Varieties of citizenship in regional organisations: A cross-regional comparison of rights, access, and belonging

This paper explores the variation of citizenship in eight regional organisations (ROs). Since the 1980s, ROs have increasingly served as spaces for developing, regulating, and providing citizenship. However, current literature primarily takes a rights-based approach and focuses on a narrow set of cases without providing an account for the variation of citizenship in ROs. This paper offers a broad, conceptual approach to the study of regional citizenship and deploys a three-tiered conceptual framework consisting of rights, access, and belonging, to analyse how citizenship varies across different ROs. It challenges the current theorisation of regional citizenship, which is primarily rooted in the study of the EU’s rights-based approach. The analysis contributes to citizenship studies and comparative regionalism. It shows how citizenship varies across ROs, thus providing the first comprehensive cross-regional comparison. The empirical findings lead to the following insights. First, citizenship in regional organisations can be conceptualised as constitutional or practice-based. Second, there are different pathways to regional citizenship where practices might precede law or where citizenship in ROs remains a practice-based concept. Third, there is variation in the link between national and regional citizenship and how ROs provide access to regional citizenship.


Introduction
How does citizenship in regional organisations vary? This paper explores the variation of citizenship in ROs from a cross-regional comparative perspective. The concept of citizenship was reassessed during the 1980s and 1990s, when ROs began establishing citizenship policies in their institutional frameworks (Meehan, 1993;Obi, 2012). This development led to an increased focus on the new role of ROs as spaces for the development, regulation, and provision of citizenship beyond national boundaries (Wiener, 1998;Hanagan and Tilly, 1999). In particular, scholars focused on the case of European Union citizenship, which arguably remains the most politically advanced type of regional citizenship (Strumia, 2017). Scholars have examined the changing structure and conceptualisation of citizenship in Europe (Soysal, 1994), European citizenship practices (Wiener, 1998), the institutional processes leading to the official establishment of European citizenship (Meehan, 1993), and practices of the EU citizenship regime (Jenson, 2007). In this literature, the scholars predominantly take a rights-based approach to regional citizenship and focus on the legal aspect of this extended membership status (Jenson, 2007;Strumia, 2017).
Beyond the EU, scholars have primarily focused on regional citizenship formation in Latin American organisations (Giupponi, 2017). However, the literature on citizenship in Mercosur, the Andean Community (CAN), and the Caribbean Community (CARICOM) predominantly examines the implications of cross-border migration on regional citizenship policies (Cernadas, 2013) or focuses on the extension of certain rights in order to protect labour migrants (Fornalé, 2017). In other regional contexts, scholars have been investigating indicators for the development of regional citizenship, such as the establishment of regional residence and mobility policies (Okom and Dada, 2012), or focused on the promotion of regional identity and legally non-binding policies as building blocks in the making of regional citizenship regimes (Cabrera and Byrne, 2021;Weinrich, 2020;Neuvonen, 2019). Such research hints at a variation across ROs but does not detail disparities or similarities.
Although most research on citizenship in ROs remains EU-centric, scholars of comparative regionalism emphasise that the EU is one case among many (Acharya, 2016;Van Langenhove, 2011). Amitav Acharya (2012) reminds us that regions other than Europe integrate, but not necessarily along the European pathway. Instead, comparative regionalism literature comprises contributions from many regions, including Latin America, the Middle East, Asia, Africa, and Europe, leading to different approaches to regionalism (Suzuki, 2021). Although regional institution-building, region-building, and regional citizenship are intertwined phenomena and processes (Wiener, 1998;Van Langenhove, 2011), comparative regionalism commonly focuses on regional integration (Murray, 2010;Paasi, 2009;Börzel and Risse, 2019), identity-building (Neuvonen, 2019;Acharya, 2012), migration policies and free movement protocols (Geddes et al., 2019), cross-border security (Suzuki, 2021), or civil society participation (Fioramonti, 2014;Junior and Luciano, 2021). Thus, scholars of comparative regionalism have refrained from engaging directly with the concept of citizenship.
The paper seeks to engage with two significant gaps in citizenship studies and the comparative regionalism literature. First, there is a prevalent focus on studying citizenship in ROs by conducting in-depth single case studies, resulting in a limited understanding of citizenship variation across ROs. Moreover, these studies take an EU-centric approach (Yukawa, 2018). Second, the research conducted on citizenship in ROs currently focuses on citizenship framed as a legal status. This means that research has focused on examining cases wherein citizenship is legally defined within the organisations' institutional framework and have abstained from including non-legal, practicebased citizenship regimes in the empirical analysis. The lack of including practice-based citizenship in the broader debate is evident despite that Élise Auvachez (2009) showed in her study on UN citizenship-building that citizenship can be practised without being legally defined. In addition, the work by Elizabeth Meehan (1993), Antje Wiener (1998), and Espen D. H. Olsen (2012) indicates that European citizenship existed through citizenship-related policies, such as visa policy, migration policy or social policy, before the concept was legally defined in the Maastricht Treaty. Thus, the research gaps merit further empirical and comparative exploration.
The paper challenges current research on citizenship in ROs, which almost exclusively focuses on regional citizenship as a legal status and does not account for the variation of citizenship across ROs. It argues that regional citizenship can be constitutional or practice-based and that legally defining citizenship is not a requirement for citizenship in ROs. The paper focuses on examining the variation of citizenship in the Association of Southeast Asian Nations (ASEAN), the Andean Community (CAN), the Caribbean Community (CARICOM), the East African Community (EAC), the Economic Community of West African States (ECOWAS), the European Union (EU), the Gulf Cooperation Council (GCC), and the Common Market of the South (Mercosur). These organisations have either legally defined a regional citizenship regime or are practising regional citizenship policies, thus establishing diverse regional citizenships.
The paper offers a broad, conceptual approach to deepen the understanding of how citizenship in ROs varies. Thus, explaining this variation lay beyond the scope of the paper. In the comparative analysis, citizenship is defined broadly as membership in a political community (Bauböck, 2017), and the conceptual framework used for the comparison constitutes the three constitutive dimensions of citizenship: rights, access, and belonging (Wiener, 1998;Auvachez, 2009). The paper draws primarily on the method of comparative document analysis and, to some extent, on primary data from semi-structured interviews conducted at the ECOWAS Commission in Abuja, Nigeria and the ASEAN Secretariat in Jakarta, Indonesia, between November 2018 and December 2019.
The paper contributes new knowledge to the field of comparative regionalism and citizenship studies. In particular, it shows how citizenship varies across ROs, thus providing the first comprehensive cross-regional comparison of citizenship in ROs. The empirical findings offer new empirical data that challenges the current theorisation of regional citizenship, which is primarily rooted in the study of the EU's rights-based approach (Wiener, 1998;Jenson, 2007). Moreover, it shows how ROs perceive and emphasise citizenship policies and how these institutional designs vary across regions, thus making an important contribution to comparative regionalism. The conceptual framework helps unpack the variation of citizenship in ROs and facilitates new findings that show regional citizenship as a practice-based concept, defined through different citizenship-related policies. This type shows similarities with the pre-Maastricht Treaty EU citizenship regime. However, the paper provides a novel contribution by showing how ROs conceptualise citizenship as a practice-based, non-legal status defined through citizenship-related policies. Therefore, the main contribution of the comparative analysis is drawing attention to the conceptual variation in this concept.
The paper proceeds in the following four steps. First, it provides a brief overview of theoretical debates in citizenship studies and outlines the current literature on citizenship in ROs. The second section outlines the three-tiered conceptual framework and specifies each of the three components of the framework, respectively. In the third section, the paper presents the case selection criteria, the research method, and the three-level classification system used for the analysis. In the fourth section, the paper presents the empirical findings on the variation of citizenship in the eight ROs and discusses the three main results.

Citizenship and the region: theoretical foundations and current literature
Citizenship is a contested concept that denotes various modes of statuses, rights and duties, identities and belonging, and participation in political communities (Turner, 1993;Wiener, 1998). Over the past decades, theoretical and empirical research has opened up to new ways of thinking about citizenship beyond the nation-state (Sassen, 2005;Auvachez, 2009). Today, citizenship is forged at sub-national and supra-national levels, and scholars have opened the debate that citizenship can be conceptualised as both a legal or informal membership. The latter conceptualisation is, for example, illustrated by Luis Cabrera and Caitlin Byrne (2021) in their research of the informal ASEAN citizenship regime. Building on the early conceptualisation of citizenship as membership in a political community (Marshall, 1950;Bauböck, 2017), scholars increasingly theorise citizenship as a multilevel concept (Maas, 2013;Strumia, 2017). This theorisation has led to increased scrutiny over the dimensions and locus of citizenship (Auvachez, 2009), and three main theoretical approaches guide the theoretical and conceptual debates on studies of citizenship: the liberal, the republican, and the communitarian (Habermas, 1994).
The liberal theory saw an increase in popularity, especially in the Global North, during the period following the Second World War. This theoretical perception emphasises the individual and rational citizen and primarily theorises citizenship as a 'contract' between an individual and the nation-state (Habermas, 1994). In this period, British sociologist Thomas H. Marshall (1950) developed his well-known three-tiered theory of citizenship, arguing that citizenship consists of civil, political, and social rights. One of the main contributions of Marshall's theory was the attention to social rights and the introduction of the welfare state as a vital granter of social rights. Thus, in criticising the capitalist system of exploitation, Marshall problematised the (un)equal access to citizenship rights in British post-war society (Marshall, 1950). Marshall's theorisation also shed light on the link between institution-building and citizenship-building, claiming that civil, political, and social rights emerged alongside various government institutions such as courts and schools. Thus, Marshall is one of the earliest scholars to shed light on citizenship-building as a social process. Despite often being seen as a liberal scholar, the Marshallian theorisation of citizenship emphasised social practice as fundamental to the formation and conceptualisation of citizenship, rather than the more passive, contract-based conceptualisation of citizenship associated with the liberal approach.
The Marshallian theorisation of citizenship consisting of civil, political, and social rights and conceptualised as a social practice became crucial for later scholars. Sociologist Charles Tilly (1996) theorised citizenship as a 'tie' in his socio-historical institutionalist approach to citizenship formation. He argued that citizenship is 'a continuing series of transactions between persons and agents of a given state in which each has enforceable rights and obligations' (Tilly, 1996: 8). However, central to his theorisation of citizenship as a dynamic social contract, Tilly emphasised that the social contract is subject to ongoing modifications dependent on context and specific citizenship practices (Tilly, 2005). Tilly emphasised that communities are not fixed geographical locations but rather social constructs or citizenship associations (Tilly, 2005).
While Tilly focused on citizenship as a national concept, the attention to practices in citizenship studies also influenced scholars who examined citizenship beyond the nation-state. In her analysis of regional citizenship formation and practices in the EU, Antje Wiener (1998) argued that citizenship is a socially constructed concept formed through practices. Thus, in principle, citizenship is about distributing certain rights to all citizens of a community, having access to participation in the community, and holding a sense of belonging to the community. However, in practice, marginalisation based on gender, ethnicity, or religious beliefs might affect the de facto experience of citizenship or access to citizenship rights and opportunities. Thus, although the Marshallian theory of citizenship problematised the unequal access to social rights in post-war Britain, it has also received extensive criticism for neglecting dimensions such as gender or religion (Kabeer, 2002).
Therefore, when conceptualising citizenship, legal status cannot necessarily guarantee equal citizenship to all. Thus, as emphasised by Wiener, studies on citizenship formation ought to consider informal criteria such as the feeling of belonging (Wiener, 1998). This aspect of citizenship, the notion of civic identity, which often derives from a sense of belonging, has mainly been associated with the republican conceptualisation of citizenship. In citizenship studies, republicanism focus on participation as a way to become a citizen, and belonging is understood as a central aspect of what keeps a community bounded. Thus, civic identity or community belonging, often two terms used interchangeably, are as essential to the republican theorisation of citizenship as rights are to the liberal (Isin and Turner, 2007). Besides discussing the constitutive dimensions of citizenship, Jo Shaw (2007) emphasised the importance of making a distinction between formal memberships (de jure citizenship) and that of having access to the benefits of membership without legal status (de facto citizenship).
The turn towards new conceptualisations of citizenship beyond the nation-state (Shaw, 2007;Wiener, 1998) and the increased focus on practices and context in shaping citizenship are part of the broader scholarship on citizenship reconfiguration. In the EU, citizenship was profoundly reconceptualised when the Maastricht Treaty established the EU citizenship regime. Although the EU had practised citizenship-related policies since the 1970s (Meehan, 1993), the Maastricht Treaty formally established this additional legal status (Jenson, 2007). The status grants the regional citizen certain additional rights, the most important of which is the right to move and reside in another member state. However, EU citizenship is also based on having access to the political community and forming a regional sense of belonging (Jenson, 2007;Strumia, 2017).
Similarly, the Economic Community of West African States (ECOWAS) established a legal citizenship regime with the 1982 Protocol Relating to the Definition of Community Citizens (ECOWAS, 1982), which outlined the specific rights and duties of the ECOWAS Community citizenship. The case of ECOWAS Community citizenship shows that citizenship in ROs has developed independently from the EU. Moreover, the Andean Community (CAN) and the Caribbean Community (CARICOM) have practised citizenship-related policies such as voting rights, free movement, and regional healthcare for almost two decades. That said, regional citizenship is not legally defined in the organisations' institutional frameworks (Giupponi, 2017). Although the literature on regional citizenship regimes within and beyond the EU provides interesting insights into how citizenship is forged through citizenship-related policies, it still treats EU citizenship as the preferred model to which ROs should compare themselves (Strumia, 2017).
The different theorisations of citizenship highlight citizenship as a contested concept and show that 'a complete or elaborate theory of citizenship does not exist' (Turner, 1993: viii). However, based on the conceptual and theoretical debates, two overarching statements can be made about citizenship. First, citizenship denotes a set of rights, access to participation, and a sense of belonging. Rights are predominantly divided into civil rights, political rights, and social rights, although scholars increasingly emphasise other rights such as economic and cultural rights (Isin and Turner, 2007;Kabeer, 2002). Access refers to the possibility of having access to political participation, and belonging refers to a rootedness to and identification with a community (Wiener, 1998). Second, the locus of citizenship is continually reassessed, and scholars generally agree that citizenship is a multilevel concept that denotes membership in communities within and beyond the nation-state and in the nation-state itself (Maas, 2013;Strumia, 2017).

A conceptual framework for comparing the variation of citizenship in ROs
This section outlines the conceptual framework deployed for the comparative analysis, designed to explore the variation of citizenship across the eight ROs. The three core dimensions of citizenship, rights, access, and belonging serve as the conceptual framework. The three dimensions are helpful as analytical concepts for exploring how citizenship varies across ROs. The conceptual framework allows for a comparative analysis that goes beyond the legal status as a precondition for regional citizenship. Thus, it is broad because it pays equal attention to the three different dimensions of citizenship. It is inclusive because it can be applied as a lens to analyse citizenship in ROs that have not legally defined the concept of regional citizenship but are practising citizenship through citizenship-related policies.

Rights
Citizenship rights commonly refer to a set of civil, political, and social rights granted to citizens of a political community. According to Marshall (1950), civil rights comprise the right to liberty, property and residence, freedom of speech, movement, faith, and equal treatment. Political rights include the right to participate in political procedures, such as running for office, establishing associations, and assembly. The third aspect of rights refers to social rights, which, according to Marshall, emerged in European nation-states after the establishment of civil and political rights. They amount to the right to social welfare and security, such as healthcare and education, including cultural rights (Marshall, 1950;Wiener, 1998). In the comparative analysis, the paper explores the three different types of rights initially proposed by Marshall (1950).
In the analysis, the paper focuses on the right to movement and residence as the primary civil rights, the right to political procedures, association and assembly as the primary political rights, and the right to social protection, healthcare, and education as the primary social rights. However, the comparative analysis also pays attention to other citizenship rights when included in the regional policies. The rationale for focusing on these specific rights is: first, they have been emphasised as important regional rights in previous research (Benhabib, 2004;Ukaigwe, 2016). Second, the eight organisations include these specific rights in their policy documents, either as part of regional citizenship policies or in independent, separate protocols (Ukaigwe, 2016;EU, 2000;Giupponi, 2017;ASEAN, 2013;GCC, 2001).

Access
The participatory dimension of citizenship is mainly referred to as access (Auvachez, 2009), political participation or participatory practices (Schlenker and Blatter, 2014). Contrary to the set of civil, political, and social rights granted to citizens in a community, access determines whether or not a citizen can participate in the community's political life (Wiener, 1998). For example, in liberal democracies, rights are considered universal and equally granted to all citizens of the same community. However, in countries formerly colonised by European colonial powers, nationality laws are still today based on the colonial legal system of direct and indirect rule, which created an oppressive hierarchy of citizens and non-citizens. Therefore, rights are still not necessarily equally accessed by all citizens of a community, and factors such as ethnicity, gender, or religion might affect the de facto access to citizenship rights (Soysal, 1994). Therefore, access matters for the opportunity to realise one's status as a citizen of a political community.
In the comparative analysis, access is employed as an analytical lens to examine to what extent ROs allow for political participation in their policies and protocols. In the paper, this dimension is divided into constitutional access and practice-based access. The first type refers to access to participation through direct voting and has a legalistic character. The second type refers to participation through political forums, civil society organisations (CSOs) or non-governmental organisations (NGOs). The latter is characterised as practised-based, and although both types are commonly regulated by law, the latter can also develop through social practices and norms within a community or institution (Wiener, 1998).

Belonging
Belonging encompasses two intertwined conceptualisations. First, belonging can be conceptualised as legalistic identity, referring to the legal linkage between people within a community and between citizens and a polity (Wiener, 1998). This type of belonging has also often been conceptualised as legal status -that of being legally recognised by a polity. The second type refers to practice-based belonging, meaning that belonging can also be understood as a shared, collective identity created through practices, norms, and interactions among people (Jenson, 2007). In this paper, the dimension belonging is conceptualised as a legal status, regulated by law and documented by ID cards or passports, and as a practice-based concept generated through interaction among people over time.
In the comparative analysis, this dimension is analysed by examining how organisations emphasise belonging, or the often interchangeably used term identity, within their institutional policies and protocols. Thus, the comparative analysis draws on policy documents referring to both types of belonging, the legalistic and the practice-based. By exploring visa, passport, and ID policies, the paper examines to what extent the organisations emphasise the first type of belonging, which commonly takes a legal and formal character. Moreover, belonging is equally created by facilitating cultural and social events or establishing common symbols such as flags and anthems. These organisational initiatives are, therefore, also examined in the comparative analysis. In Table 1, the three-tiered conceptual framework is outlined and summarised.

Case selection
The comparative analysis explores citizenship in eight ROs located across Europe, South America, Africa, the Middle East and Asia. The organisations are the Association of Southeast Asian Nations (ASEAN), the Andean Community (CAN), the Caribbean Community (CARICOM), the East

Rights
Access Belonging

Definition:
Civil: the right to movement and residence.
Political: the right to political procedures, association, and assembly. Social: the right to healthcare, social protection, and education. Operationalisation: Analysed by examining ROs' policies on the extensions of civil, political and social rights to citizens of all member states.

Definition:
Having the possibility to participate in the political life of a community.

Operationalisation:
Analysed by examining ROs' policies on the possibility of accessing political decisionmaking and discussions through direct voting or engagement in political forums, CSOs or NGOs.

Definition:
Having a legal or practice-based linkage to other people in a community and the governing polity.

Operationalisation:
Analysed by examining ROs' policies on identity formation and construction of regional belonging through material sources (passports or symbols) or social engagements (cultural events).
African Community (EAC), the Economic Community of West African States (ECOWAS), the European Union (EU), the Gulf Cooperation Council (GCC), and the Common Market of the South (Mercosur). A common critique comparative regionalism scholars meet is that the ROs differ too much to be compared in any meaningful way (Haas, 1970). Of course, this critique applies to this comparison of the eight organisations, as they differ extensively in their institutional design and cultural and political composition (Spandler, 2018;Krapohl, 2017). However, the eight ROs have been selected on two main case selection criteria, making it possible to explore the variation in citizenship across the cases. First, in citizenship studies, comparative research on citizenship in ROs is limited, and in comparative regionalism, there is an absence of studies that explore variation in regional citizenship regimes. Only one other comparative study includes more than two cases (Strumia, 2017). Strumia provides an overview of citizenship in ROs and trade agreements, which does not include all cases wherein citizenship has been established in ROs but instead outlines some of the most known examples from the literature. She establishes three groups of supranational citizenship, a lowcapacity, an intermediate, and an advanced, based on three criteria for her case selection: (a) the organisation must, to some degree, work towards the projection of establishing citizenship beyond a single nation-state; (b) this beyond-state citizenship should be articulated within the boundaries of a supranational entity pursuing a collective purpose; and (c) the reconfiguration of citizenship should take place through a dynamic of mutual recognition of national citizenships (Strumia, 2017).
In her categorisation, Strumia excludes the EU and views it as a case in itself, and she emphasises the importance of free movement as a key marker for regional citizenship. Her overview provides valuable initial insights into how citizenship is characterised in ROs beyond the EU. However, in her categorisation, she includes trade agreements such as NAFTA and excludes other organisations such as the EAC (Mengisteab and Bereketeab, 2012). Moreover, according to her case selection criteria, ASEAN, like NAFTA, does not realise a regional citizenship regime because it 'does not go as far as conceiving a form of belonging across national borders' (Strumia, 2017: 685). However, current research challenges her perception of ASEAN (Weinrich, 2020;Cabrera and Byrne, 2021). Thus, the first case selection criteria are rooted in the wish to build on existing research and broaden the case selection to include all ROs where citizenship is evident. The conceptual framework based on research that has examined especially the EU citizenship regime (Wiener, 1998) helps advance the overview by Strumia and includes new cases that make it possible to examine the variation in citizenship in ROs, such as the EAC, ASEAN, and EU.
Second, the eight ROs have been selected as they make up the universe of cases. They can all be characterised as multipurpose regional organisations, which, according to Fredrik Söderbaum (2016: 106), are organisations that 'have a multitude of aims and activities across many sectors in combination with a centralised/comprehensive organisational structure'. Thus, each of the eight organisations is pursuing the objective of building citizenship -commonly intertwined with a broader regional integration agenda -albeit with various constitutional and practice-based approaches. Thus, as previous research has focused on citizenship in ROs with a legally defined status (Strumia, 2017;Wiener, 1998;Jenson, 2007;Ukaigwe, 2016;Alasfoor, 2007), the paper makes a comparative analysis of all possible cases where citizenship is present in multipurpose regional organisations. In other words, although scholars have shown how citizenship is forged within trade agreements (Strumia, 2017) or international organisations (Auvachez, 2009), the paper has chosen to focus on all cases of citizenship, constitutional or practice-based, in organisations specifically characterised as multipurpose regional organisations. The comparative analysis thereby consists of similar and dissimilar cases (Krapohl, 2017).

Methods
The paper draws on primary and secondary data for the comparative analysis. In particular, it focuses on the content of citizenship policies and citizenship-related policies of the eight ROs. These documents are divided into two main groups: legally binding and legally non-binding policy documents. The first group refers to charters, treaties, agreements, protocols, statutes, conventions, acts, and supplementary acts. The second, legally non-binding group, refers to plans of action, general meeting documents, statements, reports, declarations, announcements, and information from the organisations' webpages. The analysis draws on both citizenship policies and citizenshiprelated policies. Examples of citizenship policies are the 'Economic Agreement between the GCC States' (GCC, 2001), the 'Charter of Fundamental Rights of the European Union' (EU, 2000), or the 'Protocol Relating to the Definition of Community Citizens' (ECOWAS, 1982). These are policies that either define, explain, or describe citizenship in the ROs' institutional framework. Examples of citizenship-related policies are the 'Protocol Relating to the Community Parliament' (ECOWAS, 1994), the 'ASEAN Agreement on the Movement of Natural Persons' (ASEAN, 2012), or the 'Protocol on the Establishment of the East African Community Common Market' (EAC, 2009). Thus, as argued by Wiener (1998: 42), citizenship-related policies consist for example of 'social policy, security policy, migration policy, visa policy, and labour market policy and consumer policy', as these relate to the three dimensions of regional citizenship: rights, access, and belonging.
Moreover, the paper draws on secondary literature on citizenship in ROs (Strumia, 2017;Alasfoor, 2007;Mengisteab and Bereketeab, 2012). Furthermore, to gain a nuanced perspective of regional citizenship as constitutional and practice-based, the paper includes primary data obtained during fieldwork at the ASEAN Secretariat in Jakarta, Indonesia, and at the ECOWAS Commission in Abuja, Nigeria from November 2018 to December 2019. Citizenship in ECOWAS and ASEAN has been neglected in contemporary literature, which has been focused on the EU, the Latin American organisations, and the Middle East (Strumia, 2017;Alasfoor, 2007;Jenson, 2007). Thus, the primary data obtained through semi-structured interviews complement the document analysis where secondary research is scarce.
The primary data used in the analysis has primarily been established by the different organisations' Commissions, Supreme Councils and Secretariats. The organisations' design and governance structure vary, but most citizenship policies or citizenship-related policies have been agreed upon by the Heads of State (GCC, 2001;ECOWAS, 1982). Concerning the citizenship-related policies, these are often documents forged by different bodies under the commissions or secretariats, of which the ASEAN's Socio-Cultural Community Blueprint 2025 (ASEAN, 2016) is an example. Moreover, these can also be the results of political decisions in judicial or legislative bodies of the organisations, as seen in the 'East African Legislative Assembly Elections Act 2011' (EAC, 2011) or the 'Protocol Relating to the Community Court' (ECOWAS, 1991).

Classification
In order to structure the analysis of the policy documents, the paper developed a three-level classification system. The classification system makes it possible to provide an overview of the empirical findings, which shows how the eight ROs emphasise the three citizenship dimensions. The three classifications deployed are 'weak', 'moderate', and 'strong'. This classification exemplifies the variation of citizenship in ROs and shows the differences and similarities between the three citizenship dimensions across the cases. During the analysis, the classification 'weak' is used when a policy document does not, or rarely, emphasise one of the three citizenship dimensions.
The classification 'moderate' is used when a policy document includes the citizenship dimension examined, but the main objective of the policy document is not primarily related to the dimension. The third classification, 'strong', is used when a citizenship dimension is the main objective of the document. Before outlining the empirical findings, the paper first provides examples of the conceptual definition of the three-level classification system and how it has been deployed in the comparative analysis. Thus, it shows how rights, access, and belonging are operationalised and how each of the classifications is deployed empirically.
In ASEAN, the classification 'weak' is allocated to the rights dimension because the organisation does not have a regional parliament and does not facilitate voting (part of the political rights). Moreover, the civil right to movement is restricted (ASEAN, 2006). Another example of the classification 'weak' is the dimension belonging in the GCC, as the organisation has almost entirely refrained from including this dimension in policy documents. Despite that the member states share close similarities, such as 'a common language, a common religion, comparable social structures, very similar systems of government, a collective culture and a shared geography' (Legrenzi, 2002: 27).
An example of the classification 'moderate' is found in the EAC, where rights are classified as 'moderate'. This result derives from the EAC having extended the right to movement to all EAC citizens under the 'Protocol on the Establishment of the East African Community Common Market' (EAC, 2009). However, at the same time, social rights are almost absent from the institutional framework. In several organisations, access is classified as 'moderate'. The comparative analysis shows that these organisations increasingly open the political decision-making up to NGOs and CSOs (Acharya, 2009). Another example of the classification 'moderate' is the right to vote. Contrary to Mercosur, where member states allow all Mercosur citizens the right to vote (Modolo, 2011), the 'East African Legislative Assembly Elections Act 2011' (EAC, 2011) informs that positions into the EAC parliament are distributed according to the number a political party has in the national assembly. In this case, national parties usually nominate potential members, rank them and forward their names to the national assembly for further selection and approval.
The classification 'strong' is given in cases like the GCC, where the organisation has legally defined regional citizenship as 'Gulf Citizenship' (Alasfoor, 2007;Kinninmont, 2013), and where explicit regional citizenship rights are defined in the 'Economic Agreement between the GCC States' (GCC, 2001). In both CAN and CARICOM, rights are classified as 'strong' because the three different types of rights are strongly emphasised in the recently signed 'Protocol on Contingent Rights' (CARICOM, 2019) and the 'Instrumento Andino de Migración Laboral' (CAN, 2003), which extend the right to free movement, and the 'Decision 504' (CAN, 2001), which created the Andean passport. Furthermore, access has been allocated the classification 'strong' in the EU and ECOWAS, as the two organisations grant access to participation through legally binding documents. They both allow voting in regional parliament elections and providing access to participation through civil society engagement. Thus, both types of access, the legalistic and the practice-based, are emphasised in the policy documents of the EU and ECOWAS (ECOWAS, 1994;EU, 2012).

Empirical findings
The empirical findings are summarised in Table 2, where the three classifications, 'weak', 'moderate', and 'strong', describe how each of the citizenship dimensions is emphasised in the eight organisations. The eight organisations are listed in the vertical row, and the three citizenship dimensions are listed in the horizontal row. Based on the comparative analysis, the eight organisations are organised into two different groups. In the first group, the organisations have legally defined citizenship in the institutional framework. In the second group, citizenship is not defined legally but rather practised through citizenship-related policies. The comparative analysis is conducted based on content analysis of the institutional documents, primary material, and secondary literature. Thus, the comparative content analysis helps analyse the primary and secondary documents and sources outlined in the method section (Weber, 1990).
In the following three sub-sections, the paper focuses on three main empirical findings, which are divided into three arguments: first, two main groups of citizenship in ROs emerge from the analysis, a constitutional (de jure) and a practice-based (de facto). Second, there is a variation in how citizenship is institutionalised in the two groups. Thus, there are different pathways to regional citizenship where we see that practice precedes law. Third, the link between national and regional citizenship varies, and the two groups show a disparity between how regional citizenship is obtained.

Constitutional and practice-based regional citizenship
The analysis shows that two types of citizenship exist in the universe of cases. In Table 2, these two types are shown by dividing the eight organisations into two groups. In the first group, citizenship is conceptualised as constitutional, legally defined, and based on formal rules and regulations. In the second group, citizenship is conceptualised as practice-based and primarily founded on informal institutional norms and documents that are primarily not legally binding and on citizenshiprelated policies. In the paper, the conceptualisation of citizenship as practice draws on previous literature. Tilly (2005) and Wiener (1998) emphasise the importance of practices when studying citizenship. They remind us that citizenship is modified and defined by different social, institutional, formal, and informal practices in a community. Drawing on their conceptualisation of citizenship as practice-based, the term practice-based citizenship refers in this paper to a type of citizenship founded on various citizenship-related policies and institutional norms and behaviours (Lawy and Biesta, 2006;Cabrera and Byrne, 2021). In other words, in the second group, citizenship is defined not as a legal status but as a concept based on a set of various citizenship-related policies, such as visa, healthcare, or mobility, which creates citizenship practices at the regional scale (Wiener, 1998). Thus, common to the organisations in the second group, regional citizenship is conceptualised as an ongoing practice developed by different citizenship-related policies that are often legally non-binding. On the contrary, the organisations in the first group have established a legal, institutional definition of citizenship and established this type of regional membership in legally binding laws.
The two different conceptualisations of citizenship in ROs illustrated by the two groups have theoretical significance. The constitutional conceptualisation of group one shows a likeness to the liberal theorisation of citizenship as a legal status, which strongly emphasises individual rights and a 'contract' between citizens and a polity (Bauböck, 2017). However, the second, practice-based conceptualisation of group two challenges the liberal theorisation of citizenship and the prevalent assumption that citizenship in ROs is a legally defined status (Strumia, 2017). In group two, the conceptualisation of citizenship shows similarities to the republican and communitarian perceptions, which emphasise access to participation and belonging -or civil identity -as the constitutive dimensions of citizenship (Habermas, 1994;Turner, 1993). Thus, the second group and the conceptualisation challenge the prevalent theorisation of regional citizenship and illustrate how citizenship in ROs varies from being defined through legal status to practices established by citizenship-related policies.
Moreover, the communitarian conceptualisation of citizenship rejects liberal individualism and the idea of a 'contract' between the citizens and a polity (Turner, 1993). This theorisation is important as it helps understand how regional citizenship is established in ROs within the second group. There, citizenship is framed and perceived through legally non-binding documents and established on a set of institutional practices. Moreover, republicanism and communitarianism share their emphasis on access to participation as a means to realise one's citizenship in a community. This furthermore links to the conceptualisation of citizenship in the second group, where access is highly emphasised. This is not to say that access to participation is not crucial to the constitutional conceptualisation of citizenship. Instead, it shows that although the organisations in group two have not legally defined citizenship, there are apparent similarities between the two groups. Thus, the first empirical finding shows that citizenship in ASEAN, CAN, CARICOM, and EAC is a de facto status. Despite conceptual differences, there are similarities between the two groups, such as the emphasis on access to participation that is not necessarily dependent on the legality of the citizenship regime.

Pathways to regional citizenship
The comparative analysis shows that there are different pathways to establish regional citizenship. In group one, the four ROs have developed citizenship as an almost 'step-by-step' process, where citizenship was first introduced through citizenship-related policies and later defined legally in the institutional frameworks. This way of thinking about citizenship in regional organisations is commonly known, especially from the European context (Meehan, 1993). However, the findings indicate that the organisations in group two are establishing regional citizenship differently. Although there are similarities between the organisations in the two groups, the second group has refrained from legalising the concept. As a regional policy, citizenship is in all ROs established to obtain closer regional integration between member states (Obi, 2012;Strumia, 2017). That said, the analysis shows that in a majority of the ROs, the initial political steps towards establishing citizenship were done by extending rights to mainly migrant workers in order to accommodate the extensive intra-regional migration and workflows within the regions and to benefit economically from the movements (Maas, 2013;Giupponi, 2017;Chung, 2017). This is similar between the organisations in both groups. For example, the GCC citizenship 'rests on the pillars of four fundamental freedoms; the free movement of goods, persons, capital, and freedom to provide services' (Alasfoor, 2007: 59), which similarly is a key policy focus of the ASEAN citizenship (Weinrich, 2020;Cabrera and Byrne, 2021), where the ASEAN Economic Blueprint seeks to 'transform ASEAN into a region with free movement of goods, services, investment, skilled labour, and freer flow of capital' (ASEAN, 2015). However, the data shows that over time, some organisations initiated individual rights beyond the nationstate and legalised the status of regional citizen, while others began promoting regional belonging to increase regional integration.
The constitutional and the practice-based conceptualisations of citizenship in ROs thus show a variation in the legality of the concept, as discussed in the previous section. However, it also shows how citizenship and law are perceived in the two different groups. At an interview at the ECOWAS Commission, the legality of ECOWAS citizenship was discussed with Dr Tony Elumelu, Head of Division of Free Movement and Migration. He explained that: The protocol on citizenship from 1982 was made to re-enact the relationship that was there before colonialism, and it made something that was already practised and felt by the people official so that citizens are protected legally. It needed to be established as a legal agreement between our countries to matter and to be enforced by ECOWAS and the countries (Interview, Dr Tony Elumelu, Head of Division of Free Movement and Migration, ECOWAS Commission, Abuja, 06.11.2019).
Dr Elumelu emphasised that to enforce citizenship, it needs to be defined legally. This perception consolidates the prevalent perception of regional citizenship as a legal status. That said, the Lagos Treaty from 1975 includes free movement of persons as one of the core objectives, later codified in the Revised Treaty of 1993 as a right of the 'common citizens of the Community' (ECOWAS, 1993). Thus, an examination of the citizenship in ECOWAS shows, similar to the EU, that regional citizenship developed in steps. The first step was abolishing visas, followed by the right to residence and the right to establishment. These steps are outlined in the 1979 'Protocol on Free Movement' (ECOWAS, 1979), which came before the legal definition of the ECOWAS community citizenship in 1982. Thus, the extension of citizenship rights, such as movement and residence, developed primarily because they were seen as beneficial to the economic prosperity of the member states and the RO, which is a tendency shared between the organisations in each of the two groups (Giupponi, 2017).
Contrary to the organisations of the first group, where citizenship has been legally defined, the four organisations in group two have not provided a legal definition of citizenship in the institutional frameworks. During an interview in Singapore, Ambassador Berry Desker, who held the position as Singapore's representative to ASEAN's Intergovernmental Commission on Human Rights (AICHR), explained the perception of citizenship and law in ASEAN: There is a fundamental difference between the citizenship approach of the EU and the approach taken by ASEAN. However, since the ASEAN Charter, there has been a greater trend towards forming legal documents, and I think what ASEAN has tried to do is move in the direction where they establish more legally binding documents on different issues related to citizenship. If ASEAN is to progress on this matter, the leaders and citizens in ASEAN will have to make sure that it develops its own sense of purpose (Interview, Ambassador Berry Desker, Singapore's Representative to the ASEAN Intergovernmental Commission on Human Rights, Singapore, 31.01.2019).
As explained by Ambassador Desker, and shown in the data, rights are emphasised to a lesser extent than belonging, and contrary to the EU, ECOWAS, and CAN, ASEAN has a strict intra-regional mobility policy currently allowing for visa-free travel up to 14 days (ASEAN, 2006). However, the ASEAN Agreement on the Movement of Natural Persons (ASEAN, 2012) enables temporary movement for high-skilled labourers. Furthermore, as emphasised by Ambassador Desker, the ASEAN Charter and the ASEAN Declaration on Human Rights (2013) illustrate the increased attention to extending specific rights to all ASEAN citizens.
Another example of the 'step-by-step' development of citizenship is detected in the building of Mercosur citizenship. In the organisation, 'labour citizenship' was used as an 'in-between term' in the institutional framework to capture the extension of rights to workers before regional citizenship became officially defined (Giupponi, 2017). In 2002, the right to free movement was adopted in the 'Residence Agreement', which entered into force in 2009. In 2010, the route to establish formal regional citizenship was clearly defined as Mercosur signed the 'Citizenship Statue', providing a plan for implementing Mercosur citizenship by 2021. Thus, similar to the development of citizenship in ECOWAS, the Statue includes three main objectives for the realisation of the Mercosur citizenship: the right to free movement, equal social, cultural, civil, and economic rights, and equal condition for accessing work, education, and health (Mercosur, 2010). However, it is unclear whether this 'step-by-step' process will automatically result in all ROs legally defining citizenship. This question implies examining the variation of citizenship in ROs and goes beyond the scope of this paper. However, the current data shows no indication that regional citizenship will necessarily always be legally established.

The relationship between national and regional citizenship
The empirical findings show a variation in the link between national citizenship and regional citizenship. In other words, how ROs build citizenship is dependent on how ideas about personhood and the relationship between citizens and a polity are understood (Kabeer, 2002). Regional citizenship is never meant as a replacement for national citizenship (Strumia, 2017). However, the analysis shows a variation between how the national-regional link is perceived and how regional citizenship is obtained. In group one, regional citizenship is provided based on first having national citizenship in a member state. For example, in ECOWAS, national citizenship is a precondition for regional citizenship.
Moreover, in the GCC, where Gulf citizenship is also built on having national citizenship (GCC, 2001), another tension influences the realisation of regional citizenship. Access to national citizenship in the GCC is limited and often restricted to certain families and clans. This creates tensions in accessing citizenship rights regionally, as national citizenship is a precondition for Gulf citizenship (Kinninmont, 2013). Similar tensions are evident in other ROs in group one. In ECOWAS, accessing national citizenship is also challenging, and current political discussions focus on ensuring national citizenship for citizens to provide a basis for accessing regional citizenship rights. However, contrary to the GCC, ECOWAS took serious steps to overcome statelessness and create a regional citizenship regime for all. In 2015, the organisations ratified the 'Abidjan Declaration on the Eradication of Statelessness' (ECOWAS, 2015), which has resulted in changes to national citizenship laws, ensuring the right to nationality based on which the ECOWAS citizenship status is granted.
In group two, regional citizenship is not necessarily granted through having national citizenship. Instead, the data shows examples where regional citizenship status is defined through citizenship-related policies, which does not necessarily target only member state citizens. For example, the visa policy of ASEAN is also extended to all permanent residents who might not be national citizens of a member state (ASEAN, 2006). Moreover, CARICOM provides a very different perspective on how regional citizenship is obtained. In CARICOM, access to regional citizenship benefits is also granted to non-member state citizens. This is done by broadening the definition of what it means to belong to a member state. In Article 32, 5 (a) of the 'Revised Treaty of Chaguaramas', it states that 'a person shall be regarded as a national of a member state if such person -(i) is a citizen of that State; (ii) has a connection with that State of a kind which entitles him to be regarded as belonging to or, if it be so expressed, as being a native or resident of the State for the purposes of the laws thereof relating to immigration' (CARICOM, 2001: 20). Thus, CARICOM bases the regional citizenship status on formal and informal belonging, either legally recognising a person as a national citizen or informally recognising a person as belonging to the nation-state.
In group one, obtaining regional citizenship is based primarily on having national citizenship, and regional citizenship mainly denotes legal status and individual rights. This might be the case as the four organisations, EU, ECOWAS, GCC, and Mercosur, have legally defined regional citizenship and how to obtain the status. On the contrary, the ROs in the second group, ASEAN, CAN, CARICOM, and EAC, show more openness to allowing non-member state citizens some regional citizenship rights and benefits based on their identification with the countries in the region. In some of the examined ROs, the nation-states are in ongoing decolonisation and state-formation processes. Thus, struggles over national citizenship and nationality laws have not yet been fully resolved, which could influence how regional citizenship is obtained, the legality and non-legality variation between the two groups, and account for the open and different approach that especially CARICOM has taken.

Conclusion
This paper has assessed how citizenship in ROs varies by examining citizenship rights, access, and belonging in eight ROs located in five different regions across the world. The empirical findings lead to the following three insights. First, citizenship in regional organisations can be conceptualised as constitutional or practice-based. Thus, the eight ROs have been divided into two separate groups, one that legally defines regional citizenship and where citizenship is based on formal rules and regulations, and a second group where citizenship is conceptualised as practice-based and founded on informal institutional norms and citizenship-related policies. Second, there are different pathways to regional citizenship. In group one, citizenship practices preceded law, but in group two, citizenship remains a non-legal status, established and institutionalised through citizenshiprelated policies. In particular, the analysis shows that in most ROs, the initial steps towards establishing citizenship were done by extending certain rights. In group one, the citizenship regime eventually became legalised, whereas the organisations in group two refrained from legalising the concept and focused on establishing a sense of belonging or access to participation through citizenship-related policies.
Third, there is variation in the link between national and regional citizenship and how ROs provide access to regional citizenship. For example, the organisations in group one are granting regional citizenship based on having national citizenship in a member state. However, group two shows that citizenship can be accessed through other means, and especially CARICOM challenges the assumption that regional citizenship is an 'added-on' status based on national citizenship by granting regional citizenship to a person based on a sense of belonging (CARICOM, 2001).
The empirical findings shed light on the variation in citizenship across the eight cases by applying the conceptual framework and the three-level classification system, a variation not previously accounted for in citizenship studies or comparative regionalism. In particular, the paper shows new and different conceptualisations of regional citizenship as practice-based and the similarities between the legal and non-legal type of citizenship in ROs. The empirical findings thereby emphasise the nuances, similarities, and differences across all cases of regional citizenship in multipurpose ROs. Furthermore, we find that contrary to findings of the established literature rooted in the European context, we need to rethink the theoretical and conceptual understanding of what constitutes regional citizenship. In particular, citizenship is not necessarily a legal status based on specific citizenship policies but can also be formed through institutional norms and citizenship-related policies. Citizenship in the ROs in group two are examples of the latter. Thus, the analysis shows that ROs' development, regulation, and provision of citizenship happens in various legal and non-legal ways.
Finally, the comparative analysis points to a need for more research on citizenship in ROs, primarily in organisations that do not define such a concept. In particular, the research should account for why regional citizenship varies and what institutional factors might explain the variation, mainly focusing on the second, practice-based type of regional citizenship, which is still underresearched today. The paper hints at these research topics, but in-depth single and comparative case studies hold the possibility of explaining the different types of citizenship, which would lead to a better and empirical grounded knowledge of how ROs develop people-centred regionalism and integration policies different from the European pathway. Going beyond the European experience helps further the research in citizenship studies and comparative regionalism and provides much needed new empirical insight that challenges the prevalent EU-centrism in both citizenship studies and comparative regionalism.

Declaration of conflicting interests
The author declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.