Federalism and Ethnic Relations

Abstract

This contribution explores how federalism is employed for multinational state-building and sheds light on the topic of federalism and ethnic relations at large. The specific objective is to explain fundamentals as to the functioning of some federal systems subsumed in the category of multinational or ethnic federalism. This contribution demonstrates that no global theory of federalism exists and that federal systems encompass also states that formally are not called federations. It explains how federal theory interprets ethnic and territorial federalism, and ethnoterritorial federalism, and it discusses state formation and its implications for multinational state-building. This framework is applied to different world regions and to the analysis of rationales and peculiarities of federal and decentralized systems with ethnic elements in Asia, Africa, and Latin America. In conclusion, this contribution identifies three common and novel traits among ethnic federal systems, namely, that most are a recent creation and have yet to give sufficient attention to the shared-rule element of federalism, that they have only recently shifted from mono-ethnic conceptions of the state and nation to a multiethnic or plurinational recognition, and that this recognition of difference and the associated holding-together federalization process tend to result in an asymmetrical federalism.

Introduction

Federalism is anything but a simple and static concept. While there is general agreement that it entails self-rule and shared rule of different levels of government, the practice of federalism remains a complex and nuanced exercise. Worldwide federalism carries many meanings that have developed differently under different historical circumstances and in relation to different power balances. The same is true of multinational state-building. While there is general agreement that it entails governance in respect of more nationalities, the practice of multinational state-building remains a complex and nuanced exercise.

John Stuart Mill (1861, 296) argued that “Among a people without fellow-feeling, especially if they read and speak different languages, the united public opinion, necessary to the working of representative government, cannot exist.” Mill, however, also conceded that multinational state-building might succeed by means of federalism. He points to three conditions under which a federal government is possible: first, mutual sympathy among the populations and a community of interest; second, power balances, i.e., no constituent unit of a federation must be so powerful as to solely rely on individual strength; and third, no marked inequality must exist among the member states. After having expounded on the way US federalism works, Mill points to the multiplication of Federal Unions as a benefit to the world. He gives two reasons. First, federations make the weaker stronger. Second, federations prevent aggressive and petty policies and wars.

The general purpose of this contribution is to explore how federalism is employed for multinational state-building and to shed light on the topic of federalism and ethnic relations at large. Its specific objective is to explain fundamentals as to the functioning of some federal systems subsumed in the category of multinational or ethnic federalism. To do so, the analysis is structured as follows: Part B conceptually reflects on federalism, federation, and federal systems. It demonstrates that no global theory of federalism exists and that federal systems encompass also states that formally are not called federations. They however take on federal features in practice. Part C discusses ethnic and territorial federalism and ethnoterritorial federalism. It explains how federal theory interprets these and similar terms, limits such discussions for the purpose of the contribution, and refers to worldwide examples. Part D discusses state formation and its implications for multinational state-building. It differentiates between coming-together and holding-together federalism. In doing so, it refers to worldwide federal systems, and it also explains Canada, Switzerland, and Indonesia more in detail. Parts E, F, and G zoom in to different world regions and analyze rationales and peculiarities of federal systems with ethnic elements in Asia, Africa, and Latin America. Part H offers final remarks.

Federalism, Federation, and Federal Systems

The concept and practice of federalism has existed for centuries. So have discussions on the issue of a universal agreement on what federalism means. Although well entrenched in social, political, and legal sciences, federalism remains a contested term. There is no globally accepted theory of federalism (Gamper 2005). Some scholars stress the importance of advancing a new dynamic federal theory (Popelier 2021), while others opt for a more pragmatic approach by referring to the toolbox of federalism (Palermo and Kössler 2017, 2).

A strict legal definition of federalism would often follow the early works of Kenneth C. Wheare, where federalism exists when more than one level of government has final say in one at least competency or area of authority. He defines federalism as the “method of dividing powers so that the general and regional governments are each, within a sphere, co-ordinate and independent” (Wheare 1947, 11). He reduces federalism to its state-related dual governments’ dimension and equates the term federalism with terms such as federation, federal state, and federal government. While this definition, part of the classical Anglo-American conceptualization of federalism, can be strictly applied, it means that countries that are generally accepted to have federalism, such as India and Canada, would be unitary states because of the override and intervention powers their systems have. Indeed, Watts (2008, 9), discussed later, categorizes India and Canada (by its formal constitution) as hybrids. Wheare’s static, legalistic approach to federalism has been challenged by scholars within the Anglo-American tradition and by scholars within the continental European tradition.

Carl J. Friedrich (1968, 7) referred to federalism as a process more than as a fixed design and pointed to the fact that “federal relations are fluctuating relations in the very nature of things.” Indeed, federalism is in essence a hybrid, originating through the combination of confederal units and a strong central authority as an outcome of the US constitution-making process (Galligan 2007) at the one side. At the other side, federalism in its actual operation has always been exposed to complex, multidimensional, and dynamic de/centralization phases, as scholars continue to demonstrate (Dardanelli et al. 2019).

William A. Livingston also warned against a static approach to federalism. According to him, “the essence of federalism lies not in the constitutional or institutional structure but in the society itself. Federal government is a device by which the federal qualities of the society are articulated and protected” (Livingston 1952, 84).

William H. Riker discredited Friedrich’s work because of its vagueness (Riker 1969, 137). He construed federalism as a constitutional bargain between future leaders of governments (Riker 1964, 1). He stressed that ultimately it is the political culture at large that determines the operation of federalism.

Since the time of these discussions, many different hybrid compound systems – combinations of autonomy and centralization – have been established, although the US federation has become somewhat of a standard. Together with the many different hybrids, the approach to defining federalism by pointing to its state-related dimension has been abandoned in favor of broader categories. They all differentiate between federalism as a normative principle and federation as an organizational one. Transatlantic flows of ideas between European continental and Anglo-American scholars have contributed to this (Palermo and Kössler 2017, 14–20).

In 1982, the conceptual distinction between federalism and federation was finally established in an explicit way. Preston King divided the discussion of federalism as a political ideology from the discussion of federation as a particular institutional relationship. Accordingly, federalism is the multidimensional driving force and federation its institutional, structural, and systemic counterpart. Hence, “there may be federalism without federation,” but there can be “no federation without some matching variety of federalism” (King 1982, 76).

Ronald Watts’ comparative federal studies build on this and on Daniel Elazar’s (1987, 5) interpretation of federalism as self-rule and shared rule. Watts (2008, 8–18) distinguishes between federalism as a normative objective or principle and federal systems as the institutional arrangement. He categorizes a federation as a species of a larger genus, the one of federal political systems. Such a genus comprises a variety of species of political organizations that somehow embody a combination of self-rule and shared rule. Hence, Watts (Table 1) categorizes not only federations but also regionalized/decentralized unitary systems as well as confederations and federacies under the term of federal political systems.

Table 1 Types of federal systems

Michael Burgess, building on the works of Elazar and Watts, has reiterated the distinction between federalism and federation (Burgess 2006). He also introduced the concept of federality, i.e., the capacity of federal systems to sustain federalism as a principle that animates a federation in the face of challenges and constant change (Burgess 2012, 139–140). Federality is especially important in federal systems that regulate ethnic relations and that arise out of conflicts. Such systems oftentimes consist of constitutionally decentralized unions, where powers are (asymmetrically) allocated to more than one level of government with, as a rule, the central government retaining important intervention powers (Watts 2008, 16).

Ethnic and Territorial Federalism

Federalism can be more or less accommodating of ethnic diversity, or not at all, depending on the underlying rationale of its design. Hence, federal systems can be categorized depending on whether they have internal boundaries and (plural) legal systems that account for ethnicity or whether they are neutral. The contribution adopts Yash Ghai’s definition of ethnicity (2019, 53). He refers to ethnicity as “a situation where a community goes beyond a mere consciousness of what binds it together and distinguishes it from other communities (such as language, religion, and race) to claim that these differences mark it as a ‘people’ or ‘nation’ entitled to political recognition.”

Building on this, the contribution calls federal systems that are ethnically based as having “ethnic federalism.” Other terms that are also widely used include ethnofederalism, plurinationalism, and multinational federalism. The contribution refers to them too and, at this point, makes clear that it does not aim to stipulate a definition of the various terms used nor to thoroughly problematize the issue whether constitutional federalism has failed or not in the systems it refers to. It gives evidence on the topic of federal systems and ethnic relations, with references to worldwide case studies. As a rule, it adopts terminology regarding federal theories/practices as commonly used in the doctrine and in the respective federal systems and official documents. Of course, all terms used in the contribution are differently interpreted and distinguished from each other in academia and by practitioners in the different world regions.

Gagnon (2021, 102), for example, stresses that “multinational federalism assumes the existence of several demoi in contrast with the notion of ethnofederalism which suggests that ethnicity is the main driving force … [and that ethnofederalism is] potentially unstable as narrow ethnic identities would prevail over broad civic achievements.” Though not being a panacea, democratic multinational federalism, instead, best regulates ethnic relations. He argues that it does so by “(1) decoupling and distinguishing the notions of ‘nation’ and ‘state’, (2) strengthening a sense of identity through the implementation of politics of recognition, and (3) developing a healthier equilibrium between self-rule and shared rule through the implementation of a multiplicity of collaborative initiatives while respecting the principle of political autonomy” (ibid).

Federal systems that, instead, are “neutral” have territorial federalism, also known as regional or mono-national federalism (e.g., Germany, the United States). In contrast with multinational federalism, territorial federalism does not explicitly recognize the presence of multiple identities. It does not reflect cultural and societal characteristics of a given nation-state in the overall system’s institutions and political arrangements, and it only exceptionally promotes group-related sociolinguistic minority rights in the form of territorial autonomy or non-territorial autonomy (on minority protection and autonomy arrangements, see Malloy and Palermo 2015; on the nature of different minorities [national, linguistic, religious, indigenous, etc.] and for an overview of approaches to dealing with the constitutional recognition of minorities, see Anderson et al. 2014).

Often the key factor in federal “ethnic” systems is whether or not internal boundaries that designate the subnational units (states, regions, provinces) are drawn in order to incorporate as many of one ethnic group as possible and to exclude as few people from that group as possible. In other words, boundaries reflect the clustering of ethnic groups or traditional homelands, with all challenges that such a strong ethnic-territorial link reserves regarding the recognition of the internal diversity within the constituent unit. In such cases, the respective groups oftentimes see federalism as ownership over a territory, and this normally leads to conflicts (Palermo 2015). The name of a subnational unit in an ethnic federal system is usually reflective of the given ethnic group. This is the case of Tamil Nadu in India. Ethiopia serves as another example in this regard. Article 47 of the 1995 Ethiopian Constitution first draws the boundaries of most constituent states to the benefit of a specific ethnic group and then facilitates their redrawing along ethnic lines by granting “nations, nationalities, and peoples within the states … the right to establish, at any time, a state of their own.” An ethnic federal state would also include or enable laws and policies that give special rights to members of the designated ethnic groups. Such group-related rights can be in the form of preferential treatment or political prerogatives (e.g., to leadership positions), language rights, natural resource rights, or other (for a legal analysis of minority protection in Europe that includes the deconstruction of primordial theories of ethnicity and a sociologically informed political theory for how to govern cultural diversity, see Marko and Constantin 2019).

Opposed to this, federalism can be ethnically and legally neutral, comes in the form of territorial federalism in the sense that internal boundaries are based on geographical or economic features, or may even be designed to cut across different ethnic groups. In such a model, there are no special rights and cultural matters are treated as private matters. This type of federalism may do little to accommodate different ethnic groups. However, it may support the participation and development of more culturally sensitive policies simply by virtue of creating new critical masses and by bringing government closer to the people in line with the principles of vertical (territorial) and horizontal (functional) subsidiarity (i.e., relations between the central authorities and subnational levels of government at the one side and collaborative governance patterns between the authorities and civil society at the other side).

Of course, federal systems are rarely purely ethnic or territorial types. They are often a combination. Liam Anderson (2013) conceptualizes “ethnoterritorial federalism” as another type of federalism where the dominant ethnic group is carved up across several subnational units, but the smaller and territorially concentrated groups are accommodated with their own subnational unit. Ethnic federalism has been the subject of considerable criticism and has a patchy record. Roeder (2009), for example, argues that privileging some groups increases the likelihood of conflict; Cornell (2002, 225) warns that institutionalizing identity makes ethnic groups more cohesive and likely to act; Bunce and Watts (2005, 254) claim it increases mistrust, isolation, and competition; Hale (2004) shows that “bipolar” federations are more likely to collapse; and Basta (2021) points to the relevance of institutional symbolism in multinational federal systems, i.e., to the fact that institutions may be more important for what they mean than for what they do. Liam Anderson (2013) accounts more specifically for institutional design and demonstrates that ethnofederalism since 1945 has a better track record than most critics (and even advocates) would concede given it is often implemented in situations where unitary and nonethnic federalism have failed or are otherwise not viable.

For example, commonly cited ethnofederal “failures” Yugoslavia and Czechoslovakia introduced ethnic federalism as a means of holding-together. Despite the collapse of these communist-era ethnofederations – extremely violent in the case of Yugoslavia and peaceful by mutual consent in the case of Czechoslovakia – the concept of federalism persists in multiethnic state-building in Eastern Europe (Hughes 2018). It has been implemented by elite successors of the communist era in Russia and Bosnia and Herzegovina (BiH). It also continues to be proposed and implemented by Western Europe as a conflict resolution device in Eastern Europe and elsewhere. Federalism, decentralization, and regional (asymmetric) autonomy have, for example, been the basis of many (EU/European-mediated) peace talks to resolve (frozen) conflicts (e.g., Transnistria-Moldova, Kosovo-Serbia, Cyprus), and it also has been used in multiethnic state-building (e.g., Gagauz Yeri autonomy in Moldova; forms of non-territorial autonomies in Central, Eastern and Southeast Europe, to protect minorities and to favor relations between minorities and their kin-states).

In addition, many partial or “ethnoterritorial” solutions including forms of consociational and regional autonomy have been successful at addressing specific conflicts in Western Europe (Gray et al. 2021; Swenden 2006). Three of six Western European countries with federal traits are clearly multinational (Belgium, Spain, the United Kingdom), while one, Switzerland (discussed later), grounds its federalism in multilingualism with sociolinguistic cleavages that are crosscutting. Italy is a hybrid. Its constitutional design, a highly asymmetric regional state, combines moderate regionalism in most regions with peaks of advanced federal traits in some special regions (Arban et al. 2021). Beyond Europe, Indonesia/Aceh (discussed later) and Philippines/Mindanao serve as examples for highly decentralized systems with autonomous constituent units in which territorial and ethnic elements coexist and interact.

To conclude, there are very few examples of full ethnofederal systems (where every subnational unit is premised on ethnicity, exceptions include Belgium, BiH, Ethiopia, and Pakistan), and most are partial ethnofederal systems (e.g., Canada, India, Russia, Switzerland) (Anderson 2014). Ethnoterritorial federalism, which relies on the division of the dominant group, also overcomes the design flaw highlighted by Hale (2004).

Federal State Formation

An important distinction in federal studies relates to the method of formation of a federation. Watts (2008, 65) outlines three models: First, a federation can be formed through the aggregation of former separate subunits. Second, a federation can come into being through devolution from a former unitary state. Third, a federation’s birth is a combination of aggregation and devolution.

Alfred Stepan (1999, 22) reduces these formation processes into two and differentiates between coming-together and holding-together. While coming-together federalism refers to federations that are born through aggregation of prior sovereign states, holding-together federalism refers to federations that are born through devolution from the preexisting constituent units of a unitary state. This distinction is important because it determines, to a certain extent, the balance between the central and subnational governments, as well as whether or not the subnational units have an ethnic basis. Specifically, when a holding-together federal system is negotiated (or imposed), the powers are being devolved downward from the center. The central government/key actors will tend to be in a stronger position because they hold the existing powers. This means that holding-together federal systems will tend to be more centralized and, quite often, the center will retain intervention rights. Conversely, coming-together federal systems will tend to result in stronger subnational units, because those units come to the table with their existing powers and give up/transfer to the center only those necessary.

The reasons for forming a federation are multiple. There are four traditional explanations discussed in literature: first, the aspiration to have a democratic system that ensures individual liberty, freedom, and prosperity (Elazar 1987, 83–91, building on the Federalist Papers, a series of 85 essays written by Alexander Hamilton, John Jay, and James Madison between 1787 and 1788, published anonymously under the pen name Publius, with the aim to urge New Yorkers to ratify the proposed US Constitution of 1787); second, in line with Riker’s reasoning (discussed earlier), the will to “bargain” for a union because of the existence of external threats and for trade and economic reasons; third, the necessity to form a federation because of a disjuncture between political identity and a given geographic, democratic model of governance, i.e., “conflicts in political identity” (Feeley and Rubin 2009, 39); and fourth, the will to recognize state-society dynamics in constituent units and their governance capacity (Ziblatt 2004). These explanations, or a combination of them, can only partially capture the original logic of many of the federal systems the contribution refers to. Therefore, a fifth explanation is put forward: the accommodation of competing demands of ethnonational groups as an inevitable path for territorial integrity, i.e., the use of federalism/regional autonomies as a tool of conflict resolution (Keil 2019; Keil and Alber 2021, with case studies on Spain, Moldova, Ukraine, Iraq, Papua New Guinea; Anderson and Keil 2021, analyzing BiH, Spain, Myanmar). Whenever international actors impose a federal setup on a country as part of a conflict resolution process, this is also known as putting-together federalism, though fundamentally this keeps being holding-together federalism (as the difference is solely in the role and influence of external actors, not in the purpose as such).

Empirical evidence shows that the longevity of systems born out of conflicts depends on many context-bound factors. In the main, these are three: first, the political strength leaders of ethnonational groups have during the constitutional moment; second, the will of political leaders, ethnic groups, and the citizenry at large to commit to democratic federalism (Inman and Daniel 2020), i.e., to constitutional federalism according to the federal spirit (Burgess 2012); and third, the extent to which such a system is affected by dynamics ascribable to third-party intervention, geopolitics, and geoeconomics. Geoeconomics is a relatively new term compared to geopolitics. It stands for resource conflicts as the outcomes of geoeconomic power games that are no longer grounded on the traditional concept of geopolitics that goes back to imperial concerns only (Wigell et al. 2020).

Of course, there is no magic formula that can serve as a blueprint for federalism as a tool of conflict resolution. Different arrangements offer diverse solutions. They all have advantages and drawbacks. Two criteria seem, however, to be key to successful conflict resolution through the provision of federalism and other forms of territorial autonomy. First, any arrangement must be suitable for the country and the territory in question. This means that historical legacies matter as much as governmental capacity to implement and safeguard formulas of self-rule and shared rule. It also implies that any settlement to an ongoing conflict needs to arrive through an inclusive process that ties in key actors and potential spoilers to any settlement. The territorial autonomy of South Tyrol in Italy’s asymmetric regional state is an example in this regard (Alber 2021; Woelk et al. 2008). Second, whatever arrangement is chosen at a particular moment in time to end a conflict needs to be flexible enough to adapt and change as time passes, and the initial issues related to the outbreak of violence become less important (Keil and McCulloch 2021). The inflexibility of arrangements in BiH (Banović et al. 2021), the unwillingness of Iraqi elites to revisit their constitutional framework (Belser 2021), and the inability of Lebanese politicians to move beyond an outdated power-sharing framework (Fakhoury 2014) all highlight that while peace agreements certainly need to be robust and stable, they also cannot be engrained forever. Ultimately, power-sharing arrangements between elites, territorial autonomy decisions, and institutional choices evolve in countries that have not faced violent conflict, and they do so as well in countries with a legacy of ethnic conflict.

Coming-Together Federalism and Evidence from Canada and Switzerland

Historically, federalism followed a coming-together of sovereign entities to form a federation. Such processes of federal formation, though in essence voluntary and gradual, also met the resistance of individual states (e.g., in the case of the US federalism and Switzerland) and, in part, were a process lasting several decades. This model of coming-together federalism is primarily seen in the eighteenth and nineteenth centuries and includes the United States (1787), Switzerland (1848), Canada (1867), and Australia (1901). Although significant differences exist in the original design and evolution of these federations, they, when born, all shared the logic of relatively autonomous units that “come together to pool their sovereignty while retaining their individual identities” (Stepan 1999, 23).

The Example of Canada

The union of the British North American colonies of New Brunswick, Nova Scotia, and the Province of Canada (what is now Ontario and Quebec) was the first step in a slow and difficult nation-building exercise in multinational Canada. Today, Canada is one of the most decentralized federations worldwide that, although well-established, somehow is still in search of the federal spirit (Mathieu and Gagnon 2021; Bakvis and Skogstad 2020).

In Canada, English and French are both official languages: The Official Languages Act 1969 granted language rights to all Canadians, in particular the right to service in the official language of their choice from the federal government. It also sought to make the federal public service more representative of Canada’s French-speaking population while respecting the merit principle (Turgeon and Gagnon 2013). In 1988, the Act was repealed and replaced by the Official Languages Act 1988, mainly to promote the official rights of linguistic minorities in accordance with the most recent developments in Canadian language policy, with provinces and territories playing a significant role in the recognition and promotion of multilingualism through education programs (Cardinal and Léger 2019). For example, in 2008, Nunavut declared English, French, and Inuktitut to be official languages. In 2010, Manitoba became the first province to recognize indigenous languages officially. The Aboriginal Languages Recognition Act states that Cree, Dakota, Dene, Inuktitut, Ojibway, and Oji-Cree are “spoken and used in Manitoba.” The Act, however, does not include provisions to promote Indigenous languages. The case of Quebec, instead, has the distinction of being bilingual on constitutional and federal levels while officially allowing only French in its provincial institutions. Another distinctive case is the province of New Brunswick. It is Canada’s only officially bilingual province. Indeed, the 1982 Canadian Charter of Rights and Freedoms, part of the Canadian constitution, in its sections 16 to 20 makes clear that official language rights apply in respect of both the federal government and the provincial government of New Brunswick. That means that Canadians can use English or French when they deal with the federal authorities and that the residents of New Brunswick also have the right to use English or French when they deal with their provincial authorities.

When considering group dynamics at large, there are three main axes to be identified in “ethnic relations” in Canada (Breton 2013): First, relations between indigenous and nonindigenous people, for now characterized by marginality and dependence. In 1982, the Canadian Charter of Rights and Freedom Canada affirmed “existing Aboriginal and treaty rights” of the Aboriginal Peoples of Canada (Section 35). These rights, however, continue to require elaboration and clarification through Canada’s judiciary system. Language as an existing right, for example, was not assumed until its inclusion in the Indigenous Languages Act of 2019. This Act clarifies the government’s role in the current state of language crisis of indigenous languages and its responsibilities to reverse language decline. The four key provisions within the Act are (a) the affirmation of indigenous languages as a right protected under section 35 of the constitution; (b) the commitment to adequate, sustainable, long-term funding; (c) the commitment to support interjurisdictional cooperation; and (d) the institution of an indigenous languages’ commissioner. Whether or not the implementation of the Act and the revision of provincial language policies will lead to substantive equality is yet to be seen (Cardinal and Williams 2020).

Canada has also adapted its federal setup to better recognize indigenous territories. In 1999, the Northwest Territories was split to create a Nunavut Territory for Inuit Peoples. Although the territory is represented in the federal legislature, it is established by legislation and does have the same rights as the provinces. The creation of the Nunavut Territory was an outcome of treaty-making processes. Additional indigenous self-governing regions have been negotiated within other subnational regions, giving rise to the concept of “nested federalism” (Wilson et al. 2020).

Second, relations between English and French speakers. The competition between the two groups has always been manifested in the struggle over the control of government institutions, the distribution of powers, and the symbolic recognition of the language and culture of both groups (see the Quebec Act of 1774; the 1791 Constitutional Act; the Act of Union of 1841; the British North America Constitution Act 1867 and its modifications; the 1982 patriation of the Constitution; the failure of the Meech Lake Accord of 1987 and the Charlottetown Accord of 1992 that led to the 1995 referendum on Québec sovereignty in which the No side achieved a victory by a narrow majority of 50.58%).

Third, relations between the colonizing groups and other immigrants and their descendants. Such relations are linked to issues as the unequal access to employment and education, as well as unequal security and social respect. Canada’s multiculturalism policies and programs of various levels of government have been a step in recognizing diversity, although their aim has not been to restructure institutions to reflect the composition of the population. It is worth noting that Canada, as the first country worldwide, adopted its first policy on multiculturalism in 1971. In a speech in the House of Commons in April of 1971, Prime Minister Pierre Trudeau introduced it as a policy of multiculturalism within a bilingual framework, a policy that would complement the Official Languages Act 1969 by facilitating the integration of new Canadians into one or both of the official language communities. The origins of Canada’s multiculturalism policy can be found in the Royal Commission on Bilingualism and Biculturalism (1963–1969). Multiculturalism was intended as a policy solution to manage both rising francophone nationalism and increasing cultural diversity across the country. It recognized the rights of Canadians to preserve their cultural heritage as an essential dimension of its federal system. Achieving this policy objective depended on government funding. The Ministry of Multiculturalism was created in 1973 to monitor the implementation of multicultural initiatives within government departments. In addition, formal linkages between the government and ethnic organizations were established to provide ongoing input into the decision-making process. In 1988, when the Canadian Multiculturalism Act was adopted by the Parliament, Canada was the first country in the world to pass a national multiculturalism law (Brosseau and Dewing 2009).

The Example of Switzerland

Switzerland, the world’s second federation in time, is based on liberalism, by and large a Protestant phenomenon sparked by the French Revolution of 1830. The 1847 Civil War against the alliance of seven Catholic conservative cantons (Sonderbund) marked the end of the Old Swiss Confederacy and showed that cantonal identities, stretching back centuries, had to be considered in the rebuilding of Switzerland. The Swiss constitution of 1848 thus created a sovereign central government but, in deference of the defeated Catholic cantons of the Sonderbund (five rejected the constitution), the cantons retained a very high degree of substantial autonomy. The powers of the central level of government were limited to a few essential areas (Dardanelli and Müller 2019). In 1874 (and 1999), the 1848 constitution was fully revised, though not regarding the polity organizational framework that is still to a large extent based on the 1848 constitution. The 1874 constitutional reform and subsequent legislation introduced changes with significant effects for the Swiss political systems. Most importantly, the introduction of direct democracy, i.e., the possibility to submit each law to a popular vote in case of sufficient signatures, allowed Catholic conservatives to thwart the legislative agenda of the liberals. Simple majority politics therefore became impossible. The Catholics (and their state-society views) had to be integrated into the system, through participation in the government. Proportional representation was step by step introduced in all central authorities. Hence, the Swiss model of liberal power-sharing democracy contributed to cooling down the cultural struggle (Kulturkampf) between Protestants and Catholics, by now faded away. Federalism was key to this. It permitted procedural “in-between solutions” to the conflict: “Cultural issues were less settled than given time to cool down” (Linder and Müller 2021, 33–34).

Another aspect, constitutive of the very foundation of Switzerland, is Swiss multilingualism (Stotz 2006, 248–249). At the federal level, Switzerland is multilingual with German, French, Italian, and Romansh as recognized national languages (since 1848, except for Romansh that was added in 1938). At the subnational level, instead, it is each canton that decides autonomously on its own language policy. By choosing the official language, cantons determine the language used by their authorities, schools, and public services. Although they have the possibility to provide special language regimes for certain districts or municipalities, also with regard to nonofficial languages, they rarely make use of it (Belser and Waldmann 2015, 25). The cantonal competence in language policy is, however, constrained by Article 70(2) of the Swiss Constitution that requires cantons to take two criteria into account. First, they must respect the traditional territorial distribution of languages. Second, they must take account of traditional linguistic minorities. In actual practice, only few cantons are concerned by these constraints. Twenty-two of the total 26 cantons are traditionally monolingual (17 German-speaking, 4 French-speaking, one Italian-speaking) and, consequently, recognize only 1 language as the official language and, in most cases, school language. In monolingual cantons, the official language of the canton is simultaneously the official language of all local governments of the canton (Glaser 2020, 6). Three cantons are bilingual with German and French as official languages. One canton is trilingual, and the only canton recognizing Romansh as an official language. The application of the principle of territoriality, one pillar of the Swiss language policy next to the ones of individual language rights and support measures for the promotion of internal cohesion and cultural diversity (Article 70(3) Swiss Constitution), leads to a rather linguistic compartmentalization, with monolingual regions and identities (Lüdi 2007, 160). Such a compartmentalization, though affected by power imbalances, never led to major political problems (with one exception, the Jura case: formerly part of the canton Bern, the Jura canton was born following a cascade system of popular votes; Linder and Müller 2021, 38–40, 93–96). The reason for this is that federalism is grounded in the nonethnic concept of the Swiss nation, combined with strongly enforced proportional representation of all language groups at federal level that is ensured by (informal) quotas and that makes sure that the cantons’ identities are equally guaranteed.

Moreover, and this from a comparative viewpoint is the most important element of Swiss federalism, crosscuttingness has favored the longevity and success of Swiss federalism (Linder and Müller 2021, 39): “religious, linguistic and socio-economic cleavages do not coincide with the geographical boundaries of the cantons – instead, they cut across each other.” Political majorities vary from issue to issue. All groups experience being part of both a minority and a majority. For example, among French speakers, there are both Catholic and Protestant cantons, and among the economically poor cantons, there are both German- and French-speaking cantons.

Holding-Together Federalism and Evidence from Indonesia

Contemporary federalism follows a holding-together model. Holding-together federalism explicitly serves the purpose of accommodating ethnonational diversity. Following World War II, new sets of federal systems emerged across the globe as a result of the decolonization process and out of conflict resolution.

The collapse of European colonial empires led to the creation of numerous ethnically based federations. Permanent federal systems emerged in India (1950), Pakistan (1947), Malaya (1957)/Malaysia (1963), and Nigeria (1963). Federalization in the course of decolonization was, however, not always crowned with success. The federation with probably the shortest lifespan was the United States of Indonesia. The Indonesian case (discussed later) bears certain similarities to the federation of Ethiopia and Eritrea between 1952 and 1962, which was imposed on the former Italian colony of Eritrea by the United Nations and Ethiopia under the rule of Emperor Haile Selassie. The federation was dissolved by Ethiopia in 1962, and Eritrea was brought under the full rule of Ethiopia. This ultimately led to Eritrea’s independence in 1993. In addition, there are smaller insular micro-federations that also emerged in the 1970s and 1980s in the wake of decolonization, such as The Comoros, Micronesia, and St. Kitts and Nevis. Interestingly, smallness has not tempered secessionist urges. The case of the Union of the Comoros proves this. The Comoros is a Muslim country between Madagascar and Mozambique with Arab, French, and Comorian as official languages. Interestingly, the Comorian society was not divided over ethnicity, as in Nigeria and Ethiopia. It has been the mistrust among the elites of the three islands that is the reason for forming the Union, i.e., the incapability and incompetence of the islands’ leaders to share power (Garcia-Rojas 2014, 1890–1891). The fourth island of the archipelago is Mayotte, a French Overseas Territory that voted for remaining part of the French Republic when in 1974 the independence referendum was held in The Comoros (Alber 2020). The two-island federation of Saint Kitts and Nevis is one of the few federal states contemplating a right to secede (the Ethiopian constitution does so too). Interestingly, it allows only Nevis to secede. Since the 1990s, further federal systems arose out of conflict resolution processes, for example, South Africa (1996), Ethiopia (1995), Iraq (2005), and Nepal (2008/2015: Nepal established a unitary constitution in 1990 that did not recognize multiethnicity; in 2008, the Interim Constitution was amended to include the term federalism, but the federal structure was not established until the 2015 constitution).

In the main, these holding-together federal systems are ethnically based. They are premised on a need or will to prevent secession or the disintegration of the state otherwise. Indeed, without a risk that the state will break up, there is no holding-together.

The Example of Indonesia

The United States of Indonesia was proclaimed in 1949 and transformed into an authoritarian unitary state just 1 year later under Suharto (Goumenos 2008). After the fall of Suharto’s regime in 1998, Indonesia, one of the most diverse states in the world with over 30 provinces covering more than 18,000 islands that are home to over 300 ethnic groups and 6 official religions, adopted a “quasi-federalist approach” in dealing with its many ethnonationalist groups by following the leitmotif of “accommodation amidst strong integrationist tendencies” (Bertrand 2007).

Today, Indonesia is the country with the most extensive exercise in decentralization in Southeast Asia and one of the most extensive in the world, with a complex overlap of pluralities in society and in law (Wiratraman and Shah 2019). Centralized power was shifted to a decentralized model of politics, to the lowest governmental levels – cities and regencies – as opposed to provinces, where increased powers might fuel secession (Butt and Lindsey 2012, chapter 6). This became necessary because of interethnic violence and for the sake of cooling down secessionist claims.

The recognition of Indonesia’s plural character is also manifested in the special autonomies, almost constitutions, of Aceh and Papua. Aceh is a strongly Islamic province. Its autonomy goes much further than the general decentralization. The Law on the Government of Aceh 2006 grants Aceh asymmetric powers in many subject matters, especially religion, customary law, education, and natural resources. It also regulates central-local relations (Heintze 2018). Of Indonesia’s provinces, only Aceh has powers over religion, and only Aceh has the right to veto national legislation in its application to the province. Aceh’s sharia court jurisdiction goes beyond the rest of Indonesia and the Southeast Asian region, embracing both criminal and commercial law as well as the more obvious subjects of family law and succession, also resulting in laws that may not be in conformity with human rights as expressed in the Indonesian constitution (Butt and Lindsey 2012, 183).

The Papua special autonomy relates to Indonesia’s easternmost provinces of Papua and West Papua, where a low-level insurgency for independence has simmered for decades within the plural indigenous Melanesian communities. Two are the most significant features of the 2001 Law on Special Autonomy for Papua: first, the provision for return of 70% of oil and gas revenues to the province for 25 years (the proportion then reduces to 50%) and, second, as distinct from the Aceh law providing for religious powers, the creation of a Papuan People’s Assembly (Majelis Rakyat Papua, MRP) that must be consulted whenever the interests of indigenous people are concerned (the provincial government is responsible for the protection of customary law and indigenous land rights). On July 15, 2021, the Indonesian House of Representatives unilaterally passed the extension of the Autonomy Law for another 20 years (Supriatma 2021). It did so without involving the MRP and the provincial parliament. It proposed the revision of 16 articles and the addition of 2 new articles, with a total of 20 amendments that were made to the 79 articles of the 2001 Law on Special Autonomy for Papua. The revised law has serious implications for issues of decentralization and autonomy. According to Article 76, the central government can now decide on the creation of new regencies and districts. Two sections of Article 28 were omitted, resulting in the removal of the right to form local political parties. A new rule is now also in place. The vice president will have an office in the provinces to oversee the implementation of the Special Autonomy Law. Discontent and uprisings persist.

Evidence from Asia

Malaysia, India, and Pakistan each established a federation following the decolonization process. Both India and Malaysia have what can be described as ethnoterritorial federal systems, where the dominant group is split across several subnational units, while other units are designated – by name and/or special rights – for particular ethnic groups. Pakistan has an ethnic federal system, although it too has considered amending its structure in a way that would split the dominant group across more than one province (Mughal 2020). All of Asia’s federal systems were formed through a holding-together process, although Malaysia is an example of both aggregation and disaggregation. This means that they tend to be centralized, ethnically based, and asymmetrical (Breen 2022).

India’s federal system has been celebrated as an example of a successful multinational democracy in a large state (Stepan et al. 2011; Lijphart 1996; Adeney 2017; Tillin 2017). Established out of the decolonization process, and shaped by the scars of partition, India’s federal system has, in the main, managed to accommodate many of its different ethnic groups. India comprises 28 states and 8 territories. Further, India has constitutionally entrenched local governance and the largest set of participatory and deliberative institutions in the world through the village assemblies (Gram Sabhas) (Sanyal and Rao 2019). This local level enables for more targeted and culturally sensitive policies and programs to be developed.

One important feature of India’s constitution that sets it apart from most other federal constitutions is the ease by which its constitution can be amended to alter, abolish, and create new states and territories. Although conventional federal theory would have it that the rigidity of a constitution, in particular, the protection of the numbers and boundaries of the subnational units (which normally can only be changed with the consent of the unit(s) affected), is an important if not essential protection of subnational rights, the case of India suggests otherwise. In India, a new state can be unilaterally created, or boundaries amended, by the central parliament. In addition, the upper house of parliament is population-based, and there is not an equal representation of states, which weakens the protections that smaller states tend to have in bicameral federal systems. In 1956 India reorganized its states and territories on the basis of language and then progressively added more states and territories, which was considered an essential step to recognize diversity (Singh and Kukreja 2014; Shneiderman and Louise 2015; Breen 2022).

But there are still many groups that do not feel accommodated by the federal arrangements. Conflict has been ongoing in both the northeast and the northwest of the country. In the northeast, groups such as the Naga have been in conflict with the state for much of its postcolonial existence. New states, for example, Assam in 1950 and Nagaland in 1963, have been established to accommodate tribal communities, but these are not considered to be sufficient and have been coupled with militarization and elite co-option (Kikon 2005). In the northwest, the region of Kashmir remains disputed between Pakistan and India. Further, under Prime Minister Modi (2014–), India has lurched to the right. The Modi government has sought to reshape India as a kind of Hindu nation-state, through a series of education, immigration, and other legal reforms (Tillin 2017; Adeney and Bhattacharyya 2018). This imperils the future of federalism in India, although it would appear strong enough to withstand this assault and flexible enough to address existing grievances into the future.

Pakistan is an ethnic federation with a bicameral parliament and two levels of constitutional government. Like India, it was formed out of decolonization and partition, and it too changed its provincial makeup accounting for ethnic diversity. In 1956, 9 years after independence, Pakistan enacted its first constitution and merged the four provinces and one tribal region in West Pakistan. This was ostensibly in order to create parity with the Bengali-speaking East Pakistan (which was one province) and a more unified ethnically blind national identity, establishing a kind of bipolar federation (Adeney 2007). But a military coup in 1958 and the entrenchment of division between East and West resulted in a catastrophic war between its East and West regions and ultimately the secession of East Pakistan, now Bangladesh, in 1971. This was not before the provincial model was again restructured as the original four ethnic provinces and one tribal area were reestablished (1970). This structure has been retained with the addition in 2010 of new regions into the tribal area and the renaming of one province to reflect its ethnic identity (Khyber Pakhtunkhwa) (Adeney 2012) and the merging of the Federally Administered Tribal Areas with Khyber Pakhtunkhwa in 2018. The federal government has, however, been dominated by the Punjab province which supplies the majority of MPs to the lower house of parliament. The Imran Khan government, elected in 2018, has a policy to split the Punjab province, to better reflect its ethnic diversity, establishing a new Seraiki province.

Malaysia is a bicameral ethnoterritorial federation, comprising 14 states, 10 that are majority Malay, 1 that is plurality of Chinese ethnic group (Penang), and 2, Sabah and Sarawak, that are majority Dayak (indigenous peoples of Borneo) (Harding 2022; Teik 2005). Malaysia was formed by the merging of the Federation of Malaya with Singapore, Sabah, and Sarawak. Singapore was expelled after less than 2 years, in part because of the significantly different ethnic makeup, which was perceived to be a threat to Malay hegemony (Barr 1997). Under the constitution, Sabah and Sarawak have a range of special autonomy rules and asymmetrical provisions, including special language rights (e.g., Article 161A), special grants and revenue (Parts IV and V), and a range of authorities that are supplemental to those allocated to the other states according to the schedules of the constitution (Lists IIA and IIIA, Ninth Schedule), including customary law (Article 76). Notwithstanding, although the authorities of the states are secure, the division of powers is heavily weighted toward the center, with few substantial authorities allocated to the states (Harding 2019).

Indonesia, the Philippines, and China all have a constitutional protection of subnational units, but they remain fundamentally unitary. Notwithstanding, both Indonesia and the Philippines can be classified as constitutionally decentralized unions as the subnational governments are constitutionally protected, and each has special autonomy for one or more special regions (Aceh, Papua, and West Papua in Indonesia and the Bangsamoro Autonomous Region in the Philippines) (Breen 2020). China has constitutionally protected provinces, including five nominally autonomous regions that are ethnically based. However, China does not have constitutionalism, in that the Chinese Community Party is above the constitution, and many of the autonomy provisions for the ethnic provinces remain unimplemented (Buhi 2018).

A set of three “emerging federal systems,” Nepal, Myanmar, and Sri Lanka, are at various stages. Each can/could be classified as an ethnoterritorial federation. Nepal in 2015 established a federal constitution (a federation) and held successful central, provincial, and local government elections in 2017. Nepal has established three levels of government, each with their own constitutional division of powers. There are seven provinces. One can be characterized as ethnic, while another (three) have a plurality of minority ethnic groups. There are 753 local governments. Nepal is described as a kind of hourglass federalism, whereby the provinces are poorly resourced and weakly empowered in comparison with the central and local levels of government (Payne and Breen 2022). In addition, Nepal has incorporated the principle of self-determination in its constitution, and it has a process defined by which a territorially clustered ethnic group may establish an autonomous region (see Article 56(5); 295(3)). As of 2021, no autonomous regions have been established.

Myanmar and Sri Lanka are, on paper, ethnoterritorial constitutionally decentralized unions. Both constitutions establish subnational units (7 in Sri Lanka and 14 in Myanmar) that have a constitutional division of powers. Myanmar’s constitution of 2008 also includes self-administered zones (a kind of special autonomy) for smaller clustered groups and a form of non-territorial ethnic representation through ethnic or “national race” affairs ministers. In both cases, there exists considerable pressure for further federalization, which has been the source of a significant social and political cleavage, contributing to “backward steps” (Breen 2018). In Sri Lanka, the Rajapaksa family was returned to power in presidential elections in 2020, on the back of an “uncompromising” commitment to a unitary state and a series of constitutional crises in 2019 (Breen 2019). In Myanmar, a military coup in February 2021 stalled plans to strengthen the country’s federal system (Breen and Decobert 2021). The new military claims to be acting in accordance with the constitution (which it authored) and has promised an election by 2023. The future of federalism in Myanmar and Sri Lanka remains uncertain.

The Pacific Islands region has also faced considerable challenges associated with ethnic division and conflict. Papua New Guinea is highly diverse and a federal state, comprising 20 provinces, 1 special autonomous region (Bougainville), and the capital district. Both provincial and local government levels are protected by the constitution. In 2019, a nonbinding referendum on independence was held in Bougainville, and the federal parliament has since agreed to a timetable for independence by 2027. Bougainville is both culturally and geographically distinct, and it had been in conflict with the state until reaching a peace agreement in 2001. The Solomon Islands has also considered establishing a federal system to reflect its ethnic and geographic diversity.

There are two federal systems in the Middle East (United Arab Emirates and Iraq), one of which is an ethnic federal system. Iraq is a holding-together ethnic federation (some might say putting-together, given that the federal constitution was introduced by the US-led provisional authority in 2005; see, e.g., Alkadiri 2020). It was “clearly designed as an ethnofederal state” (Danilovich 2016, 3). However, as Belser (2021) convincingly argues, there are significant shortcomings that mean that it is both de jure and de facto a unitary state with one autonomous region. There are 18 governorates, which operate at the provincial level. The 2005 constitution provides an option for governorates to join together and form a semiautonomous region. To date, the Kurdistan Region is the only semiautonomous region, consisting of three governorates, and there has been considerable resistance from the center to moves by some governorates to establish additional regions. This is important because, although the constitution allocates residual powers to regions and governorates, the powers of the governorates not incorporated into a region are to be regulated by law (which was passed in 2008). Thus, it is argued that governorates are reduced to mere administrative units of the central government (Belser 2021). The Kurdistan Region has its own constitution and national assembly and has jurisdiction on all matters except foreign relations, security, defense, and fiscal matters. The Federation design is a compromise between Kurdish demands for independence or, at the least, a confederal structure, and the desire from the then dominant elite for a more integrated structure, but it remains a divided state and its federal future is tenuous (Alkadiri 2020; Fleet 2019).

Evidence from Africa

Federal constitutions have been established in several states in Africa as a part of ethnic conflict resolution and to hold-together the state, including Ethiopia (1994), South Africa (1996), the Democratic Republic of the Congo (2006), Kenya (2010), and Somalia (2012) (Steytler 2016; Fessha 2016, analysing South Africa and Ethiopia). Like in the Asian region, federalism was initially established in several states by departing colonial powers (e.g., Mali Federation, The Federation of Arab Republics, Rhodesian, Nigeria) to respond to ethnic diversity. However, this meant that federalism was regarded with hostility by many postcolonial leaders (Kefale 2019), and, besides Nigeria, none of these federal systems sustained (Fessha 2012). After the 1990s, South Africa and Ethiopia established federal systems, setting off a wave of new devolutionary reforms (Steytler 2016). But even then, most states, including South Africa and Kenya, avoided using the term federal because of its historical connotations – “federalism is regarded as the ‘F’ word and it is carefully avoided” (Fessha 2012, 273). Although the management of ethnic diversity has been high on the agenda in each case, not all these new federal systems have been designed to accommodate ethnicity. Several can be regarded as ethnic federal systems, including Ethiopia, Somalia, and the Democratic Republic of the Congo, while Nigeria’s and South Africa’s federalism has been aimed more at integrating and cutting across ethnic division. Sudan and South Sudan have also been facing pressure to federalize. Federalism, and especially autonomy in the South, has long been a controversial matter in Sudan and South Sudan. Sudan became a formal federation following a military coup in 1989. A federal constitution enacted in 1994 established 23 states. But this period was punctuated by conflict and the centralized and autocratic rule of Omar al-Bashir (Johnson 2014). A peace agreement in 2005 included an agreement for a referendum on independence for southern Sudan (comprising ten states). The referendum, held in 2011, was approved by 98% of voters and South Sudan became independent later that year. This risk of secession that is often associated with ethnic federalism played out in this case. Both countries (Sudan and South Sudan) remain committed to federalism but are mired in conflict and are debating the particulars of further reforms (e.g., see Elhaj 2021).

Nigeria has maintained a federal system, in one form or another, through several constitutions, military coups, and substantial territorial restructuring. Nigeria’s system offers a different approach to ethnic federal states elsewhere in Africa and Asia. It is territorial in design, even anti-ethnic, with the aim of integrating different ethnic groups and creating a territorial pluralism that cuts across different ethnic rivalries (Imuetinyan 2017; Suberu 2009; Horowitz 2001 [1985]). In some respects, it has been successful, limiting secessionism and fostering integration (Suberu 2009). But it has also been subject of continued demands for, and creation of, new states to better accommodate diversity. Beginning with three heterogeneous regions that were otherwise dominated by one large ethnic group, Nigeria evolved to reach 36 states. Most of the provinces are clearly associated with an ethnic group, but none are named as such, and major groups are distributed across several states (Dent 2000).

Nigeria has also enacted a “Federal Character” principle in its 1979 and 1999 constitutions and established a Federal Character Commission to oversee its implementation. The principle has contributed to there being broadly equal provinces and some ethnic distribution of official appointments (Dent 2000). However, overall, the principle and the commission are argued to have made little impact due to the path dependence of the ethnic character of institutions (Benson and Audu 2021; Demarest et al. 2020). Babalola (2017) further argues that Nigeria has become a “federation without federalism,” because of the increasing centralization and large vertical fiscal imbalance that is based around oil revenue and which makes states almost completely reliant on the center. Suberu (2013) characterizes it as a federal-unitary hybrid, for these reasons and due to the dominance of a single political party and the integrated administrative system.

Ethiopia is another federal country where ethnicity has played a major part in its constitutional reformation – “ethnicity is the foundational, organizational, and aspirational principle” (Arban and Dirri 2021, 372). Ethiopia is formally a federation comprising 11 ethnically based regional states (kililoch) and 2 self-governing administrations (astedaderoch). It has a bicameral parliament, with a directly elected lower house and an indirectly elected upper house (House of the Federation) that is representative of the state assemblies/councils (the state councils are directly elected and select from among their members a president who then forms an executive committee, from among the council members). The upper house is quite limited in its function, in comparison with other federal upper houses of parliament (Arban and Dirri 2021). Significantly, the upper house is representative of the “Nations, Nationalities, or Peoples” rather than the state per se. Article 61(2) of the constitution states that “Each Nation, Nationality and People shall be represented in the House of the Federation by at least one member. Each Nation or Nationality shall be represented by one additional representative for each one million of its population.”

Ethiopia’s constitution provides for its states to enact their own subnational constitutions and systems of government (Article 52 (2)). Notwithstanding, the states’ constitutions are mostly uniform, though some states have established administrative and political arrangements that account for their own internal diversity, for example, the Southern Nations and Nationalities and Peoples state (Vaughan 2013, 432–436). Article 39(1) states that “Every Nation, Nationality and People in Ethiopia has an unconditional right to self-determination, including the right to secession.” Article 47(2–3) establishes the process, including the requirement for majority support in a referendum held “in the Nation, Nationality or People that made the demand” (Article 47(3b)). This process was tested recently with the establishment of two new regions, one in 2020 (Sidama) and one in 2021 (Southwest Ethiopia Peoples’ Region), following referenda that gained more than 95% support. Nevertheless, tensions between states, and states and the center, remain very high. In 2020, a civil war erupted between the Tigray Region and the federal government.

South Africa had one of the most studied constitution-making processes, which reformed the country’s apartheid system, and established a hybrid federal constitution establishing three levels of government, incorporating 9 provinces and 278 municipalities, comprising 8 metropolitan, 44 districts, and 226 local municipalities. As Steytler (2013, 443) argues, the “hybrid nature of the country was the result of political compromise to end white minority rule” as well as to head-off Zulu secessionists. So, while most holding-together ethnic federal systems have been established to the benefit of minorities, South Africa has had different ends in mind. The constitution allocates powers to provincial levels of government, but most of them are concurrent with the center. The provinces have no constitutional taxation powers (Steytler 2017, 444), and local governments have a right to govern local affairs and to raise revenue. Indeed, Steytler (2017) has labeled South Africa’s federalism as being an “hourglass,” with weakly empowered and resourced provinces “squeezed” between a dominant center and a strong local level.

South Africa’s federalism has also been labeled “watermark ethnic federalism.” In other words, internal boundaries were territorial – based on geography, economic viability, etc. – but, in practice, were aligned with the basic structure of ethnolinguistic communities (de Visser and Steytler 2018, 3). This is argued to have worked to defuse ethnic mobilization while still retaining some sense of ethnic belonging. The South African constitution also includes a right to self-determination, but this has never been enacted in law, as is required before it becomes operational. Arban and Dirri (2021, 369) argue that the idea of national solidarity is underpinned/supported by the principles of cooperative federalism (“cooperative government”) that requires all levels to work together and that South Africa is “seeking to build a common national identity…, but not at the expense of ethnic diversity.”

Kenya, since its 2010 constitution, is a formally federal country comprising 47 counties. One key way that this has been approached has been through flexibility. Kenya’s federal system is unique in its subnational structure. Instead of devolving power to large provinces or small local governments, Kenya established counties as the second-tier level of government. The counties have devolved constitutionally protected powers from the center, a directly elected governor, and a unicameral legislature. They are responsible for administering agriculture, health, facilities, sanitation, transport, and trade licenses. County governments receive revenue from the central government and have responsibility for raising their own revenue. There are special provisions to ensure that legislatures are inclusive of marginalized groups and genders. These arrangements were aimed on the one hand to accommodate diversity by dividing up major ethnic groups but, on the other, to undercut secessionist ambitions by making subnational units that are fiscally reliant on the center and too small to accumulate significant power and resources (Steytler 2016; Cheeseman 2016). Nevertheless, the boundaries of the 47 counties were based on old colonial boundaries, which in turn were based on ethnicity (de Visser and Steytler 2018, 4). Kenya also has a bicameral legislature. The upper house, the House of Federation, is directly elected and representative of the counties.

Somalia has also established an ethnic federation, but the country is war torn and divided. Somalia is a formally federal country comprised of five federal states established after 2012, plus the claimed territory of Somaliland. It is also divided into administrative regions and districts. On paper, Somalia has a parliamentary system with an indirectly elected president and a prime minister commanding a majority of the lower house of its bicameral parliament. However, there have not yet been elections under this system (the current president was indirectly elected in 2017, and a direct election was planned to be held in 2021, but it has been postponed repeatedly), and not all federal states recognize the central government. The upper house is representative of the different peoples, indirectly elected by an electoral college selected by 136 traditional clan elders. In Somalia, federalism exists on paper but not in practice.

Evidence from Latin America

In Latin America, a region that is marked by deeply hyper-presidential forms of government and in which very few constitutions last over a hundred years, since the late nineteenth century, four countries have consistently identified themselves as federations in their constitutions: Argentina, Brazil, Mexico, and Venezuela (for an overview regarding the allocation of powers in these federations, see Serna de la Garza 2000; in short: although the residual clause is in favor of the states/provinces, the actual allocation of legislative powers is in favor of the federal legislature, making all systems highly centralized).

Historically, Latin American federations are the result of long-term and complicated processes in the wake of the achievement of independence from Spain and Portugal that, unlike the British colonial power in Canada and the United States, tightly governed Latin America by a sophisticated hierarchy of secular and religious controls. From a comparative viewpoint, Mazzuca (2021) contrasts Latin American state formation with Western European and other experiences and offers helpful explanations for a better understanding of the troublesome nature of Latin American federations. Unlike in Europe, where (parts of) countries throughout the nineteenth century had to defend themselves to survive, Latin American countries did not face the same level of external threats. Instead of building internal (cooperative) state capacities, state formation in postindependence Latin America focused on international trade. The benefits of international trade, however, led to states with chronic weaknesses, notably patrimonial administrations and dysfunctional regional combinations.

When dynamic commercial hubs in Argentina, Brazil, and Mexico finally annexed backward areas that were governed by patronage elites that chronically sapped national treasuries, “genuine state formation” in Latin America came to a stillstand. This is the reason why initial federal constitutions were followed by periods of instability and a reestablishment of a federal state after some decades. Concretely, the Argentine federal state has been created in stages since 1831 via the constitution of 1853 and further amendments throughout the nineteenth century. Brazil was established as a federal state in 1891 only, after a lengthy process beginning with independence in 1822. Venezuela was constituted as a federal state immediately after independence in 1811 but was finally reestablished as a federation in 1864. Mexico established itself as a federal state after independence in 1810. A first federal constitution in 1824 was followed by periods of instability and the subsequent reestablishment of a federal system in 1857, which still exists today based on the 1917 constitution. The short-lived phase of the United States of Colombia from 1863 to 1886 is also worth mentioning (Colombia is discussed later).

When considering the four countries that identify themselves as federations in their constitutions, three issues are noteworthy: First, these federations are aggregative in nature, and, in their conception, they are clearly modeled on the US dual federalism. However, and this is the second and most important issue, they have been subject to strong centralization. Though to a different extent from one to another case, within the federal practices of these countries, constituent units throughout the twentieth century fell victim to either military-led (in the case of Argentina, Brazil, and Venezuela) or civilian authoritarianism (in the case of Mexico). This, however, does not necessarily mean that central governments are strong in actual politics. Scholars differently discuss the balance of power between national and subnational governments and the issue of the relative leverage that each wields over the other (Eaton 2019). The third issue worth noting in Latin America is the role of the local government. Unlike for the classical federal systems of the United States and Canada, one can speak of a tripartite federalism in Latin American federal practices: The local level of government plays a relevant role. This bears similarities with the establishment of recent federal systems in other world regions (e.g., South Africa, Kenya, or the Philippines and Indonesia). From a conceptual viewpoint, it is worth recalling the definition Keith Rossen (1994, 5) gives on federalism in Latin America: “a form of government in which sovereign powers are constitutionally divided between a central government and geographically defined, semi-autonomous levels of government” (emphasis added by the authors).

Whether or not Latin American federations qualify as federal or not, this of course depends on how federalism is defined. As a rule, the federal nature of states in which constitutionalism, and the rule of law, has not developed strong roots is, strictly speaking, questionable. The primary aim of this contribution, however, is not to problematize the issue whether federal constitutionalism has failed or not but to give evidence on the topic of federal systems and ethnic relations. And Latin America has interesting insights to offer in this regard. Since the late 1980s, many Latin American systems have adopted new constitutions or introduced major reforms that address the exclusion of indigenous peoples from governance (rooted in colonialism, Altmann 2019).

According to Rodrigo Uprimny (2011, 1599), the recent constitutional reforms prove that a first true wave of (multicultural) constitutionalism is somewhat ongoing. Raquel Z. Yrigoyen Fajardo (2015) has identified three cycles of recent Latin American constitutional reforms. First, the multicultural cycle from 1982 to 1988: As a forerunner, the 1988 Brazilian constitution in Chapter VIII of title VIII specifically covers indigenous communities and, essentially, deals with the recognition of certain rights (no other Latin American constitution had dedicated a whole chapter to the subject beforehand). Second, the pluricultural cycle from 1989 to 2005: The recognition of indigenous rights was deepened in policy fields such as languages and bilingual education. The main innovation, however, came from the recognition of the principle of legal pluralism. Examples of constitutional reforms include the ones in Mexico (1992) but also Paraguay (1992) and Peru (1993). Third, the plurinational cycle from 2006 to 2009: Indigenous peoples are recognized as original nations, as collective subjects of rights, with autonomy and self-government. Examples are the 2008 Ecuadorian and the 2009 Bolivian constitutions (discussed later).

Overall, the recent constitutional reforms have established highly decentralized systems that, though still badly implemented in practice, are of high interest for federal scholars because of their institutional and democratic innovations. More specifically, they offer interesting examples for comparative studies dealing with federalism as a tool of conflict resolution at the one side and comparative studies dealing with federalism and democratization processes at the other side. Though yet understudied, different scholars in recent years confirm this. They of course acknowledge the per se unitary nature of systems such as Bolivia, Peru, and Ecuador, but they also conclude that the results of recent constitutional reforms, i.e., significant decentralization, provide an ideal starting point for making federal-like multicultural/pluricultural/plurinational constitutionalism a reality. They refer to the relevance of democratically elected plural regional and local identities at the one side and to plural, inclusive good governance processes at the other side (Gordin 2021; Gordin and Renno 2018; Eaton 2020 on Colombia; Bertel 2014 on Peru). Of course, at large, these reforms serve the purpose of good governance, i.e., development by democratization (Barrera et al. 2010). However, as empirical evidence suggests, this cannot be decoupled from multicultural/pluricultural/plurinational constitutionalism as a means to deal with territorial and societal pluralism. Indeed, the constitutional reforms are to be considered as the epitome of what has been termed new Latin American constitutionalism, multicultural constitutionalism, or diversity constitutionalism (Uprimny 2011, 1589). Viewed positively by most scholars, such constitutionalism is, however, also critically discussed (as a toxic cocktail of cultural relativism and accommodation, see Aguilar Rivera 2014).

Of course, when analyzing the systems in more detail, there are important differences from one system to another. If, however, the constitutional reforms are viewed through the lens of ethnic-indigenous relations, then all reformed systems somehow include at least three of the following five elements (Tomaselli 2021a): “the recognition of the multicultural nature of Indigenous societies and the existence of Indigenous peoples as distinct sub-state social collectivities, the recognition of Indigenous customary law as part of the ordinary law, rights to protect Indigenous property from collective sale, dismemberment or confiscation; the grant of the status to or official recognition of Indigenous languages [though there are more, the Bolivian constitution for example establishes 36 indigenous languages] and the guarantee of bilingual education.”

In actual practice, multicultural/pluricultural/plurinational constitutionalism, along with the recognition of equality between different religions including indigenous ones, in most contexts is far from becoming a reality. Yet, the constitutional reforms have significantly changed the understanding of national unity across the region. Rodrigo Uprimny (2011, 1589) argues that all reforms “emphasize that unity is not accomplished by a homogenization of cultural difference, …, but by a sharp appreciation of differences and a greater approval of pluralism in all its forms.” This, for example, includes giving special protection that, in part, also comes with a differentiated citizen status and with special, asymmetric forms of autonomy granted to persons belonging to indigenous and Afro-Latin communities (Wade 2010).

From a conceptual viewpoint, Rodrigo Uprimny (2011, 1590–1591) argues that the Latin American constitutional shifts go well beyond the scope of liberal multicultural constitutionalism, moving toward a different constitutional form that is multinational, intercultural, and experimental in nature. The constitutional recognition of such multinationalism, along with openness to international human rights law, has led to what he terms as “the erosion of both the traditional system of legal sources and the central role played in the past by law and government regulation within national legal systems” (ibid, 1592). Two ways led to such a paradigm shift: first, the application of ideological formulas like those of postwar European constitutionalism and, second, the development of own formulas such as the one of plurinationality in the cases of Bolivia and Ecuador. The Ecuadorian and the Bolivian constitutions of 2008 and 2009, respectively, suggest the existence of a nation of peoples or a multinational state. The two constitutions, unlike most other constitutions (e.g., the Mexican and Argentinian), widely recognize the collective rights of indigenous peoples (see Article 2 of the Bolivian constitution and Article 57 of the constitution of Ecuador). They mark the Latin American “turn to plurinationality” (Centellas 2013 speaks of a liberal-pluralism trend), a turn that is also under discussion in the participatory constitution-making process in Chile that started in 2020 with a historical referendum by asking Chileans whether they wanted a new constitution (a vast majority, 78.2%, voted yes, also for the establishment of a constitutional assembly composed of 155 members, with 17 persons belonging to Chile’s 10 recognized indigenous peoples; the assembly was elected on May 15–16, 2021) (Tomaselli 2021b; Negretto 2021).

As in Bolivia, which fought a deadly “federal war” between the regions at the end of the nineteenth century, the association between federalism and interregional conflict in Colombia has always been troublesome. In 1991, Colombia, the third largest country in Latin America after Brazil and Mexico and in terms of land mass the fourth most territorially extensive country of the region, shifted from a system that combined shared rule with limited self-rule to a system with the exact opposite configuration (Eaton 2020). In essence, shared rule was curtailed through electoral reforms that reconfigured the senate (100 senators are elected from a single national constituency and the remaining 2 in a special national constituency for indigenous communities; on the inability of second chambers to represent territorial interests due to a political-parliamentary logic that supersedes the territorial one, see Palermo 2018). By withholding shared rule, a pattern that according to Müller (2017) is widespread as “it is one thing to decentralize power to a region over its own territory and population (and thus hand over a certain degree of responsibility and the duty to self-finance the new tasks), [but] it is quite another to grant regions a say in national matters which are often deemed too important for local idiosyncrasies (e.g., defense, macroeconomic policy, social welfare),” subnational territories have been deprived of institutional levers at the national level. In other words, they do not anymore possess the levers to veto central policymaking, i.e., to secure self-rule. Thus, in the case of Colombia, democratization has undermined federalization (Eaton 2020, 4). The case proves the reverse of what literature on subnational authoritarianism in Argentina, Brazil, and Mexico has demonstrated in relation to the question of how federalization can undermine democracy (Eaton 2019). The 1991 constitutional reform, i.e., the turn to self-rule while hollowing out shared rule, along with a weak constitutional court, thus has merely to be interpreted as being part of the “global zeitgeist of decentralization” (Eaton 2020, 21).

Another paradox is noteworthy in the case of Colombia. It regards indigenous peoples. With the 1991 constitution, the rights of indigenous peoples together with those of other marginalized groups were constitutionalized. Several articles recognize the multicultural nature of Colombia and the special status of indigenous peoples (Herrera 2021): for example, Article 330 provides the establishment and functions of self-governing councils in indigenous territories. Since 1991, subsequent legislation regulates details. The constitutional court has also developed and given scope to the constitutional provisions through a rich case law, with precedents resonating beyond the country’s borders (ibid). Such a proactive judicialization of the indigenous question has, however, had adverse side effects too. As the constitutional court has established free, prior, and informed consent as a fundamental right in regard to issues concerning indigenous territories, a special type of legislation and procedures is required. This per se is positive. However, in the case of Colombia, with no real shared rule, this leads to normative and regulatory paralysis and thus to the situation that indigenous peoples do not have control over illegal activities that threaten their lands anymore.

To conclude, decentralization in Latin America may not convert unitary systems into federations anytime soon in Latin America, nor will subnational authoritarianism fade out anytime soon in Latin American federations. Constitutional reforms have, however, pushed many unitary countries in a federal direction, with two positive effects. First, subnational officials may directly challenge the policy direction of the national government. Whether they succeed, so argues Eaton (2018), depends less on the institutional design and more on the underlying strength of the central authorities to monitor and punish these kinds of subnational deviations. Second, federal systems offer multiple ports of entry for developing a truly decolonial approach between states and their levels of governments and (territorial autonomies/rights of) indigenous peoples. They thus also allow to overcome the mere recognition of legal pluralism as a constitutional principle by putting it in action.

Final Remarks

This contribution has given evidence of federal theories and practices with a special focus on ethnic relations. It has shown that the more federal states were established in the nineteenth and especially in the twentieth century, the more federalism as a governance tool was studied, proposed, and ultimately also used as a tool of conflict resolution. The reason for this is simple. In the main, federalism is an instrument that deals with territorial and societal pluralism, in the presence but also in the absence of ethnic relations. In other words, managing (ethnic) complexity is the backbone of recent federalism and similar tools such as decentralization and regional/local autonomy.

Many of the federal and decentralized systems discussed in this contribution share certain common and novel traits. This conclusion refers to three.

First, most systems have only recently committed to constitutional federalism, i.e., the constitutional allocation of powers, resources, and competences, and to the rule of law. In doing so, they oftentimes have put too much attention to self-rule and too little to shared rule (legislative and executive structures and dynamics involved in intergovernmental relations as well as the role of agencies and intergovernmental agreements are extensively discussed in Poirier et al. 2015). This makes the systems fragile, with conflicts that persist. In the absence of adequate bodies and procedures that regard the participation of constituent units in decision-making at the central level of government and their involvement in intergovernmental relations, any commitment to constitutional federalism is, in actual practice, confined to the spirit of the law. But institutional and democratic innovations are nevertheless on their way, as this contribution has demonstrated by zooming into systems in different world regions. The reason is simple: There is no other solution to solve conflicts and govern (ethnic relations) in a truly participatory and plural manner.

Second, many of the systems discussed only recently recognize their multiethnic character in an official manner. For constitutional federalism to be effective, such a recognition and the actual appreciation of ethnic groups and (indigenously guided) societal pluralism are crucial. Of course, this oftentimes involves the recognition of legal pluralism, even at the level of legal sources. Political conflicts stemming from irreconcilable underlying principles may arise, especially in the case when liberal principles of Western federal democracies collide with illiberal principles of competing tribal or religious rules (Ryan 2020, 493–494). However, as “federal systems are founded on the premise that multiple sources of sovereign authority create simultaneous normative forces on the legal actors within them,” federalism is “the realm in which legal pluralism is least controversially and most undeniably made manifest” (ibid, 495). Ultimately, it is federalism with all its structures and procedures that allows for valuable cross-jurisdictional platforms in which pluralist deliberation and innovative policymaking may occur, regardless of its per se statist nature (that tends to sidestep vexing questions such as the relationship between state and non-state sources of normative authority).

Third, asymmetry is a near unavoidable outcome of the holding-together federalization process in multiethnic states. Negotiating parties bring to the table different strengths and capacities, cultural and geographical distinctiveness, and different histories that are inevitably institutionally reflected in some way. Although asymmetry is often a compromise and an outcome of conflict resolution, it tends to have a positive effect on relations between ethnic groups and on the ability of the state to hold-together. Asymmetry is a feature of many new political systems both unitary and federal (e.g., Indonesia, Iraq), as well as those that followed the decolonization process (e.g., India). The challenge is to not use asymmetry to create additional separation between existing states and regions but to ensure that asymmetrical regions are adequately or additionally involved in shared rule at the center.

To conclude, the evidence given in this contribution proves that under conditions of ethnic complexity and division, federal states become more truly federal while many formally unitary countries become less genuinely unitary. Decentralization reforms may turn them into de facto federal systems. As the examples from Asia, Africa, and Latin America have shown, those world regions have emerged as an especially important arena in which to ask what difference federalism makes (in the presence of societal pluralism). Institutional design matters. So do political cultures and society.

Cross-References

• Democracy and Rule of Law

• Federalism

• Federalism and Democracy

• Political Ideologies and Ethnic Identity in Europe