Proceduralizing territorial conflicts in the name of constitutionalism

When conflicts arise, the binary alternative is to either fight over them, or to (try to) accommodate them. The legal challenge is to find procedural ways in which conflicts can be channelled. The paper looks at such possible ways based on the comparative practice. It also argues that the EU, while not best suited to adopt such rules, should encourage their adoption in its function of promoting the rule of law

The question and the legal answer

Territorial conflicts do arise as a matter of fact and require a response as a matter of law. As experience shows, secession takes place irrespective of its constitutional permissibility. While most constitutions are either silent on or explicitly forbid secession (normally by protecting the territorial integrity of states), and just a few contain clauses (trying to) regulate it (Ginsburg and Versteeg 2019), about 30 new states have been created since 1990 (with a particularly high number following the fall of the iron curtain), almost 60 in the last 50 years and 129 since 1941 (Baldacchino and Hepburn 2013).

Conflicts over sovereignty on a territory are politically and emotionally loaded. Whether a conflict is justified and who is right in a conflict are matters that entail moral and political arguments. At the same time, the overall context poses law in an awkward position: on the one hand, a strict implementation of positive law in most of the cases (i.e. except in the exceptional circumstances in which secession is allowed under customary international law) would prohibit any attempt to legally challenge sovereignty over a given territory, based on the innate principle of self-preservation of states (Sunstein 1991: 633). On the other hand, as advocates of the functional approach argue, “demonizing secession, turning it into a constitutional taboo, often adds fuel to secessionist claims” (Mancini 2012: 482). And in fact, as the main function of law, in each of its areas, is to solve conflicts, a mere dismissal of claims might exacerbate conflicts rather than overcoming them. As the history of constitutionalism is the story of gradually placing under the rule of law phenomena that were previously left to the rule of force (to the “normative power of the factual”, in the words of Georg Jellinek), it is not surprising that contemporary constitutional law devotes growing attention to conflicts of sovereignty, with a view to constitutionalising them (Haljan 2014, Belov 2021). Indeed, from a legal point of view, the only alternative to a conflict are procedures. Therefore, irrespective of the stand that anybody might take on each particular territorial conflict, a legal response cannot but be procedural in nature.

Procedures and their distorted implementation? The case against the monopoly of referendums

After the Canadian Reference, which inaugurated a functional and procedural approach to territorial conflicts of sovereignty, one would have expected that most of the countries experiencing secessionist requests would have adopted a similar complex legal framework to encapsulate secession into a sophisticated procedural shell. However, more than twenty years down the road, this has not been the case. Indeed, a growing number of recent constitutions include procedures to deal with secession of part of the territory with a view of taming territorial sovereignty conflicts (Haljan 2014). But overall, rather than establishing procedural and substantive checks and balances regulating secession, what has happened is a proliferation of referendums on sovereignty (Mendez and Germann 2018), very often unofficial and thus contested. The Canadian Supreme Court’s call for balanced rules that take into account the principles of rule of law, federalism and protection of minorities in determining the legal framework for possible negotiations over secession remained confined within the Canadian borders. Elsewhere, such principles, while occasionally referred to in judicial rulings on secession claims (Martinico 2017a), have been trumped by the use of the referendum as (if not the only by far) the dominant means for constitutional change, or rather for constitutional challenge, ie. for secession claims.

Following the predominant approach, the overwhelming majority of constitutions does not provide for a right to secession and very often they rule it explicitly out by calling for eternity, unity, indivisibility, indestructibility of the state (Novic and Priya 2016, Ginsburg and Versteeg 2019). In very rare and exceptional cases, constitutions do contain provisions allowing for secession. These can provide for an undefined right to freely secede — such as in the case of art. 72 of the Soviet constitution of 1977 [2]2 — This article was specified by a law as late as in 1990, when the dissolution of the Soviet Union was already in place (Bowring 2015). or of (the purely on-paper provision of) article 74 of the Uzbek constitution in the case of Karakalpakstan — or for a conditional although equally unspecified right to do so, like for Gagauzia in Moldova “in case of a change of the status of the Republic of Moldova as an independent state” [3]3 — Art. 1 Law on the special legal status of Gagauzia, 1994. . Only in the case of Ethiopia the 1995 constitution (art. 39) lays down procedural rules on how to achieve independence by some subnational unit, which also include a referendum (Habtu 2005). In a handful of islands autonomies specific provisions have been inserted in the course of time to allow for referendums on independence (or quasi-independence), such as in Greenland, Faroe Islands, Bougainville, Falkland Islands, New Caledonia (Ellis 2018).

In a comparative perspective, it results that all the few explicit—general or ad hoc—provisions on the possible separation of a territory from the state it belongs to provide for a referendum. Only in Ethiopia some additional procedural provisions are included, notably the initiative by a two thirds majority of the subnational legislature (art. 39 para 4 lit a const.), although the referendum (which is to be organized within three years from the initiative) does not require any entrenched majority (“supported by a simple majority vote in the referendum”, art. 39 para 4 lit c const.). From a historical perspective, the last separation that took place without a referendum was the split of Czechoslovakia in 1992—and interestingly enough, a referendum was possible there too [4]4 — If so requested by the Federal Assembly or by one or both subnational parliaments: art. 1 Constitutional Act 327/1991. but it was not held, also because according to opinion pools at that time, the majority in both parts of the then Czechoslovak federation opposed dissolution [5]5 — According to a poll taken for the government in summer 1992, “only 37 percent of Slovaks and 36 percent of Czechs said they would vote for a split in a referendum, but more than 80 percent said that they considered a break inevitable”. Available online. . Since then, all attempts—both successful and failed—to achieve independence resorted to a referendum, with the sole (and partial) exception of Kosovo due to its peculiar situation of (then) international protectorate [6]6 — It must be recalled that the Provincial Assembly of Kosovo declared independence in 1991, and immediately after a referendum was called, in which 99% of the voters (by a turnout of 87%) supported independence. The whole process was blocked by Serbia. After the NATO military intervention in 1999 and the following international protectorate, Kosovo unilaterally declared independence in February 2008. .

It can thus be said that since the post-1989 wave of constitution-making (for the concept see Elster 1995) the referendum has become not only the main but often the exclusive means to address sovereignty claims (Şen 2015), irrespective of the typology of constitutional referendum (for a categorization of constitutional referendums see Tierney 2012). More specifically, secession and territorial referendums have been different in scope, functions and procedure, as well of course in outcome (Qvortrup, 2014). Most of the referendums held in nearly all former Soviet and former Yugoslav republics between 1991 and 1992 were plebiscitary moments of a process that was already in unstoppable motion (Kössler 2018). In some cases they just rubber stamped the already achieved statehood, like in Georgia, Ukraine Uzbekistan (where the referendum took place three days after the official dissolution of the UdSSR) and in the more contested cases of Latvia, Estonia and Lithuania. In other cases, more nuanced issues emerged, either in terms of participation (in the 1992 independence referendum in Bosnia Herzegovina the Serbian community boycotted the vote) or of threshold: for instance in Slovenia the referendum in December 1990 was subject to a threshold set at 50% of the whole electorate, which was easily reached [7]7 — The total support for independence was 88.5% of eligible voters, 94.8% of those participating. Independence was accordingly proclaimed in 1991. . Other referendums simply remained without any consequence, at least in terms of international recognition of the formally proclaimed statehood, such as in Nagorno-Karabakh in 1991 and in Somaliland in 2001.

Several other independence referendums took place in recent years or are scheduled in a number of islands or other territories already enjoying autonomy from the mainland, normally agreed between the central government and the affected territory. Both in the Falkland Islands (2013) and in Gibraltar (2002) referendums were called in order to reiterate the expected support in favour of the territories’ belonging to the United Kingdom, and did not require any special majority as landslide support for the status quo was known. In Bougainville a (consultative) referendum took place in 2019 on independence from Papua New Guinea (Regan 2013), and while the overwhelming majority voted for independence, the situation remains in a state of flux and the final say remains with the Parliament of Papua New Guinea. In New Caledonia, after a first referendum was held in 1987 (with less than 2% of the votes in favour of independence from France), the Nouméa Accord from 1998 [9]9 — See French Organic law No. 99-209 (1999). , provided for a series of three independence referendums after twenty years, within a time interval of two years. The first two took place in 2018 (56,67% against independence) and 2020 (53,26% against) and the last one is scheduled for 2022 (Alber 2020). Another referendum based on mutual agreement between the concerned parties was scheduled for spring 2018 in the Faroe islands, however not on full independence from Denmark but on a new constitution granting a sort of quasi-independence, but it was indefinitely postponed.

No doubt, the best known example of an agreed referendum on independence has been the one that took place in Scotland in September 2014. After a complex set of negotiations following the seize of power in Edinburgh by the Scottish National Party (2007-2011) (Mitchell 2014), the so called Edinburgh agreement (2012) provided for a referendum on just two options: independence or status quo, although the political promise from London was further devolution of powers in case of a vote against independence (Tierney 2013). Especially noteworthy for the purposes of this paper are the absence of any entrenchment for the referendum, the scope of the right to vote and the exclusion of a third legal option between independence and status quo. As to the first aspect, the very concept of entrenched majorities is alien to the political nature of the British constitution and it was never really considered as an option. As to the right to vote, article 2 of the Scottish Independence Referendum (Franchise) Act 2013 extended it to persons aged 16 or over, resident in Scotland and citizen of the Commonwealth, of the Republic of Ireland or of the European Union, thus de facto increasing the chances for a “no” vote. Not least, no third option between independence and statu quo was legally provided for by the Scottish Independence Referendum Act 2013, although this was initially considered (Devo Plus Group 2012), subsequently politically promised in case of a “no” vote and eventually realized with the adoption of the Scotland Act 2016.

Several other independence referendums took place in recent years illegally, ie. without having been agreed between the affected territory and the central government, or even informally. Among them the unofficial, privately organized online consultations in Catalonia and Veneto in 2014, and the more official, albeit contested, referendums in Iraqi Kurdistan in 2017 [10]10 — Despite the fact that the Iraqi Federal Supreme Court ordered the suspension of the referendum, the President of Iraqi Kurdistan went ahead with the vote on 25 September 2017. By a turnout of 72/, 92.7 of the voters were in favour of independence. The Federal Supreme Court declared the vote unconstitutional the same day and reiterated in November 2017 that Article 1 of the Iraqi constitution prevents any region from seceding (decision in Arabic available online). and in Gagauzia in 2014 [11]11 — With a turnout of more than 70%, 98.9% of the voters supported the declaration of independence of Gagauzia if Moldova eventually lost its sovereignty, a scenario that could imply the merger of Moldova and Romania into a single sovereign state, an option that is provided for in the Law on the Special Legal Status for Gagauzia from 1994. Similarly, 98.4% of voters said they prefer to tighten ties with Russian-led Customs Union rather than moving towards EU integration . In 2014 a referendum was held in Crimea on the separation of the peninsula from Ukraine and its access to the Russian Federation. The referendum was held despite the fact that its call had been declared illegal by the Ukrainian Constitutional Court [12]12 — Decision no. 2-rp/2014, case no. 1-13/2014. and the Crimean Parliament had unilaterally declared independence three days before the vote. The referendum resulted in a plebiscite for secession from Ukraine and annexation to the Russian Federation (95,7% by a turnout of 82%, although the referendum was boycotted by Tatar and Ukrainian minorities) and was immediately followed by Russian annexation, regardless of its illegality under both international and Ukrainian constitutional law (Venice Commission 2014 and Bílková 2016). In the Bosnian Entity of Republika Srpska a referendum on independence has been often announced as a threat but never realized so far, while several identity-driven referendums took place. Common to all these unofficial referendums is their political nature: while legally irrelevant, they are used as instruments for political pressure (sometimes they are organized by political parties) [13]13 — In 2013 an independentist political party in South Tyrol that currently enjoys support of 7.2% of the electorate organized a referendum on independence from Italy, to which less than 15% of the electorate participated and which resulted in a 92% “yes” vote for independence. In Iraqi Kurdistan, an informal independence referendum was organized by the Kurdistan Referendum Movement, producing an outcome of 99% of votes for independence. and/or to make the cause for independence more popular and always result in plebiscites in favour of independence, even when the turnout is low [14]14 — Normally the turnout is low when informal referendums are organized by political parties or movements, while it tends to be high when they are called by subnational governments, such as in the case of Gagauzia (Fn 14) and Iraqi Kurdistan (72,6% turnout, 92,7%). High turnouts were achieved also in the illegal but equally locally organized referendums in Somaliland (2001) and in the Serb enclave of Kraijna in Croatia (1992). .

The most significant and best known among the unofficial referendums of this kind is the consultation that took place in Catalonia on October 1, 2017. The “referendum on self-determination” was called by a law of the regional parliament adopted less than one month before in a contested procedure [15]15 — Law no. 19/2017, of 6 September, “on the referendum of self-determination”. Together with law no. 20/2017 (“on juridical transition”), the law was adopted following a special procedure that cuts debate and amending proposal to just one day. For more details Castellà Andreu 2017. and was immediately suspended by the Constitutional Court, which subsequently declared it unconstitutional [16]16 — STC 114/2017. Subsequent ruling STC 124/2017 outlawed also the Law on transition. For the Court, both laws violated the rights of participation of the minority in Parliament and introduced a new legal order against the foundations of the Spanish Constitution without following the procedures for constitutional amendment. . The referendum was held anyway, which led to harsh confrontation with the Spanish government and also to cases of violence and repression. The Catalan Government officially announced the results of the referendum [17]17 — According to the Catalan government (22 October 2017), the turnout was 43,03% of the electorate and the support for independence was 92,01% of those who participated in the vote. Data available online. and on that basis the Catalan Parliament proclaimed, in a contested, majority vote, a unilateral declaration of independence [18]18 — Resolution 27 October 2017. . In response, the Spanish Senate enacted coercive measures de facto suspending Catalan autonomy (article 155 Spanish constitution) and called for early regional elections. After the elections the legal situation was ‘normalized’ but political tensions remained (Poggeschi 2018) and continued also after governmental changes both in Barcelona and in Madrid.

A final, sui generis example of independence referendum is represented by the United Kingdom’s vote to leave the European Union (so called “Brexit”) in 2016. The Treaty on the European Union lays down the procedure for withdrawing from the Union (article 50 TEU), although it refers to “its own constitutional requirements” as to how a State may decide on the withdrawal. The British political constitution did not contain a provision on the “own constitutional requirements” although it was clear that a decision on leaving the EU was possible only by referendum, as the British membership in the EU, provided for by the European Communities Act 1972, was confirmed by referendum in 1975 (67,2% in favour by a turnout of 64,6%). Accordingly, the European Union Referendum Act 2015 provided for a consultative referendum on “whether the United Kingdom should remain a member of the European Union”. Like for the referendum on Scottish independence, no entrenched majority was required, nor an alternative was offered between leaving or remaining, although the government had already negotiated with Brussels a new treaty redefining the terms of UK membership in the Union, that should have entered into force in case the referendum was in favour of “remain” (Martinico 2017b). The Referendum Act 2015 also defined the electorate in a significantly different manner as compared to the Scottish referendum: only persons aged 18 or over, no citizens of EU countries residing in the UK and no UK citizens residing abroad for more than 15 years. The referendum took place in June 2016 and 51,9% of the voters (by a turnout of 72,2%) supported the withdrawal from the EU, leading to the approval of the European Union (Withdrawal) Act 2018 and to negotiations with the EU according to article 50 TEU (Bradley 2020).

Beyond plebiscitarianism: constitutionalizing and pluralizing secession procedures. Some hints from comparative constitutional practice

The spreading of referendums as a means to channel secession claims has the great merit to institutionalize challenges that so far have been left outside the realm of law and thus decided either politically or on the battleground (Elster 2012). In this perspective, independence referendums represent an important step on the evolutionary path of constitutionalism, which ultimately means to regulate political processes with legal instruments (Hayek 1960, 176). Furthermore, referendums are also a manifestation of the democratic principle, at least when they are agreed and regulated by legal procedures. Against this background, it can be concluded that the post-1989 constitutional wave has represented a major step forward for constitutionalism with regard to secession, one of the areas so far immune from legal regulation.

The limit of referendums, however, is that they are too simplistic and ‘trivial’, especially when deciding on existential issues like secession and statehood, at least when they are not supported by procedural entrenchments or by additional procedures. The independence referendum held in Quebec in 1995 did not require any special majority. It was rejected by a narrow majority of 50.58 percent, fewer than 55,000 votes, and following the vote, there was significant controversy relating to the counting of the ballots, the enumeration of eligible voters and other concerns. The outcome of the referendums on Scottish independence in 2014 and on Brexit in 2016 were largely determined by the definition of the eligible voters. Furthermore, the question of structural minorities is always a challenge for referendums. If decisions are to be made by a majority (majoritarian constitutionalism), especially if simple, without a quorum and extemporary (plebiscitarian constitutionalism), how can the rights of minorities be respected, even more if such minorities are territorial majorities? In the Brexit referendum 62% of the votes in Scotland and 55% in Northern Ireland were for “remain”, representing a much more consistent support than the 51,9% for “leave” in the whole of the UK. At the same time, in the third minority nation of the UK, Wales, the majority was for “leave” (52,5%). Brexit triggered a number of controversies and put political change in motion, such as the growing demand for a new referendum on independence in Scotland and stronger support for Irish unification.

The problem of respect of minorities remains unsolvable as long as counterbalances are not provided: 85% of the UK population is English and any majority vote will inevitably reflect the dominance of the dominant group, frustrating minority positions. Furthermore, beside ethno-cultural minorities, should other minorities be considered, and how? In the London metropolitan area 59,9% of votes have been for “remain”, and so were 73% in Cambridge and 95,9% in Gibraltar. In Catalonia, both the (informal and not reliable) referendum and the results of elections (especially the one held in December 2017 after the dissolution of the Catalan Parliament by the Spanish Government, which has been politically considered an equivalent to a referendum on independence) show support for independence by little less than 50% of the electorate [20]20 — The support is constantly monitored by the poll center of the Catalan Government (Barómetro de Opinión Pública del Centre d’Estudis d’Opinió) which shows support ranging from 43% and 48% in 2018 (see data of the CEO). See also López-Basaguren 2018. , with uneven geographical distribution (more support in the periphery, less in the metropolitan areas): can a fundamental decision be made by a tiny majority where society is split? More generally, if referendums aim at legitimizing constitutional discontinuity, they fail to do so when legitimacy is conferred by simple majority.

To reduce the risk of majoritarian and even plebiscitary referendums (Sáenz Royo and Garrido López eds. 2017, Martinico 2019), the Canadian Supreme Court’s Quebec Secession Reference (followed by the Parliament with the Clarity Act) has taken the natural step suggested by constitutionalism: it designed a legal framework for secession claims. These must include a referendum as unavoidable democratic expression by the people, but its shortcomings make the referendum unfit to represent the only decisional moment, especially if not assisted by some procedural entrenchment.

Moving from the Canadian experience, legal scholars have started to pay growing attention to this phenomenon and also to put forward proposals on how to combine constitutionalism and secession (among others Haljan 2014, Martinico 2018). The common denominator of any formula in this regard cannot but be the combination of procedures producing an effect that goes beyond rudimentary, plebiscitary majoritarianism and combines popular decision-making with the rule of law, including respect for minorities. Based on some of the greatest achievements of constitutionalism such as federalism, constitutional amending procedures and participatory democracy, on comparative case-law on secession and constitutional amendment, and on the “soft-jurisprudence” of international bodies such as the Venice Commission, a few elements for a comparative procedural framework on how to address secessionist claims can be sketched.

The first and simplest entrenchment is the provision of quorums both for the turnout and for the approval in a referendum on independence, as was provided for in the 2006 referendum for Montenegrin independence. While quorums are generally problematic in referendums as they can easily be misused by opponents to torpedo the vote by encouraging abstention (as suggested by the Venice Commission’s Code of Good Practices on Referendums 2006) [21]21 — CDL-AD(2007)008, esp. paras. 51 and 52. See also the Compilation of Venice Commission opinions and reports concerning referendums, CDL (2017)002. , the situation is partly different when it comes to existential issues such as secession and statehood. In these cases abstention is unlikely and not due to lack of interest or of information, and the criterion of the “clear majority” put forward in the Quebec Secession Reference is of utmost importance and can be ascertained only if entrenched majorities are provided for..

The potentially unequal position between supporters of independence and those against it in referendums with special entrenchments can also be addressed following the procedural way. A fundamental example is provided by the experience of the Swiss canton of Jura. It was established in the 1970ies splitting from the canton of Berne, following a series of steps and referendums that safeguarded the positions of both advocates and critics. Prior to the series of referendums, the constitution of canton Berne was amended, allowing for referendums on the exercise of the right to self-determination by the population of Jura. The constitutional amendment was itself accepted in a referendum in 1970, supported by a large majority (87%). The amendment introduced a secession procedure in the cantonal constitution which largely took over the recommendations of a commission set up a few years before by the federal government and composed by former politicians, with the explicit task to develop a specific procedure for a possible secession at cantonal level. The amendment provided for three referendums, one for each segment of the affected population: regional, district and municipal (Hänni and Iseli 2014). The referendum for the Jura region should determine whether or not to form a new canton. Based on the outcome, a second referendum should take place at district level: if the majority of the Jura region was in favour of creating a new canton, the districts that voted against could decide whether to remain with the canton of Berne or to join the new canton; if instead the majority at regional level was against the establishment of the new canton, the districts voting in favour could decide whether to leave the canton anyway. Finally, the municipalities bordering on a district that decided to separate from or remain with the canton of Berne could vote on their territorial affiliation. All referendums had no quorum.

At the level of national constitutional law, the mentioned experience of New Caledonia is of paramount importance. Three referendums within six years represent far more than a snapshot of the occasional will of the people and are likely to be really representative. In New Caledonia, the vote in favour of independence in just one of the three would trigger separation from France — and interestingly enough the trend suggests that support for independence is growing over time (from less than 2% in 1987 to 46,7% in 2020). One may also consider a weighted majority, summing the votes in the successive referendums and providing that the average will be the definitive decision.

A third way to balance the majoritarian attitude of a referendum is to put more than just two options to the vote, thus allowing for a more nuanced deliberation, as it was initially proposed in Scotland (so called “Devolution-max”) and eventually realized with the Scotland Act 2016. In fact, a referendum is a black-and-white instrument only if it is made one such, but can be more nuanced if more alternatives are offered. Such more nuanced option could be combined with the pluralization of referendums following the Jura example: for instance, a first referendum could take place between independence or status quo, and if the status quo prevails various options such as no change, more autonomy, further negotiations within a given time, etc. could be put to a vote.

A fourth element of a procedural framework for approaching territorial conflicts of sovereignty is to avoid that independence referendums be the only decisional instruments. In fact, for every (legal) referendum on independence or secession an act of Parliament is required, but it often merely reflects political agreements made elsewhere. The involvement of Parliament could however be made more ‘resistant’ to possible abuses by occasional majorities. Article 39 of the Ethiopian constitution requires a request for starting the procedure by at least a two thirds majority of the legislature of the entity concerned. The referendum is not subject to entrenched majority (article 39.4 lit c), but the constitution requires a subsequent law that inter alia transfers assets to the “nation, nationality or people which has opted for secession” (article 39.4 lit e, Bihonegn 2015). This way, parliaments of all affected levels of government must be included in the decision-making, in some instances with entrenched majorities. These can counterbalance majorities in other levels of government, for example in case a minority group demanding secession is a majority in one particular territorial entity. In such a case, to balance between majorities at different levels, which will always prevail in the respective area and thus create an irreconcilable conundrum of opposite legitimacies, it could be provided that, if a subnational parliament adopts a motion for secession with a qualified majority, the national parliament can stop it only by an equally qualified majority.

Similarly, in secession processes also courts are normally involved (Martinico 2017). The role of courts can be merely negative, ie. affirming that secession of parts of the national territory is simply not admissible. This was the case, inter alia, of the pivotal US Supreme Court decision in Texas v. White (1869, just after the civil war) [23]23 — Texas v White, 74 US 700 (1869). , of the ruling of the Italian Constitutional Court in 2015 on the request of the region of Veneto to hold an independence referendum [24]24 — Ruling no. 118/2015. , of the 2016 decision by the German Federal Constitutional Court on the individual complaint of a political group supporting Bavarian independence [25]25 — BVerfG, 2. Kammer, 2. Senat, decision 16 december 2016 – 2 BvR 349/16. The decision is as short as 3 sentences. Two of them simply affirm: “in the Fundamental Law there is no place for secession claims of individual Länder. These violate the constitutional order”. , of the Iraqi Federal Supreme Court decision in 2017 [26]26 — See Fn 13. , etc.

Courts can however also clarify important aspects in terms of procedure and constitutional principles that must guide a request for independence. This notably happened in the case of the Quebec Secession Reference, or of the seminal ruling of the Spanish Constitutional Court that clarified the terms for a possible independence claim under the Spanish Constitution, specifying that these had not been met in the case of Catalonia [27]27 — STC 42/2014 and subsequent rulings – see Castellà Andreu 2016. . Also the seminal Miller judgment by the UK Supreme Court in 2017 [28]28 — R (on the application of Miller and another) v Secretary of State for Exiting the European Union, [2017] UKSC 5. , is of significance for our purposes, as it clarified the process that the UK must follow to withdraw from the European Union. An early involvement of courts in determining paths and limits of secessionist processes would in any case help clarify the legal contours of requests, the principles to be respected and the modalities of decision-making. Such an involvement is less unthinkable than was some years ago the inclusion of constitutional/supreme courts in constitution-amending processes, which is now a not uncommon practice (Chen and Poiares Maduro 2013: 103). In some cases, courts can give advisory opinions (as it was the case for the Canadian Reference on the Secession of Quebec), which is an additional way to involve the guardians of the law in these complex processes. For sure, an early involvement (ie. prior to a vote on a referendum) is more helpful and less politically divisive than ex post rulings.

A fifth guarantee against plebiscitary abuses in independence processes is the ‘double-check’ by the electorate through elections after a referendum. Elections and referendums are different categories of vote, especially as the latter are mediated by political parties and produce a longer-term effect in shaping the political composition of parliaments. Constitutionalism offers several examples of complex procedures for the total revision of constitutions, which is a similar attempt to regulate regime changes that historically used to take place in revolutionary ways (see the distinction between “revolutionary” and “institutionalized” constituent power by Burdeau 1985). Such procedures often require dissolution of Parliament in order for the total revision of the constitution to be voted (by qualified majority) by two different parliaments and supported by a referendum. Article 168 of the Spanish constitution provides that total or other revisions affecting existential principles of the constitution [29]29 — The Spanish Constitutional Court convincingly explained the rationale of the total revision. In its ruling 48/2003 it stated that the Spanish constitution does not contain any substantive limitation to the amending power (unlike most other European constitutions) and precisely for that it contains a detailed (and extremely rigid) procedure to legally regulate changes that are so significant to substantially modify the basic traits of the current constitution. have to be approved by two thirds majority in both chambers of Parliament; subsequently, Parliament is dissolved and the new chambers have to approve the same text by the same majority. After that, the revision is to be ratified by popular referendum. In Switzerland article 191.3 of the federal constitution provides that “if the people vote for a total revision, new elections shall be held to both Chambers”. In Bulgaria, total o structural amendments to the constitution can only be adopted by a special ‘Grand Assembly’ specifically elected for this purpose, by a “majority of two-thirds of the votes of all members, in three ballots on three different days” (article 161 Bulgarian constitution). The idea is that decisions made in referendums are supported also by parliaments, and secession could take place in a more legitimate way if there is convergence between direct democratic decisions in referendums and representative democratic support through parliaments.

A sixth possible way, which is a variation of the last mentioned approach, could be to first ask the electorate to make the fundamental decision on yes or no to independence, and then involve parliaments in defining the more detailed terms of the issue. An interesting example is the most recent referendum that took place in New Zealand together with the general elections in October 2020. One of the two referendums was on whether or not to legalize cannabis. Had the voters chosen to legalize it (which they didn’t, although by a thin majority of 50.7%), the question would have been introduced to Parliament to make fundamental decisions on, for instance, what exactly “legalisation” means as opposed to decriminalisation or on the workability of the provisions (McLean 2020).

Concluding remarks: is there a role for the European Union?

In sum, the issue at stake is the provision of procedural rules to encapsulate territorial conflicts of sovereignty into a legal framework, taking them out from both the rule of force and the rule of either occasional (thus unreliable) or structural (thus unchangeable) majorities. This trend seems unstoppable as it reflects the overall attitude of constitutionalism. The terrain in which such procedures are mostly being developed is inevitably the domestic constitutional level, both national and subnational. However, to the extent in which constitutionalization also affects the supranational level (Schütze 2012), such level also plays (or might play) a significant role in experimenting with and in promoting procedures addressing territorial conflicts of sovereignty.

At the same time, being still an international organization, albeit sui generis, the European Union is conditioned even more than other levels not only by meta-legal factors, but also by structural elements pertaining to its institutional setting and the way it operates. Some fundamental decisions of the EU, such as the accession of new states, are still subject to the rule of unanimity. This means, in plain words, that the Member State from which a territory might secede would retain the veto over the accession of such a new state and will use it unless the separation is consensual. And so would probably do other countries fearing secessionist threats in their own territories or in those of neighbouring countries. The example of the recognition of Kosovo by EU Member States is telling, with five countries not recognising Kosovo as an independent state [30]30 — Spain, Slovakia, Cyprus, Romania and Greece. . At the same time, the very institutional structure of the EU might favour secessionist claims: despite some role attributed to subnational entities (Toniatti, Palermo and Dani 2004), the institutions and, consequently, the decision-making procedures of the Union are based on states. Statehood matters for representation in the EU institutions and this might be an incentive for strong subnational entities to pursue independent statehood rather than seeking stronger autonomy within the states they belong to.

While it seems unlikely, for institutional and political reasons, that the European Union could play an active role in developing a procedural framework to solve territorial conflicts of sovereignty within its Member States, its role in this area is not necessarily secondary. Art. 2 of the Treaty on the European Union lays down the values on which the Union is founded, which include, inter alia, the rule of law in a prominent position. Other relevant values listed in the same article are respect of rights of persons belonging to minorities, pluralism and tolerance. The rule of law is also one of the guiding principles of the EU’s external action (article 21.1 and 21.2 TEU), which would cover the accession of new Member States, but also an internal principle for which the Court of Justice has jurisdiction (article 263.2 TFEU). This means that at the very least the EU should support and encourage the development, at domestic level, of provisions enforcing the rule of law including in the area of territorial conflicts of sovereignty, such as those mentioned above.

Furthermore, the EU has also been proactive in facilitating such solutions, although with different degrees of success and of care. A rather negative example is represented by its engagement in Cyprus prior to the accession of the island to the Union in 2004. On that occasion, the EU supported the referendum on the UN-plan for unification. The referendum would have produced its effects only by gaining a majority in both the Greek and the Turkish part of the island. Otherwise, only the Greek part would have joined the EU (Ker-Lindsay 2005). This engagement turned out to be a torpedoing factor against the plan, as the Greek Cypriots had no incentive to support the plan and actually saw an incentive to boycott it, knowing that they would have joined the EU in any case and the referendum provided the opportunity to leave the “enemy” out. It thus came by no surprise that the Turkish Cypriots supported the plan in the referendum and the Greek Cypriots did not. Far more positive is the example provided by the EU engagement on the occasion of the independence referendum in Montenegro in 2006 (Džankić 2014). The EU was the main driver behind the agreement on a complex referendum which required a double threshold: 55% of the population voting for independence by a turnout of minimum 50% of those entitled to vote [31]31 — The referendum took place in May 2006 and the threshold was met by just 2000 ballots out of about half a million votes. . In more recent times, the EU has been mediating agreements between Serbia and Kosovo (in EU terminology, between Belgrade and Pristina, in order to avoid the status issue) since 2011. This role has produced remarkable results, which were hard to expect when the engagement started, and succeeded in addressing practical aspects to the advantage of both sides, wisely using “creative ambiguity” (Bieber 2015).

• Notes and references

  1 —

  Reference re Secession of Quebec, [1998] 2 S.C.R. 217

  2 —

  This article was specified by a law as late as in 1990, when the dissolution of the Soviet Union was already in place (Bowring 2015).

  3 —

  Art. 1 Law on the special legal status of Gagauzia, 1994.

  4 —

  If so requested by the Federal Assembly or by one or both subnational parliaments: art. 1 Constitutional Act 327/1991.

  5 —

  According to a poll taken for the government in summer 1992, “only 37 percent of Slovaks and 36 percent of Czechs said they would vote for a split in a referendum, but more than 80 percent said that they considered a break inevitable”. Available online.

  6 —

  It must be recalled that the Provincial Assembly of Kosovo declared independence in 1991, and immediately after a referendum was called, in which 99% of the voters (by a turnout of 87%) supported independence. The whole process was blocked by Serbia. After the NATO military intervention in 1999 and the following international protectorate, Kosovo unilaterally declared independence in February 2008.

  7 —

  The total support for independence was 88.5% of eligible voters, 94.8% of those participating. Independence was accordingly proclaimed in 1991.

  8 —

  Referendum Bill, 2006. The then Prime Minister Milo Đukanović promised that he would declare unilateral independence if the votes passed 50%, irrespective of the double majority requested by the law mediated by the European Union. The double threshold was however reached, albeit barely.

  9 —

  See French Organic law No. 99-209 (1999).

  10 —

  Despite the fact that the Iraqi Federal Supreme Court ordered the suspension of the referendum, the President of Iraqi Kurdistan went ahead with the vote on 25 September 2017. By a turnout of 72/, 92.7 of the voters were in favour of independence. The Federal Supreme Court declared the vote unconstitutional the same day and reiterated in November 2017 that Article 1 of the Iraqi constitution prevents any region from seceding (decision in Arabic available online).

  11 —

  With a turnout of more than 70%, 98.9% of the voters supported the declaration of independence of Gagauzia if Moldova eventually lost its sovereignty, a scenario that could imply the merger of Moldova and Romania into a single sovereign state, an option that is provided for in the Law on the Special Legal Status for Gagauzia from 1994. Similarly, 98.4% of voters said they prefer to tighten ties with Russian-led Customs Union rather than moving towards EU integration

  12 —

  Decision no. 2-rp/2014, case no. 1-13/2014.

  13 —

  In 2013 an independentist political party in South Tyrol that currently enjoys support of 7.2% of the electorate organized a referendum on independence from Italy, to which less than 15% of the electorate participated and which resulted in a 92% “yes” vote for independence. In Iraqi Kurdistan, an informal independence referendum was organized by the Kurdistan Referendum Movement, producing an outcome of 99% of votes for independence.

  14 —

  Normally the turnout is low when informal referendums are organized by political parties or movements, while it tends to be high when they are called by subnational governments, such as in the case of Gagauzia (Fn 14) and Iraqi Kurdistan (72,6% turnout, 92,7%). High turnouts were achieved also in the illegal but equally locally organized referendums in Somaliland (2001) and in the Serb enclave of Kraijna in Croatia (1992).

  15 —

  Law no. 19/2017, of 6 September, “on the referendum of self-determination”. Together with law no. 20/2017 (“on juridical transition”), the law was adopted following a special procedure that cuts debate and amending proposal to just one day. For more details Castellà Andreu 2017.

  16 —

  STC 114/2017. Subsequent ruling STC 124/2017 outlawed also the Law on transition. For the Court, both laws violated the rights of participation of the minority in Parliament and introduced a new legal order against the foundations of the Spanish Constitution without following the procedures for constitutional amendment.

  17 —

  According to the Catalan government (22 October 2017), the turnout was 43,03% of the electorate and the support for independence was 92,01% of those who participated in the vote. Data available online.

  18 —

  Resolution 27 October 2017.

  19 —

  See the interesting arguments put forward in the UK Supreme Court’s landmark “Brexit” case R (Miller) v Secretary of State for Exiting the European Union [2017] UKSC 5.

  20 —

  The support is constantly monitored by the poll center of the Catalan Government (Barómetro de Opinión Pública del Centre d’Estudis d’Opinió) which shows support ranging from 43% and 48% in 2018 (see data of the CEO). See also López-Basaguren 2018.

  21 —

  CDL-AD(2007)008, esp. paras. 51 and 52. See also the Compilation of Venice Commission opinions and reports concerning referendums, CDL (2017)002.

  22 —

  On top of several constitutions establishing time limits for, in particular, constitutional reforms, it is worth mentioning Schedule 1 of the Good Friday / Belfast Agreement of 1998. As to the possibility to call for a referendum on unification of Ireland it states: “1. The Secretary of State may by order direct the holding of a poll for the purposes of section 1 on a date specified in the order. 2. […] a majority of those voting would express a wish that Northern Ireland should cease to be part of the United Kingdom and for part of united Ireland. 3. The Secretary of State shall not make an order under paragraph 1 earlier than seven years after the holding of a previous poll under this Schedule” (emphasis added).

  23 —

  Texas v White, 74 US 700 (1869).

  24 —

  Ruling no. 118/2015.

  25 —

  BVerfG, 2. Kammer, 2. Senat, decision 16 december 2016 – 2 BvR 349/16. The decision is as short as 3 sentences. Two of them simply affirm: “in the Fundamental Law there is no place for secession claims of individual Länder. These violate the constitutional order”.

  26 —

  See Fn 13.

  27 —

  STC 42/2014 and subsequent rulings – see Castellà Andreu 2016.

  28 —

  R (on the application of Miller and another) v Secretary of State for Exiting the European Union, [2017] UKSC 5.

  29 —

  The Spanish Constitutional Court convincingly explained the rationale of the total revision. In its ruling 48/2003 it stated that the Spanish constitution does not contain any substantive limitation to the amending power (unlike most other European constitutions) and precisely for that it contains a detailed (and extremely rigid) procedure to legally regulate changes that are so significant to substantially modify the basic traits of the current constitution.

  30 —

  Spain, Slovakia, Cyprus, Romania and Greece.

  31 —

  The referendum took place in May 2006 and the threshold was met by just 2000 ballots out of about half a million votes.

• Literature

  • Alber, Elisabeth (2020), New Caledonia Rejects Once Again Full Independence, but the Die Is Not Yet Entirely Cast, IACL Blog, 6 d’octubre de 2020. Disponible en línia.

   

  • Baldacchino, Godfrey and Hepburn, Eve (2012), “A different appetite for sovereignty? Independence movements in subnational island jurisdictions”, a Commonwealth and Comparative Politics, 50 (4), 555-568.

   

  • Belov Martin (Ed.) (per al proper 2021), Territorial Politics and Secession – Constitutional and International Law Dimensions, Londres, Palgrave.

   

  • Bieber, Florian (2015), “The Serbia-Kosovo Agreements: An EU Success Story?”, a Review of Central and East European Law, 40:3-4, 285-319.

   

  • Bihonegn, Tesfa (2015), “Federalization with a Constitutional Guarantee to Secession: Controversies, Paradoxes and Imponderables in Ethiopia”, a Regional & Federal Studies, 25 (1), 45-70.

   

  • Bílková, Veronika (2016), “Territorial (Se)Cession in Light of Recent Events in Crimea”, a Law, territory and Conflict Resolution, Matteo Nicolini, Francesco Palermo i Enrico Milano (Eds.), Law as a Problem and Law as a Solution, Brill-Nijhoff, Leiden-Boston, 194-218.

   

  • Blokker, Paul Ed. (2018), Constitutional Acceleration within the European Union and Beyond, Routledge, Londres.

   

  • Bowring, Bill (2015), “From Empire to Multilateral Player: The Deep Roots of Autonomy in Russia”, a Minority Accommodation through Territorial and Non-Territorial Autonomy, Tove Malloy i Francesco Palermo (Eds.), Oxford University Press, Oxford, 133-157.

   

  • Bradley, Kieran (2020), “Agreeing to Disagree : The European Union and the United Kingdom after Brexit”, a European Constitutional Law Review, 16 (2020), 1-38.

   

  • Burdeau, Georges (1985), Traité de Science Politique, Tome V, Les régimes politiques, 3a ed., Librairie générale de droit et de jurisprudence, París.

   

  • Castellà Andreu, Josep Maria (2016), “Tribunal constitucional y proceso secesionista catalán: respuestas juriídico-constitucionales a un conflicto político-constitucional”, a Teoría y Realidad Constitucional, (37), 561-592.

   

  • Castellà Andreu, Josep Maria (2017), “Sobre el encaje constitucional del pretendido referéndum de secesión en Cataluña”, a La funcionalidad del referéndum en la democracia representativa, Sáenz Royo, Eva i Garrido López, Carlos Eds., Tirant lo Blanc, Valencia, 129-153.

   

  • Chen, Albert HY i Poiares Maduro, Miguel (2013), “The judiciary and constitutional review”, a Routledge Handbook of Constitutional Law, Mark Tushnet, Thomas Fleiner and Cheryl Saunders (Eds.), Routledge, Londres, 97-109.

   

  • Delledonne, Giacomo i Martinico, Giuseppe (Eds.) (2019), The Canadian Contribution to a Comparative Law of Secession, Londres, Palgrave.

   

  • Devo Plus Group (2012), A New Union. Third Report of the Devo Plus Group. Disponible en línia.

   

  • Džankić, Jelena, “The role of the EU in the statehood and democratization of Montenegro”, a Keil, Soeren i Arkan, Zeynep (Eds.), The EU and Member State Building: European Foreign Policy in the Western Balkans, Londres, Routledge, 83-102.

   

  • Ellis, Andrew (2018), Research Report “Administration of Referendums”. A Comparative Study of Independence Referendums, Institut Nacional de Recerca, Boroko.

   

  • Elster, Jon (1995), “Forces and Mechanisms in the Constitution-Making Process”, a Duke Law Journal 45(2), 364-396.

   

  • Elster, Jon (2012), “Constitution-making and violence”, a Journal of Legal Analysis (4), 1, 7-39.

   

  • Ferraiuolo, Gennaro (2016), Costituzione federalismo secessione. Un itinerario, ES, Nàpols.

   

  • Fraenkel-Haeberle, Cristina, Kropp, Sabine, Palermo, Francesco i Sommermann, Karl-Peter Eds. (2015), Citizen Participation in Multi-Level Democracies, Brill-Nijhoff, Leiden-Boston.

   

  • Ginsburg, Tom i Versteeg, Mila (2019), “From Catalonia to California: Secession in Constitutional Law”, a Alabama Law Review, vol. 70:4 [2019], 923-985.

   

  • Habtu, Alem (2005), “Multiethnic Federalism in Ethiopia: A Study of the Secession Clause in the Constitution”, a Publius 35(2), 313-335.

   

  • Haljan, David (2014), Constitutionalising Secession, Hart, Oxford.

   

  • Harzl, Benedikt (2018), The Law and Politics of Engaging De Facto States: Injecting New Ideas for an Enhanced EU Role, Center for Transatlantic Relations, Washington D. C.

   

  • Hänni, Peter i Iseli, Tamara, “Territoriale Veränderung im Bundesstaat: Das Beispiel des Kantons Jura“, a Jahrbuch des Föderalismus 2014, Baden-Baden, Nomos, 2014, 344-361.

   

  • Hayek, Friedrich A. (1960), The Constitution of Liberty, Chicago University Press, Chicago.

   

  • Ker-Lindsay, James (2005), EU Accession and UN Peacemaking in Cyprus, Palgrave McMillan.

   

  • Kössler, Karl (2018, properament), “Territorial Referendums from a Constitutionalist Perspective: Functions, Justifications and Legal Design”, a Claims for Secession and Federalism. A Comparative Study with a Special Focus on Spain, Alberto Lopez Basaguren i Leire Escajedo San Epifanio (Eds.), Springer, Berlín.

   

  • Levrat, Nicolas, Antunes, Sandrina, Tusseau, Guillaume i Williams, Paul (2017), Catalonia’s Legitimate Right to Decide. Paths to Self-Determination, informe encarregat pel Ministeri d’Afers Exteriors, Relacions Institucionals i Transparència de la Generalitat de Catalunya, Barcelona.

   

  • López-Basaguren, Alberto (2018), “The Secession Issue and Territorial Autonomy in Spain: Bicameralism Revisited”, a Perspectives on Federalism, vol. 10 (2), E-238-267.

   

  • Mancini, Susanna (2012), “Secession and Self-Determination”, a The Oxford Handbook of Comparative Constitutional Law, 481-500, Michael Rosenfeld i András Sajó (Eds).

   

  • Martinico, Giuseppe (2017), “Identity Conflicts and Secession before Courts: Three Case Studies”, a Revista general de derecho público comparado, (21), 1-30.

   

  • Martinico, Giuseppe (2017b), “Considerazioni comparatistiche sul referendum alla luce del caso britannico”, a Diritto Pubblico (2), 429-467.

   

  • Martinico, Giuseppe (2018), How can Constitutionalism deal with Secession in the Age of Populism? The case of Referendums. Treballs d’investigació STALS 5. Disponible en línia.

   

  • Martinico, Giuseppe (2019), “Preserving Constitutional Democracy from Populism – The Case of Secession Referendums”, a National Journal of Constitutional Law, 39:2, 139-157.

   

  • Mendez, Fernando and Germann, Micha (2018), “Contested Sovereignty: Mapping Referendums on Sovereignty over Time and Space”, a Revista britànica de Ciències Polítiques, 48 (1), 141-165.

   

  • Mendez, Fernando, Mendez, Mario i Triga, Vasilika (2014), Referendums and the European Union: A Comparative Inquiry, Cambridge University Press, Cambridge.

   

  • McLean, Janet (2020) “Referendums on Public Policy Questions: The Case of New Zealand” Blog IACL-AIDC (novembre de 2020). Disponible en línia.

   

  • Mitchell, James (2014), The Scottish Question, Oxford University Press, Oxford.

   

  • Novic, Elisa and Priya, Urs, “Secession”, a Max Planck Encyclopedia of Comparative Constitutional Law, Oxford University Press, Oxford.

   

  • Palermo, Francesco i Alber, Elisabeth Eds. (2015), Federalism as Decision-Making. Changes in Structures, Procedures and Policies, Brill-Nijhoff, Leiden-Boston.

   

  • Poggeschi, Giovanni (2018), La Catalogna: dalla nazione storica alla repubblica immaginaria, ES, Nàpols.

   

  • Qvortrup, Mads (2014), Referendums and Ethnic Conflict, University of Pennsylvania Press, Filadèlfia.

   

  • Regan, Anthony J (2013), “Autonomy and Conflict Resolution in Bougainville, Papua New Guinea”, a Practising Self-Government. A Comparative Study of Autonomous Regions, Yash Ghai i Sophia Woodman (Eds.), Cambridge University Press, Cambridge, 412-448.

   

  • Schütze, Robert (2012), European Constitutional Law, Cambridge University Press, Cambridge.

   

  • Şen, İlker Gökhan (2015), Sovereignty Referendums in International and Constitutional Law, Springer, Berlín.

   

  • Sunstein, Cass R. (1991), “Constitutionalism and Secession”, University of Chicago Law Review 58(2), 633-670.

   

  • Tierney, Stephen (2012), Constitutional Referendums: The Theory and Practice of Republican Deliberation, Oxford University Press, Oxford.

   

  • Tierney, Stephen (2013), “Legal Issues Surrounding the Referendum on Independence for Scotland”, a European Constitutional Law Review, 9(3), 359-390.

   

  • Toniatti, Roberto, Palermo, Francesco i Dani, Marco (eds.) (2004), An Ever More Complex Union. The Regional Variable as A Missing Link in the EU Constitution?, Baden-Baden, Nomos.

   

  • Comissió de Venècia (2014), Opinió sobre “si la decisió del Consell Suprem de la República Autònoma de Crimea a Ucraïna per organitzar un referèndum per convertir-se en un territori constituent de la Federació de Rússia o restaurar la constitució de Crimea del 1992 és compatible amb els principis constitucionals”, CDL-AD (2014) 002-e.

   

  • Weinstock, Daniel (2001), “Constitutionalizing the Right to Secede”, a The Journal of Political Philosophy (9), 2, 182-203.

   

  • Weller, Marc (2008), “Why the Legal Rules on Self-Determination Do Not Resolve Self-Determination Disputes”, a Settling Self-Determination Disputes: Complex Power-Sharing in Theory and Practice, Marc Weller i Barbara Metzger (Eds.), Brill, Leiden-Boston, 17-46.