Building Slavery-free Communities: A Resilience Framework

ABSTRACT There is growing interest in the use of community-based approaches to address the causes of modern slavery and the related goal of building anti-slavery ‘resilience.’ However, the concept of resilience is often poorly understood and applied without attention to the specific challenges of anti-slavery policy and practice. This paper provides a conceptual framework for understanding the process and outcomes of building resilience against contemporary forms of slavery within place-based communities. Inspired by established ecological models of resilience, we propose an adaptive ‘resilience cycle’ that activists and policymakers can draw upon to inform the process of designing and delivering policy interventions. This process is combined with a review of evidence about the multi-level social determinants of modern slavery to suggest a framework of topic areas for local review and measurement, as a means to assess existing gaps and assets, enable comparative learning, and measure progress toward goals. We also outline a future research agenda exploring locally grounded perspectives on modern slavery risk and resilience, to improve understanding of the factors underpinning resilience across different social and economic contexts. This article will assist policy-makers by clarifying the concept of anti-slavery resilience, which can in turn inform policy design and implementation, and help to make connections between disparate initiatives from multiple actors. By combining a process for building resilience with an overview of social determinants underpinning a slavery-free community, we offer a basis for gap-analysis and ongoing measurement. The research agenda that we outline to better understand factors underpinning resilience would make a valuable contribution to improving anti-slavery governance and assist in developing a better understanding of the linkages between achieving Sustainable Development Goal target 8.7 1 1 Take immediate and effective measures to eradicate forced labour, end modern slavery and human trafficking and secure the prohibition and elimination of the worst forms of child labour, including recruitment and use of child soldiers, and by 2025 end child labour in all its forms. and wider sustainable development goals.

grounded development of a slavery-free communities index, which could assess and guide public policy responses both sub-nationally and trans-nationally, serving as a mechanism for future learning and comparison.

The Utility of Conceptions of 'Resilience' in Relation to Anti-slavery Policy
The term 'resilience' has been utilized in many diverse settings. From physical material properties in engineering (Hoffman, 1948); to individual characteristics in psychology (Luthar et al., 2000;Tryon & Radzin, 1972); to community, city, and systems level resilience in ecology (Holling, 1986) and disaster management (Godschalk, 2003), the concept has been adopted (and perhaps, stretched) for application in an array of fields. This section explores the meaning and application of resilience, discusses limitations of transfer of the concept of resilience to social systems, and highlights the "additional conceptual layer[s]" (Davidson, 2010(Davidson, , p. 1142 that are necessary to apply the concept in relation to anti-slavery policy. To achieve this we ground our analysis in the eco-systems resilience of Holling (1986Holling ( , 2001 who argued that resilience is the adaptive capacity of a system, and "can be thought of as the opposite of the vulnerability of the system" (Holling, 2001, p. 394). Holling's (2001) diagram of the adaptive cycle shows four stages (exploitation, conservation, release, and reorganization) of an infinite loop, plotted against two dimensions of potential and connectedness (see Figure 1). This adaptive cycle is a continuous process that 'alternates between long periods of slow accumulation and transformation of resources with shorter periods that create opportunities for innovation' (Holling, 2001, p. 394). In Holling's model, 'release' occurs when there has been a breakdown in the system, causing a period of uncertainty. 'Reorganisation' comes about as the result of pulses of innovation, and novel re-combinations of system elements. At this point, potential for change begins to increase, but must also be nurtured through a period of growth or (in Holling's original terminology) 'exploitation.' If harnessed effectively, that growth can accumulate in a period of conservation, before the cycle re-starts with a fresh disruption. Gunderson and Holling (2002) go on to situate the adaptive cycle as one of many 'levels' within a more complex system or 'panarchy.' A panarchical system nests the cycle of resilience with other linked adaptive cycles, so that rather than looping endlessly, change in one part of the cycle can be connected with influence in other, related layers. A practical illustration of this concept might be to consider the linkages between the UN's sustainable goals: increased resilience in quality education (goal 4) for example, is also likely to have an iterative effect on resilience against poverty (goal 1) and decent work (goal 8) (Project Everyone, 2015).
The concept of resilience has been applied across international organizations such as the United Nations Development Programme and European Union (Mikulewicz & Taylor, 2019). However, the vast array of multi-disciplinary interpretations has also prompted criticism of the concept. These range from accusations of de-politicizing, techno-managerialism (Mikulewicz & Taylor, 2019) and arguments about the deterministic structure (Folke, 2006) of such an approach, to the supposed "nihilism of the underlying ontology of vulnerability" implicit in the assumption that communities must accept and prepare for inevitable disaster (Evans & Reid, 2013, p. 85). Others have questioned the ability of this eco-system theory to be adequately applied to social systems (Adger, 2000) due, in part, to the added complexity that arises from the additional conceptual layers of agency, power, and collective action when dealing with human interactions (Davidson, 2010). There is also a critique of resilience in relation to social policy as a justification of 'responsibilization' of individuals and communities in reaction to neoliberal reforms which have fragmented and rolled-back social support systems. Commentators such as Amery (2019) argue that the term is often used without definition and pays insufficient attention to structural drivers such as poverty.
However, others find the concept inspiring. Carl Folke (2006) argues that resilience enables people: to think about structures and processes in a dynamic fashion, to move away from a steady-state world where change is looked upon as an exception, to confront complexity and uncertainty, and move further into patterns and processes that you cannot directly observe and quantify with available data.. (p. 259) Whilst Holling's resilience framework was designed in relation to ecosystems, the concept has also been used in much broader socio-ecological research. There are limited examples of specific social science applications, including as an alternative approach to the war on drugs (Chandler, 2015); in peacebuilding (De Coning, 2016); highlighting the power of Indigenous communities (Thomas et al., 2016); and both macro (Cavanagh, 2017) and micro (Thiede, 2016) economics. There is a more extensive body of work on community and urban resilience, which combines some work on local social dynamics and power structures (Lerch, 2015;Mulligan & Rogers, 2018), with a greater focus on local physical infrastructures, particularly post-natural disaster or in the context of preparing for climate change (Tyler et al., 2016). Shaw (2012, p. 286) also points to the role of resilience in transformation of UK local authorities and communities under austerity-related spending cuts, highlighting the difference between "bouncing back" (recovery) and "bouncing forward," adapting dynamically to new conditions. Many of these applications link to attempts to deliver sustainable development, and Holling (2001) expressly recommends the use of a resilience framework toward a better understanding of this topic. Given that contemporary slavery is itself a development issue, outlined in the goal 8.7 of the UN's (2015) Sustainable Development Goals (SDG), and situated within the wider goal of providing 'decent work and economic growth' as one of seventeen over-arching SDGs, it is a short step to see utility in the concept of resilience for anti-slavery interventions.
However, in adapting the resilience cycle for application to the anti-slavery agenda, it is important not to reinforce existing narrow preconceptions of appropriate paths to intervention. Chandler (2019) raised concern that many of the applications of resilience are too focused on sustaining existing systems, and favored a greater focus on transformation: Resilience as a policy framework of adaptation appears to be drawing to a close as it lacks an adequate agential or transformative aspect: it is always too oriented to adapting to feedbacks and modulating around sustaining what exists. When what exists is the problem itself . . . then it is clear that critical thought and policy practice need to go beyond imaginaries of resilience. (p. 2) In addition, the importance of locally determined measures of success means that resilience must be constituted differently in every community, and prompts the questions of 'resilience for whom, what, when, where, and why?' (Beymer-Farris et al., 2012;Meerow & Newell, 2019). These questions highlight the hierarchies of power that exist within (and without) a community, meaning that a defining feature of a resilience-based framework would be its ability to "work outside of and against dominant power hierarchies" and to "[resist] epistemological closure" (Walsh-dilley & Wolford, 2015 pp.176, 174). Although establishing resilience also requires acknowledging areas of vulnerability (Cannon & Müller-Mahn, 2010;Miller et al., 2010), there is an emphasis on 'what communities can do for themselves and how to strengthen their capacities' (Twigg, 2009). This requires a flexibility for adaptation, and implies the need for community-led and inspired systemic and policy change to form part of the resilience cycle. In addition, a resilience framework for a slavery-free community must be ready to critique existing actors and institutions, whether they are individual consumer choices; business model decisions; local policy implementation and enforcement; or national and international regulation and systemic infrastructures.
With this in mind, we have adapted Holling's model to outline a resilience cycle which could inform the policy journey toward building a community free from contemporary slavery. In setting out this cycle we draw on Holling's original four conceptual phases, but adapt them to incorporate the insights gained from the literature above (and change the confusing use of the term 'exploitation'). This cycle is explained below, beginning at problem diagnosis (number 1 bottom right) and moving through to normalization and sustained resilience (see Figure 2.)

Diagnosis of Problems and Potential Solutions
Although this paper argues against measures of risk and vulnerability as a stand-alone focus of antislavery intervention, such assessments form an important first stage of a resilience cycle. Local manifestations of exploitation can be rendered more obvious by analysis of risk factors, and by highlighting geographic, demographic, or sectoral weaknesses can enable particular threats to be identified. This process can also recognise assets within the community that can help to address different forms of exploitation. By using different tools and datasets, a cross-sectional picture can emerge that reveals locally specific patterns, as well as potential levers for change.

Challenging Hierarchies and Systems
Having identified potential risks and assets at the community level, it becomes important to involve a wide range of actors in discussing and validating risk and vulnerability findings, identifying assets for change, prioritising and implementing action, sharing best practice, and learning from both external networks. Survivor voices are crucial to informing this process, and challenging existing systemic imbalances and weaknesses. Action-research and consultation with other communities (for instance professionals or stakeholder groups not involved in the original problem identification) can also assist in bringing new insights.

Changing Cultural and Institutional Landscapes
In order to nurture change and ensure that innovation is not lost, attention should be paid to the cultural and institutional changes that are needed to support resilience. This is particularly relevant to some of the structural issues that underpin contemporary slavery, and may involve extended work with significant institutions, such as the media or businesses. Without attention to challenging the cultural and institutional causes of exploitation, interventions run the risk of endlessly cycling between stages one and two, problem diagnosis and local validation, prioritisation and action-planning. However, it is also recognised that the process of influencing cultural and institutional change is likely to require long-term activism.

Normalize and Sustain Practice
This phase of the cycle considers what changes to governance, legislation and policy are needed to embed positive changes identified and achieved. It involves monitoring and evaluation of progress to date, as well as sharing learning widely, and initiating further governance change where necessary. As innovation shifts the cultural landscape and contributes towards liberation, a theoretical 'freedom dividend' can be realised (in terms of greater prosperity, and reduced opportunity for crime) as practices that normalise exploitation are switched for practices that promote greater resilience (Ramburn et al., 2018). The point of this adaptive cycle is not to see resilience as a fixed and finite endpoint, but to be continually adjusting and re-evaluating the local context to enable continuous improvement towards creating a community where contemporary slavery cannot easily be established or flourish. This framework has a number of advantages. Firstly, it is dynamic: it recognizes that the pursuit of a slavery-free community will be an ongoing endeavor which is continually adapting to new environments, pressures, and information. However, it is not nihilistic, in that it proposes a process to deal with emerging problems and build resilience on a continuing basis. Secondly, the framework accepts the value of studies of risk and vulnerability in the diagnosis stage, but provides a mechanism to address results with attention to systemic and structural assets that can become drivers for change. Thirdly, the framework systematizes the engagement of survivor and community voices in the diagnosis of problems and search for solutions, a factor agreed to be missing within much existing anti-slavery policy formulation. Fourth, it underlines the importance of engaging with power structures and institutions to nurture and consolidate change.
Multiple programmes and tools already exist at local, national, and trans-national level which can contribute to this cycle. For instance, Delta 8.7 is an online international programme that gathers data at a global level, enabling comparison between countries which can help in challenging local norms and stimulating debate. We are starting to see organizations emerging that provide additional capacity to support champions and empower survivors to be involved in policy development, such as the UKbased NGO 'Survivor Alliance.' Free the Slaves has developed a community liberation model (Ramburn et al., 2018) which has implemented and tested a community-based approach to nurturing entrenched, institutionalized forms of exploitation and achieving change in high prevalence countries. Such interventions have developed independently, but using the framework above, we can also view the contributions of such tools as parts of a wider system that increases resilience against modern slavery, rather than as isolated and abstract initiatives.
However, what this process-focussed cycle lacks is a mechanism for local review and measurement, both as a means to assess existing gaps and assets, to enable comparative learning and as a mechanism to measure progress toward goals. The next part of this paper will explore how this can be addressed by combining the resilience cycle with an improved understanding of social determinants of contemporary slavery.

Identifying the Social Determinants of Contemporary Slavery
To date, measurement, monitoring, and evaluation have been acknowledged to be a problematic area for the anti-slavery sector. A meta-evaluation of 'promising practices' has shown that the quality of evidence from existing programme evaluations is low, with a lack of clarity in relation to the independence of the evaluation, data sources, and methodologies, and a failure to focus on impact, as opposed to project outputs or outcomes (Bryant & Joudo, 2018). Other studies on monitoring and evaluation in anti-slavery work have also found that most anti-slavery and anti-trafficking programmes 'bean count' activities and direct outputs, to the detriment of producing meaningful evidence of which interventions are effective (Davy, 2016;Harkins, 2017;Van Dyke, 2017;Yea, 2017).
Even in countries with a relatively high-profile response to modern slavery, such as the UK, existing statistics -such as prevalence measures or referrals into victim support frameworks -are often viewed by NGOs and statutory agencies as unreliable or ineffective measures of progress. For instance, a National Audit Office (NAO) report on reducing modern slavery stated that the UK's Modern Slavery Strategy "does not have a measure of success for its objectives nor a definition of what success looks like" (NAO, 2017, p. 18). Studies have also criticized the UK government for a lack of meaningful evidence (Cairney, 2017;Van Dyke, 2017). Responsibility for monitoring and evaluation of local interventions also frequently falls to local actors working with minimal resources and few reference points for evaluating performance (Gardner et al., 2017).
This lack of monitoring and evaluation information presents a challenge for the operation of the resilience cycle: without an understanding of the factors which underpin effective governance, institutions, and practices to challenge contemporary slavery and exploitation, problem diagnosis will be difficult. The engagement of communities of practice will be less fruitful if there is no common basis for comparison of strengths and gaps in systems and performance. And understanding whether or not desired change is occurring will also be problematic without a means of quantifying and monitoring that change.
For this reason, we suggest that it is critical to develop a mechanism for measuring and monitoring assets and gaps in resilience to modern slavery. This implies a need for an understanding of core elements of resilience that can be tested at community-level through an audit or similar tool. In order to identify the key factors comprising this index, it is necessary to engage with systemic and institutional factors that influence the local anti-slavery landscape, in addition to the local context. This is where identification of social determinants of resilience can play a role.
The concept of social determinants originated in public health (Braveman et al., 2011;Gareth H. Williams, 2003;Marmot, 2005) and has been defined by the World Health Organization (WHO) as: "The conditions in which people are born, grow, live, work and age. These circumstances are shaped by the distribution of money, power and resources at global, national and local levels" (World Health Organisation, 2019). There is also a growing body of literature (see, for example, George et al., 2016;Such et al., 2018;Zimmerman & Kiss, 2017) linking public health approaches to effective policy responses to modern slavery and human trafficking, particularly in relation to population-level prevention initiatives and the provision of improved support services for victims and survivors. As Dahlgren and Whitehead (1991) devised a 'rainbow' model of social determinants of health, taking as their starting point the contention that socio-economic, cultural, and environmental conditions, living and working conditions, social and community networks, and lifestyle factors all contribute toward the health and wellbeing of individuals. Their framework also highlights key influencing factors, including the role of education, work environments, housing, water and sanitation, and unemployment. Critically this framework has an asset-based focus, in that it focusses on health rather than sickness. Asset-based approaches are common in public health (see, for example, Foot & Hopkins, 2010;Rowett & Wooding, 2014;The Health Foundation, 2015), where they have been used to guide public policy interventions that focus on strengths rather than needs. Key elements include a philosophy of seeing citizens and communities as co-producers, rather than the recipients of services; a focus on community networks, relationships, and friendships; attention to what is already working well; and empowering individuals and communities to control their futures and create tangible resources such as services, funds, and buildings (Foot & Hopkins, 2010). This philosophy, coupled with the multi-layered and systemic perspective, makes it a good conceptual fit for our resilience cycle.
Inspired by public health perspectives, particularly Dahlgren and Whitehead (1991) model, it is possible to start to identify how legal and regulatory, social and cultural, and personal factors relating to anti-slavery resilience interact with structural-level social determinants (see Figure 3). The diagram above includes a dynamic element of progression (stages of prevention, discovery, respite and recovery and promoting a sustainably resilient local context) reflecting the systemic and temporal interconnection of anti-slavery policy and processes. We consider that these stages make intuitive sense and give greater clarity than the UK's version of the '4P's' paradigm. 3 In suggesting elements to populate the framework we have drawn on a wide range of academic literature as well as a series of action-research workshops with anti-slavery partnerships within the UK for the local and cultural determinants (see below for sources of evidence). However, at this point, these elements are offered as the basis for further research rather than as a definitive, tested list. A number of studies (notably Cho, 2015;Perry & McEwing, 2013;Williams et al., 2010) have previously explored social determinants of modern slavery and human trafficking, but generally The international framework utilized by the UN and the US focusses on prevention, protection, prosecution, and partnership, whereas in the UK, the government's modern slavery strategy is based on a counter-terrorism strategy of 'prepare, pursue, protect and prevent.' from the perspective of factors increasing or decreasing the exploitation associated with migration. Whilst these studies are helpful in identifying indicators of interest, they do not directly address what factors enable individuals to avoid exploitation in a settled community context (although there is considerable overlap between the most commonly cited policy recommendations in Perry and McEwing (2013, p. 151) study and the framework below). It is also probable that key factors underpinning resilience will vary between different social and economic contexts. The diagram below is designed as a starting point for discussion on how and where that variation occurs.

Structural Factors
Poverty was the most cited determinant of trafficking in Perry and McEwing's (2013) Cochrane review. Both this study and the work undertaken by Burns et al. in Tamil Nadu emphasize the importance of access to credit and capital as a means to avoid debt and enable resilience in times of crisis (Burns et al., 2016). A linked issue is the provision of healthcare, as poor health is a frequent cause of debt (Burns et al., 2015). Formal education is also a key protective factor in the majority of studies examined by Perry and McEwing (2013) particularly maternal education. Although forms of contemporary slavery affect both men and women, women are disproportionately affected, and societies which placed a low value on women and girls were also noted by as carrying additional risks for those groups.
Strong institutions, facilitating anti-corruption initiatives, justice, and enforcement of the rule of law (with lower crime rates) diminish the push factors associated with human trafficking, according to Cho (2015) and Williams et al. (2010). Finally, the availability of employment has been identified as a constraint on trafficking outflows by Cho (2015) although Burns et al. (2015) and Choi-Fitzpatrick (2017) also note the importance of responsible business practices and norms, particularly in relation to health and safety and labor rights, to limit abuse.

Legal and Regulatory Factors
Although it is commonly repeated that slavery is universally illegal, Schwarz and Allain (2019) have recently demonstrated that many states have not passed criminal provisions to enforce the international norm, particularly in relation to servitude and forced labor. Such provisions are therefore desirable assets to promote resilience. However, from a social determinant perspective, many other areas of legislation are equally important. For example, the protection of basic rights associated with citizenship, and importance of safe migration pathways was mentioned in 41% of the studies reviewed by Perry and McEwing (2013) with 16% highlighting the significance of citizenship documentation to protecting rights. Zimmerman and Kiss (2017) document forms of exploitation and harm that affect migrants in the pre-departure, destination, and return phases of migration, from which it is possible to extrapolate supportive public policy interventions. For example, legislation on a legal minimum wage, regulation of employment agencies, procurement law, and regulations on recruitment and supply chain transparency also contribute both to preventing exploitation and creating a long-term sustainable business culture (although the lack of adherence to UK 2015 Modern Slavery Act legislation also indicates that enforcement is essential if they are to be effective.) For the benefit of victims and survivors, advocacy organizations furthermore require secure but flexible approaches to data sharing and a legal defense for those forced to act criminally under coercion. In addition, our action-research with a group of frontline service professionals in a UK context showed that statutory entitlements (and access) to support services and compensation for survivors are important to break cycles of exploitation and facilitate prosecutions.

Culture and Locality
The culture and locality layer of the framework covers the assets which may be available through local state or NGO service providers, or provided collectively through the natural support networks of communities. As we move toward the center of the diagram there is likely to be more variance in the resilience assets identified, according to the nature of the locality. Areas included in our social determinants are deliberately general topics that are likely to be relevant in a variety of contexts. In terms of prevention, they include support for those who are particularly vulnerable: in the UK groups perceived as carrying particular risks include homeless people, adults with learning disabilities, and young people vulnerable to being caught up in 'county-lines' drug trafficking. Support to promote safe and stable families is also critical, especially given that children and young people make up 25% of trafficking victims worldwide, and vulnerability increases in contexts of family disfunction (Williams et al., 2010, p. 137). Multi-agency partnership working is being widely used to target joint enforcement (Foot, 2016;Gardner et al., 2017) whilst both short and long-term housing and advocacy services are crucial aspects of local survivor support. Such local action also benefits from local civic leadership, both in terms of political commitments mandating engagement from mainstream statutory services governed at the local level, and commitment to high standards from the local business community (Gardner, 2018).

Personal
Recognizing that communities are also comprised of individuals, the framework also includes elements influencing individual behavior which could be promoted to whole populations. The social determinants addressed by the Perry and McEwing (2013) study highlight two factors impacting at the personal level: ignorance and demand. The assets mobilized in response to these determinants, therefore, include education and awareness campaigns, and interventions to alter consumer behavior. These also impact on an individual's ability to recognize signs of trafficking -essential for discovery of victims -and their sympathies toward victims and survivors (although recent research on the impact of media and campaigns at local level indicates that awareness does not automatically lead to increased reporting (Birks & Gardner, 2019)). However, it is important to acknowledge that building resilience for specific individuals, for instance, tailoring local services to the needs of survivors to address the impacts of trauma and prevent re-victimization, could demand attention to a much wider and more nuanced set of individual-level interventions. Strategies at a personal level are also likely to vary according to the local cultural and familial context, and the types of exploitation being targeted.
Viewing these factors together as a multi-layer overview of social determinants helps us in a number of ways. Firstly, it acknowledges the systemic linkages between individual action, social and cultural activities, legal and regulatory assets, and structural factors. Structural factors surrounding the local context are connected with the dynamic nature of the system in a broad sense, with most having a relationship to all four of the process-stages of prevention, discovery, respite and recovery, and sustainable freedom. For instance, welfare and anti-poverty provisions are a critical factor in preventing individuals from falling into debt or hardship precipitating exploitation (Burns et al., 2015). The availability of anti-poverty measures can promote discovery through providing a means for individuals to highlight and exit exploitative relationships, as we have seen in the case of Indian bonded laborers assisted by state-subsidized food supplies (Choi-Fitzpatrick, 2017). Access to social support and a means of income are critical to promoting survivor recovery and preventing reexploitation (Clayton, 2017). Additionally, welfare rights build sustainable resilience by providing individuals with greater freedom of choice in the employment options they select, bolstering labor rights, and sustainable business practices (Choi-Fitzpatrick, 2017). Taking another example, educational interventions have been used to help raise awareness of modern slavery and human trafficking amongst young people at risk, and are also used as part of community development interventions to heighten awareness of existing exploitation (Ramburn et al., 2018). Education is also frequently an important aspect of survivor re-skilling and community integration (Williamson, 2018), and is being used to promote a sustainably slavery-free environment through teaching materials on topics such as fast fashion, which inform consumer choices. Similarly, many of the other structural factors can be seen as having a relationship with all four dynamic stages of the system. Secondly, this set of social determinants shows us that the legal and regulatory landscape is much broader than specific anti-trafficking and anti-slavery legislation, which complements but does not complete minimum requirements. In particular, regulation that promotes swift resolution of citizenship, safe migration, a minimum wage, support, non-criminalization, and compensation for survivors, non-criminalization of survivors, and ethical business practices in terms of health and safety and procurement are all essential to avoid undermining policy objectives.
The diagram (Figure 3) also connects these wider determinants to personal-level assets, such as recognition of the existence of modern slavery and human trafficking, knowledge of the signs of exploitation, individual support for survivors (and more widely, encouraging communities to help create a supportive cultural and social context for them) and personal recognition of how consumer habits and lifestyle choices contribute to both the problem and potential solutions, with accompanying behavior change. Such individual responsibility sits at the root of a community response to modern slavery and human trafficking.
As it stands, we recognize that this selection of determinants has limitations. For instance, it currently reflects a view which is strongly influenced by a western and global north perspective. This view is unlikely to be comprehensive, and the factors underpinning community resilience are likely to vary across different social and economic contexts. It is also relatively complex and requires translation into an audit tool, or similar, for operationalization in the field. Nonetheless, it provides a useful basis for conversations about what resilience factors are most important in different social and economic settings. The next step is to identify data sources and measures that can be used to assess these assets and to contribute toward a deeper understanding of anti-slavery resilience.

Combining an Understanding of Resilience Process and Social Determinants: An Illustrative Case Study
The two diagrams presented above are potentially useful for retrospective policy analysis and review and as a means to inform forward-planning and progress monitoring on new areas of policy initiative which aim to build resilience. However, more work is needed to explore what insights they can provide to empirical research and how process and measurement can best be combined. An illustrative case study of how these concepts can inform review of a UK initiative is presented below, but we also hope to test these ideas in other social and economic contexts over the next two to three years.

Case Study: Building Resilience through Local Long-term Support for Survivors of Modern Slavery
A local authority in a UK city was keen to support a Home Office project exploring options for supporting survivors of modern slavery when they exited the UK NRM. As one of the six pilot areas selected, the authority was able, with central government funding, to establish a team to assist survivors to settle locally. The project was enabled by political support from Councillors and the local Police and Crime Commissioner, as well as a supportive local context for anti-modern slavery work including established multi-agency partnerships.
Despite safe-housing in other parts of the country being over-stretched, the local authority received few referrals. This was partly because there were no safe houses in that area of the UK, meaning any survivor accessing the service had to relocate. Additionally, the offer was initially restricted by central government to those who were entitled to welfare benefits and had received a 'conclusive grounds' decision confirming them as victims of modern slavery. Because this constituted a relatively small proportion of survivors exiting the NRM, it proved too narrow to enable the pilot, and the criteria were eventually broadened.
Housing was the biggest challenge. All referrals were identified as requiring social housing, which is in very short supply. Project referrals were accepted as a result of a goodwill from both social housing providers and the local authority. One unanticipated difficulty was that the first two referrals came with dogs, excluded by many housing providers but deemed essential to survivors' wellbeing. Again goodwill from partners led to a solution.
The pilot established that modern slavery survivors' needs are complex and that understanding and meeting this requires considerable patience and the recognition that behaviors may be inconsistent with institutional expectations of 'victimhood' and include, for example, substance misuse or domestic abuse. Alongside this, justifying provision for survivor needs had the potential to be politically contentious in the context of constrained resources. The team often found its key role was in facilitating access to mainstream services that were also difficult to access by the general population. Without advocacy and support, survivors would be at the back of a very long queue despite having higher levels of need. Whilst strong local relationships and a desire to support the project were effective, this approach would not be sustainable as a long-term solution without legislation and further resources.
Relating this case study to our framework of social determinants, we can see that although many anti-slavery resilience assets exist in the UK (for example, welfare benefits, the NRM, and anti-slavery legislation), modern slavery victims, and survivors are frequently denied access to those benefits through immigration or social exclusion, or are simply unable to access them for reasons of geography. Individuals outside the safety net remain vulnerable without advocacy and intervention, and action also benefits from civic leadership and a supportive cohesive community.
The case can also be related to the ability to learn and improve through a resilience cycle. In this example, initial problem identification (stage 1) had been nuanced by engagement with practical case management and conversation with survivors, raising a core policy challengehousing -that appeared indirectly related to the issue of modern slavery, but was actually a key determinant of wellbeing (stage 2). Whilst there was learning and innovation at the local level in terms of changes needed for local institutions and practice (stage 3), this learning has not yet been consolidated and normalized through changes in national legislation, policy, and practice. Thus, the potential increase in resilience that could be gained from this learning has not yet been fully realized.

Conclusion: Operationalizing Community Resilience as a Means to Address Contemporary Slavery
In a context where responses to slavery are characterized by 'polycentric' governance -multipolar and multiactor in nature, with few clear accountabilities and no obvious chain of command -the resilience cycle offers an inclusive framework for action, to which multiple actors can contribute. When combined with the social determinants rainbow, as a guide to the areas that might be probed in more depth, and measured to track progress, we have a practical basis for investigating and comparing the assets that underpin resilience in specific settings. In addition, the combination of the two tools ensures that the resilience cycle moves beyond responsibilization of particular actors to take in a perspective that connects the interaction between local and individual initiative and wider social and legislative drivers.
One potential criticism of this approach, as a pathway to inform local change, is that it will highlight some factors that cannot be influenced at a local level. However, such information is still valuable to focus attention upon legislation and policy that counter-acts stated policy aims, and to underpin anecdotal frustrations (encountered mainly at the frontline) with evidence that can influence policymakers. In addition, the process of understanding interactions between social determinants will assist in identifying and repurposing local assets to mitigate negative impacts.
It is also true to say that in framing our argument to address a broad spectrum of exploitation, and choosing to use a systems-based approach, less emphasis has been given to recognizing the nuances that will be present in relation to different manifestations of Modern Slavery and Human Trafficking, or through variations in individual's circumstances. We hope that development and operationalization of this framework will bring some of these subtleties to the fore and enable us to further refine our understanding of how individual, community, and structural resilience connect.
Further work is needed to identify indicators and proxies that can help us to understand how effectively we are closing gaps in resilience. For instance, as we have seen, states may have a broad framework of legislation in place but still be failing to effect real change due to multiple negative impacts from structural externalities or a lack of community engagement to facilitate implementation. Understanding these dynamics through rich qualitative case-study examples will be essential. In order to assist with this process, and to test and develop our initial framework, we are working to establish a global network of community practitioners which will connect academic partners, law enforcement, local government, NGOs and faith organizations working in communities, with the aim of building and improving place-based antislavery interventions. Initially being tested in four communities in Brazil, UK, Mozambique, and Thailand, this programme will establish the infrastructure for a wider network and will trial a slavery-free communities audit tool.
In addition, there is the opportunity to pursue the research aim of a slavery-free communities index as a means to compare and benchmark progress more widely. As a starting point, this index could draw on common resilience factors identified within our framework and cross-reference, where appropriate with other existing datasets such as the International Standards Organization (ISO) standard on 'Sustainable cities and communities -indicators for city services and quality of life' (International Organization for Standardization, 2018). Whilst such a dataset would inevitably require difficult choices in terms of the assumptions made, aggregation of data, weightings given, and how the index will be visualized, it could also assist in remedying the data and evidence gap in relation to local anti-slavery interventions. It may furthermore provide a useful comparator to existing judgments of the effectiveness of governance response promoted through the US Trafficking in Persons Report and the Global Slavery Index.
Eventually, and acknowledging very specifically the complexity and uncertainty that an investigation into resilience necessitates, our aim is to add more levels of interaction and parallel systems to be able to further define a full "panarchical" system of anti-slavery resilience. This panarchy would theoretically be able to situate modern slavery as a proxy system for much broader sustainability, highlighting the commonalities with other multi-dimensional challenges, such as decent work and sustainable cities. If the connections to multiple systems -and the potential to interact and change broader social determinants -can be more clearly illuminated, we will be closer to achieving the sustainable development goal to eradicate modern slavery.