concepts tackled in this diverse, interdisciplinary literature (Hsieh and Shannon, 2005).

Definitions and framing of agency

The majority of the studies invoke agency without defining it, with only five papers providing explicit definitions (Arnall et al., 2019; Arroyo and Åstrand, 2019; Bohle and Marone, 2022; Gibb, 2024; McMichael et al., 2019). Some of these were done by citing other scholars: Arnall et al. referred to sociologist Anthony Giddens, known for his structuration theory, Arroyo and Åstrand referenced philosopher Roy Bhaskar’s approach of causal agency, and lastly, Gibb cited political geographer Merje Kuus. While other papers also make reference to Giddens (McMichael et al., 2019; Wiegel et al., 2021), we overall observe a relative heterogeneity in the definitions provided and theoretical legacies invoked, confirming the lack of definitional clarity that prompted this review.

A wide range of frames and types of agency were unearthed in this scoping review. We found that there were two primary conceptualisations: (a) agency as the capacity for action and decision-making; and (b), agency as a wider process that is both relational and subject to constraints.

Within the former, agency is understood as ability or capacity to act, a property to be had (Arnall et al., 2019; Bertana and Blanton, 2023; Bohle and Marone, 2022). The concept of autonomy comes through: Bohle and Marone for example propose that agency “indicates to accentuate the ‘autonomy’ (of the human agent), namely to act on own initiative to face (diverse) local circumstances” (Bohle and Marone, 2022, p. 8). This orientation prioritises the question of voluntary relocation, stressing individual autonomy and logics of decision-making. Furthermore, much of the literature examines these questions, including via choice experiments, regardless of whether they explicitly frame them in terms of agency. For example, Rashid et al. explicitly applied a stated preference choice model to analyse possible relocation decision-making in Dhaka, Bangladesh (2007) as do Gatiso et al. in their study of preferences regarding cultural heritage in relocation in Fiji (Gatiso et al., 2025). This approach utilises random utility theory to analyse the role of incentives, relying on the assumption that individuals make decisions in-line with utility maximisation. This approach, with a heritage in economics, oversimplifies choice and preference, often overlooking wider contexts and human realities. While this is not the predominant observed approach within the surveyed literature, it is essential to understand that rational utility paradigms are common, especially in economics and behavioural public policy. In line with this more discrete understanding of choice and agency, we see an overall focus on the question of the desire to relocate or not, as a highly individual consideration.

There is however a more comprehensive approach which understands agency not as solely centred on the individual, or as a property to be had, but instead as continuous and relational. Agency emerges through a negotiation of wider constraints. We can consider here both material constraints, the various physical, economic and technical limitations that circumscribe action; and larger questions of social and structural contexts and perceived moral obligations. Agency from this perspective is therefore inseparable from the power structures, institutional rules, historical legacies, and even the natural environment that condition what is possible, permissible, or imaginable. This relational view sees agency as distributed and emergent, where agency is negotiated rather than an expression of free will. This wider approach argues that agency persists in spite of restrictive contexts, such as forced relocation. Agency is present through improvisation and reinterpretation of rules (Wu and Lu, 2024), through creativity (Cajilig, 2017) and through imagination (Dickinson, 2021). As Gibb describes “agency as not just the capacity for progressive change, but also the capacity to endure, suffer and persist” (Gibb, 2024, p. 129). This perspective was particularly present within studies that employed critical lenses, such as decolonial or feminist ones—sensitive to the power dynamics and wider structural and historical contexts.

Lastly, the literature revealed recurring analytical frameworks such as human rights and environmental justice beyond the approaches mentioned above. Firstly, agency as a question of justice—whether under the banner of social or climate justice (Farbotko and McMichael, 2019; Ajibade and Siders, 2022; Ajibade and Siders, 2022; Nagra, 2024). A second framing, often used in tandem, was human rights (Arnall et al., 2019; Farbotko and McMichael, 2019; Bergmann, 2021; Bohle and Marone, 2022). Lastly, and perhaps the most common framing within the selected literature was understanding agency through empowerment, which was found across both of the approaches outlined (McMichael et al., 2019; Hanna et al., 2020; Hadlos, 2021; Ngulube et al., 2024).

Understanding the landscape of agency

The literature differentiates agency along questions of scale and orientation, some papers opting to focus their analysis on one angle of agency such as everyday agency (McMichael et al., 2019) and others not defining a scope. These distinctions are useful for undertaking a comprehensive analysis of how agency works in practice; individual level, everyday agency possibly poses different theoretical or methodological questions compared to collective political agency, for example. Overall, we adopt a wider, relational understanding of agency in our analysis that is sensitive to the wider interactions of agency and structure as we navigate these diverse categorisations.

Scalar considerations: individual agency and plural agency of collectives and communities

The literature proposes different ways to see agency in terms of scale, but it is important to consider that these scales interact and are often culturally-mediated. Individual agency focuses on the granular: taking action at the individual level, such as relocation acceptance or refusal, and livelihood strategies, for example (Dickinson, 2021; Miller, 2020; Nine, 2016). Collective agency occurs with a group of people acting together towards a shared goal (Bergmann, 2021; Farbotko and McMichael, 2019; Manda et al., 2023; Matti et al., 2023). This collective can be at many levels, from informal groups, to entire towns, to specific advocacy associations—the primary consideration being that there are multiple agents instead of one (Ludwig, 2016). A multi-scalar approach to agency is crucial, as accounts of agency are often reduced to overly individualised frameworks that neglect collective conditions of action. Furthermore, certain forms of environmental relocation, such as whole-community planned relocation, inherently require a consideration of plural agency. Collective and individual agency can be co-dependent and dynamically intertwined, as are political and everyday agency. Individual agency can be exercised and have collective consequences and vice versa: as political philosopher Ludwig states: we can understand “collective agency in terms of our capacities as individuals to engage in intentional behaviour, still we could not understand our own forms of life without understanding it as being grounded in collective agency” (2016, p. 3).

Political agency

Political agency is the ability to assert or influence decisions, political processes and wider power structures (Renkamp et al., 2025; Grace-McCaskey et al., 2021; Jain and Bazaz, 2020; Herrmann, 2019; Marlow and Sancken, 2017). In relocation, this could resemble the agency of individuals or collectives within organised processes such as consultation, or the ability to influence the relocation programme and wider agendas, or self-organisation, protest and advocacy. For example, the village of Kivalina in Alaska took matters into their own hands when, despite long-standing demands for relocation, they were met with government inaction and barriers to funding (Marlow and Sancken, 2017). Kivalina used avenues such as claims to legal remedies through courts, fostering media coverage and organising with non-state actors, to name a few, in order to draw attention to the need for relocation and claim local leadership over the process. Political agency can also be the development of tactics to leverage desired outcomes, whether through institutional or non-institutional channels. In Indonesia, faced with a government resettlement programme due to volcanic risk, the Pelemsari community self-organised to construct a list of demands such as land ownership certificates. While the government reached a conclusion, some community members purchased the land they wanted themselves and relocated independently (Depari and Lindell, 2023).

Political agency can also manifest as resistance whether at an individual or collective level (Ahearn, 2001). Organised resistance as political agency can represent protest, grassroots mobilisation, campaigning, and political lobbying (Arnall et al., 2019; Tsai et al., 2022); individual resistance can be observed as rejection of offers or the rejection of housing sites (Gibb, 2024). A further element of consideration is the role of representation as a tool for political agency, whether that be for individual or community interests. Bavinck et al. describe in a study of post-tsunmai relocation for fisher settlements in India that caste representatives—the panchayat—played a significant role in relocation outcomes. This form of local governance meant that when the panchayat had more influence or say, the villagers felt better represented and supported; when they were weaker or seen as less legitimate within the resettlement process, the village was often relegated or had sub-par conditions in housing allocation (Bavinck et al., 2015).3

Political agency is not limited to the relocatees: Arroyo and Åstrand observe the role of the “Filipino power elite” in using what the authors call “intentional causal agency” to impact bureaucratic and land-use policies post Typhoon Haiyan to their advantage (Arroyo and Åstrand, 2019, p. 161). Groups such as the host or receiving communities also ought to be considered (Geekiyanage et al., 2021; Li and Spidalieri, 2021). Lastly, we cannot separate the concept of political agency from a wider understanding of the political context: is political agency being enabled?

Everyday agency

However, agency is not necessarily strategic or occurring at large scales. Here, the concept of everyday agency is pertinent. McMichael, Katonivualiku, and Powell define this concept as “expressions of agency in daily life, events, practices, and relations; to the routine, ordinary, familiar, and often mundane ways in which people and communities act to gain control over their lives” (McMichael et al., 2019, p. 326). The authors argue that these day-to-day decisions are key for multi-level challenges such as climate change. Daily, individuals make active decisions about their future and negotiate power—these may be lower profile but not any less important. These can also be political in nature, such as “microprocesses of resistance” in everyday life—demonstrating the overlaps in these categories (Ahearn, 2001 p. 119).

The framing of everyday agency allows us to consider a range of different manifestations of agency at the micro-level. For example, we can consider the role of creativity and imagination: Cajilig proposes that we can see “creativity as an improvisational process (which) highlights the agentic potential of design in even the bleakest and most quotidian of settings” such as post-disaster recovery (Cajilig, 2017, p. 79). Cajilig argues that how households improvise and alter their new housing post-resettlement requires creativity, especially due to restrictive conditions of government-provided housing. This allows individuals to make-meaning and interpret their circumstances in such a way that they can shape their choices (Dickinson, 2021). Furthermore, certain papers invite us to consider the role of religion through this frame. There are conflicting view points on whether religion enables or constrains agency: religion can be understood as a way of meaning-making that contextualises how individuals understand their capacity to act, whether through moral frameworks that justify action, the accompanying community or the sense of identity and purpose it engenders (Bertana, 2020; Gibb, 2024). Gibb argues that religion “shapes everyday socio-spatial realities and embodied experiences of religious beliefs and practices” making it “a demonstration of post-disaster agency” (Gibb, 2024, p. 128; in reference to Aijazi and Panjwani, 2015). However, others disagree, instead finding that higher powers reduce agency by leaving action to the discretion of a higher power. For example, Bertana describes how the belief of Noah’s Ark in Christian communities in Fiji created a sense that they would be protected from sea level rise4 (Bertana, 2020; Wiegel et al., 2021). This debate is sensitive to how an individual or collective would conceptualise their own agency in respect to a higher-power, and wider socio-cultural contexts.

Agency for whom?

With this landscape of agency in mind, we can further delve into the question of who is assumed to hold agency. Individual and collective experiences are embedded in larger social structures and power dynamics; it is vital to critically explore how different social characteristics of individuals or households shape agency and determine possibly differentiated or unequal capabilities and aspirations (De Haas, 2014).

Gender

Firstly, climate mobility is “inherently gendered” yet gender was found to be sparsely addressed in the surveyed literature (Hoffmann et al., 2020, p. 593). Relocation can exacerbate minority gender groups’ vulnerability and gendered power dynamics, shaped by questions of cultural norms or economic dependence, at play in household decision-making. For instance, a study in Senegal found that the traditional polygyny, in which men can marry multiple wives, created a hierarchy wherein younger wives have the least say in relocation (Renkamp et al., 2025). Renkamp et al. state that “singles, younger individuals, non-household heads, and second or third wives were especially vulnerable due to limited agency and decisional power” (Renkamp et al., 2025, p. 15). This gender exclusion can also translate into policy processes. Bertana and Blanton found that women in rural villages in Fiji were not asked for input directly, despite nearly half of the women attending the community meetings. The authors note that government officials were aware of this imbalance and yet continued with the relocation (Bertana and Blanton, 2023). Disparity in decision-making for relocation is further reflected into research practices where women specifically were excluded. In a study of post-flooding relocation in Pakistan, researchers did not include any women in their survey because they focused on heads of households (Akbar et al., 2024). Similarly, a study in informal settlements in Bangladesh purposefully excluded female participants as women were not allowed to talk to male strangers and the researchers uncritically state “no attempt was made to hire an alternate female assistant because, in this case, the focus of the study was on economic incentives for relocation by the household heads, which happened to be male” (Rashid et al., 2007, p. 97).

Race

Race is also a key consideration, and was predominantly discussed in an American context within the surveyed literature (Aidoo, 2021; Grace-McCaskey et al., 2021; Nagra, 2024; Nelson and Molloy, 2021). Many communities currently facing relocation are situated on high-risk land as a direct consequence of historical marginalisation and racial discrimination, which effectively forced them into vulnerable locations. Princeville, in North Carolina, is the first incorporated “Black town5” in the USA and is located on flood-prone land due to it being historically “unwanted” (Grace-McCaskey et al., 2021, p. 347). This unique and significant history of “resistance to White supremacist forces who worked to diminish Princeville’s political agency and abolish the town’s charter” entrenches a deep, and political, sense of place and identity that is at risk should the town take up buyout offers (Grace-McCaskey et al., 2021, p. 343). Overall, race is a key lens through which to understand differentiated access to resources, exposure to risk, and systemic discrimination. Therefore, race does not just influence who is affected by environmental change but also how much control people may have over their possible responses to it.

Perspectives of indigenous communities and knowledge systems

Another important theme in the literature was the relocation of indigenous communities (Bailey-Winiata et al., 2024; Hsu et al., 2019; Jarillo and Barnett, 2022; Maldonado, 2014; Marlow and Sancken, 2017; Simms, 2021; Wu and Lu, 2024). For example, Bailey-Winiata et al. argue for the need for Māori -led relocation in New Zealand, drawing on historical analysis of past case studies of Māori communities relocating due to natural hazards as early as pre-1840 (Bailey-Winiata et al., 2024). Relocation programmes responding to land degradation in Tibet run into conflict with local nomadic lifestyles, changing these ways-of-living and rural economies, to name a current example (Chies, 2018). The importance of self-determination and ownership of the relocation process for indigenous communities facing displacement is fundamentally linked to historical injustices, specifically the enduring legacies of colonialism, profound historical marginalisation, and cases of forced displacement (Farbotko and McMichael, 2019; McMichael et al., 2025; Simms, 2021). These communities have often been subjected to decisions made by external powers without their consent, leading to forcible loss of land, culture, and social cohesion. Recognising and facilitating agency and self-determination is a way to avoid repeating these historical injustices.

Some of the literature invites us to consider non-human agency. For example, Gini explores how the Cardoso Island local indigenous knowledge-system “acknowledges the agency of more-than-human entities, whether these are water, memories, or fish, and incorporates them into the broader narrative of mobility and displacement” (Gini, 2025, p. 3). Furthermore, Jarillo and Barnett (2022) critique the “oceanocentrism” in Pacific sea level rise discourse, which overemphasises the ocean’s agency at the expense of recognising the atoll’s agency as an assemblage of nature and humans. These approaches are situated within a world-view that fundamentally differentially conceptualises agency and the human-nature relationship, which in turn impacts ethical frameworks and possible governance implications. Though beyond the scope of this paper’s focus on human agency, these are significant considerations and an avenue for future research.

Class and intersectionality

Furthermore, class was a cross-cutting issue, underpinning many of the dimensions mentioned previously but not often taken as its unique angle of analysis within the identified literature. Socio-economic factors shape individuals’ capabilities by influencing their access to the material, educational, social, and political resources necessary. In environmental relocation, the question of wealth and financial resources is crucial: securing alternative housing, being able to absorb the economic disruption of relocation or even putting money aside in the face of growing climate risk are all factors that affect households’ decision-making around relocation. Here, there is room for future exploration of the Capability Approach. For example, Kleine has explored the relationship between choice and agency in development through a discussion of the resources that enable agency: from education resources, to financial, cultural, social resources and more, including the dimensions of information, time and health (2019). Given the significant socio-economic implications of environmental relocation, as well as the significant financial loss it consistently presents, regardless of the level of compensation, a class-informed and capability-informed reading of agency in environmental relocation is a pertinent avenue to pursue.

As a whole, it is not a question of who holds agency but rather how different groups and individuals’ agency may be constrained or enabled due to wider structures: multiple levels interacting. Political agency, as discussed in section 4.1, is crucial; its exercise is shaped by broader structures and legacies like colonialism. This calls for an increased use of intersectionality as a lens to understand environmental relocation, as well as “paying attention to spaces and scales previously deemed private or apolitical” so that “the conception of who qualifies as a political agent is broadened, and new ways to rethink agency, structure, and action are created” (Gibb, 2024, p. 129).

Agency when?

Relocation research often treats agency as occurring at a single moment of decision. But by focusing primarily on environmental relocation as a question of acceptance or refusal, the literature risks overlooking how agency unfolds across multiple stages of relocation processes. Agency is shaped by temporal conditions, including the time available for decision-making, prolonged periods of waiting within bureaucratic processes, and the possibility of revising earlier decisions. These dynamics point to the complex temporalities of relocation, where agency evolves over extended processes of deliberation, delay, and adaptation. We have to consider how the past and the future inform people’s agency, as Emirbayer and Mische state: individuals are “always living simultaneously in the past, future, and present, and adjusting the various temporalities of their empirical existence to one another (and to their empirical circumstances) in more or less imaginative or reflective ways” (Emirbayer and Mische, 1998, p. 1012). These temporal considerations require us to understand the contextual histories within which these relocations are taking place, with legacies of historical injustice or land dispossession, for example, informing the present reality or interpretation of relocation (Simms, 2021). Lastly, while we follow the temporal proposal by Emirbayer and Mische, there are diverse ways of understanding time and temporality that need to be recognised and valued—for example, Gini discusses how the local population of Cardoso Island has a non-linear understanding of time which informs their engagement with relocation, understood as thick time “where past and futures are entangled in the present” (Gini, 2025, p. 1).

Additionally, time is a key dimension of relocation as hazards impose varying time-scales (Yarina and Wescoat, 2023). Relocation related to sudden onset events tend to have shorter timelines than slow onset ones. Emergency post-disaster resettlement can often forgo steps such as community consultation which are deemed too time-consuming. For example, following the 2009 earthquakes in L’Aquila in Italy, the top-down post-disaster resettlement “closed the door to grassroots involvement, justified in the urgency” of the situation (Contreras et al., 2017, p. 86). Disasters have also been found to be used as windows of opportunity to put forward relocation programmes (Chies, 2018; Thaler et al., 2020). While this is sometimes observed empirically, the speed at which a hazard develops does not, however, warrant the omission of crucial considerations such as consultation.

Conversely, some relocation processes, such as buyouts, can be found to be too drawn out. Waiting is common because of the scale of these processes, which include bureaucratic processing, funding approval, property assessment, and possibly consultative processes. This presents somewhat of a paradox: people may have little control over when relocation occurs but must keep their lives in suspension while waiting for a range of administrative processes to run their course (Hanna et al., 2020; Peterson et al., 2020). Participants in a New York home buyout programme following Hurricane Sandy characterised the protracted process as a “constant state of waiting,” indicating this was the most challenging aspect (Binder and Greer, 2016, p. 102). We should also consider that extended periods of waiting open the door to people changing their minds (Binder et al., 2020; Farbotko and McMichael, 2019; Iuchi, 2024). In extended time frames, personal circumstances, such as finances, can change and subsequently impact capacities to engage in relocation programmes. While a precise “Goldilocks” point for the optimal timeframe to respect individual and community agency cannot be prescribed, nor is it substantially engaged within the literature, it is nevertheless a valuable question.

At the level of the affected populations, the implications of relocation extend beyond fixed programme end-points resulting in misaligned timelines (ten Brink et al., 2025). We can consider how people behave in the aftermath of relocation: multiple studies underline how people return to their previous homes (Akbar et al., 2024; Balachandran et al., 2022; Gussmann and Hinkel, 2021). We can also consider modifying houses, adapting livelihoods, maintaining ties with former places or navigating land-use changes (Cajilig, 2017; Cottar et al., 2025). These are all dimensions within which we can embed an analysis of agency.

Agency how?

What mechanisms or practices can create the conditions for agency within relocation? Here we explore key dimensions from a relocation design and policy perspective: the provision of choice, the participation of the community through consultation or co-production, and the ability to negotiate, make-claims and self-organise. Overall, compared to a relatively clear understanding of how agency is limited or constrained, the avenues for enabling agency are notably under-discussed. Before exploring the perspectives provided by the reviewed literature, it is important to note that the disciplinary and cultural context, as well as the type of environmental relocation, shape how agency is operationalised. Additionally, this is a non-exhaustive exploration of what the inclusion of agency could resemble. While it was not proposed in the reviewed literature, we can also consider that questioning how policy may create the conditions for respecting or enhancing agency are likely to be far wider-reaching than simply the relocation programme. A capabilities approach, as proposed by Sen and explored by Kleine, envisages that agency is shaped by the diverse resources animating them and therefore, in this logic, policy considerations would have to include far wider questions of respecting or enhancing educational resources or psychological resources for example.

The provision of choice

Providing choice is one way to enable agency within the constraints of governed relocation, reflecting a core question in the literature: do people want to move, and if so, under what conditions? The literature highlights a wide range of approaches in providing (or not providing) a choice in relocation: from take-it-or-leave-it-offers to the provision of structured choices to the use of lottery systems for housing provision (Akbar et al., 2024; Li and Feng, 2021). Providing options can enhance individual agency but the quality of these choices is crucial (Straehle, 2023). For example, individuals and communities can feel pressure and even be threatened, turning relocation coercive (Ajibade, 2019). For example, Binder and Greer found that some individuals in New York felt they had little choice in accepting buyout offers: “participants described several more nuanced ways in which fears and consequences (real and threatened) made the buyout feel like less of a choice and more like their only real option.” (Binder and Greer, 2016, p. 102). Coercion is not black and white, and can be hard to pinpoint. Namely, relocation programmes can be framed voluntarily while being involuntary in practice: Arnall et al. describe a relocation process in Mozambique where the government described the programme as voluntary but the use of strategic investment in target communities shifted how much choice some people felt they had (Arnall et al., 2019). Furthermore, the role of financial incentives in terms of coercion is difficult to discern, indeed “it is unclear at what point economic incentives cross the line between nudges and coercion” (Siders, 2022, p. 285). On the one hand, the provision of compensation is often a necessary inclusion in relocation as it is the main tool to minimise households’ loss. Nevertheless, monetary support is rarely neutral: whether compensatory, enabling relocation transitions and livelihoods; or actively encouraging or accelerating relocation—these roles imply distinct questions of coercion and agency, that remain largely underexplored.

Lastly, the literature also pointed to the issue of community pressure or “collective coercion” (Supekar, 2019, p. 1326). For example, in buyout programmes, individuals may have an individual choice but their entire community is targeted. Consequently, the choices made by other households can influence their own.

Consultation, participation and access to information

A reoccurring recommendation emphasised within the reviewed literature is that relocation programmes are more effective and respect agency when the affected households or communities are involved in the process: from risk evaluation and risk perception (Balachandran et al., 2022; Rulleau et al., 2017), to planning (Lund, 2021; Siders, 2019), to wider calls for locally-led adaptation (Yee et al., 2025). This helps ensure that key needs of households and communities are considered. Too frequently, relocation programmes, especially post-disaster resettlements, are not designed in culturally sensitive ways and can lead to social dissatisfaction and even abandonment of new sites or returns to old sites (Bavinck et al., 2015; Jain and Bazaz, 2020). Furthermore, a lack of consultation breeds frustration and is often paired with unfulfilled promises and lack of clear communication (Dalla Fontana, 2025) and can also erode trust (Schwaller et al., 2022; Weber et al., 2025). The cultural appropriateness of such measures, and the governance of relocation more broadly, plays a role in how far local communities, households and individuals feel they can voice their needs; as Iuchi and Mutter observed in a comparison of post-disaster relocation in Japan, the Philippines and Indonesia (2020).

An important consideration is that consultation is useful only insofar as people are capable of engaging meaningfully. Therefore, we need to ask: how do people get access to information and how is relocation presented? Adequate information is a prerequisite to decision-making (Mayr et al., 2020). Lack of information, or only provision of oral information with little written trace, is linked to high uncertainty and the spreading of rumours, intensifying negative perceptions (Iuchi, 2024; Matti et al., 2023). Additional considerations include if information is only presented using expert, inaccessible language (Thaler et al., 2020) and whether community participation is culturally sensitive (Iuchi and Mutter, 2020). To this point, information can occur in a hierarchy of knowledge in which official perspectives are prioritised, limiting community agency and reducing the diversity of viewpoints—this is discussed by Nicholson who explored the concept of “sensitisation” in flood-affected communities in Malawi, a top-down process of informing or persuading communities rather than engaging in two-way dialogue, despite governments and NGOs claiming to use participatory approaches (Nicholson, 2023). As Nicholson aptly points out, “this sensitisation presents a classic conundrum in development discourse: the desire to empower and give agency to vulnerable populations—but only if it coincides with the NGO and government view” (Nicholson, 2023, p. 445). In parallel, this also highlights the importance of cultural-appropriateness within approaches to procedural participation, sensitive to local knowledge systems. With these points in mind, a useful framework to consider is that of co-production (Tubridy et al., 2022). Decision-making, under this approach, is viewed as a collaborative process. It emphasises the value of diverse ways of knowing, and non-expertise, thereby integrating both knowledge-creation and the decision-making process itself. Overall, we can further consider a Capability Approach perspective here by underscoring that formal inclusion does not translate into effective participation if there is no consideration of inequalities in capabilities, and the role of resources from financial to educational resources, or even the availability of time.

Resistance, negotiation, and accountability

Resistance to relocation is nothing new and represents a concrete manifestation of agency (Dalla Fontana, 2025; Depari and Lindell, 2023). There are a plethora of reasons for resistance, from place attachment and historical injustice (Grace-McCaskey et al., 2021), to the role of vested interests (Gussmann and Hinkel, 2021) or even climate scepticism (Lo, 2014). In the Philippines, Gibb described agency being “exercised as the rejection of official evacuation camps, transitional housing sites and permanent housing sites in favour of utilising resources from personal networks or clear religious spaces” (Gibb, 2024, p. 141). Acts of resistance can also be attempts to elicit support for relocation. Bergmann describes a case in Peru where government inaction pushed communities to take on many of the roles necessary in implementing relocation as well as using tactics, to more or less effective ends: “frustrated villagers refused to let go of state representatives after a meeting, attempting to press the regional governor to fulfil long-standing promises. The act of resistance, which the state perceived as hostage taking, hardened lines” (Bergmann, 2021, p. 370).

Resistance often stems from opposition to how a programme, or its absence, is governed. Feelings of injustice or procedural opacity can cause dissatisfaction and anger. It is therefore important to consider pathways to accountability for those governing relocation. One avenue is what Arnall, Hilson and McKinnon call “claims-making” which is both being able to advocate for desires, make demands and condemn the relevant authority, based on perceived injustice (Arnall et al., 2019, p. 6). This therefore is a form of political agency, it is strategic. Relocation programmes can also facilitate this by incorporating pathways for accountability through formal grievance or appeal processes, as well as maintaining transparency throughout. Furthermore, an overall approach of processes of ongoing feedback and dialogue with the community is important (Usamah and Haynes, 2012).

To conclude, we conducted a scoping review of 163 papers to understand how literature on environmental relocation understands and explores facets of agency. We find that most of the selected studies invoke agency without defining it, and that it is often one term of many value-laden ones such as empowerment and resilience. There were two broad main framings in the literature: one firstly focused on the capacity for action and decision-making, where agency was highly individual and discussed in terms of voluntary or involuntary relocation, and another that viewed agency as an expansive, relational process that exists at multiple levels including that of the everyday, and that is pervasive in contexts of constraints. Despite this ensuing wider approach, agency was most frequently operationalised as acceptance or refusal of relocation, overshadowing other necessary considerations. Collectively, these studies emphasise the critical role of participation and community consultation which was the largest overarching recommendation within the literature, possibly at the expense of other approaches. Overall, this review has highlighted the insufficiency of current conceptualisations of agency, and the subsequent complex and open-ended task of considering the application of this concept to localised cases of environmental relocation. The demands of integrating both the need to understand how individuals, households and communities exercise agency and make choices, while recognising that this agency is shaped, enabled, and delimited both by governing policies and programmes as well as by wider societal norms, culture and expectations, opens up a range of avenues for future research. Six gaps are identified and suggested below.

Gap 1: limited considerations of intersectional questions at both systemic and household levels

A critical weakness identified in the approach to relocation programmes is the insufficient, or otherwise superficial consideration, of intersectional dimensions, both at the level of the relocation programme and at the granular level of individuals, households and communities. As to the latter, there was little on how these intersectional facets shape agency and decision-making. Certain choice experiments or survey-based methodologies paid attention to demographic considerations and whether they impacted relocation preferences but there is insufficient attention at how decisions are taken, and how agency is negotiated, in light of diverse marginalisations. Another angle of consideration would be how dynamics internal to the household unfold in relocation, and how these dynamics may be coloured by questions of gender or age, for example. Research delving into these household-level dynamics and the coinciding micro-politics would be a rich addition to the scalar picture we are drawing of environmental relocation. Across these considerations, there is room for wider critical, and decolonial, analysis of relocation.

Gap 2: linking risk perception and action—the “Sense of Agency”

Continuing this question of micro-level considerations, there were studies that considered risk perception in relocation desires (Alexander et al., 2012; Dachary-Bernard et al., 2019) but perhaps not the relationship between risk perception, the perceived realm of actions and the relative feeling of control over actions. This can be understood as a sense of agency, a subject of study in psychology and philosophy (Moore, 2016). For example, what set of choices do individuals or households perceive when presented with the possibility of relocating? What sense of agency do they have over the relocation process, especially against the backdrop of environmental or climate change related risks? This echoes one of Simpson et al. research priorities for climate mobility linking public understandings of climate risk with research on how individuals weigh harms and benefits when making mobility decisions (Simpson et al., 2024).

Gap 3: limited study of temporal considerations

While there are a few longitudinal studies (Iuchi, 2024; Mcmichael et al., 2025; Seebauer and Winkler, 2020; Shrestha et al., 2025), there is room for studying how temporal questions—such as the duration of displacement, the timing of interventions, the perceived speed of environmental change, and the historical context of a community’s relationship with its environment—come into play to shape both the experience of environmental relocation and the exercise of individual and collective forms of agencys. When specifically considering agency and the question of voluntary movement, we argue it is necessary to move beyond static snapshots. Concepts such as sequential or iterative decision-making would be of interest (Castro, 2026). Furthermore, following Gap 2, it would be valuable to understand how an individual’s past experiences and future projections shape their own perceived realm of agency and action.

Gap 4: theorising coercion in environmental relocation

The importance of understanding involuntary relocation is consistently emphasised in the literature. However, there is room for more nuance by critically considering how there may be more subversive forms of coercion in relocation programmes, and how these are facilitated by different mechanisms. This is not always overt physical force, but can be a subtle, systemic, or structural pressure that effectively limits genuine choice. As mentioned previously, there is a particular black box around the relationship between financial incentives and coercion (Siders, 2022). Additionally, further research could investigate the concept of “collective coercion” examining pressures either internal to communities or those resulting from partial community relocation (Supekar, 2019). Finally, it would be pertinent to explore how the background conditions of risk and threat associated with climate change create circumstances of duress that condition decision-making and complicate the notion of freely chosen action.

Gap 5: plural epistemologies of agency and the non-human

A limitation of this paper has been the narrowed focus on human agency despite the fact that environmental relocation is inseparable from a wider question of the environment and risk. The identified literature brought up agency of atolls against wider oceanocentric perspectives (Jarillo and Barnett, 2022) and the “the agency of more-than-human entities, whether these are water, memories, or fish (...) into the broader narrative of mobility and displacement” through (Gini, 2025, p. 3). Nonetheless, this was widely not considered in the selected literature, and coincided with a lack of Indigenous or non-Western conceptualisations of agency. However, it is insufficient to discuss non-Western epistemologies only through the lens of considering the non-human (Rosiek and Adkins-Cartee, 2023). Diverse knowledge systems hold diverse and fundamental conceptualisations of how (and whether) the human is situated in relation to the wider environment and therefore how agency is produced or understood (Rosiek et al., 2020). For example, the perspective of ‘Place-Thought6’ described by Watts in Anishnaabe & Haudenosaunee cosmologies is “based upon the premise that land is alive and thinking and that humans and non-humans derive agency through the extensions of these thoughts” (2013, p. 21). Overall, the literature globally insufficiently addresses the diversity of Indigenous epistemologies specifically. While we have discussed certain aspects that surfaced in the papers, such as thick time (Gini, 2025), this remains a significant and crucial open-ended research gap, especially given the recurring assertion that relocation policies must be localised, as therefore must conceptualisations of agency (Rosiek and Adkins-Cartee, 2023).

Gap 6: enabling agency? Insufficient consideration of policy design implication

Lastly, as we have argued throughout, the link between agency and what it implies for environmental relocation policy and governance remains vague. This gap is two-fold, firstly, the academic literature highlights well how agency is limited or constrained, however the discussion and subsequent implementation strategies for fostering agency remain under-developed. Policy design in relocation is complex, facing many trade-offs (Siders, 2022). In terms of policy design that enables agency, the current literature heavily emphasises participation and consultation, at the detriment of other possible avenues and often without sufficient sensitivity to the cultural specificity of such processes; this is essential as enabling agency will not be a one-size-fits-all approach. As mentioned in this review, there are proposed frameworks from co-production to claims-making but the literature provides limited analysis of the institutional arrangements through which agency can be embedded and sustained.

An additional future research avenue within policy design is continuing careful exploration of the relationship between choice and agency, and how this can be integrated into environmental relocations. A possible framework would be Kleine’s Choice Framework which applies the Capability Approach to comprehensively consider “(a) the existence of choice—whether the different possibilities exist and are, in principle, attainable for the individual if the combination of their resource portfolio and the structural conditions would allow it; (b) sense of choice—the subset of available choices an individual is aware of and believes that they have; (c) the use of choice—whether or not an individual actually makes the choice and (d) the achievement of choice—whether the outcome matches the choice expressed” (Kleine, 2019, pp. 6–7). Lastly, with this Capability Approach in mind and the wider challenge of public policy, this opens the door to considering how resources affect the ability to be agents in the context of climate change, as also identified in (a) of Kleine’s framework. Here, an additional layer of methodological and epistemological complexity emerges, since policy design is inseparable from questions of evaluation and measurement (Ibrahim and Alkire, 2007). Overall, a significant area for future research is the complex and open-ended challenge of determining how agency can be effectively and meaningfully integrated into both the design and implementation of relocation programs.