Family reunion policy for resettled refugees: Governance, challenges and impacts

The past decade has seen renewed efforts to establish resettlement as a durable solution for refugees, both as a protection tool and a mechanism to equitably distribute them among countries. Although the right to a family life is widely recognised as a fundamental human right, whether refugees can arrive with their family or be reunited with family once resettled varies across receiving countries. Little is known about family reunion policies in countries leading the resettlement efforts, and about the impact of these policies on the lives and experiences of resettled refugees. This paper addresses this gap though a systematic scoping review of academic and policy literature on family reunion policies for resettlement refugees, and on the impact of such policy on their lives. Based on a review of 42 papers published between 2010 and 2021, we outline the policies implemented in different receiving countries to enable resettled refugees to reunite with family, documenting at the same time the challenges refugees face in the process, as well as the impact of policy on their experiences. The findings evidence a tension between the refugees' own understanding of family and definitions of family in policy in receiving countries, which often results in family separation or reconfiguration. Additionally, high costs and other administrative barriers, as well as long waiting times associated with family reunification, lead to delayed or denied reunion, having detrimental effects on resettled refugees' well-being in the present and on their future prospects.

The past decade has seen renewed e orts to establish resettlement as a durable solution for refugees, both as a protection tool and a mechanism to equitably distribute them among countries. Although the right to a family life is widely recognised as a fundamental human right, whether refugees can arrive with their family or be reunited with family once resettled varies across receiving countries. Little is known about family reunion policies in countries leading the resettlement e orts, and about the impact of these policies on the lives and experiences of resettled refugees. This paper addresses this gap though a systematic scoping review of academic and policy literature on family reunion policies for resettlement refugees, and on the impact of such policy on their lives. Based on a review of papers published between and , we outline the policies implemented in di erent receiving countries to enable resettled refugees to reunite with family, documenting at the same time the challenges refugees face in the process, as well as the impact of policy on their experiences. The findings evidence a tension between the refugees' own understanding of family and definitions of family in policy in receiving countries, which often results in family separation or reconfiguration. Additionally, high costs and other administrative barriers, as well as long waiting times associated with family reunification, lead to delayed or denied reunion, having detrimental e ects on resettled refugees' well-being in the present and on their future prospects. KEYWORDS refugees, resettlement, family reunion, policy, forced displacement

. Introduction
The number of forcibly displaced people has doubled over the past decade, surpassing 80 million in 2020 (UNHCR, 2021a), while the return to power of the Taliban in Afghanistan in 2021 and the ongoing conflict in Ukraine have drastically increased the number of people seeking international protection. Some 86% of the worlds' refugee population reside in Global South countries (UNHCR, 2021a). Despite a tendency towards increasingly restrictive migration policies designed to keep most migrants, including refugees, out of Northern countries, resettlement continues to be an important tool for protecting refugees and a mechanism to ensure the equitable distribution of refugees among countries (Labman, 2007). Resettlement is the transfer of refugees to a country which has agreed to admit them and offer permanent residence (UNHCR, 2022b). UNHCR considers resettlement one of three durable solutions for forced migrants, alongside local integration and return, intended to enable them to live safely, peacefully and in dignity (Jubilut and de Lima Madureira, 2016).
While the UNHCR identified millions of refugees of concern in 2020, less than 1% of them were actually resettled (UNHCR, 2022b). In 2021, of the 1.4 million refugees cleared for resettlement only 34,044 were indeed resettled (UNHCR, 2021b). The COVID-19 pandemic has severely impacted the resettlement of refugees, with some countries reducing or temporarily halting their resettlement programmes. In 2021, 23 states accepted UNHCR resettlement submissions, down from 34 in 2017 (UNHCR, 2022a). Nonetheless, refugee resettlement is viewed by many, including politicians, policymakers, and the public, as the preferred solution for those who cannot access protection and refuge closer to home, many of whom might otherwise resort to lengthy, dangerous, and irregular journeys in search of safety. In addition to UNHCRreferred resettlement and special resettlement programmes to address specific humanitarian crises (such as the Syrian Vulnerable Persons Scheme in the UK), recent years have seen new approaches to resettlement, such as community sponsorship and humanitarian corridors introduced to supplement existing resettlement programmes. As middle-and lower-income countries accommodate the vast majority of displaced people, resettlement is mainly taken on by high-income countries (UNHCR, 2022a). The US and Canada usually accept the highest number of refugees annually, accounting together for almost 30,190 cases or 50% of resettlement in 2019 (UNHCR, 2022a) with the US recently returning to its position as the leading resettlement nation following a reduction in the number of refugees admitted during the Trump administration (UNHCR, 2022b). In recent years, Germany and Sweden have also resettled increasing numbers of refugees receiving about 5,000 each in 2021 (UNHCR, 2022a).
Forced migrants are faced with extreme and often lethal forms of violence, border violence notwithstanding, during their lengthy and dangerous border crossing journeys to safety. They are often separated from family members or form new family units along the way. The Universal Declaration of Human Rights recognises the right to a family life as fundamental (UN, 1948), andUNHCR (1983) stress the importance of family (re-)union for the wellbeing and integration of resettled refugees. Some countries prioritise the resettlement of families, while others offer refuge to individuals. Whether refugees can arrive with their family members or be reunited with them once resettled varies greatly depending on family reunion policies in countries of resettlement (UNHCR, 2015). Reunion policies determine who can be considered family, and outline the procedures, conditionalities and temporal frameworks of reunification. In this way, heteropatriarchal and hegemonic Norther European/North American white, middleclass, understandings of family are reinforced, resulting in what has been described as a form of intimate management of family life (Turner and Vera Espinoza, 2021). However, little is known about family reunion policy and how it impacts the lives and experiences of resettled refugees.
This paper addresses these gaps in knowledge though a systematic scoping review of academic and policy literature on the nature of family reunion policies for resettlement refugees, and on the impact of such policy on their lives. Based on 29 papers and 13 reports published between 2010 and 2021, we outline the policies put in place in different receiving countries to enable resettled refugees to reunite with family, documenting at the same time the challenges refugees face in the reunification process and the impact of policy on their experiences. The papers included focus on a wide-range of resettlement countries including Australia, Canada, the US, the UK, Iceland, New Zealand, Sweden, Finland and Germany with some papers covering multiple countries. The paper starts by outlining the methodology used for our systematic scoping review of the literature, following Arksey and O'Malley's (2005) method. The rest of the paper is organised along the three main strands that we identified: (a) the governance of family reunion: policies, practises and regulations in the countries leading resettlement efforts globally; (b) the challenges resettled refugee face as they try to reunite with family; and (c) the impact of family separation on the well-being of refugees from a gendered and intragenerational perspective. Finally, the paper discusses the gaps in knowledge that we have identified and outlines potential areas for further research.

. Methods
In order to map the literature on family reunion concerning contemporary refugee resettlement policy, we conducted a systematic scoping review using Arksey and O'Malley's (2005) methodology. We focused on literature published since 2010, a period that has seen: the doubling of the number of displaced people globally (UNHCR, 2021a); a renewed policy interest in resettlement as a response to the Syrian War and other crises (Hashimoto, 2018); and increased restrictions on migrant mobility. Originating in health sciences, "a scoping review or scoping study is a form of knowledge synthesis that addresses an exploratory research question aimed at mapping key concepts, types of evidence and gaps in research related to a defined area or field by systematically searching, selecting, and synthesising existing knowledge" (Colquhoun et al., 2014(Colquhoun et al., , p. 1293. In contrast to a systematic review, it emphasises the breadth and depth of a field but doesn't evaluate the quality of available studies (Arksey and O'Malley, 2005;Levac et al., 2010). We followed the five steps proposed by Arksey and O'Malley (2005): (1) identifying research questions; (2) identifying relevant studies; (3) study selection; (4) charting the data, e.g., grouping the data according to regions, focus areas and methods; (5) collating, summarising and reporting the results. As Levac et al. (2010) note, these five steps were not necessarily chronological but often overlap and proceed in parallel, especially during the initial stages of identifying and selecting studies for inclusion and exclusion, as the research focus evolves with the studies identified.

. . Stage : Research question
This scoping review on family reunion was originally conducted as part of a broader review of research literature and evidence on international refugee resettlement policy and integration of resettled refugees, conducted on behalf of a UK Government Ministry. The scoping review initially focused on refugee resettlement policies across the world and how they impact resettled refugees. Family reunion as a specific subtopic of interest was defined at the outset, given previous research experience and interests of the authors on the importance of family for refugee integration. The narrower research question was: What do we know about family reunion policy in refugee resettlement and what are .
/fhumd. . the impacts of family reunion policy for resettled refugees and those who arrive under family reunion?

. . Stage : Study identification
Studies were identified in two stages: first (November 2020-February 2021), we searched academic citation indexes (Web of Science, SCOPUS), using the keywords refugee * PLUS (sponsor * OR resettle * ) PLUS one of 36 other search terms, including resettlement, private sponsorship, family reunion/reunification and legal pathways. We identified papers that spoke to all aspects of refugee resettlement without limiting the search to family reunion because we assumed that family reunion would also be mentioned in literature on other areas of refugee resettlement. Papers were included/excluded after reviewing the title and abstract based on the following criteria: material had to address resettlement and/or community sponsorship, be published after 2010, and in English. As refugee resettlement is significantly shaped by policy, we also scanned selected stakeholder websites (e.g., UNHCR, government websites) and grey literature search engines (Social Care Initiative for Excellence, Open Grey and ProQuest) where grey literature is defined as any literature published outside of traditional commercial publication channels (Bonato, 2018) (see Table 1). This search yielded 4,043 results. 3,573 documents were excluded after reviewing the title and abstract, leaving 470 papers for reviewing.
Exclusion criteria at the second stage were: duplication of papers between resettlement and sponsorship; no specific focus on resettlement or sponsorship; broader focus on refugees and asylum seekers' integration experiences; and duplicate work from the same authors. A second search (December 2021) was specifically conducted on family reunion in order to update the original search, resulting in 12 additional papers. In total, 488 papers were identified.

. . Stage : Study selection/Screening
At the third stage, from these 488 we excluded papers when the full text wasn't available in major libraries or when a paper referred to a historical resettlement programme, to keep the focus on the state of the art. We divided literature according to type of resettlement programme (private sponsorship or resettlement programmes) and country. We added several papers traced back from the reference list of the papers reviewed but did not do forward searches. Overall, we identified 242 papers for in-depth reviewing. Finally, we excluded papers with no findings on family reunion. This resulted in a list of 42 papers.

. . Stage : Charting the data
In their description of a scoping review, Arksey and O'Malley (2005) suggest conducting an initial "charting" of key information to get an initial sense of the available data. While in their description there is overlap with the next stage (collating, summarising and reporting the results), we understood this step to aggregate  the available studies according to several important categories such as regional coverage, type of paper and methodology. This helped us get a sense of the breadth and content of available on family reunion. The largest number of non-excluded papers focused on Australia (10), with other single-country cased studies on Canada (9), the US (6), the UK (4), and one each from Iceland, New Zealand, Sweden, Finland and Germany. Eight papers were multicountry studies or had an international focus (see Table 2). Most papers were published after 2018, with very little research before then (see Figure 1).
About two thirds (29) of the papers were published in peer-reviewed journals, the rest (13) were reports and policy papers from NGOs, think tanks and international organisations. Most peer-reviewed articles were identified in interdisciplinary journals within the broad area of Social Sciences, including Migration Studies, Refugee Studies, Social Policy and Social Work. A smaller number of articles were published in Health Sciences and Psychiatry, Economics, Geography, Education and Law. Most journals published single papers relating to family reunion; exceptions were Refugee Survey Quarterly with 5 papers and Refuge with 4 papers, respectively. All papers focused on resettlement refugees with some comparing their experiences to those of Convention refugees and others looking specifically at Community Sponsorship programmes.
In terms of methods, 16 of the 42 papers used qualitative methods, 6 quantitative methods and 4 used a multi-method approach. The remaining 16 papers were based on desk research and reviews of policies. Five of the non-peer reviewed publications and reports also conducted primary data collection, mostly qualitative interviews. Taken together, 36 papers were based on primary data collection and analysis (see Figure 2).
. . Stage : Collating, summarising, and reporting the results Following the papers' overview, we divided them amongst the co-authors and summarised each paper. We identified themes and gaps in dialogue with previously identified literature on family separation and the importance of family for refugees in general, focusing on the effects of family reunion policy. We discussed emerging themes in repeated meetings, identifying and refining them into three broad strands. The first is concerned with the governance of family reunion, the policies, practices and regulations put in place in the different countries. The second strand of research examines the main challenges faced by resettled refugees in their effort to reunite with their family members. Finally, the third strand focuses on the importance of family reunion and the impact that family separation has on the mental health, overall well-being and future prospects of resettled refugees.

. . Limitations
The time limit (publication after 2010) may have excluded earlier publications that have shaped contemporary research. Earlier influential papers that we identified through back-tracing and research on the theoretical background were included in the background section. In terms of content, we focused on the relationship between family reunion policy and refugee resettlement only; there may be more research published on the effects of family reunion (or lack thereof) for refugees and asylum seekers who have arrived independently. While some countries handle family reunion for resettled refugees differently to other refugees (e.g., Germany), many countries do not distinguish between different categories of refugees. There may also be overlap with family reunion policy for migrants in general.
. The governance of family reunion: Policies, practices and regulations . . Background Family is recognised as society's predominant social unit by Article 16 of the Universal Declaration of Human Rights and Article 23 of the International Covenant on Civil and Political Rights (UNHCR, 1983). Family reunification/reunion is usually understood within the context of international migration, and in particular, labour migration. The International Organisation for Migration (IOM) defines family reunification as "[t]he right of non-nationals to enter into and reside in a country where their family members reside lawfully or of which they have the nationality in order to preserve the family unit" (IOM, 2019, p. 72). In post-WWII Europe, family migration, often assumed to be a dependent female spouse and minor children reuniting with a male labour migrant, was generally permitted. This practice was both rooted in humanitarian principles and seen as suiting demographic and economic interests (Lahav, 1997). This changed during the economic slowdown of the 1980s, because of concerns over "chain migration" via family reunion (John, 2003). As the demand for labour dropped and the number of displaced people increased, family reunion came to be seen as problematic, in particular in Northern Europe. As a result, "immigration regulations have sought to contain [family migrations], their geographical reach and structures; they define the composition of the family and restrict its flexibility, frequently reinforce gender inequalities and truncate the cohabitation of generations" (Kofman et al., 2011, p. 13). Family reunification policy and restricting reunion rights (Lahav, 1997) reflect the tension between respecting refugees' right to family life and the states' aspirations to control migration (John, 2003). Internationally, family reunification relates to "two sets of legal principles: those relative to freedom of movement and those linked to the family as a unit of society" (Lahav, 1997, p. 355). Even though legal and policy instruments are more likely to protect the family rights of citizens than of migrants, there are examples where migrants' right to family reunification is recognised, although not guaranteed, and is only limited to specific groups, e.g., labour migrants (Lahav, 1997). Other legal provisions do not distinguish citizens from migrants yet shape opportunities for reunion, such as the Convention on the Rights of the Child, where it is stated that children cannot be separated from their parents. However, "they remain limited regarding to whom their provisions apply, the ambiguity of their text, and their ultimate deferment of implementation to the state" (Lahav, 1997, p. 355).

. . Policies and regulations
The 1951 Refugee Convention does not mention family reunification. It is only mentioned in the Final Act of the United Nations Conference of Plenipotentiaries on the Status of Refugees and Stateless Persons, stating that member-states "take the necessary measures for the protection of the refugee's family, especially with a view to [. . . ] ensuring that the unity of the refugee's family is maintained particularly in cases where the head of the family has fulfilled the necessary conditions for admission to a particular country" (UN, 1951; as cited in UNHCR, 2015, p. 1). Who is defined as "the head of the family" is not set out, which is important given that the designated individual may have priority in future applications for family reunion (Wilmsen, 2011). Furthermore, not all individuals who are resettled have full refugee status. People with subsidiary protection or asylum seekers often have different rights to family reunion than recognised refugees.
In the European Union, the right to reunification is often linked to Article 8 of the European Convention on Human Rights, recognising the right to "private and family life, his [sic] home and his correspondence", and to the best interest of the child (European . /fhumd. . Court of Human Rights, 2021, p. 11). However, this article is frequently used to justify the non-expulsion of migrants, rather than to justify the right to reunion (John, 2003). Twenty-five EU countries have adopted the Family Reunification Directive (Council of the European Union, 2003), which includes refugees, but with restricted criteria compared to UNHCR's. Under this directive, refugees have 3 months to apply for reunion after gaining status, but there is considerable flexibility, with many member-states extending or removing the time limit, while thirteen memberstates are currently implementing changes for beneficiaries of international protection (EASO, 2021). Others have extended the definition of who qualifies as family-which has generally focused on immediate, dependent, family members, including spouses and minor children-to potentially include other dependents who are not able to provide for themselves (UNHCR, 2015).
In the last decade, many member-states have developed community/sponsorship programmes offering displaced people an alternative route to access refuge (Fratzke et al., 2019). In some cases, such schemes represent the only possible option for refugees to be reunited with their family members. Family reunificationbased sponsorship schemes developed in Ireland, Germany and France have, for instance, allowed for the resettlement of 27,000 individuals in 2 years. However, these schemes impose a heavy burden on family member sponsors, including financial responsibilities (ICMC Europe and Caritas, 2019). In Canada, where private sponsorship was first introduced during the 1970s, the possibility for sponsors to name the person to resettle has turned the scheme into a de facto family reunification programme, raising criticism that it is no longer used for helping the most vulnerable refugees, as initially intended (Labman and Cameron, . /fhumd. . 2020). Arguably, such complementary pathways to safety ought to increase available solutions for refugees, not substitute family reunion. Additionally, having the possibility to reunite with family potentially reduces secondary migration, since one of the main reasons why refugees move from where they are initially resettled is joining family (Lumley-Sapanski, 2020). In Europe, limited family reunification procedures result in minors, who could meet the legal requirement for family reunion, turning to smugglers and undertaking unaccompanied irregular and dangerous routes, with many "going missing" en route (Allsopp, 2017).
. Main challenges in family reunion: Conditionalities, temporalities and definitions

. . Bureaucratic and financial conditionalities
Research has shown that the conditionalities attached to the right to family reunion restrict possibilities for resettled refugees, impacting on their life and future prospects. In many resettlement countries, refugees face high administrative burdens to be eligible for family reunion, such as the requirement to have permanent residence or even citizenship, in some cases. Reunion in Australia requires sponsorship from an Australian citizen or permanent resident, an approved organisation or an eligible New Zealand citizen (Okhovat et al., 2017). Some German Federal States implement their own "humanitarian admission" (read: resettlement) programmes; in most of these, the applicant must have either German or Syrian citizenship (Tometten, 2018). In the US, refugees must be at least eighteen in order to apply for reunion (Carrera and Cortinovis, 2019). Brunner et al. (2014) highlight the case of male Acehnese refugees resettled in Canada who wished to marry and reunite with their partners in Aceh. Because they feared persecution instead of obtaining travel documents from the Indonesian embassy, they had to wait until they qualified for citizenship before they could move on with their plans for a family life. With citizenship dependent on demonstrating language competency in countries such as Canada, Australia and the UK, refugees face lengthy waits before they can be eligible for family reunion via that route (Brunner et al., 2014;RCOA, 2019).
Additionally, in many resettlement countries there are financial conditions for "host" refugees, including minimum income requirements (Innes and Steele, 2015), and high visa, travelling and long term costs (Refugee Council of Australia, 2021), excluding many from being able to apply for reunion (Brunner et al., 2014;Innes and Steele, 2015). In Australia, while there is a family stream to the overarching migration programme, this is of little use to refugees given prohibitive costs, especially for women (Lumbus et al., 2021). Associated costs for visa, travelling and longer-term caring may impoverish applicants or even lead to destitution. In Australia sponsoring a partner costs at least Aus$6,865, while the sponsor is required to provide a bond as an Assurance of Support (Aus$10,000 for the main applicant and Aus$4,000 for any additional adult applicant) which will be withheld for 10 years (Refugee Council of Australia, 2021). Okhovat et al. (2017, p. 275) refer to the case of a refugee with psychological and physical disabilities who had to pay $22,000 to bring his wife and children to Australia. Under a Community Proposal Pilot (CPP), which allows individuals and community groups to apply to sponsor a person for resettlement, proposing family members needed to spend around Aus$40,000 to bring just one relative to Australia, with a further Aus$5,000 for each additional relative (Okhovat et al., 2017).
Another financial barrier is the competing priorities faced by resettled refugees, such as sending remittances to family abroad. In Canada, family reunification and remittances appear to be at odds with each other, as refugees face difficulties earning enough to pursue both (Johnson and Stoll, 2013), while those arriving through the Canadian resettlement programme are also subject to travel loan repayments (Brunner et al., 2014). Since 2016, Finland has applied income-testing to refugees with subsidiary protection who wish to reunite with family (Tervola, 2020). Germany has income and time requirements for family reunification (Tometten, 2018), while in the UK, families go to extreme lengths to reunite with loved ones, often becoming destitute as a result of covering reunification costs (Beaton et al., 2018). Rung (2015) describes the placing of reunion's costs on resettled refugees as a shift from state to private responsibilities.
Finally, while most resettlement refugees have access to welfare and integration support once resettled, the same does not necessarily apply to individuals arriving under family reunion. In Germany, the individual sponsoring family reunion has to take full financial responsibility for the family member, entering a legally binding contract covering their accommodation, livelihood and even deportation costs if their stay is terminated for security reasons, although some Federal States will cover medical costs (Tometten, 2018). In Canada, the sponsoring party must cover all transport, housing and subsistence costs after arrival (Bevelander and Pendakur, 2014), while family reunion refugees do not have access to integration programmes except for language classes, potentially accounting for their worse financial outcomes compared to resettled refugees, a particular problem for women (Bevelander and Pendakur, 2014). In Sweden, where all refugees, regardless of status and arrival path, have access to approximately the same assistance (eighteen months of settlement training, language and labour market training, credential recognition/assessment), family reunion refugees do better than resettled refugees (Bevelander and Pendakur, 2014) indicating that policy does influence outcomes.

. . Complicating temporalities
Time is a key factor shaping refugees' ability to reunite with family. Firstly, there are time restrictions to the right of refugees to apply for family reunion, while, once an application is filed, the time spent waiting for a decision can be lengthy. In some countries, resettled refugees need to apply for reunion within a set period after arrival. This limited window, combined with the often-unknown location of relatives dispersed across different countries, means that refugees may fail to apply in time. In Canada, Johnson and Stoll (2013, p. 61) found that Sudanese refugees could, in theory, have accessed reunion without meeting financial requirements by sponsoring non-accompanying family members in the "one year .
/fhumd. . window of opportunity" policy, but many lacked this knowledge. Family reunion is enabled (and prioritised) in the US resettlement system, although not for everyone. The window of opportunity is wider than in Canada, allowing refugees to apply for family reunion within 5 years of being granted asylum or admitted into the US (Carrera and Cortinovis, 2019). In Germany, conversely, refugees (resettled or having arrived independently) are subject to lengthy waiting periods, as they need to gain refugee status before being eligible to apply for family reunion (Tometten, 2018). Once status is granted, there is a three-month window in which they can apply for family members to join them without being subjected to minimum income requirements. For Syrians in particular, Federal States run separate humanitarian programmes, with specific and varying deadlines and requirements (Tometten, 2018), generating confusion around timings and eligibility. In Australia the process can take up to 20 months for a partner visa, 8 months for a dependent child, and three and a half years for an orphaned relative (Refugee Council of Australia, 2021). Finally, in Canada, research has shown that not all families are treated equally (Coker, 2017): the reunification process is slower for refugees than for other migrants, and slower for refugees from Africa, than those from the rest of the world (Kaduuli, 2020). In Canada in particular, the combination of high cost and lengthy waiting times associated with family reunion has had a displacement effect. Labman and Pearlman (2018, 135) argue that the Private Sponsorship Programme has become a de facto family reunion programme as sponsors, often community or religious groups, are approached by sponsored refugees to facilitate family reunion under the "naming" initiative. Morris et al. (2021) also note that private sponsorship is used as a way around the 1 year window of opportunity. They highlight a case in which a group initially sponsored a Syrian family of six, and then managed, through eight more sponsorships, to bring over thirty-five family members. They argue that this sponsorship approach disrupts power relations in the decision-making about who constitutes family and is worthy of sponsorship. Others express concern that the scheme "is being used for family reunification and not protection of the world's most vulnerable refugees as intended" (Ilcan and Connoy, 2021, p. 135). To address this problem, Morris et al. (2021) make the case for Canada to open a special family reunification programme with quotas additional to sponsorship citing the example of such a programme in New Zealand.

. . Definitions of family, dependency, and vulnerability
One consistent theme across several papers was that family reunion (re)configures refugee families through eligibility and vulnerability criteria, and through definitions, shaping refugees' own strategies to navigate these processes. Most countries favour narrow definitions of family when setting out criteria around who is eligible for reunion. While UNCHR in principle promotes a wider definition beyond the "nuclear family", family reunion in most resettlement programmes is limited to spouses and children under eighteen (Morris et al., 2021). In the UK, adult refugees can reunite with (married or civil) partners and underaged children, but not with parents, grandparents, siblings and adult children (Beaton et al., 2018). In Finland, only in special cases can relatives other than children, spouse or parents of children under eighteen be regarded as family members (Tervola, 2020). Family reunification is possible for eligible resettled refugees in Germany, for "core family" (spouses, minor children and parents of minors), and only in highly exceptional circumstances, parents of adults in need of specific personal care, or other extended family members (Tometten, 2018). In the US, there are two pathways to family reunion: (a) the Affidavit of Relationship Programme, available to relatives already recognised as refugees outside the US, including spouses, unmarried children under the age of twenty-one, and parents; and (b) the Family Reunification Programme, available to the spouse or children of a "principal" refugee up until 2 years after resettlement (Bruno, 2015).
Beyond the nuclear family, narrow criteria of "dependency", mostly the age until which a child is considered dependentgenerally 18 years old (unless married)-also reconfigure refugees' family composition (UNHCR, 2011). Such criteria reflect white middle-class Northern European/North American social norms of financial, emotional and physical dependence and kinship (Wilmsen, 2011;Ilcan and Connoy, 2021) and vary significantly across socio-economic class, culture and personal circumstances. In Australia, family members are limited to immediate family (spouse, child or parents if the applicant is underaged) with significant pressure placed on refugee families to evidence "dependency" in their family reunion applications (Okhovat et al., 2017) when applying for non-eligible relatives. A separate case can be made for family reunion under the family stream of Australia's Migration Programme to sponsor aged parents, children, partners, remaining relatives, carers, or aged dependent relatives, but as noted above this is very costly (Wilmsen, 2011). Lumbus et al. (2021) argue that the Australian reunification policy, with its focus on the conventional nuclear family, is based on heteronormative values naturalised in Australia's colonial history. UNHCR rejects polygamy and child marriage "enforcing the boundaries of socially acceptable or desirable family constellations" (Welfens and Bonjour, 2021, p. 227). In Canada, the definition of family, restricted to spouses, minor children and parents, does not reflect the refugees' reliance on extended family networks such as siblings, cousins, and nieces once resettled (Johnson and Stoll, 2013). Moreover, Wilmsen (2013) highlights the ways in which conceptions of family can shift according to context or experience, for example by refugees recognising the importance of non-biological children adopted during conflict or displacement. In some cases, resettled refugees have pushed back against these restrictive definitions of family. For example, Yazidi refugees, many of whom had no nuclear family left as a result of the atrocities and deadly persecution committed by the Islamic State, have mobilised to expand the family reunification criteria (Ilcan and Connoy, 2021, p. 134).
Research has also shown that narrow definitions of family can push refugees to risky strategies seeking to negotiate their more complex family ties. Wilmsen (2011) highlights cases of applicants leaving family members off their application, believing that resettling a large family with several children would decrease their chances of admission. Such tactics lead to long-term or even permanent separation, as family members can rarely be recognised retroactively. Requirements for access to healthcare can also end .
/fhumd. . up separating families who end up being admitted to different resettlement countries if a country is willing to admit refugees in need of particular treatments, but not their family (Wilmsen, 2011). In the US, reunion applications from certain African countries were suspended "due to indications of extremely high rates of fraud in claimed family relationships identified through pilot DNA testing" (Bruno, 2018, p. 7). The programme was eventually resumed but with an added requirement for DNA evidence to prove parental relationships (Carrera and Cortinovis, 2019), reinforcing the biological conception of family. There is also evidence to suggest that resettled refugees whose application for family reunion is ineligible or unsuccessful may put themselves or their families at risk by returning to visit family in conflict zones, or paying smugglers to bring over their family (Beaton et al., 2018). Finally, family reunion is highly gendered and heteronormative, often reinforcing existing harmful stereotypes (Welfens and Bonjour, 2021). For example, UNHCR's (2011) preference for resettling families or vulnerable single women with children over single men, by default reproduces the figure of the single men as "risky" and women as vulnerable. The heteropatriarchal family bias extends to integration policy, since being part of a family is considered to increase the capability of family members to fit in with socio-cultural expectations in receiving countries and look after each other (Welfens and Bonjour, 2021). Often, family reunion policy is implicitly structured around the idea of a lead applicant, usually a male breadwinner, with a dependent (female) spouse, arriving as homemakers, inevitably reproducing heteronormative family relations (Turner and Vera Espinoza, 2021). Overall, conceptions of family are not neutral but highly political, and are shaped to a large extent by the policies determining who has the right to (re)unite with family.
. The impact of family separation on refugees' well-being . . Background Family reunification has been argued to be particularly important for refugee welfare. In 1999 UNHCR highlighted a range of family protection issues, including the importance of family for refugee's emotional, social and economic well-being, and the need to prevent separation (EXCOM, 1999). According to Schweitzer et al. (2006) separation from family members is the most commonly mentioned traumatic event experienced by refugees, with 85% of those questioned experiencing such trauma (see also Rees et al., 2013;Tay et al., 2015;Morris et al., 2021). Recognising the constraints associated with many states' definitions of family, UNHCR (2015, p. 1) has recommended the adoption of wider definitions, which should consider the "strong and continuous social, emotional or economic dependency between family members" and should extend to members of the wider family. Allowing broad family reunification pathways could limit dangerous journeys and reduce spontaneous asylum seeking, a key aim of many states (Havinga and Böcker, 1999;UNHCR, 2015).
In turn, family separation has considerable implications for the refugees' wellbeing and integration prospects. Focusing on refugee survivors of torture and/or violence, Gorst-Unsworth and Goldenberg (1998) found that the lack of affective social support in exile was more important in determining depressive morbidity and PTSD than trauma itself (see also Hauff and Vaglum, 1995;VFST, 1998;Rousseau et al., 2001). Family acts as an anchor for both emotion and identity in exile, where refugees often feel estranged and alienated, with the presence of family having a protective effect on wellbeing (Gonsalves, 1990;Rousseau et al., 2001). . . Impact on well-being in the present and future prospects Family reunion is critical for the mental health and overall well-being of refugees. Morris et al. (2021) found that resettled refugees had an increased desire to belong and more commitment to life in the new country after family reunification, while those separated from family experienced social isolation and distress (Nickerson et al., 2010;Miller et al., 2018;Lumbus et al., 2021). Focusing on West Papuan refugees in Australia, Rees et al. (2013) found that separation from family was the greatest post-migration stressor they faced. In some cases, refugees reported poor health, such as headaches and stomach-aches, which they associated with separation (Nickerson et al., 2010;Wilmsen, 2013;Liddell et al., 2021). Ilcan and Connoy (2021) pointed to other unexpected findings such as delaying surgery due to lack of family members to help with childcare. The precarious and even dangerous conditions faced by family members left back home is another stressor, accentuated by smartphones and other information and communication technologies (Savic et al., 2013;Beaton et al., 2018). Miller et al. (2018) report that many refugees suffer from "ambiguous loss", a relational disorder described by Boss (2010), in which a loved one is either physically present but emotionally absent, or the opposite. As a result, refugees separated from family are plagued by worry and guilt, struggling to focus on their present (Savic et al., 2013;Choummanivong et al., 2014).
Studies from New Zealand (Choummanivong et al., 2014) and Australia (Wilmsen, 2013, p. 241) found that the impossibility to reunite with family was the biggest barrier to successful resettlement and had detrimental effects on refugees' capacity to plan their future. These effects and the resulting vulnerability to mental health disorders extended into the long-term, worsening over time (Miller et al., 2018). Feeling stressed, guilty and powerless when failing to secure family reunion or send remittances can result in harmful behaviours, such as excessive alcohol consumption (Savic et al., 2013). Additionally, such financial responsibilities to family members abroad and the associated stress prevent refugees from focusing on the present and investing in their future. Morris et al. (2021) report that private sponsors in Canada raised concerns that the remittances sent by sponsored refugees to family members with whom they are unable to reunite, undermined the refugees' quality of life. Some sponsors tried to help with these remittance costs, while others discouraged remittances, urging refugees to focus on their lives in Canada. On the whole, concerns have been raised that the high cost of family reunion push refugees into short-term precarious work and away from training, which would ameliorate their future prospects (Brunner et al., 2014). Anjum et al. (2012) found that West African refugees in Sweden .
felt responsible for supporting wider family remaining in poor conditions in Africa. They felt guilt at separation and were unable to concentrate on learning. The longer they were in Sweden, the more pronounced concerns became. Gardener and Costello (2019) argue that separation takes a toll on the well-being of individuals, families and the wider community. This claim is also supported by Choummanivong et al. (2014) in New Zealand, where refugees reported separation impacted on their ability to participate in local community activities.

. . Gendered and generational impact
There is clear evidence from analysis of administrative datasets in Canada and Sweden that men arriving on family reunion visas find it easier to access and maintain employment, than those without family (Bevelander, 2011;Bevelander and Pendakur, 2014;Tervola, 2020). Research has shown that this is due to the uncertainty many face of whether they will be able to remain in the country if family reunion fails (Beaton et al., 2018;Morris et al., 2021). For example, in the UK and Australia, evidence suggests that refugees are unable to settle and concentrate on language learning and education more broadly, as well as on finding work while separated from family members (Wilmsen, 2013;Beaton et al., 2018;Gardener and Costello, 2019). Family separation also leads to pressures to take on unfamiliar and challenging family roles in the absence of wider family, such as young Sudanese refugees in Australia feeling pressured to adopt adult roles (Savic et al., 2013). Similarly, research on refugees in the UK separated from family showed that undertaking caring responsibilities for a family member prevented them from working or studying (Wilmsen, 2011;Beaton et al., 2018).
In turn, qualitative research with resettled women refugees in Australia (Lumbus et al., 2021) and New Zealand (Choummanivong et al., 2014) found that the presence of family led them to feel supported in their everyday life. However, family reunion can also reinforce gendered dependencies, as it often assumes a male breadwinner with a female homemaker spouse (Turner and Vera Espinoza, 2021). Indicatively, research has shown that in Finland women resettled through family reunion tend to have lower levels of labour market engagement than refugee women arriving independently through resettlement routes, suggesting that women are confined to roles associated with their "dependent" status (Tervola, 2020). Widespread concerns have been expressed about enforcing dependency on women joining their family via spousal visas and policies that restrict their access to welfare support and hinge residence on maintaining the relationship for a probationary period of several years (Innes and Steele, 2015). Policies which enforce dependency can leave women refugees at risk of abuse, but more evidence is needed to establish how common this is. Reuniting after many years of separation is in itself stressful, requiring effort to re-establish continuity and common ground, while it can also kindle intergenerational and culture clashes (Choummanivong et al., 2014). However, little is known about the kind of family reunion policy that is effective in supporting integration processes for refugees and their families.
. Discussion: The state of knowledge and an agenda for future research Our scoping review revealed 42 papers with relevant findings on family reunification policy for resettled refugees published since 2010 (in English), a time during which both resettlement and reunion have been of increased interest to policymakers. We identified a growth in the literature since 2018 and relatively little activity before that, suggesting that resettlement refugee family reunion as a topic is beginning to attract more scholarly attention. The literature we identified was a mix of articles in peer-reviewed journals and reports from NGOs, think-tanks, and international organisations concerned with forced migration. While peer-reviewed articles represented most papers reviewed, we find that some of the most useful insights come from NGO/policy reports, for example identifying which policies have which effects on resettled refugees. These are, however, less rigorous in terms of descriptions of methodology and theoretical background and may be biassed towards highlighting problems and minimising successful or less harmful policies. This underlines the need for greater academic attention and high-quality evidence on the topic.
Most of the literature originates from just a handful of country contexts, chiefly Australia and Canada, followed by the US and the UK, as these are the countries that lead resettlement efforts globally. We could not find any research on how family reunion is implemented in emerging resettlement countries in Latin America or East Asia. This restricted regional scope limits both the diversity of types of family reunion policies and the diversity of refugee groups studied. Furthermore, to our knowledge, so far research has focused entirely on countries of resettlement, with no research conducted specifically on families waiting for reunion with their resettled relatives. The phenomenon of family reunion (and separation) would also lend itself well to multi-sited, transnational research, looking at families in multiple countries.
In terms of methods, qualitative research was most prevalent, with quantitative or mixed-methods studies being much rarer; a third of the relevant papers found were desk research or larger literature reviews. Overall, most published research on family reunion policy in refugee resettlement is based on small, exploratory and single-case qualitative studies. There is limited systematic quantitative analysis of processes or outcomes of family reunion, which confines our understanding of overall dynamics and wider patterns. We note a lack of comparative research, which would shed light on different patterns across countries and policies. The reasons why certain countries adopt particular family reunion policies are not currently clear, nor are the ways in which different policies lead to different outcomes. More research in this area would help identify what a (more) successful family reunion policy could look like, a focus currently entirely lacking from the literature. Similarly, there is no longitudinal research at country or family level wherein families can be traced from separation to reunion and beyond, which would allow for examining and comparing the effects of various policies and the causal effects of family separation and reunion. More broadly, there is a dire need for more research into how and why certain types of family reunion policy is passed into state legislation, and how it is implemented by resettlement agencies, in order to identify more clearly which .
/fhumd. . factors contribute or hinder successful reunification. Much of the literature reviewed also tends to be more descriptive than analytical, meaning there is little conceptual advancement on the different forms of policies or processes of family reunion, further highlighting the need for more in-depth scholarly attention on the topic. In terms of content, reviewing the existing literature, we identified some key themes. On the one hand, the literature outlines several factors shaping family separation/reunion, including: policy definitions of family; bureaucratic barriers to family reunion; and temporal factors such as windows of opportunity and waiting. Another strand of literature focuses on the impact of family reunion policy, in particular the effects of family separation on refugees' well-being in the present and their future prospects.
Some of the key insights concerning these themes include the fact that narrow or rigid policy definitions of family are in tension with refugees' own understanding of who constitutes their family, ultimately resulting in family separation and/or family reconfiguration. Furthermore, the governance of refugee resettlement, including bureaucratic, financial and temporal requirements and the cost of reunification procedures, substantially shape the refugees' experiences of family reunion. These regulations and costs act as obstacles that, again, ultimately lead to delayed or denied family reunification. Delayed or denied family reunification has several detrimental effects for resettled refugees: it affects their psychological well-being as a major post-migration stressor and physical health, often resulting in sadness, stress, worry, and guilt concerning family members left behind. It also seems to undermine integration into labour markets or education systems, for both emotional and practical reasons.
In sum, there is thus considerable, though often small-scale, evidence concerning the negative effects of family separation. While refugees are separated from their families when fleeing conflict and violence, policies that restrict comprehensive family reunion, high financial and administrative barriers and long and inefficient bureaucratic procedures compound, extend and exacerbate family separation and its harmful effects. Future research should consider different effects of separation on various groups of the heterogeneous group of "resettled refugees", such as refugees from different countries of origin and countries of first settlement, children reuniting with their parents, sibling reunification, women of various ages reuniting with partners, unmarried couples or non-heteronormative and LGBTQIA+ refugees. More research is also needed into the dynamics and processes of these bureaucratic processes, including how bureaucrats and policymakers view and implement policies and practises of family reunification, and how refugees cope, navigate or resist these bureaucratic systems. There is a huge gap in literature on how being reunited affects refugees' well-being and integration in the short and long term in terms of effects on employment, childcare arrangements, gender relations, mental, emotional, and physical health, etc. On the other hand, there is also very little evidence on the impact of reunion on the reuniting individuals, or the impact of separation on family members who are not eligible for reunification under the very narrow policy definitions of family. More research is needed to study the intersectional impact of separation/reunification on different families based on gender, age and occupation of main applicant, size and composition of family, country of origin or country of first asylum. Building on some interesting but so far small-scale findings, the gender dimensions of family reunion should be examined more closely, focusing on its impact on gender dynamics and roles, and on patterns of who can and does apply for family reunion and who is more likely to be approved. There is evidence to suggest that the desire to be reunited with family affects decision-making regarding migration options and asylum destination countries, while refugees may undertake significant risks returning to visit or bringing over family through irregular routes. However, more large-scale evidence is needed to examine the prevalence of these phenomena.

Data availability statement
The original contributions presented in the study are included in the article/supplementary material, further inquiries can be directed to the corresponding author.