Introduction: Resilience, objectivity and conviction
Judicial review of consultative secession referenda is a form of existential constitutional contestation and a test of the conditions for the democratic credibility of any subsequent deliberative settlement. In one of his less theoretical essays, titled Objectivity and Truth: You’d Better Believe It, Ronald Dworkin argues that pragmatism, among other considerations, is the fundamental reason why ‘we’ trust in judicial objectivity. According to Dworkin, there is no external Archimedean point from which moral or judicial ideas can be neutrally judged.Footnote 1 Dworkin locates its authority in the practical resilience of ‘conviction’. If, after sustained reflection, a belief withstands every reasonable challenge and no superior alternative appears, he argued, we are entitled to rely on it.Footnote 2 In other words, the point of judicial objectivity is not that conviction makes a claim True with a capital ‘T’ but that its resilience against reasonable doubt renders further scepticism unwarranted.
In a multinational system, a reasonable doubt about the court’s legitimacy may act as the harbinger of a deeper constitutional rupture, placing negotiation under strain and unsettling the institutional balance among regional identity groups, the central state and the rule of law. In a short period of time, three cases assessed the legitimacy of consultative referenda on secession in Italy,Footnote 3 Spain,Footnote 4 and the UK.Footnote 5 These judgements were selected because they represent apex judicial reasoning across three separate constitutional systems spanning civil and common law traditions.Footnote 6 They are existential for the future of the state’s territorial integrity, and they might indeed be a political proxy for secession. However, they are not part of violent conflict, which denies the legitimacy of state institutions a priori . Instead, these cases are examples of the processes that ensure referenda meet the standards of reasonable deliberation, political negotiation and the protection of human dignity.Footnote 7 They are thus distinguished from post-conflict precedents such as Kosovo and Bougainville, where the legitimacy of the process was driven, among other factors, by the inability of Serbia and Papua New Guinea to guarantee peace and protect rights.Footnote 8 Handed down within a relatively short period, these three cases offer a concentrated snapshot of how supreme courts understand judicial objectivity when faced with structural changes related to secessionist claims. Crucially, all three courts operated under the additional scrutiny of international norms: the European Court of Human Rights’ appearance-of-bias doctrine and, for Spain and Italy, the Court of Justice of the European Union’s requirements for judicial independence.Footnote 9 This dual filter exposes the strengths and structural limits of domestic adjudication in managing secessionist pressures.
Scholars have argued that constitutional courts are ill-suited to the task because their interpretive frameworks often embed a foundational commitment to state unity, at times shading into constitutional casuistry when the normative basis of that premise is left insufficiently explained.Footnote 10 Tom Ginsburg and Mila Versteeg reviewed 200 constitutions, and only five allowed a secessionist referendum.Footnote 11 However, there is a grey area surrounding the consultative referenda. In advanced deliberative democracy, which relies on individual and mass commitment to the practice of deliberation, any debate over future political negotiations should, in theory, be good for an open deliberative democracy.Footnote 12 However, that is an idealistic assumption.Footnote 13
There are good reasons for a constitutional court to protect institutional stability, safeguard democracy, foster a culture of human rights and uphold the rule of law.Footnote 14 Erin Delaney and Gabrielle Appleby explain that courts also need to be accepted by ‘the society they serve’.Footnote 15 That multidimensional activity includes, among a plurality of activities, preventing any ethnic group from tilting the balance between an open political arena and the protection of rights and liberties.Footnote 16 This difficulty arises when substantial segments of a region perceive the court as advancing a nationalist ideology. These regional perceptions, or convictions à la Dworkin, of central state nationalist bias are also exploited by neighbours with imperialist aims. Led sometimes by demagogues, these neighbours employ such perceptions as a pretext for invasion or, bizarrely, as grounds for an offer to be bought.Footnote 17
There is little, sceptics might argue, that courts in the pursuit of objectivity can do to restrain demagogues. In this essay, the assumption that nationalism has the potential to divide us is treated as a historically settled question.Footnote 18 The issue explored here is, instead, what final appellate courts should do in addressing what Ginsburg and Versteeg call the grey area surrounding consultative referenda on secession.Footnote 19 Courts must balance two claims. On the one hand, they must protect the constitutional system as a shared project for the entire political community. On the other hand, they must take seriously a clearly expressed majority claim by a substate community to decide its constitutional future. I argue that when courts decide legal questions arising from the tension between these two claims, they must not only reason objectively but also appear objective, using universally accepted standards for the communities whose constitutional future is at stake. The comparative evidence evaluated in this essay suggests that three courts, in three different cases, failed to meet international standards of judicial objectivity.Footnote 20 They are the Italian Constitutional Court’s Decision n.118 of 2015 regarding the Venetian secession,Footnote 21 the Spanish Constitutional Court’s decision, Sentencia n.90/2017 of 2017,Footnote 22 concerning Catalonia and the Reference by the Lord Advocate, which discussed a bill preparing a consultative referendum in Scotland.Footnote 23 In all these instances, a review of the court cases shows evidence that at least one judge had a personal or relational connection to the outcome of a consultative referendum on secession. For example, two Scottish judges in the Reference by the Lord Advocate simultaneously decided on Scotland’s future constitutional status and the fiscal status of their current or future financial assets in Scotland and in the rest of the UK.
A few preliminary issues need to be dealt with before I articulate my claim. First, the essay does not seek to prove an actual judicial bias. This is, indeed, not a report of an investigative activity. The essay asks whether, in these three disputes over consultative secession referenda, publicly available information could lead a fair-minded observer to conclude that established international and national norms concerning the appearance of objective adjudication were not met. This principle is expressed in the trope: ‘justice must not only be done but must also be seen to be done’.Footnote 24 This standard, originating in R v Sussex Justices, ex parte McCarthy, was subsequently adopted by the European Court of Human Rights (ECtHR) and Court of Justice of the European Union (CJEU).Footnote 25
Second, the article does not argue that courts should refuse to consider the admissibility of consultative referendums simply because questions of state dissolution are political or because courts, as part of the national state apparatus, may be affected by ideological nationalist or unitary bias.Footnote 26 The claim is more precise. Courts have a legitimate role in assessing the legality of such initiatives and promoting a statewide communal project, but these two roles become constitutionally fulfilled only if the court can satisfy the standard of apparent objectivity. Modern democracy rests on the interplay between the rule of law and popular decision-making, a bond that James Tully identifies as the defining feature of modern indirect democracy. He explains that ‘[a] political association is legitimate if and only if it is equally constitutional and democratic’.Footnote 27 Courts occupy a pivotal position in this regard by sustaining the link between indirect democratic practices and a modern liberal constitutional system that prioritises liberal values and human dignity. They operate under the understanding that democratic activities occur within legal and deliberative boundaries, in which the force of the better argument should prevail. Jürgen Habermas’s analysis is helpful here, distinguishing statutory policy, concerned with facts, from public deliberation, concerned with norms and assigning courts the task of guarding deliberative freedom while enforcing outcomes.Footnote 28 Accordingly, a regional government that initiates a consultative referendum on secession does not override the normative values embedded in the rule of law and the role that the court must have in assessing the legitimacy of these political initiatives. Richard Bellamy, in assessing the justifiability of political questions, argued that overtly political questions should lie beyond judicial review.Footnote 29 Bellamy explained that courts have a poor record of protecting regional minorities and often appear beholden to a foundational commitment to state unity, which entrenches social divisions. He concluded that ‘[a]rrangements that institutionalise community divisions […] entrench the positions of elites, thereby removing their incentives to either better serve the public interest or accommodate others’.Footnote 30 There is merit in Bellamy’s scepticism, but in this article, it is assumed that courts can, and indeed must, question the legitimacy of consultative referenda over secession.Footnote 31 One of constitutionalism’s central achievements lies in its capacity to mediate the relationship between the individual and the political community. This essay proceeds from the axiomatic premise that, at present, there is no practical, viable alternative to constitutionalism that can protect this mediation process and, at the same time, ensure respect for human dignity.
Third, this essay will not engage with the extensive literature on regional self-determination and constituent power, as it remains orthogonal to the present inquiry. The question of where absolute sovereignty lies and how it connects to secession or federalism belongs to a longstanding and well-rehearsed debate.Footnote 32 Pau Bossacoma Busquets provides a clear articulation of the dilemmas of accommodating multiple claims of sovereignty.Footnote 33 For a review of the relationship between constitutional law and constituent power, Nick Aroney, George Duke and Stephen Tierney have produced one of the most authoritative recent accounts.Footnote 34 Regarding the practical challenge of bootstrapping legitimacy outside formalised deliberation, a concise summary can be found in Ron Levy’s article titled Right to a Referendum, or Duty to Deliberate? Rethinking Normative Entitlements to Secession. Footnote 35
Fourth, for clarity, the phrase ‘judicial review of consultative/secession referenda’ is abbreviated to CRiS, and the ‘fair-minded, informed external observer’ standard is referred to as the IEO test. Fifth, consultative secession referenda in Spain and Italy could deprive residents, including members of the judiciary, of EU citizenship, which might affect a large section of the judiciary. A full doctrinal and empirical exploration of that consequence would exceed the article’s permitted length and divert attention from its central focus: the role of the perception of objectivity in a CRiS.
Courts and secession referenda
In disputes over consultative secession referenda, judicial objectivity carries wider constitutional significance because the court intervenes in a live moment of constitutional contestation. The stakes of such disputes should elevate, rather than dilute, the standard expected of an institution whose primary function is to uphold the rule of law.Footnote 36
Recent scholarship on multinational constitutionalism engages the role of courts and the question of whether identity-based claims can be accommodated within a constitutional system. Héctor López Bofill’s article is particularly useful on this point because it shows how constitutional design and judicial doctrines favouring territorial integrity can reduce the room for negotiated outcomes.Footnote 37 The same pattern may appear in systems in which ideology and wider political strategy shape the terms on which distinct national communities seek recognition, accommodation or constitutional change. Pau Bossacoma’s essay suggests that there is still theoretical room for improvement,Footnote 38 and Brown, Swan and Anderson show that centralised and hierarchical state responses may intensify the distance between central institutions and regional political aspirations rather than help build a democratic settlement that strengthens a multinational constitutional system.Footnote 39
A recurring theme in comparative public law is that relations between central and regional institutions are normally constitutionally unsettled. They are shaped through an ongoing multilayered interaction in which political negotiation, constitutional design and judicial intervention each play a role. Francesco Palermo and Karl Kössler explain, in one of the leading comparative works on federalism, that constitutional design and political negotiation are closely intertwined with adjudication.Footnote 40 This dynamic is most visible when one side refuses further engagement and disputes over autonomy, referendum design or constitutional status are reframed as legal questions for courts or other constitutional actors. In that setting, the legitimacy of adjudication depends not only on legal authority but also on the expectation that courts will be, and will appear to be, objective and impartial.Footnote 41
However, there is relatively little clarity regarding the precise kinds of processes that deliver a convincing, à la Dworkin, objective decision.Footnote 42 The most persuasive account is offered by Jules Coleman and Brian Leiter, who associate judicial reasoning with the idea of modest objectivity. They frame this as an epistemic method that dismisses a commitment to strong forms of justification due to the interpretive indeterminacy inherent in working with legal texts.Footnote 43 Courts assess the persuasiveness of legal arguments and evidence using robust reasoning standards and reach consensus around the epistemic method that guides textual interpretation;Footnote 44 however, there is no unified theoretical account of how these are integrated into judicial practice.Footnote 45 Instead, judges have national practices that partially overlap across national and international jurisdictions.Footnote 46
Most liberal democracies, both within and beyond the group of signatories to the European Convention on Human Rights, apply the IEO test as the standard for assessing judicial impartiality.Footnote 47 Whether through domestic constitutional law or alignment with Article 6 jurisprudence, the principle that ‘justice must be seen to be done’ measured by the IEO has become a foundational element of judicial legitimacy across these systems.Footnote 48 Therefore, judges must remain vigilant regarding the public perception of their monetary interests and extrajudicial activities, as both institutional legitimacy and public trust in judicial decisions rely on this perception. For example, Lord Hoffmann ought to have recognised that the implications of his position on the board of Amnesty International Ltd would be perceived as overlapping with the allegations discussed against the former dictator of Chile, Mr Pinochet. Footnote 49 Similarly, in Hoekstra and Others v HM Advocate, Lord McCluskey’s extrajudicial comments, published in a newspaper, should have triggered immediate self-recusal.Footnote 50 In civil law systems, the IEO test is codified, but its practical impact on judicial practices remains similar to common law practices. In France, for instance, Article L. 111–6 of the Code de l’organisation judiciaire sets the general duty of recusal where legitimate doubt exists, while Articles 341 et seq. of the Code de procédure civile implement this in civil cases.Footnote 51 The Italian Codice di procedura civile (Article 51) and German Gerichtsverfassungsgesetz (§ 41) codify analogous requirements.Footnote 52
Outside the self-recusal process, there are two categories of disqualifying relationships that judges should be aware of. The first concerns a factual connection between the judge and the subject matter of the case. This is a personal connection with the legal matter. In the French legal system, for instance, a judge can be alleged to be biased if they are a creditor, debtor, presumptive heir or beneficiary of one of the parties. The second category pertains to a relational connection with the case. Again, the Code de l’organisation judiciaire provides a precise definition of this situation: disqualification applies when there is a well-known relationship of friendship (or enmity) between the judge and one of the parties.Footnote 53
The combination of the self-recusal procedure and alleged bias by the party indicates that judges are expected to have, in the majority of instances, the cognitive capacity to recognise potential biases directly or as part of peer assessments.Footnote 54 If the issue is raised during proceedings and the judge considers it insubstantial, the concerned party may request an incidental assessment of the alleged bias. Where the problem is discovered after judgement, the party may seek a review. Regardless of whether the connection is personal or relational, or whether the question is raised before or after the trial, both scenarios are assessed according to the perspective of an IEO. If the evaluation occurs prior to the trial, the judge may be dismissed. If it occurs post-trial, the substantive issue is reviewed again by a different panel of judges.
In this setting, visible impartiality is one of the institutional conditions for trusting in the lawful management of all legal matters, including referenda that might divide a polity.Footnote 55 A review of judicial bias in the constitutional review of the CRiS should be in line with other ancillary reviews of bias. The difference lies in the practical consequences because secession has an extremely wide range of implications. For instance, a judge may hold a tangible stake in the resolution of the dispute, such as a direct financial interest or ownership of property in a seceding region. Alternatively, following secession, the judge might no longer meet the citizenship requirements to practise in the country.Footnote 56 In Italy, for instance, access to the judiciary requires Italian citizenship, and the process is governed by the concorso in magistratura, which is regulated under the Testo unico delle disposizioni concernenti lo statuto degli impiegati civili dello Stato. Footnote 57 In Spain, judges are required to be Spaniards, and the Constitutional Tribunal likewise confines membership to Spanish citizens.Footnote 58 In the case of Scottish independence, for example, the Scottish government has declared its intent to repatriate judicial competences, which makes the future role of Scottish justices sitting in London personally connected to the case.Footnote 59 In practical terms, judges such as Lord Reed and Lord Hodge, in the Reference to the Lord Advocate, also weighed a cascade of personal life choices, including where to reside and whether to remain in a court that may no longer engage with Scottish legal disputes.Footnote 60
The second category of alleged bias is known as relational bias (or association bias), which occurs when a judge maintains a personal connection with the individuals involved in the case. Assessing this risk involves two steps: first, the judge’s relationship with the party must be close enough to raise doubts about impartiality; second, the party’s interests must be directly engaged in the outcome of the dispute.Footnote 61 The connection and its relevance need not be proven as a matter of fact. It is sufficient that a fair-minded and informed observer perceives elements as potentially detrimental to the judge’s objectivity.Footnote 62 In secession cases, it is sensible to assume that judges involved in reviewing a CRiS will know junior colleagues whose careers, and whose future citizenship status, may be dependent on the state’s secession negotiations. That might be someone whose job prospects would be curtailed in a successor state (e.g. via citizenship-based disqualification like in Italy and Spain).Footnote 63 It is even more probable that a judge will have familial or commercial ties to entities that stand to benefit or lose from the economic consequences of secession. For instance, in the context of the Veneto independence referendum, publicly available information shows that Judge Marta Cartabia’s husband, Giovanni Maria Grava, serves as managing director of Tutela Legale S.p.A., an insurance company whose services are distributed across Italy, including by multiple brokers across the Veneto region.Footnote 64 As secession does not guarantee automatic admission to the EU, the ‘service’ included in the insurance products offered by Mr Grava’s Milan-based company might cease to be legally available in an independent Veneto.Footnote 65
I will return to some of these examples later in the text, but there is already enough to show that selecting who should decide litigation over a CRiS must be a particularly careful exercise. This is not because judges sitting on a national court are incapable of deciding such cases, as they plainly are competent to do so, but because the constitutional review of a CRiS carries unusually far-reaching effects. If that concern is not taken seriously, it brings a broader erosion of the court’s perceived legitimacy by a large group of individuals whose identity is often associated with the history of discrimination and, in turn, of the constitutional system’s long-term capacity to sustain a multinational deliberative constitutional system. The comparative analysis that follows shows how this standard operates when constitutional conflict reaches a structurally transformative level and when adjudication shapes the credibility of any future constitutional settlement.
The jurisprudence of bias: A pragmatic solution to a normative issue
The current judicial practice, as developed by the ECtHR, the CJEU and domestic courts, requires judges to be perceived as impartial and objective. In the previous pages, I explained that judicial objectivity is fundamentally rooted in pragmatic judicial practices. In this section, I will review the European jurisprudence that establishes the standard for an objective court and then proceed to analyse whether the panel compositions in the three CRiS cases appear to meet this test.Footnote 66
Article 6(1) of the ECHR provides that ‘everyone is entitled to a fair and public hearing within a reasonable time by an independent and impartial tribunal established by law’. The ECtHR has drawn from this text a well-settled orthodoxy that judges must not only be impartial but also appear to be so.Footnote 67 In Piersack v Belgium, the court held that a tribunal might be perceived as biased even in the absence of proof of actual unfairness, and in De Cubber v Belgium, it affirmed that apparent bias alone could render a decision invalid. Their articulation of the IEO leaves little space for doubts: ‘[E]ven appearances may be important; in the words of the English maxim quoted in, for example, the Delcourt […] “justice must not only be done: it must also be seen to be done”’.Footnote 68 In Hauschildt v Denmark, the court clarified that the IEO test focuses on public perception rather than proven prejudice.Footnote 69 In other words, public perception alone would render a judge unsuitable to decide a case.Footnote 70 In Langborger v Sweden, the ECtHR provides further confirmation that the ideal external observer should be informed of all relevant circumstances that might support or refute the allegation of bias.Footnote 71 This review can go on, but there is enough in the analysis to support its claim: the ECtHR’s interpretation of Article 6 is that judges must appear to be objective.Footnote 72 The criteria that fulfil the requirement are also applied directly to Spanish and Italian courts.Footnote 73
Until 1998, the UK developed its IEO test without authoritative reliance on ECtHR jurisprudence. Following the enactment of the Human Rights Act 1998, which incorporated a selected number of convention rights into domestic law, British courts began to draw expressly on the authority of the ECtHR, and they have since treated that line of reasoning as authoritative when applying the IEO standard.Footnote 74
Earlier Scottish cases, such as Bradford v McLeod and Hoekstra v HM Advocate, had already adopted an appearance-based approach.Footnote 75 The leading case, however, is Porter v Magill. Lord Hope cleared the debate over the role of the IEO in a lucid narrative. ‘I would however delete from it the reference to “a real danger” […] The question is whether the fair-minded and informed observer, having considered the facts, would conclude that there was a real possibility that the tribunal was biased’.Footnote 76 Lord Hope deprecated references to ‘a real danger’ and affirmed, with his traditional acuity, that the true question is whether a fair-minded and informed observer would perceive a real possibility of bias.
Northern Irish case law also adopts the fair-minded and informed observer standard set out in Porter v Magill but with a distinctive contextual difference. For a period of time, the region’s public institutions, including courts, have been perceived as captured by a sectarian group. Northern Ireland is a post-conflict society that endured roughly three decades of violence, also known as the Troubles. Footnote 77 During the Troubles, state institutions, including the judiciary, were perceived as dominated by unionists, and several judges were assassinated by paramilitary groups. This legacy of sectarian division means that public sensitivity to potential bias remains acute. Courts have therefore been explicit in their application of the IEO standard.Footnote 78 In Re O’Connor’s and Broderick’s Application for Judicial Review, which considered alleged bias on a police disciplinary panel, the Queen’s Bench Division highlighted the distinction between perception and proof of prejudice, stating in a clear narrative: ‘Apparent bias does not depend upon proof of prejudice but is concerned with perception’.Footnote 79 Again, I can continue to provide evidence on how the IEO test is woven into the fabric of the British, Italian and Spanish legal systems, but there is sufficient clarity on the issue to suggest that the appearance of bias is a default standard adopted within these jurisdictions.
A Scot, a Venetian and a Catalan Enter a Courtroom
The Ultra Vires Scottish Government Referendum Bill
This section analyses the details of the Reference by the Lord Advocate, the Sentencia 90/2017 and the Sentenza 118/2015. Footnote 80 The previous section explained that judicial objectivity is assessed through personal and relational factors from the perspective of the IEO. European and UK doctrine converge on the idea that what matters is not a judge’s actual ability to reach a fair decision but whether a fair-minded and informed external observer would think there is a possibility of bias.
The three cases differ significantly in context and legal history. In the UK Supreme Court, the issue was not whether a referendum was possible in principle but whether a devolved institution had the legal competence to introduce legislation whose practical effect was to prepare for such a referendum. Sentenza 118/2015, in contrast, formed part of a broader clarification of what the 2001 constitutional reform, often presented as having strengthened regional autonomy and federalising tendencies, actually entailed in relation to the principle of the unity of the Republic.Footnote 81 The Spanish case was different again. It was inserted into a longer process of escalating institutional confrontation between Catalan institutions and the central state, a process that had been developing at least since the Spanish Constitutional Court’s partial invalidation of the Catalan Statute of Autonomy.Footnote 82
These differences have two consequences for the present analysis. First, a close textual comparison of the three judgements has limited explanatory value because the constitutional settings, legal questions and political contexts were substantially different. Second, and more importantly for the argument advanced here, those differences bring the objectivity question into sharper focus. The central issue is not whether the courts reached the same doctrinal conclusion – or even whether their legal reasoning followed comparable paths. The question is whether the composition of each panel could withstand scrutiny under the international standard of judicial impartiality that all three legal systems endorse. Put simply, the question is whether each court satisfied the IEO standard.
In the Reference by the Lord Advocate, the Scottish justices determined their professional futures and those of all Scottish judges in two fundamental respects.Footnote 83 First, they ruled on matters affecting the jurisdiction of the Supreme Court, which had multiple implications for their current legal status as justices and their future professional lives.Footnote 84 Second, as a corollary of the first, secession is likely to have a putative financial impact on judges’ assets and pensions.Footnote 85
The Reference by the Lord Advocate clarified a grey area of the UK legal system by confirming that the Scottish Parliament lacks competence to debate a bill that would have prepared for a CRiS. The constitution of the UK of Great Britain and Northern IrelandFootnote 86 is flexible, uncodified and plurinational; it contains no explicit separation of political powers because they are understood to be fused in the absolute legislative sovereignty of the King in Parliament.Footnote 87 The Scottish Parliament has the power (ex-Section 2 (a) Schedule 5) to hold referenda and consultations and to call for citizen assemblies.Footnote 88 While holding advisory referenda on secession is not a matter that is directly reserved to the UK Parliament in Schedule 5 of the Scotland Act 1998,Footnote 89 the ordinary meaning of Paragraph 34 of Schedule 6 1(f) empowered the Supreme Court to argue that the question posed in Section 2 (2) of the Scottish government’s bill (that is, ‘Should Scotland be an independent country?’) had the potential to alter the UK’s governance system and impinged on the prerogatives of the UK Parliament.Footnote 90 The Supreme Court stated that this is a situation even if neither the bill nor the referendum related to it has any legal repercussions. The reason for intervening in the Scottish Parliament activity is explained in a plain narrative: ‘That legislative framework is put in place because the result of a lawfully held referendum is a matter of importance in the political realm, even if it has no immediate legal consequences’.Footnote 91 Thus, even without an express statutory ban, the reference mechanism empowers the court to prevent a devolved deliberative exercise that undermines the meaning and purpose of the act, thereby preserving the constitutional order. This conclusion is a good example of the type of balanced exercises the court should do, as mentioned earlier, and it is technically sound.Footnote 92 As Ginsburg and Versteeg’s analysis would have predicted, the UK Supreme Court has assumed the role of guardian of constitutional stability.Footnote 93 Brown Swan and Anderson show that the UK Supreme Court’s intervention formed part of a broader pattern through which central institutions managed and constrained the prospect of a second referendum, ultimately reinforcing Westminster’s authority over the constitutional question posed by the Scottish government.Footnote 94
I will return to this point later, but it is already clear that the court’s assertion that the referendum’s effects were not merely putative, but carried real practical implications, is also one of the reasons why the assessment of judicial objectivity of the panel falls short of the IEO test. In relation to the contention advanced in this article (that is, CRiS has multiple implications and that extreme care must therefore be taken in selecting the judicial panel that assesses its legality), two issues arise concerning the objectivity of the panel that decided the Reference by the Lord Advocate. Both concern the Scottish justices. First, two Scottish justices participated in resolving a question that affected the ongoing role of their tenure and the future role of Scottish judges at the UK Supreme Court. A long-standing convention requires at least two Scottish judges to be in the UK’s final appellate court. Neil Walker notes that since the late 1920s, there has been ‘no significant period of time […] when there have been fewer than two Scottish judges in the House of Lords’.Footnote 95 Although the Acts of Union 1707 maintained the separation between the English (and Welsh) and Scottish legal systems, final appealable civil matters moved to London, where Scottish judges sustained the Scottish civil law presence.Footnote 96 The practice began with the Appellate Jurisdiction Act 1876, whose creation of Lords of Appeal in Ordinary saw one of the first two appointees drawn from the Scottish Bar.Footnote 97
Scottish independence was not discussed in the case, but that could be one of the second-order consequences of the bill under review. Another of the second-order consequences of the case is the future redundancy of Scottish justices. The point is delicate: secession would not necessarily have brought the immediate removal of the current Scottish justices from the Supreme Court (who also decided on the case), but it would have materially altered the institutional rationale for their continuing role.Footnote 98 For instance, if an independent Scottish Parliament were to restore final legal authority to Scotland’s own courts in a way comparable to the UK’s reassertion of legal autonomy after Brexit, the institutional rationale for retaining Scottish judges in the Supreme Court would be substantially reduced.Footnote 99 It is reasonable to assume that after a settling period in which legacy issues might still be pending, there would be no more Scottish cases for Scottish law experts to decide in London.Footnote 100 The Irish secession in 1922 is an analogous example here.Footnote 101 Irish law lords of the era, such as Lord Atkinson and Lord Carson, continued to serve as Lords of Appeal in Ordinary until their respective retirements in 1928 and 1929, but the establishment of the Irish Free State ultimately led to the cessation of Irish representation in the House of Lords.Footnote 102 Based on this precedent, it should be assumed that in the Reference by the Lord Advocate, Lord Reed and Lord Hodge ruled on a matter that could be seen as affecting their future judicial roles, a circumstance that squarely engages the standard of perceived bias set by the IEO test.Footnote 103
Furthermore, in determining the Reference by the Lord Advocate, it is logical to suggest that the justices may also have been making determinations about their financial interests.Footnote 104 The publicly available materials of the UK Supreme Court do not include any justice-specific register of property, land or shareholdings because the court does not publish a register of interests for any justice. The published expenses tables do, however, record recurring official travel claims by Lords Reed and Hodge across several quarters. Those claims are mainly for flights, rail travel, taxis and car hire, with occasional subsistence or hotel-related items, while the available tables contain no entries for rent, mortgage interest, council tax or other standing London housing costs.Footnote 105 Again, this is a speculative narrative. In the event of Scottish independence and assuming Scottish justices might wish to move to London, their Scottish homes would become foreign-situs assets under Section 275 of the Taxation of Chargeable Gains Act 1992, and any gain on a later sale would be returned as a foreign chargeable gain while they remain UK residents.Footnote 106 Conversely, if they hold on to their residence in Scotland and later dispose of a putative London property (this is again a speculative second-order consequence of the result of the CRiS), the UK non-resident capital gains tax would apply to that sale under Sections 1A(3) and 1C and Schedules B1 and C1 of the same act, provisions introduced by Section 37 and Schedule 7 of the Finance Act 2015.Footnote 107
The analysis could be further extended by considering judges’ pensions. The Scottish government’s official plan for independence is in a 2013 white paper titled Scotland’s Future. Footnote 108 The document makes clear that in the event of independence, responsibility for paying the UK State Pension to anyone living in Scotland would pass to the Scottish government, while entitlements already accrued in Scotland would become part of a new Scottish State Pension, and rights earned elsewhere in the UK would remain with the UK government’s prerogatives.Footnote 109 However, there is no clear guidance on whether such a transfer would be permissible in circumstances like those examined here, where the parties might have, it is reasonable to infer, accumulated substantial capital in London or Hong Kong.Footnote 110
This point is delicate. This information is taken from publicly available sources, and no evidence suggests that either judge was consciously biased; however, the factual background is such that an independent and informed observer could perceive a significant career and financial interest in the outcome of the case. Read together, the Scottish government’s independence proposals, the data from the judges’ weekly expenses for travel between Scotland and London, and the second-order potential effect of independence on their present or future Scottish and London properties create a substantive personal interest. Thus, a fair-minded and informed observer applying the objective standard articulated in Porter v Magill could conclude that there was a real possibility that those substantive personal interests might be perceived as influencing the judicial determination of the case.Footnote 111
The unconstitutional referendum in Veneto
Sentenza n. 118/2015, in comparison to the UK case, shows different types of potential bias. It focuses on the relational dimension of the IEO standard, showing how constitutional adjudication in autonomy disputes may be shaped by a hostile publicly visible history of connections that sit close to the subject matter. In Sentenza n. 118/2015, the Constitutional Court examined Veneto Regional Law n. 16 of 19 June 2014, which authorised a consultative referendum on whether the region should become an independent and sovereign Republic, and declared it unconstitutional.Footnote 112 As discussed throughout this essay, the Italian decision is technically well reasoned. The panel consisted of 12 judges, and this section focuses on the involvement of Justice Amato. The decision is collegial, and there is no indication that any judge, including Amato, contributed to a particular part of the text. This does not affect the present analysis. The IEO test does not assess the reasoning behind the decision or the contribution of any individual judge to it. It asks whether a fair-minded and informed observer looking at the public record would perceive a real possibility of bias in a judge who sat on the panel.
Italy has a rigid constitution and an asymmetric system of territorial governance.Footnote 113 However, the constitution largely overlooks the country’s socioeconomic diversity.Footnote 114 The Italian state was formed through uneven processes of political consolidation marked by episodes of regional resistance.Footnote 115 The constitution establishes an asymmetric territorial model dividing the state into twenty regions, grants special autonomy to five (out of the twenty) and formally recognises twelve minority languages. Secessionist claims have been voiced in German- and French-speaking areas, and a strong sense of distinct identity persists in Sicily and Sardinia.Footnote 116 Veneto, situated in Northeastern Italy and home to roughly five million residents, ranks among the wealthiest regions in Europe, and for the past two decades, it has been politically controlled by the Lega Nord.Footnote 117 In scholarly literature, support for Lega Nord, a regionalist political party with a troubling history in relation to human rights protection, is often linked to a polity with low public trust in central state institutions.Footnote 118 David Haljian refers to these types of political interaction as supporting constitutional disorder, a situation in which the rule of law is formally present but substantively hollowed out by social factors.Footnote 119 These are also, he explains, situations where the demand for secession thrives.Footnote 120
In Sentenza n. 118/2015, the Constitutional Court examined the most recent attempt by the Veneto Regional Council to hold a consultative referendum on secession and concluded that such a referendum would be incompatible with, among other provisions, Article 5 of the Constitution.Footnote 121 The court reviewed two statutes: Veneto Regional Laws n.15 and n.16 of 2014. Footnote 122 Law n.15 prepared the ground for five consultative referenda.Footnote 123 The court declared four of them unconstitutional. It upheld only the proposal inviting Veneto’s electorate to mandate negotiations with the national government, under Article 116(3), for an extension of the region’s legislative competences.Footnote 124
In Sentenza n. 118/2015, however, one of the judges had a public, long-standing, unamicable relationship with the leadership of Lega Nord, a circumstance that would ordinarily engage the test of IEO.Footnote 125 Justice Amato, a constitutional scholar and senior member of the former Italian Socialist Party, served twice as prime minister (1992–93; 2000–1).Footnote 126 The adversarial relationship between the Socialist Party and Lega Nord predated Sentenza n. 118/2015, most notably during the 1993 referendum, which sought to replace proportional representation with a majoritarian voting system.Footnote 127 That referendum, which was supported by the Lega, contributed to the collapse of the Amato government.Footnote 128 In media commentary, the Lega Nord leader likened collaboration with Amato to ‘smoke in the eyes’.Footnote 129 Later, during Amato’s second term as prime minister in 2000, the insults escalated, with Lega Nord’s leader mocking him as ‘the little man’, a pointed and derogatory reference to his Sicilian origins.Footnote 130 A year later, Amato, as Italian prime minister, lost the general election and his office again to a coalition led by Lega Nord and Forza Italia.Footnote 131
This unamicable relationship of the 1990s would probably be sufficient to fail the IEO test; however, the antagonism continued for the next two decades. In April 2010, the president of the Republic appointed Judge Amato as chair of the Committee of Guarantors for the 150th Anniversary of the Unification of Italy.Footnote 132 This statutory organisation was established to oversee and monitor the initiatives commemorating the 150th anniversary of Italy’s unification, celebrated in 2011. Under Amato’s chairmanship, the committee recommended that the Italian Parliament adopt Law-Decree n. 5 of 22 February 2011, which designated 17 March 2011 as a national holiday to commemorate unification. In this role, Amato criticised the leader of the Lega Nord again and dismissed the substate nationalist movement’s theatrical expressions of identity as both ‘pitiful and misguided’.Footnote 133
The irony of supporting nationalist narratives that divide in the name of unity is not missed here, but this is a trivial point. Judge Amato’s public record shows, beyond doubt, a commitment to social liberal principles. The above review shows, however, how the IEO standard remains constitutionally salient when relational proximity becomes publicly visible. Put differently, Amato’s decision not to recuse himself from Sentenza n. 118/2015 appears to embody what Richard Arneson calls the moral logic of patriotic partiality.Footnote 134 He had twice lost public office following electoral contests involving coalitions that included the party in Sentenza n. 118/2015. Based on the available evidence, Justice Amato was a public official with a long political career shaped by a contingent but durable conflict with a party connected to the legal dispute. The relevance of this political history is not political but constitutional. It matters only because the appearance-of-bias standard is assessed through the public record of the individual judge, so a durable and visible antagonism between a sitting judge and political actors associated with the legislation under review engages the IEO test regardless of the collegial form of the decision. On that historical record, a fair-minded and informed observer could conclude that his visible alignment with central state positions on questions of national cohesion, combined with overt public hostility, gave rise to a real possibility of apparent bias.Footnote 135
The Catalan referendum and professional historical connections
The judicial saga surrounding the Catalan referendum brings clearly into view the broad constitutional effects of CRiSs once territorial conflict has entered a phase of acute institutional confrontation.Footnote 136 In this case, the Spanish Constitutional Court reaffirmed that national sovereignty rests with the Spanish people as a whole and that any consultation affecting the integrity of the state must follow the formal constitutional amendment procedure.Footnote 137 It further held that Article 149(1)(32) of the constitution assigns exclusive authority over referendums to the central state, so the Catalan Parliament lacks the competence to legislate for, organise or fund such a vote.Footnote 138 A wider cluster of decisions is associated with the Catalan referendum; Héctor López Bofill’s comparative analysis of Catalonia, Scotland, Quebec and the Basque Country situates those decisions within the broader legal and institutional barriers confronting secessionist movements.Footnote 139 This essay, however, focuses on Sentencia n. 90/2017 because, within the limits of this comparative analysis, it provides the clearest setting in which to examine whether the composition of the panel satisfies the IEO test.
Similar to the Italian case, the court’s decision is delivered collegially, and there is no indication in the text that any individual judge played a special role in it. As with the Veneto decision, this does not bear on the present analysis because the IEO test looks to the public record of the individual judge rather than to the reasoning of the collegial decision. The doctrinal foundations for the court’s position had been laid several years earlier. In Sentencia n.103/2008, the Constitutional Court had already clarified that renaming a vote a ‘popular consultation’ could not hide its nature as a referendum, a field reserved for the state.Footnote 140 In February 2015, the court annulled the consultation law and decree in Sentencia n.31/2015 and Sentencia n.32/2015. Footnote 141 Furthermore, in Sentencia n.138/2015, it declared the regional statutes associated with the secession endeavour unconstitutional.Footnote 142 The Catalan authorities nevertheless pressed on, prompting, among other consequences, the proceedings that led to Sentencia n. 90/2017, which struck down the Fortieth Additional Provision of Catalonia’s Budget Law 4/2017 that, among other elements, allocated money to the referendum initiative.Footnote 143
As López Bofill explains, the Catalonia secession legal saga and Sentencia n. 90/2017 are not part of an isolated incident.Footnote 144 Spain’s post-1978 decentralisation has been far from peaceful, and multiple elements make the relationship between central and regional administration tense.Footnote 145 The Constitutional Court reaffirmed that national sovereignty rests with the Spanish people as a whole and that any consultation affecting the integrity of the state must follow the formal constitutional amendment.Footnote 146 When the referendum proceeded, the Spanish government invoked Article 155 of the constitution to dismiss the Catalan administration, impose direct rule and criminally prosecute its leaders.Footnote 147 Ley Orgánica n. 1/2024 confirms that the 2014 and 2017 Catalan referendum litigation formed part of a longer constitutional conflict whose legal consequences persisted well beyond 2017, especially in the criminal sphere. That later development is relevant here only as a subsequent context, not as part of the material considered in our putative IEO test.Footnote 148
Historically, the current constitutional text is the result of a negotiating process between the new democratic political parties and the departing military junta, which was concerned with the possibility of regional secession.Footnote 149 The result of the mediation contributed to a system of territorial governance with multiple inconsistencies.Footnote 150 That indeterminacy explains, among other factors, why the Constitutional Court issued successive rulings on the same CRiS, whereas in the UK and Italy, CRiS legal and constitutional indeterminacies were resolved by a single definitive judgement.Footnote 151
It is within this social and constitutional context that this article assesses the composition of the panel in Sentencia n. 90/2017 and asks whether, had the decision been open to review, it would have satisfied the IEO test. The enquiry does not proceed, as in the Reference to the Lord Advocate, by reference to future personal or fiscal consequences drawn from precedent or government white papers.Footnote 152 None of that material is available here, nor does the argument rest on the judges’ expertise in regional law as such. Instead, it turns on publicly visible professional links relevant to the appearance of objective adjudication.
Judge María Encarnación Roca Trías was, from 1968, Chair of Civil Law at the University of Barcelona. Her service on the Generalitat’s Legal Advisory Commission,Footnote 153 between 1983 and 2005,Footnote 154 further integrated her into the institutional fabric of the region’s autonomous administration by the president of the Generalitat. These facts do not establish actual bias, but they form part of the objective circumstances that a fair-minded and informed external observer could consider when assessing the composition of the panel.Footnote 155
At the time, historical or institutional involvement was not treated as a serious ground of concern. The Spanish Constitutional Court had developed a recusation doctrine that required parties to show objectively justified doubts supported by concrete and verifiable facts. In practice, this approach narrowed the scope of appearance-based objections and set a stricter threshold than the broader IEO standard applied in this article.Footnote 156
The narrowing of the IEO test was revisited in 2025 by the Constitutional Court.Footnote 157 The court then accepted that prior institutional involvement could, in some circumstances, affect the appearance of objectivity strongly enough to justify recusal. That decision has limited direct relevance here because it postdates Sentencia n. 90/2017 and the other decisions examined in this article. Its value is different. This shows that the court later adopted a less restrictive approach to objections based on apparent bias. This shift suggests that had a more careful analysis of appearance-based standards been applied earlier, the composition of the panel in Sentencia n. 90/2017 might have attracted closer scrutiny.
Conclusion: Constitutional contestation, the Penumbra of doubt and visible impartiality
In a multinational democracy, the binding force of legal norms is closely connected to the belief that citizens participate in a shared constitutional project. This project includes the protection of human dignity through a system of substantive rights and democratic values. The debates among political parties, the operation of public administration and citizens’ participation in public debate each contribute to sustaining a belief that the constitutional framework, which we all inherited from the Enlightenment, has limited alternatives.
At the outset of this essay, I referred to Dworkin’s article Objectivity and Truth: You’d Better Believe it. In the article, he explained that exploring the nature of objectivity and truth requires us to analyse how legal narratives are constructed. The objectivity of judges is not secured by any external metaphysical guarantee but by the capacity of a community’s shared interpretive commitments to withstand rigorous sceptical scrutiny. Delaney and Appleby bring this insight to bear on debates concerning the internal legal foundations of judicial authority and the external social perception of courts in federal systems. In relation to judicial objectivity, resilience is evaluated through the IEO test. This epistemic standard underlies modern requirements of judicial objectivity across the UK, the CJEU and the ECtHR jurisdictions.
This study limned three apex court decisions drawn from jurisdictions with distinct constitutional traditions and historical contexts and argued that in each case, the IEO standard was not fully satisfied. The review of these cases relies exclusively on openly available documentary sources, including official judgements, published expense claims, parliamentary records and submitted biographical disclosures, in line with judicial-conduct rules in Italy, Spain and the UK. The Reference by the Lord Advocate shows the role of the court in which a route to renewed negotiation has been closed at the threshold of legal competence. The precedents from the Irish Free State, alongside statutory provisions on pensions and tax residence, created circumstances in which an observer might perceive that the Scottish justices’ financial and residential interests could be affected by the outcome. The Italian case shows how relational proximity may become constitutionally salient in autonomy disputes in which the public visibility of personal connections shapes the appearance of impartiality. In Spain, instead, the recorded institutional relationship between a justice and regional institutions affects the perception of objectivity.
Indeed, judges do not operate in a vacuum, and the presumption of impartiality can be challenged when their personal or professional interests appear to intersect with the issues before them. It is also reasonable to infer a range of relational biases intricately linked to CRiS cases, as it is evident that secession affects large segments of the population. However, the idea of a contextual judge cannot be extended to bring a penumbra of doubt to the institution. If the court is reasonably and visibly perceived as structurally partial, judicial intervention may settle the immediate dispute while simultaneously weakening the trust necessary for subsequent constitutional engagement. From that perspective, the three cases examined in this article were missed opportunities to refine the justices’ role in multinational systems under conditions of a potential acute constitutional contestation.
Funding statement
This work benefitted from a University of Brescia FAACI 2025 short international research visiting grant and from participation in the International Visiting Teaching programme at the Eurac Research School of Federalism 2026.
Competing interests
The author declares no conflict of interest. For transparency, the author notes that he previously lived and worked in Scotland for the Scottish Parliament and serves as a research adviser to the Centro Studi sulle Istituzioni Livio Paladin, University of Padua. The author has no political affiliation with any political party, campaign organisation, court, government or public institution discussed in the article. The Centro Studi sulle Istituzioni Livio Paladin had no role in the research, writing or submission of this article.
Open access
