Brexit and the union: Territorial voice, exit and re-entry strategies in Scotland and Northern Ireland after EU exit

This article examines the effects of Brexit on the internal boundaries and territorial future of the United Kingdom (UK). Divergent Brexit preferences, coupled with the process of negotiating and preparing for Brexit, have raised new questions about the ability of the UK to remain united. Focusing upon Scotland and Northern Ireland, where territorial challenges are most acute, the article draws upon and adapts Hirschman’s voice, exit and loyalty framework to examine the strategic choices and options faced by institutional actors in each case when determining constitutional and political options in response to Brexit. The article contends that, while credible exit threats were used to bolster voice in the Brexit negotiations, the Brexit vote and subsequent process have exposed the limitations of territorial voice, and unleashed new exit dynamics. We enhance the framework by exploring the concept of re-entry, exploring the contrasting paths these territories would face to re-enter the European Union (EU).


Introduction
While the United Kingdom's (UK) external borders with the European Union (EU) have been the focus of much Brexit scholarship, the actual and symbolic significance of the UK's internal territorial boundaries is the focus of this article. The narrow 52% majority who voted Leave in the 2016 referendum masked deep territorial divisions in Brexit preferences, raising new questions about the ability of the 'kingdom' to remain united. We focus on Scotland and Northern Ireland, where the territorial issues raised by Brexit are most acute.
The plurinational character of the UK is at least partially recognized and embodied in its system of asymmetrical multi-level government. Since 1999, Scotland and Northern Ireland, along with Wales, have had their own elected legislatures and governments, with varying degrees of devolved political autonomy, but there is no legislative devolution in England. As a strategy of accommodating national minorities and, in Northern Ireland, of conflict resolution, devolution has been largely successful, despite periodic disruptions in Northern Ireland. Yet, devolution never eliminated disagreement over the UK's constitutional future. It also generated an opportunity to govern for those seeking a new constitutional order, including the pro-independence Scottish National Party (SNP) and Sinn Féin, the most radical pro-Irish reunification party. The 2014 Scottish independence referendum, spearheaded by the SNP government, represented the union's biggest test since devolution. Arguably, Brexit poses even stiffer territorial challenges.
The Brexit process was initiated after a simple majority in a UK-wide ballot voted for the UK to leave the EU. That two of the four constituent territories voted Remain had no bearing on the outcome. In both Scotland and Northern Ireland, the grievances resulting from having to leave the EU, despite a clear majority in each case voting Remain, have been reinforced by the frustrated efforts of key political and institutional actors to influence the Brexit process. The UK government's pursuit of a 'hard' Brexit, which entails leaving both the EU customs union and internal market, heightened territorial anxieties, and contributed to the resurgence of nationalist claims for a new independence referendum in Scotland and a border poll (unity referendum) in Ireland.
In examining the strategic responses of territorial actors to Brexit grievances, we have drawn upon and adapted Hirschman's voice, exit and loyalty framework. Hirschman's original study focused on how consumers adapt to dissatisfaction, but his framework has been frequently deployed to aid understanding of political behaviour (Dowding et al., 2000), including by Hirschman himself (Hirschman, 1993). Most of these studies focused on the behaviour of individuals or groups of individuals, with exit associated, for example, with member or voter decisions to leave a political party, service users leaving a service provider or citizens leaving a country (Sharp, 1984;Weber, 2011). Instead, we examine collective exit and voice as strategic constitutional and political options facing sub-state territories. We also adapt Hirschman's framework by adding a category of re-entry, with a particular focus on the cases under discussion of re-entry into the EU. Our unit of analysis is thus the territorial community. For the most part, this is examined in the actions and discourse of the governing administrations and political elites representing these communities, but we also consider the aggregate views of the communities themselves, expressed in voting behaviour and survey data. Our key aim was to determine the effectiveness of these territorial communities in using the credible threat of exit from the state to enhance their voice and influence over the Brexit process.
Although the challenges we discuss are set within the particular context of the UK and Brexit, they provide broader insights into the accommodation and marginalization of territorial voices within a plurinational state. In particular, the case studies reveal how the existing balance of voice, exit and loyalty can be disrupted in the face of an exogeneous or endogenous shock. In the UK's asymmetrical setting, where Brexit represents an endogenous shock, our analysis portrays a story of deep complexity in the articulation and influence of voice, and multifarious challenges and incongruities in relation to Scotland's and Northern Ireland's ability to effectively employ an exit strategy. Exit is not inevitable in either case, and so any challenges of re-entry may not arise. But their contrasting exit paths and routes to re-entry into the EU shape the dynamics of contemporary exit debates, and the options associated with a new constitutional order.

Territorial dissatisfaction after Brexit: Voice, exit and re-entry
In his seminal essay, Hirschman posited exit and voice as alternative strategies facing consumers or members of an organization who become dissatisfied with the product or service on offer. Exit is the act of leaving, in expectation that a better offer can be found elsewhere. With voice, consumers or members of an organization air their grievances from within, 'kicking up a fuss' in the hope of improving products, practices or outputs that are the source of dissatisfaction (Hirschman, 1970: 30).
Hirschman principally regarded exit and voice as having an inverse relationship: when exit is readily available, it undermines the use of voice as a strategy to achieve change. Conversely, enhancing the channels through which voice can be exerted can reduce the incentive to exit. In contrast to voice, which requires time, effort and collective organization, Hirschman regarded exit as the easier option. This rested on an assumption that exit, whether applied to a consumer seeking a better product or service elsewhere, or a citizen seeking to emigrate to a new country, is an act of individuals (Hirschman, 1993: 194). Voice, by contrast, is a public activity that thrives on collective action. Indeed, Hirschman regarded exit as an option more readily available to consumers in a competitive marketplace than to social or political contexts. In the latter context, voice may be 'the only way in which dissatisfied customers or members can react whenever the exit option is unavailable' (Hirschman, 1970: 33).
Applied to the context of territorial politics, however, exit can be considered a plausible, albeit difficult, collective enterprise to exit from a system of state government. These communities may strive for a partial exit that expands their territorial self-government within the state. Alternatively, exit ambitions might include the more radical option of leaving one state to establish or, exceptionally, join another.
Hirschman regarded the state, alongside the family and the church, among those organizations where 'the virtual unavailability of the threat of exit' will incentivize voice (Hirschman, 1970: 97-98). Certainly, exit options are hardly readily available; few states make provision for secession in their constitutions and, even where they do, the risks associated with exit are likely to be very high (Lachapelle and Qvortrup, 2020;Lechner and Ohr, 2011). Nevertheless, in plurinational states, which house multiple nations, often with overlapping boundaries, the territorial integrity of the political community is more likely to be challenged. For Rokkan (1974), one of the first to adapt Hirschman's framework to a political setting, boundary-setting and closure are essential prerequisites that enable a state to claim sovereignty over its territory and population. Internal cultural, territorial, and/or political-institutional boundaries associated with plurinational political systems can make it more difficult for states to stake that claim. Exit options, especially when the subject of intense party competition, become more viable.
For Hirschman, the choice to exercise voice or exit is conditioned by loyalty; 'loyalty holds exit at bay and activates voice ' (1970: 78). Even where the exit option is available, dissatisfied consumers or citizens may be more inclined to use voice to seek change from within when they feel a sense of attachment to the organization in question. The attitudes and behaviour of citizens, in turn, shape and constrain those institutional actors, including governments, who act in their name. Loyalty to the state, in the form of national identity or attachment, is more powerful than loyalty to a product or a brand. For one, the alternative options are fewer. Moreover, national identity is forged through decades and generations of shared institutions, experiences, customs and narratives, and the expectation of a shared future. Ferrera (2019: 177) described this as a process of 'bonding', resulting in 'a set of we-feelings that spread throughout the population in the wake of continuous mutual interaction, which generates material interdependence and cultural/emotional ligatures'. Yet, in plurinational states, the development of territorial identity is more complex, with multiple nation-building processes at play, generating bonds to more than one territorial community. Loyalty to the state as a political community cannot be presumed, especially when set against loyalty to a nation or community within the state. In such contexts, the role of loyalty to the state in constraining exit in favour of voice may carry less force.
However, voice is at its most potent when combined with a credible threat of exit. In plurinational states, where a sub-state territorial community is capable of mounting a challenge to the state's capacity to exercise legitimate sovereignty over the entire territory, a credible threat of exit can enhance voice. That could give sub-state territories and their representatives opportunity to express their dissatisfaction and to influence policy choices. From both the perspective of the territory seeking influence, and the state seeking to maintain its territorial integrity, the success of voice as a strategy may require that voice to be heard and responded to, with evidence of influence. As Hirschman observed, 'the barrier to exit constituted by loyalty is of finite height ' (1970: 79). At some point, faced with continued deterioration and widening disagreement despite increased use of voice as a form of protest, some of those loyalists (save those for whom exit would never be contemplated) may opt for exit (Hirschman, 1970: 87-89).
In the field of territorial politics, or indeed any public context, exit is never final: 'one remains a consumer of the output or at least of its external effects from which there is no escape' (Hirschman, 1970: 104). Although briefly mentioned by Hirschman, re-entry after exit was not a significant component of his framework, but we contend it is more important when it comes to territorial politics.
Re-entry has two dimensions. Rather than 'separation' from the state, exit for modern-day nationalists is often presented as an opportunity to re-enter the relationship with the 'home' state on the basis of equality. However much the exiting party may wish to negotiate an equal partnership, they are (usually) the weaker partner in an asymmetric relationship, and they can no longer access the threat of exit from the state to enhance influence in negotiations with their counterparts. For some disgruntled territorial communities with irredentist features, the ambition is not to expand self government or establish a new state, but to join a country that many regard as their kin. These ties may be the legacy of an historic partition of a formerly unified territory, or the result of cultural affinities and social interactions that have nurtured a national identity that spans state boundaries (Brubaker, 1996;Lecours, 2020;Rokkan, 1999). For irredentists, the credibility of the exit threat will be conditioned by the attitudes of the kin-state and its willingness to redraw its own territorial boundaries to facilitate the re-entry of the territorial community in question.
Re-entry also has an external dimension. Although the EU carried the potential to enable sovereignty and political authority to be shared and pooled above and below the state without the necessity of independent statehood (Keating, 2001), it has made exit options for Europe's sub-state nations more viable. Despite the absence of a clear legal route to a territory becoming a new member state by exiting another, the EU represents a secure external framework through which a transition to independence can take place. For the most prominent pro-independence movements in Europe, re-entry to the EU as an independent member state is an integral part of their ambition.

Applying Hirschman's framework to Scotland and Northern Ireland
Power allocation across the three devolved territories in the UK conforms to a reserved powers model. The devolved institutions are free to legislate on matters that are not explicitly 'reserved' or (in the Northern Ireland settlement) 'excepted' to the UK parliament. Across all of the UK's devolution settlements, external affairs are an exclusive competence of the UK parliament, leaving the devolved institutions without a constitutional voice in the Brexit process. However, the legally non-binding intergovernmental Memorandum of Understanding and concordats that accompanied devolution have long recognized the legitimate interests of the devolved institutions in external relations, and included commitment to involve them 'as directly and fully as possible' when decision-making on EU matters touched upon devolved matters (UK Government et al., 2013).
We expected the interaction of voice, exit and loyalty to work in different ways between the two cases. In Scotland, the SNP had been in government since 2007, albeit as a minority since 2016. It had a coherent narrative and clear exit ambition: Scottish independence within the EU. The UK constitution does not include a constitutional right to secession for Scotland, but nor does it explicitly prevent it. When the SNP was re-elected with an overall majority in 2011, with a manifesto commitment to hold an independence referendum, the UK government respected the mandate, facilitated the referendum and committed to respecting the outcome. In the 2014 referendum that followed, 55% voted against the exit option. But, in the immediate aftermath of the 2016 referendum, the Scottish First Minister underlined that Brexit represented a 'material change of circumstances' in which that decision was made, and thus justified a new vote (Sturgeon, 2016a). As the clearest institutional voice of Scotland in the UK, the SNP government assumed the role of the voice of a Scotland that faced the injustice of 'being taken out of the EU against our will' (Sturgeon, 2016a).
In Northern Ireland, a credible prospect of exit in the form of Irish reunification is provided for in the 1998 Belfast/Good Friday Agreement. This agreement, endorsed by the UK and Irish governments and most political parties in Northern Ireland (excluding the Democratic Unionist Party (DUP)), and backed by popular consent in referendums north and south, paved the way for an end to sectarian conflict. It led to the creation of a legislative assembly alongside new north-south and east-west institutional relations. It also recognised that it is 'for the people of the island of Ireland alone. . . to exercise their right of self-determination on the basis of consent, freely and concurrently given, North and South, to bring about a united Ireland', placing a 'binding obligation' on both the UK and Irish governments to respect and give effect to their decision (UK Government, 1998).
Northern Ireland's coherence on the exit option, however, is undermined by deep and enduring political divisions between unionists and nationalists, with the former supporting retention of the constitutional link between Northern Ireland and Britain, and the latter aspiring to a united Ireland. The power-sharing Northern Ireland Executive, which seeks to manage and moderate tensions between the two communities, includes both nationalist and unionist political parties. The divisions between the two political traditions were mirrored in Brexit debates. This made Brexit a highly contested and constitutionally sensitive issue that militated against the articulation of a coherent Northern Ireland voice. This situation was further impaired by the absence of a functioning Executive during the critical phase of EU withdrawal negotiations. The credibility of the exit threat has also been hampered by the Irish government's support for existing territorial arrangements enshrined in the Belfast/Good Friday Agreement (Irish  and its, at best, lukewarm support for the Irish unity agenda.
We thus had contrasting expectations regarding the interplay of voice and exit in our two cases. We expected the Brexit vote to have reinforced the threat of exit in Scotland, in turn bolstering Scotland's voice in Brexit negotiations. By contrast, we expected the internal divisions exposed in the Brexit process, combined with the lack of devolved government for a 3-year period, to have weakened Northern Ireland's ability to convincingly wield the exit option to influence the UK's Brexit policy. Despite these internal divisions, we anticipated that Brexit's interplay with border politics and Northern Ireland's identity conflict would have intensified nationalist exit narratives, with a clear route to re-entry in the form of Irish reunification and EU membership. By contrast, we expected added complexity in Scotland's routes to re-entry, both to a renewed partnership with the UK after independence and accession to the EU.

Exit from Brexit: Exit strategies after the EU referendum
Scotland recorded the highest remain vote (62%) of the four constituent territories of the UK. Every local authority counting area recorded a majority for Remain, adding to the symbolic significance of the outcome, ranging from 50.1% in Moray to 74.4% in the City of Edinburgh (Electoral Commission, 2016). The 56% Remain vote in Northern Ireland also masked significant geographic variations, ranging from 78.3% for Remain in the Foyle constituency to 62.2% for Leave in North Antrim. Remain support was strongest in Belfast and in constituencies close to the Irish border (Murphy, 2018: 53). Both Scotland and Northern Ireland displayed similar demographic trends to those identified in England and Wales: those on lower incomes, those with fewer educational qualifications, older voters, and those opposed to immigration were more likely to vote Leave wherever they lived in the UK (Curtice, 2016;Garry, 2016). There were simply fewer Leave voters overall.
Where Northern Ireland stands out is in the strong ethnonational basis of the vote. Surveys suggested that 85% of Catholics voted Remain compared with 40% of Protestants, and 88% of those identifying themselves as nationalists voted Remain, compared with only 34% of those identifying as unionists (Garry et al., 2017). In Scotland, by contrast, identity was related to Brexit preferences, but in a less divisive way. Henderson et al. found strong Scottish identity, relative to British identity, to be a predictor of the Remain vote in Scotland, but with far less effect than the much stronger explanatory power of English identity among Leavers in England (Henderson et al., 2020). The prevalence of Scottish identity across the political spectrum has traditionally made it a relatively weak predictor of vote choice (McCrone, 2017).
The results of the referendum generated fresh and rather distinctive 'exit' debates in Scotland and Northern Ireland. Scotland's exit strategy is most evident in the SNP government's demand for a new independence referendum, with some evidence to suggest growing support among the wider population. Opinion polls have tracked the dynamics in support for independence since the 2016 Brexit referendum (see Figure 1). By the time the UK formally left the EU at the end of January 2020, the split between Yes and No in a hypothetical independence referendum stood at around 50:50 (excluding those undecided). The rise in support for independence in the year leading to 'exit day' occurred almost entirely amongst those who backed Remain in the Brexit referendum (Curtice, 2020). Against the backdrop of Covid-19, support increased further, with polls recording sustained independence majorities. The demand for Scots to have a right to determine their future is central to the SNP's ambitions as it heads toward re-election.
There are considerable legal barriers confronting the exit option in Scotland's case, however. Under the devolution settlement, the authority to hold a referendum intended to generate constitutional change lies with the UK parliament. In 2012, after months of negotiation, an intergovernmental agreement saw the Scottish Parliament temporarily empowered to legislate for a referendum that would be beyond legal doubt. Similar requests for a transfer of power have been rejected by Boris Johnson and his predecessor, Theresa May, on the basis that the issue had already been settled in 2014. Despite its electoral dominance in Scotland, the SNP cannot force concessions from a government not reliant on its votes in the House of Commons. The party thus puts its trust in continued electoral success and growing support for independence making the case to hold a new referendum irrefutable.
The politics of exit play out rather differently in Northern Ireland. Nationalist claims are not secessionist but irredentist, and are made within the context of a deeply divided community with a recent history of sectarian conflict. The 1998 Belfast/Good Friday Agreement has been central to peace and relative stability in Northern Ireland, but Brexit has threatened that stability and led to a new discourse on Irish unity north and south of the Irish border (Connolly and Doyle, 2019: 228). In the immediate aftermath of the 2016 referendum, Sinn Féin party chairman, Declan Kearney, noted: 'We have a situation where the north is going to be dragged out on the tails of a vote in England . . . Sinn Féin will now press our demand, our long-standing demand, for a border poll' (thejournal.ie, 24 June 2016).
On the question of Irish unity, survey data is sensitive to methodology, wording and political context (Connolly and Doyle, 2019). As indicated in Figure 2, polls conducted since the Brexit referendum have put support for Irish unity between 21% and 52%. Online polls identify a smaller proportion of 'don't knows' and suggest stronger support for Irish unity than those conducted in person, but it is not possible to determine which method is most representative of voters' views. In analysis derived from both survey data and a specially designed one-day citizens' assembly, Garry et al. (2020) observed increased sympathy for reunification amongst participants, linked to the effect of Brexit on the North-South border.  Recent elections have also sharpened the focus on constitutional issues. Although Sinn Féin's vote share fell in the 2019 UK general election, Northern Ireland elected more nationalist than unionist MPs for the first time (Tonge, 2020). Sinn Féin competes north and south of the Irish border and, in the 2020 Irish general election, it won 24% of first preference votes, more than any other party. During the campaign, the party had called for a border poll within 5 years, with a manifesto pledge to set up a parliamentary committee and an all-island Citizens' Assembly, and produce a White Paper on reunification (Sinn Féin, 2020). Following the election, the new Irish government created a 'Shared Island' unit in the Department of the Taoiseach. It touches on some of the issues linked to the united Ireland narrative, but does not take a position on any future constitutional scenario (BBC, 12 July 2020). In 2019, over 1,000 civil society representatives addressed an open letter to then Taoiseach (Prime Minister), Leo Varadkar, calling for a new conversation about the constitutional future of the island after Brexit (Irish Times, 4 November 2019). Despite some unionist leaders, including former DUP leader and First Minister, Peter Robinson, urging participation, unionist engagement with the exit discussion has been negligible.
In contrast to Scotland, the Belfast/Good Friday Agreement provides a legal route to exit for Northern Ireland. The mechanism, however, is not without ambiguity. The agreement gives responsibility to the Secretary of State for Northern Ireland to initiate a referendum on Irish unity when it appears 'likely' to them that a majority of Northern Ireland's electorate would support a united Ireland. However, neither the agreement nor the legislation which gave it domestic legal effect (the Northern Ireland Act 1998) specify precisely how this assessment should be conducted; for example, whether opinion polls or election results should provide the measurement of support for Irish unity. Nor is there clarity in relation to the timing of referendums north and south. The calling of a border poll would undoubtedly be a moment of high political drama that would be potentially destabilising, illicit strong, contested reactions and possibly judicial review (Murray and O'Donoghue, 2019). The route to exit is also dependent on support in the Republic of Ireland. Article 3 of the Irish constitution states that Ireland can only be reunified with the consent of the north and the south.
Thus, whereas Scotland is divided politically over independence, the divide in Northern Ireland is ethnocultural. Yet, despite contrasting and ambiguous routes to exit, electoral success and rising support in both cases have enhanced the credibility of these respective exit threats. Alongside the prospect of increased instability in Northern Ireland, these exit threats have been used in efforts to influence the Brexit process.

Territorial voice and the Brexit process
Voice strategies were evident in both Scotland and Northern Ireland, but in different ways and with different effects. In Scotland, the credibility of the threat of exit in the wake of Scotland's strong support to remain was used to bolster voice, in this instance as a complementary rather than an alternative strategy to exit. First Minister, Nicola Sturgeon, called on the UK government to prove the worth of the union by finding ways 'to demonstrate that Scotland's voice can be heard, our wishes accommodated, and our interests protected within the UK' (Sturgeon, 2016b). Arlene Foster and Martin McGuiness, then First Minister and Deputy First Minister, respectively, in the Northern Ireland Executive, also wrote to the Prime Minister to underline the key issues affecting Northern Ireland and their expectation of being 'fully involved and represented in the negotiations' concerning the Irish border (Northern Ireland Executive, 2016).
In the months after the Brexit referendum, Prime Minister Theresa May seemed committed to enhancing the voice of devolved governments in the Brexit process. During a visit to the Scottish First Minister, she declared a willingness to listen to options, insisting that she wanted 'the Scottish government to be fully engaged in our discussion. . . I have already said that I won't be triggering Article 50 until I think that we have a UK approach and objectives for negotiations' (BBC, 2016).
The four administrations agreed to create the Joint Ministerial Committee (EU Negotiations) (JMC EN) with terms of reference designed to provide an opportunity to the devolved administrations to voice their interests and shape the UK's approach (see Joint Ministerial Committee [JMC], 2016). It represented a new institutional channel through which the devolved administrations could exercise voice in UK policymaking and promised intergovernmental cooperation on a scale not previously seen since devolution (McEwen, 2017). Twenty meetings of the JMC (EN) were held between its inception and the signing of the UK-EU Withdrawal Agreement.
However, opportunities for influence were largely limited to the domestic implications of Brexit rather than the negotiations over the UK's withdrawal (McEwen, 2020). Short of the UK remaining in the EU, the Scottish government advocated for the whole of the UK to remain within the European single market and the EU customs union. In Scotland's Place in Europe, it also suggested an alternative option where Scotland could be given special status to maintain its place within the European single market even if the UK government sought a different outcome for England and Wales (Scottish Government, 2016). There is little evidence that these proposals were given serious consideration by the UK government, nor were they subject to meaningful discussion within the joint ministerial committee (McEwen, 2020). The Scottish government was largely excluded from the process of formulating the UK's Brexit policy and negotiating priorities. According to Scottish ministers, the Prime Minister's 'Article 50' letter to President Donald Tusk, which triggered the formal exit process, was neither shared nor agreed with devolved governments prior to being sent (Scottish Parliament, 2017).
Although the JMC (EN) continued to meet throughout most of the negotiations over withdrawal, it became largely a forum in which devolved administrations voiced their frustrations. The gap between the declared UK objectives and the Scottish government's preferences underlined the limited impact of the voice strategy in advancing what the Scottish government regarded as Scotland's territorial interests. In a letter to the Chancellor of the Duchy of Lancaster, Michael Gove, on the eve of the UK's formal departure from the EU, the Scottish Minister with responsibility for Brexit negotiations with the UK government, Michael Russell, asserted that: the views of the Scottish Parliament, and indeed the people of Scotland, have been consistently ignored by the UK Government throughout the process of EU exit. . . there has been no meaningful engagement with the Scottish Government either over the terms of the UK's withdrawal from the EU or the future relationship (Scottish Government, 2020).
If the Scottish government's voice was diminished in the Brexit process, an institutional voice for Northern Ireland was notable for its absence. Since 2007, the executive, when it functioned, has been led by the DUP and Sinn Féin. These parties were on opposing sides in the Brexit referendum, when the DUP championed a Leave vote while Sinn Féin strongly supported Remain. In the 7 months between the EU referendum and the collapse of the executive in January 2017, there was only one substantial Brexit debate, on a motion calling for recognition of 'special status' for Northern Ireland in the EU withdrawal process. It was narrowly defeated by 46 to 47 (NIA, 2016), exposing opposing nationalist and unionist responses to Brexit. After January 2017, there were no ministers and no opportunities to pursue a strategy of institutional voice. Brexit also shifted and realigned Sinn Féin's strategy, from support for power-sharing in the north to a clearer exit strategy that simultaneously leveraged for a border poll and Irish unity.
An inability to coherently articulate a shared Northern Ireland voice was all the more striking given the centrality of the Irish border in Brexit negotiations. The Northern Ireland Affairs Committee in the House of Commons noted in its inquiry into the collapse of devolved government that it 'has been particularly unfortunate since it has meant the voice of Northern Ireland has not been heard in London and Brussels as loudly as that of Scotland and Wales during the Brexit process ' (2018: 12). The political vacuum created by the suspension of the devolved institutions was partially filled by civil servants, but officials were severely constrained and unremittingly cautious in how they approached policy during this period. Intergovernmental relations on the island of Ireland were also disrupted as the collapse of the Northern Ireland Executive automatically suspended the North-South Ministerial Council (NSMC). The Irish government's attempt to provide an alternative all-island forum for Northern Ireland voices to express their views on Brexit was met with resistance by the DUP and other unionists, who feared that it lent credence to the nationalist exit strategy.
Northern Ireland was not entirely devoid of representation and voice during the first phase of Brexit negotiations though. Following Theresa May's failure to secure a majority at the 2017 general election, the DUP's 10 MPs used their balance of power to negotiate a Confidence and Supply Agreement with the Conservative government. This meant that, insofar as a Northern Ireland voice was being articulated, it was filtered through the DUP and therefore partisan and reflective of only some interests and constituencies in this divided territory. The Confidence and Supply Agreement committed the DUP to support the UK government's Brexit policy. However, when different iterations of the UK-EU Withdrawal Agreement were perceived to threaten the territorial integrity of the UK, the DUP was vociferous in summoning its oppositional voice. This put the party decidedly at odds with nationalist and non-aligned parties in Northern Ireland who were supportive of Theresa May's withdrawal legislation, in particular, the 'backstop' designed to ensure an open border on the island of Ireland.
The final UK-EU withdrawal deal of October 2019 dropped the 'backstop' provision and also patently ignored unionist voices, and was rejected by all swathes of political opinion in Northern Ireland. On 20 January 2020, the Northern Ireland Assembly -in one of its first votes following its 3 year suspension -passed a motion withholding consent for the EU (Withdrawal Agreement) bill, the UK legislation giving legal effect to the UK-EU Withdrawal Agreement. The position was agreed without dissent and was supported by both nationalist and unionist Members of the Legislative Assembly (MLAs), albeit from contrasting motivations. Unionists objected to provisions in the protocol requiring customs and regulatory controls at the Irish Sea border, interpreted as undermining Northern Ireland's constitutional position as an integral part of the UK. In contrast, nationalists and other non-aligned political parties' opposition to Brexit and their dissatisfaction with the technical complexities involved in operationalising the protocol motivated their decision to withhold consent.
The MLAs' decision was an expression of the procedural mechanism, more commonly known as the Sewel convention, that gives the devolved institutions the opportunity to give or withhold consent for UK legislation that affects their competences. This does not amount to a constitutional veto. The UK parliament is sovereign. It can make or unmake any laws, including in devolved areas. However, the convention that it will not do so in relation to devolved matters or competences without the expressed consent of the devolved institutions has become an important principle underpinning UK devolution. Within the parameters of the UK constitution, it is the quintessential institutional expression of territorial voice. In the event, all three devolved legislatures withheld their consent, the first time in the history of devolution that they have done so simultaneously. That this did nothing to sway or influence the withdrawal legislation exposes the limitations of the voice strategy when those to whom voice is targeted are not minded to heed it.

Prospects for re-entry
Hirschman's framework centred on the pulls of exit and voice as alternative choices for those left dissatisfied by a product, service or organization. Although Hirschman briefly considered re-entry, it was as a move back to the entity the individuals had left, in an expression of loyalty or a belief that the source of grievance had been addressed (Hirschman, 1970: 89). In the territorial communities under consideration here, re-entry would not be to the state they had left. In the Scottish case, re-entry would be sought to the EU, whilst simultaneously entering a new inter-state relationship with the rest of the UK, its closest neighbour economically, politically and culturally. For Irish nationalists, re-entry would be to the Ireland from which the north was partitioned in 1921 and, as a corollary, to the EU, given Ireland's status as an EU member state. Of course, in each case, reentry is conditional upon exit, and both the prospects of, and the paths toward, re-entry belong in an uncertain future. However, Brexit has generated new challenges and opportunities associated with re-entry that merit examination for their capacity to shape ongoing exit debates.
In some respects, the route to re-entry to the EU for an independent Scotland is more straightforward than in 2014. Under such a scenario, Scotland would be compelled to follow the wellestablished process of accession under Article 49 of the Treaty on European Union, subject to satisfying Copenhagen criteria for membership and securing the agreement of existing member states. This suggests accepting all rights and obligations of EU membership, unless there was an exceptional case for securing a derogation (Börzel et al., 2017). The SNP government is committed to maintaining continued alignment with EU regulations within its areas of competence in part, as noted by the First Minister, to 'make it easier, when the time comes. . .for Scotland to return to the EU' (Sturgeon, 2020). But there are many areas where EU regulations apply to policy fields that are beyond the competence of the Scottish Parliament. If UK-EU regulatory divergence increases over time, Scotland at the point of independence would have departed from the acquis by default. Adopting the acquis in its entirety would also require a departure from the 2013 independence vision, which assumed continuation of UK opt-outs, rejected the Euro for a currency union with the rest of the UK, and rejected the Schengen arrangements for continued inclusion in the UK-Ireland Common Travel Area.
In other respects, the path to EU re-entry for Scotland is also more problematic as a result of Brexit. If Scotland was to accede to the EU as an independent member state, the Scottish-English border would become another external EU border. Even if, following the Irish precedent, an independent Scotland could secure a derogation from the Schengen arrangements permitting continued free travel throughout the UK and Ireland, it would still be required to oversee a regulatory and customs border. The extent to which this matters in practice will clearly be shaped by the evolution of the relationship between the UK and the EU. The dominance of the Irish border during Brexit negotiations demonstrated the importance the EU attaches to the integrity of the single market and unearthed the complexities of border management between distinctive regulatory and customs regimes. These border issues could represent a significant barrier to securing support for independence in any future referendum.
In the case of Northern Ireland, 'exit' from the UK implies 're-entry' into the EU via the reunification of Ireland. While the legal routes to re-entry in that context are somewhat clearer, the split loyalties in the north make both exit and re-entry fraught with difficulty. As the prospect has only begun to be given serious consideration, much about the process remains unclear. Murray and O'Donoghue (2019) noted that the principles and process set out in the Belfast/Good Friday Agreement suggest that reunification would involve stitching together two polities rather than subsuming Northern Ireland into the Republic of Ireland. This would require navigating immense constitutional, political, economic and public policy challenges, carrying the potential to breed tension and instability on the island between North and South, as well as between unionists and nationalists. This may explain why most parties, Sinn Féin excluded, have been reticent about planning for the possibility of Irish unity. Prior to the 2020 general election, the Irish government had not produced any plans or proposals. In contrast to Sinn Féin's proactive push for unity, the Social Democratic and Labour Party (SDLP) is more cautious in its support of 'progressive nationalism' (Eastwood, 2016). This caution stems from sensitivity to unionism and concerns about the disruptive impact of a unification process. As Taoiseach, Leo Varadkar voiced during a debate in West Belfast: I think it [reunification] would result in some of the mistakes made 100 years ago, when partition happened, being repeated but just the other way around -a huge number of people, those from a unionist, British, Ulster background, being brought into a united Ireland against their will (BBC, 6 August 2019).
For Northern Ireland, the legal process of re-entry to the EU is expected to be less fraught in the event of reunification. The Ireland/Northern Ireland protocol of the Withdrawal Agreement already stipulates that Northern Ireland will remain within the EU single market for goods, irrespective of the path taken by the UK government. This should make it easier to demonstrate that Northern Ireland had maintained the acquis in these areas. Moreover, Northern Ireland would not be seeking membership for itself, nor would it be required to establish the suite of new regulatory bodies necessary for an independent EU member state. Rather, it is expected that reunification, from the EU's vantage point, would see a redrawing of the boundaries of Ireland as an existing member state, much as East Germany assumed membership as part of Germany after reunification. When the European Council was agreeing guidelines for Brexit negotiations in 2017, former Irish Taoiseach, Enda Kenny, secured an additional declaration on Ireland: The Good Friday Agreement expressly provides for an agreed mechanism whereby a united Ireland may be brought about through peaceful and democratic means; and in this regard, the European Council acknowledges that, in accordance with international law, the entire territory of such a united Ireland would thus be part of the EU (European ).
This suggests a commitment on the part of the EU to facilitate Northern Ireland's re-entry in the event of Irish unity. However, this could further distance it from the rest of the UK, creating more border barriers in the Irish sea. Maintaining a close and constructive relationship with the rest of the UK would be politically (and symbolically) important for Unionists in particular, as well as economically important for businesses throughout the island.

Conclusion
Adapting Hirschman's framework to UK territorial politics, this article has identified that the Brexit referendum, and the process that unfolded thereafter, has disrupted the UK's territorial constitution, created limited opportunities for the devolved institutions to have an influential voice, and re-energised the politics of exit in Scotland and Northern Ireland.
We expected that the ethnocultural divisions that continue to dominate politics and representation in Northern Ireland would constrain both exit and voice. Certainly, the collapse of devolved government to which these tensions contributed meant that the institutional channels open to the Scottish government to advocate distinctive territorial preferences were not open to Northern Ireland. Even if they had been, finding common ground would have been difficult. Nationalist calls for 'special status' for Northern Ireland were utterly opposed by Unionists, who used their partisan voice during a critical period of Theresa May's premiership to block this outcome. The unquestioning loyalty of Ulster unionists means they lack a credible exit option that, in the Hirschman framework, might be expected to bolster voice. Their ability to influence UK policy in this period was down to parliamentary arithmetic. The Conservative Party's sizeable victory in 2019 saw the DUP lose not just seats, but also its voice. Despite the nationalists' ability to use the threat of exit to secure concessions, the suspension of Stormont and their lack of representation in Westminster during this period seriously compromised their ability to meaningfully affect the UK's Brexit policy. On the other hand, the Irish government, through EU channels, had considerable influence in withdrawal negotiations. Their support for special status for Northern Ireland approximated the nationalists' original ask. Consequently, the UK-EU Withdrawal Agreement and the Ireland/ Northern Ireland Protocol reflected some nationalist concerns. This points to the complexity of the relationship between voice and exit in Northern Ireland. It is based on a diversity of messages delivered by different parties with opposing agendas, selectively targeting (supposed) allies and actors in Dublin, London and Brussels.
In Scotland, despite the legal barriers to holding an independence referendum, the electoral dominance of the pro-independence SNP makes exit a plausible strategy. But, contrary to our expectations, while the threat of independence was used to leverage Scotland's voice in the Brexit process, its limited impact underlined that louder voices do not always lead to effective influence. The early promises of close intergovernmental cooperation to agree a UK approach to Brexit negotiations came to very little. Three years of frustrated efforts to shape Brexit priorities, negotiations and outcomes exposed the limited opportunities to use voice in the absence of constitutional veto rights. In turn, this reaffirmed the SNP government's preference for an exit route, with limited opportunities for an influential voice now used to generate support for exit, with some evidence of growing popular support.
The UK case has underlined the difficulties that can be generated by imposing majoritarian decisions in a territorially diverse society, especially when that diversity is manifested in divergent political preferences. There were no geographic thresholds to be met to legitimize the outcome of the referendum in the eyes of the Westminster elite. Like all referenda in the UK, the 2016 vote was non-binding legally, but was perceived to be binding politically despite the lack of consensus across the UK's territorial communities.
Devolution was introduced as a strategy of territorial management and, in Northern Ireland, conflict resolution. Yet, Brexit has represented a shock to the system, destabilizing the delicately balanced territorial constitution and exposing the limits of devolved power in an otherwise centralized state. In a federal system, a constituent unit faced with a similar shock or challenge to its authority may be able to exert greater institutional voice. Channels of intergovernmental cooperation and codecision tend to be more firmly institutionalized in federations, and constitutional or institutional vetoes used as leverage in intergovernmental bargaining. Devolution places no such constraints on the ability of central government to act, but our study has also underlined the limits and contingencies to which such a system can be subject. Faced with the shock of Brexit, and in the absence of institutional requirements or the political will to accommodate diverse preferences, the terms of the devolution settlement have proved vulnerable. As a strategy of territorial accommodation, devolution is always a process, not an event. It is never likely to be a permanent solution that eradicates calls for secession (or irredentism). Without careful management, secessionist goals can be re-ignited by a crisis or catalyst that undermines the political settlement.
Yet, when it comes to territorial communities, exit, were it to take place, is never final. Rather, it marks a step-change to a new relationship between the territorial community and the state it is leaving, as well as with the international community. Our concept of re-entry adds a new dimension to Hirschman's framework when applied to the study of territorial politics. Comparing re-entry across the two cases in advance of a hypothetical exit is of course impossible, but the article has illustrated the distinctive challenges that each territory would be expected to face. For Scotland, it would be seeking to rejoin the EU whilst negotiating a new relationship with rest of the UK that avoided a 'hard' Anglo-Scottish border. For Northern Ireland, whilst the route to Irish reunification and thus re-entry to the EU may be legally less fraught, the process would likely make internal divisions even more pronounced. These hypothetical scenarios colour contemporary debates. The future of the Union, and the place of Scotland and Northern Ireland within it, will be shaped fundamentally by the relationship that the UK develops with the EU. That relationship will be subject to its own dynamic. The more the UK withdraws from relations with the EU, the greater may be the strains at its internal boundaries. That creates dilemmas as well as opportunities for nationalists seeking the exit route. The appeal of exit will also be coloured by how the UK government responds to these territorial challenges, and whether it creates or thwarts opportunities for territorial voices to be heard.