Hostname: page-component-699b5d5946-lkkl4 Total loading time: 0 Render date: 2026-03-06T13:42:02.136Z Has data issue: false hasContentIssue false

Kelly v UCD [2025] IESC 6: clarifying the boundaries on judicial bias disqualification

Published online by Cambridge University Press:  02 March 2026

Laura Cahillane*
Affiliation:
Law, University of Limerick , Ireland
Rights & Permissions [Opens in a new window]

Extract

The Irish Supreme Court’s decision in Kelly v UCD1 is a significant ruling on judicial bias and disqualification. It is the first case to consider the implications of familial connections between judges and law firms representing parties since the Judicial Conduct Committee published its Guidelines on Judicial Conduct and Ethics in 2022.2 The case clarifies the legal test for objective bias, delineates its boundaries, and addresses the role of judicial conduct guidelines in disqualification decisions. It also has broader implications for other common law jurisdictions, particularly England & Wales, where similar issues of judicial propriety and public confidence in impartial adjudication arise.

Information

Type
Current Developments: Case Comment
Creative Commons
Creative Common License - CCCreative Common License - BY
This is an Open Access article, distributed under the terms of the Creative Commons Attribution licence (http://creativecommons.org/licenses/by/4.0), which permits unrestricted re-use, distribution and reproduction, provided the original article is properly cited.
Copyright
© The Author(s), 2026. Published by Cambridge University Press on behalf of The Society of Legal Scholars

Introduction

The Irish Supreme Court’s decision in Kelly v UCD Footnote 1 is a significant ruling on judicial bias and disqualification. It is the first case to consider the implications of familial connections between judges and law firms representing parties since the Judicial Conduct Committee published its Guidelines on Judicial Conduct and Ethics in 2022.Footnote 2 The case clarifies the legal test for objective bias, delineates its boundaries, and addresses the role of judicial conduct guidelines in disqualification decisions. It also has broader implications for other common law jurisdictions, particularly England & Wales, where similar issues of judicial propriety and public confidence in impartial adjudication arise.

1. Background and procedural history

The case arose from long-running litigation initiated by Patrick Kelly against University College Dublin (UCD), alleging gender discrimination in postgraduate admissions.Footnote 3 The specific issue before the Supreme Court on this occasion was whether Meenan J was disqualified from hearing an appeal because his son was employed as a solicitor in Arthur Cox LLP, the firm representing UCD. The son had no involvement in the case and worked in a different department.

The appellant raised the issue post-judgment, arguing that the judge’s failure to disclose this familial connection gave rise to objective bias. The Supreme Court granted leave to appeal on the narrow question: Does the fact that a close relative of a judge is employed as a solicitor in the firm of solicitors representing a party meet the well-established test for objective bias?

2. The legal framework: objective bias in Irish law

The leading case on objective bias in Irish law, prior to Kelly, was Bula Ltd and Others v Tara Mines Ltd and Others (No 6). Footnote 4 In Bula the Supreme Court developed the ‘reasonable apprehension’ formula as the correct approach to be taken on the question of apprehension of bias, as distinct from the ‘real likelihood’ test set out by the House of Lords in R v Gough,Footnote 5 which had traditionally been the preferred test.Footnote 6 The Supreme Court rejected the argument made by counsel for the state that it should adopt a ‘different, considerably stricter … [test] closer to the “real danger” test’.Footnote 7 Denham J set down the test as follows: ‘[I]t is whether a reasonable person in the circumstances would have a reasonable apprehension that the applicants would not have a fair hearing from an impartial judge on the issues.’Footnote 8

McGuinness J, who agreed with Denham J and also produced a substantive judgment, adopted the approach of the Australian High Court, emphasising that it is the Court’s view of the public’s opinion, not the Court’s own opinion, that is relevant.Footnote 9 Subsequently, in Ryan v Law Society of Ireland,Footnote 10 Herbert J held that the ‘likelihood’ test should be considered to have been superseded by the reasonable suspicion/apprehension test ‘and be heard of no more’.Footnote 11

The ‘reasonable apprehension’ or ‘reasonable suspicion’Footnote 12 test has been widely applied in Ireland and there is no significant judicial opinion that any other test should be applied.Footnote 13 Hogan, Morgan and Daly have suggested that there is overlap between this and the older ‘real likelihood’ approach, and note that the practical difference between them will depend on how much information the ‘reasonable person’ is assumed to have: if they are in possession of all the information, the test will move closer to assessing whether there is a real likelihood of bias.Footnote 14

The Bula case is also an important authority on the issue of a judge’s prior relationship with a party to litigation. In Bula, the applicant claimed that the involvement of two judges of the Supreme Court, while they were practitioners, in advising some of the respondents and acting for them in related litigation, constituted bias.Footnote 15 The Supreme Court found that the mere fact of a judge having acted for a former client who subsequently is a party to a matter before that judge is not sufficient to require recusal.Footnote 16 Recusal might be required where there was a long, recent and varied connection between the judge and a party, or where the judge had provided legal services on an issue that is before the court for determination.Footnote 17 The Court considered 17 alleged links between Barrington and Keane JJ and the parties, and found that none raised the issue of apprehended bias.Footnote 18

Familial relationships have, on occasion, been sufficient to raise the apprehension of bias. In O’Reilly v Cassidy (No 2),Footnote 19 an order made in the Circuit Court was quashed where counsel for one of the parties was the daughter of the presiding judge. Flood J rejected the submission that the parent-child relationship between the judge and counsel was sufficient in itself to give rise to a reasonable apprehension of bias, and said that to accept that contention would ‘derogate the oath made and subscribed to by every judge on appointment … to a totally empty formula’.Footnote 20 On the facts, Flood J found that the question of the relationship had become ‘so inextricably entangled with other factors’ that a reasonable observer could fear that the outcome had been affected by that relationship.Footnote 21

A more strict approach to family relationships was taken by the Supreme Court in Kenny v Trinity College Dublin. Footnote 22 The core of the applicant’s case was that the brother of the presiding judge, Murray J, was a partner in the firm of architects, Murray-O’Laoire Architects, who were central witnesses for the respondent. The firm noted that the brother in question was not based in Dublin and had no involvement in the project. However, the integrity of the firm was questioned, as the applicant alleged that documents had been concealed from the Court. Fennelly J held that where the Supreme Court is asked to determine whether one of its own judgments should be set aside for objective bias, it should adopt the interpretation more favourable to the plaintiff where there is a lack of clarity, and err on the side of caution.Footnote 23 He found that the test of objective bias was met, citing the UK decision of Locabail,Footnote 24 and set aside the prior order of the court.Footnote 25 Biehler suggests that the Supreme Court in Kenny went further than was necessary to uphold the perception of impartiality and notes that the decision is out of step with the approach taken in other common law jurisdictions.Footnote 26

3. The court’s findings

Of interest in Kelly was whether the Court would maintain the strict approach adopted in Kenny, or whether it considered it necessary to recalibrate that stance in favour of a more flexible standard. In particular, the appellant had argued that there should be a bright-line rule disqualifying any judge whose close relative is employed in a firm representing a party. The Supreme Court, per O’Donnell CJ, rejected that argument, noting that it ran counter to existing Irish jurisprudence as well as the case law of the European Court of Human Rights,Footnote 27 and he took the opportunity to make some important clarifications on this area of law.

The Court reaffirmed the objective ‘reasonable apprehension’ test as articulated in Bula Ltd Footnote 28 and subsequent cases, and confirmed that the observer is expected to be an informed observer:Footnote 29 ‘The same test is applicable in every case: whether a reasonable observer would have a reasonable apprehension of bias’.Footnote 30 O’Donnell CJ then explained that there are three components to the test, ‘which must be assessed together’:

  1. (a) the reasonable and informed observer;

  2. (b) who has a reasonable apprehension

  3. (c) of bias.

He stressed that it is important that ‘each of the components is given its correct weight, and in particular, that the test of bias when measured by the standard of apprehension is not further depreciated, since otherwise the balance sought by the test would not be achieved.Footnote 31 The test is therefore whether a reasonable and informed observer would have a reasonable apprehension that the judge would not bring an impartial mind to bear on the case.Footnote 32 He emphasised that the test is not subjective and does not require proof of actual bias but it is a strict and rigorous standard, designed to uphold public confidence in the administration of justice.

In relation to the particular issue before the Court, it was held that no disqualification arises where a judge’s relative is employed in a large firm, in a different department, with no involvement or interest in the case and that the mere fact of employment does not create a reasonable apprehension of bias. It was also noted that disclosure of such a relationship is not required, and non-disclosure does not, in itself, invalidate a judgment.Footnote 33 The Court distinguished Kenny, emphasising that the previous case had involved reputational stakes and a closer connection to the litigation.

The judgment reiterates that judges have a correlative duty to sit unless disqualified, as over-recusal risks undermining judicial independence and efficiency, and could encourage judge-shopping.Footnote 34 The Court also emphasised that judges are presumed to act impartially, consistent with their constitutional declaration under Article 34.6.1, and that disqualification must be based on a cogent and rational link between the alleged bias and the judge’s ability to decide the case fairly.

4. The role of the Judicial Conduct Guidelines

This was the first major case involving issues around judicial bias and recusal since the publication of the Judicial Conduct Guidelines adopted under the Judicial Council Act 2019. Of particular relevance was Clause 4.4, which states: ‘A judge shall not participate in the determination of a case in which any member of the judge’s family represents a litigant or is associated in any manner with the case.’ Thus, another major point of interest in this case was to what extent the Guidelines would impact questions of judicial bias and disqualification.

O’Donnell CJ found that Guidelines, including those on recusal, are helpful for setting expectations but are distinct from the legal test for bias and do not determine the legal test for disqualification. This distinction emphasises the separation between ethical guidance and legal standards, and thus ensures that disqualification decisions remain grounded in constitutional and common law principles.

He pointed out that interpreting these guidelines is primarily the role of the Judicial Conduct Committee and that breaching a guideline does not automatically amount to misconduct.Footnote 35 Even a finding of misconduct only leads to statutory consequences (such as a reprimand)Footnote 36 and does not affect the validity of the judge’s decision. Guidelines aim to provide clarity for future conduct, but whether a judge is legally disqualified for bias is a separate question and does not follow merely from breaching a guideline. That said, he made it clear that the guidelines can inform what a reasonable observer might think.Footnote 37 So while the guidelines are not determinative in relation to questions of bias and disqualification, they are nevertheless relevant in the context of what might be considered reasonable.

The Court also interpreted Clause 4.4 narrowly. The Court directed that the clause applies to direct involvement in the case and it did not extend to mere employment in a large firm that happens to be on record, where the family member has no involvement in the case.Footnote 38 Crucially, on the Court’s analysis, reading Clause 4.4 to cover direct representation or direct association with the case avoids an unduly broad rule that would trigger frequent recusals and disrupt judicial work. While critics may contend that this approach risks under-recusal, the Court’s reasoning reflects a pragmatic appreciation of the realities of a small jurisdiction such as Ireland, where an expansive interpretation could produce a substantial number of recusals and generate significant administrative difficulties in judicial scheduling and case allocation. The Court therefore sought to strike an appropriate balance between the need to maintain public confidence in judicial impartiality and the practical necessity of ensuring a functioning judiciary. Indeed, commentators have observed that excessive recusal can itself undermine confidence in the justice system by implying impropriety where none exists.Footnote 39 Taken together, these considerations demonstrate that the Court consciously prioritised the efficient operation of the judicial system over expanding the grounds for disqualification, which demonstrates that the balance struck in Kelly leans more heavily toward the efficiency side of the spectrum. The Court’s approach also accords with the comparative common law position, which resists automatic disqualification outside direct advocacy or involvement.Footnote 40

In making this decision, the Court acknowledged that the Guidelines reflect changing public expectations and evolving standards of judicial propriety.Footnote 41 For example, it accepted that a judge should not sit where a close relative appears as an advocate or is directly involved in the case.Footnote 42 This reflects a shift in practice from earlier Irish case law such as O’Reilly v Cassidy (No 2),Footnote 43 where such relationships were not seen as disqualifying, but is in line with the guidance from the Conduct Guidance, and indeed as noted by O’Donnell CJ, with the identical provisions in the Bangalore Principles on Judicial Conduct.

However, the Court also rejected the idea that the Guidelines impose a general legal duty to disclose non-disqualifying relationships. It held that disclosure may be prudent or desirable in some cases but that non-disclosure alone does not amount to bias or invalidate a judgment.Footnote 44 The absence of a general legal duty to disclose non‑disqualifying relationships means that a judgment is not vulnerable simply because of non‑disclosure; any questions of judicial conduct would be addressed separately under the Judicial Council regime.

The Court noted that the Irish Guidelines are drawn in large part from the terms of the Bangalore Principles of Judicial Conduct adopted by the United Nations Human Rights Commission in 2003 (the Bangalore Principles), and that similar standards are found in the UK Supreme Court Guide to Judicial Conduct and the Canadian and Australian judicial ethics codes.Footnote 45

The Court’s reliance on international standards underscores that the Irish Guidelines are firmly embedded within a global framework of judicial ethics.Footnote 46 The judgment demonstrates comparative consistency across common law jurisdictions, all of which reject automatic disqualification in favour of a context-sensitive test based on the reasonable apprehension of bias.Footnote 47 This reflects evolving societal expectations about judicial conduct, reinforcing public confidence and signalling that the Irish Guidelines are not insular but part of a dynamic, internationally-informed standard.Footnote 48

There is, however, a tension in relation to the issue of familial connections in that the Court says there is no category of automatic disqualification in Irish law, and yet it also says that judges should not sit where there is a close family connection to a party or advocate. In order to reconcile this tension, it appears that the ‘no automatic disqualification’ instruction operates at the level of law and that we must treat the ‘should not sit’ direction as ethical guidance, which would presumably carry substantial weight with the reasonable and informed observer.Footnote 49 It would have been preferable if this distinction had been discussed in more depth in the judgment.

Conclusion

This judgment is significant for several reasons. First, the Court provided a structured framework for assessing bias, emphasising: the three-part test involving the reasonable and informed observer, reasonable apprehension, and bias; the importance of context in assessing the connection between a judge and a lawyer; and the need to avoid overly broad disqualification rules that could undermine judicial independence and efficiency.

In the first Irish case to consider the Judicial Conduct Guidelines in the context of disqualification, the Court clarified their interpretive role and limits, offering guidance for future cases. The decision also aligns with the UK Supreme Court Guide to Judicial Conduct (2019) and the Courts and Tribunals Judiciary Guide (2023), which similarly reject automatic disqualification based on familial employment in law firms. The Irish Supreme Court’s detailed reasoning and comparative analysis (including references to Canada, Australia, and South Africa) make this a persuasive authority for courts in England & Wales and other common law jurisdictions.

The judgment is clearly an attempt to balance efficiency and perception in relation to judicial disqualification, though it leans perhaps more heavily on the efficiency side. The Court emphasised the duty to sit unless disqualified, moved away from the strict approach in Kenny, and rejected the suggestion of a duty of disclosure, which would significantly enlarge disqualification grounds and invite hindsight bias. The court also dismissed as unreasonable concerns about nepotism in the context of judge-lawyer connections,Footnote 50 even though critics might argue that societal expectations demand greater transparency.Footnote 51 Indeed, right at the beginning of the judgment O’Donnell CJ emphasised that:

[T]he test of disqualification and its application in any case should go so far as is necessary to ensure the maintenance of public confidence, but not further, since that deprives the successful party of the judgment to which they are entitled without any corresponding benefit in increasing public confidence. Indeed, the too-ready use of the term bias and the possibility of a consequent implication and perception that a judge disqualified has been guilty of serious wrongdoing even in cases very far removed from any real lack of impartiality, may undermine confidence rather than support it.Footnote 52

However, while much of the Court’s reasoning in Kelly v UCD reflects an efficiency-oriented approach, it remains clear that the judgment is anchored in the constitutional imperative of maintaining public confidence in the administration of justice. O’Donnell CJ repeatedly stresses that the objective bias test is ‘strict and rigorous’ precisely because it safeguards the appearance of impartiality. The Court’s insistence on an informed observer standard, its nuanced interpretation of the Judicial Conduct Guidelines, and its acknowledgement of evolving ethical norms all demonstrate a conscious effort to reconcile practical judicial efficiency with the fundamental need for transparency and trust in the judicial process.

Kelly v UCD may now be considered a foundational case in Irish law on judicial disqualification. It reaffirms the objective bias test, clarifies the non-binding role of judicial conduct guidelines, and rejects a bright-line rule of disqualification based on familial employment in law firms. The decision balances the need for public confidence in judicial impartiality with the practical realities of a small legal community and a limited judiciary. Its careful reasoning and comparative scope make it a valuable reference for other common law jurisdictions grappling with similar issues, where the principles of judicial impartiality and public perception are equally vital.

References

1 Kelly v UCD [2025] IESC 6.

3 For some detail see A Eustace ‘Jaundiced and jaundiced about judicial bias: Kelly v UCD’ (forthcoming) Irish Supreme Court Review.

4 Bula Ltd and Others v Tara Mines Ltd and Others (No 6) [2000] 4 IR 412.

5 R v Gough [1993] AC 646.

6 See G Hogan and DG Morgan Administrative Law in Ireland (Dublin: Round Hall, 4th edn, 2012) p 397.

7 Bula Ltd v Tara Mines Ltd, above n 4, at 479 per McGuinness J.

8 Ibid, at 441 per Denham J.

9 Ibid, at 486 per McGuinness J.

10 Ryan v Law Society of Ireland [2002] 4 IR 21.

11 Ibid, at 38.

12 It is often referred to as the reasonable suspicion test but as Denham J noted in Bula, this carries broader connotations and ‘unintended nuances of meaning’ and so she preferred ‘apprehension’ to ‘suspicion’.

13 Kenny v Trinity College Dublin [2007] IESC 42, [2008] 2 IR 40; O’Ceallaigh v An Bord Altranais [2011] IESC 50; O’Callaghan v Mahon [2007] IESC 17, [2008] 2 IR 514.

14 G Hogan et al Administrative Law in Ireland (Dublin: Round Hall, 5th edn, 2019) pp 14–74.

15 Bula Ltd v Tara Mines Ltd, above n 4.

16 Ibid, at [445]–[446] per Denham J.

17 Ibid, at [446] per Denham J.

18 Ibid, at [461]–[462] per Denham J.

19 O’Reilly v Cassidy (No 2) [1995] 1 ILRM 311.

20 Ibid, at 319 per Flood J.

21 Ibid.

22 Kenny v Trinity College Dublin, above n 13.

23 Ibid, at [21] per Fennelly J.

24 Locabail v Bayfield Properties Ltd [2000] QB 451.

25 Kenny v Trinity College Dublin, above n 13, at [25] per Fennelly J.

26 H Biehler Judicial Review of Administrative Action: A Comparative Analysis (Dublin: Round Hall, 3rd edn, 2013) p 280.

27 Kelly v UCD, above n 1, at [71].

28 Bula Ltd v Tara Mines Ltd, above n 4.

29 On the issue of the ‘informed observer’ see AA Olowofoyeku ‘Bias and the informed observer: a call for a return to Gough’ (2009) 68 Cambridge Law Journal 388 at 409, where it is argued that this test should be dispensed and a version of the Gough test restored.

30 Kelly v UCD, above n 1, at [162].

31 Ibid, at [110].

32 Ibid.

33 Ibid, at[139], [148], [155].

34 Ibid, at [6]–[9].

35 Misconduct is defined by statute as a serious departure from accepted judicial standards that brings justice into disrepute: Judicial Council Act 2019, s 2.

36 Judicial Council Act 2019, s 58.

37 Kelly v UCD, above n 1, at [55].

38 Ibid, at [75].

39 Olowofoyeku, abov n 29, at 409.

40 Locabail v Bayfield Properties Ltd [2000] QB 451; Porter v Magill [2001] UKHL 67, [2002] 2 AC 357; Helow v Secretary of State for the Home Department [2008] UKHL 62; Ebner v Official Trustee in Bankruptcy [2000] HCA 63.

41 B Barry ‘Judicial impartiality in the Judicial Council Act 2019: challenges and opportunities’ (2022) 6(1) Irish Judicial Studies Journal 38. See also Survey on Drivers of Trust in Public Institutions – 2024 Results (OECD, 2024), available at https://www.oecd.org/en/publications/oecd-survey-on-drivers-of-trust-in-public-institutions-2024-results_9a20554b-en.html (last accessed 11 February 2026).

42 Kelly v UCD, above n 1, at [99].

43 O’Reilly v Cassidy (No 2), above n 19.

44 Kelly v UCD, above n 1, at at [146]–[148].

45 Ibid, at [53], [76]–[79].

46 See B David ‘Ethics and judicial integrity under the Bangalore Principles’ (2023) 7(2) RAIS Journal for Social Sciences 50.

47 See for example, Porter v Magill [2001] UKHL 67, [2002] 2 AC 357; Helow v Secretary of State for the Home Department [2008] UKHL 62; Ebner v Official Trustee in Bankruptcy [2000] HCA 63. See also M Groves ‘Clarity and complexity in the bias rule’ (2020) 44(2) Melbourne University Law Review 565; A Nayerahmadi ‘Far beyond Baker: heuristics and the inadequacy of the reasonable apprehension of bias analysis’ (2022) 59(2) Osgoode Hall Law Journal 339.

48 See R Byrne ‘Judicial conduct in Ireland: a framework fit for purpose?’ The Bangalore Principles and the Judicial Council Act 2019’ (2022) 6(1) Irish Judicial Studies Journal 1.

49 Kelly v UCD, above n 1, at [55].

50 Ibid, at [115]–[117].

51 See Eustace, above n 3.

52 Kelly v UCD, above n 1, at [9].