Double-Hatting in Indian Arbitration: Addressing Pre-Dispositional Bias

Siddhant Shinde*

---

I. INTRODUCTION

In a recent Permanent Court of Arbitration (‘PCA’), dispute between RSE Holdings AG and the Republic of Latvia, the Respondent challenged the appointment of the arbitrator “given her practice as counsel in past and pending ECT [Energy Charter Treaty] arbitrations”, contending that the arbitrator, Ms. Frey had taken a “partisan position” on similar issues in other arbitrations, and hence was predisposed towards one of the parties.  They argued that a “hypothetical future conflict” between Ms. Frey’s position as counsel and arbitrator was enough to meet the threshold of “an appearance of pre-judgment of an issue likely to be relevant to the dispute”.  The Claimant argued that such cases are prevalent in investor-state disputes, and unless proven otherwise, such a situation does not impair her impartiality in any way. However, the tribunal ultimately disqualified her, citing her extensive involvement in ECT-related cases and the potential for perceived bias due to the “sheer number of cases” in which she had engaged with similar legal questions. This practice of an individual simultaneously playing the role of both an arbitrator and a counsel in substantively similar matters is called ‘double-hatting’.

Several institutions and stakeholders have recognized and addressed broad concerns regarding conflicts of interest in international arbitration. Recently and notably, in the 55th session of the UNCITRAL held in New York in 2022, a panel discussed how the ICSID-UNCITRAL Code of Conduct for Adjudicators in International Investment Disputes sought to address the issue of arbitrators concurrently acting as counsels and arbitrators for substantively similar disputes, i.e. the issue of “double-hatting”.  In fact, within the realm of investor-state arbitrations, such issues have become commonplace, with tribunals in cases like Saint-Gobain v Venezuela (2013) and  Ghana v Telekom Malaysia (2004), notably not being in favour of such practices.  This stance is also mirrored by a majority of inter-country legal instruments such as the Comprehensive and Progressive Agreement for Trans-Pacific Partnership, the United States–Mexico–Canada Agreement, the 2019 Netherlands Model Investment Agreement, and the 2012 South-African Development Community Model BIT. Despite this, there is no consensus on whether double-hatting as a practice requires to be regulated, and if yes, the way in which it can be regulated. Moreover, the discourse around double hatting is rather restricted. First, it is largely limited to Investment Arbitration, with an acute lack of literature on double-hatting specifically in the context of  International Commercial Arbitration (‘ICA’).  For instance, Khan Adnan in his article on double hatting comprehensively discusses the issue, and also suggests reforms, but restricts the scope to investment arbitration specifically. This trend is seen across the board.[1] Langford and Behn justify this lack of literature implicitly by propounding that double hatting is not as big an issue in ICA, as it functions within a close legal community where overlapping roles are historically accepted. On the other hand, heightened interest in confidentiality in double hatting can be attributed to multiple reasons. Firstly, given that the result of such arbitrations is likely to impact public policy, public resources and state finances, there is a heightened concern about impartiality and legitimacy. Secondly, investment arbitration often deals with similar broad legal frameworks, such as bilateral investment treaties, across different cases, which is less prevalent in commercial arbitration, where disputes are more case-specific and diverse. Lastly, being appointed as an arbitrator in an investor dispute usually requires one to be an expert in geopolitics, treaty law, etc. Consequently, the pool of qualified professionals who fit these criteria is small, increasing the likelihood of double-hatting.

Literature on the subject broadly conceptualizes the issue of double-hatting as a conflict of interest problem, which can potentially lead to bias. In this context, the Indian Arbitration and Conciliation Act, 1996 (‘A&C Act’) adopts a proximity-based view of bias, focusing on direct relationships with parties, leaving out cases where arbitrators unconsciously import biases from their role as counsel in similar disputes. This article seeks to argue that the A&C Act overlooks situations where an arbitrator’s impartiality is compromised not by vested interest but by subconscious biases, shaped by prior experiences, including their role in substantively similar cases. Drawing from cognitive theory, this article argues for a shift from ‘specific proximity’ to a broader ‘general proximity to the dispute’ standard, to address double-hatting comprehensively.

Further, in this article, the author exclusively focuses on ICA in the context of India, given India’s efforts in becoming a global arbitration hub. In Part II, the author examines the standard of impartiality under the A&C Act and its judicial interpretations, while critiquing the framework for its inadequacy in addressing double-hatting. Building on this, in Part III, the author highlights how the dual role of arbitrators creates a perception of pre-dispositional bias, which is insufficiently addressed under the current framework. In Part IV, the author proposes a pragmatic framework to address double-hatting, drawing from the UNCITRAL Draft Code and ICSID Arbitration Rules to develop dual obligations for arbitrators. Finally, in Part V, the author concludes by emphasizing the need to uphold arbitration's integrity while also preserving the practical viability of the process.

II.  IMPARTIALITY AND DISCLOSURE STANDARDS IN INDIAN ARBITRATION

Most arbitration centres/institutions have a broad disclosure requirement regarding “anything which may give rise to justifiable doubts as to their impartiality or independence”.[2] This uncertainty essentially means that discretion lies with the national courts to determine the exact nature and scope of this duty. In line with this, the author argues that the Indian A&C Act is inadequate to address the issue of double-hatting.

From a bare reading of the A&C Act and a strict interpretation of its provisions, if Ms. Frey’s appointment were challenged in Indian Courts, the petition would likely be rejected. The A&C Act, in clear terms, outlines the scope and standard of ‘impartiality’ in Schedules V and VII, read with Sec. 12 A&C Act, and finds no direct mention of involvement of the arbitrator in any capacity in a substantively similar matter. Even the Arbitration and Conciliation (Amendment) Bill, 2024, which reportedly promises to set the stage for a “transformative era”, fails to take the issue of “double hatting” into consideration. Schedules V and VII are non-exhaustive and only suggestive. While they do not explicitly address double-hatting, this does not necessarily prevent challenges to an arbitrator’s appointment on these grounds. However, the absence of clear guidance raises important questions at different stages of the arbitration process. Firstly, before the constitution of the tribunal, it is unclear if a potential arbitrator is obligated to disclose a conflict of such nature. Secondly, after the constitution of the tribunal, but before an award is rendered, if such a non-disclosure can lead to their disqualification. Lastly, after an award has been rendered, an arbitrator’s failure to make a disclosure can be a valid ground to set aside an award by a Court. Various other ancillary issues also arise from this, such as determining the relevant timeframe, i.e. how long ago the related case must have occurred for the arbitrator to be required to disclose it, and the extent to which a potential arbitrator is expected to conduct due diligence. 

Before the constitution of the tribunal, an arbitrator shall de jure be ineligible to be appointed if they fall under any of the categories listed in Schedule VII,[3] unless the parties, after the dispute has arisen, expressly agree in writing to waive the disqualification. Thus, if the arbitrator satisfies any criteria listed under Schedule VII, such an appointment is held to be void ab initio, under Sec. 12(5).[4] A provision addressing double-hatting at this juncture is not only unnecessary but also undesirable. Arbitrators are merely professionals offering a service, and not public officials, and thus there cannot be a blanket pre-constitution prohibition against individuals practising as arbitrators and counsel contemporaneously. Moreover, Schedule VII already addresses bias arising from an arbitrator’s previous role as an advisor (including legal advisory)/consultant/employee,[5] and thus a bright-line pre-constitution prohibition making persons who have adjudicated similar disputes will make Schedule VII too wide. De jure ineligibility under Schedule VII operates automatically, unlike Schedule V, which only raises grounds for "justifiable doubts" about independence or impartiality. An appointment afoul of Schedule VII is essentially a jurisdictional defect, rendering the tribunal improperly constituted under Section 34(2)(a)(v) and Section 12(5) A&C Act. Thus, if double-hatting is listed under Schedule VII, it risks parties misusing technical objections to derail arbitral proceedings unnecessarily. On the other hand, inclusion in Schedule V, which lays down the  "justifiable doubts" framework, may be more prudent, rather than triggering automatic disqualification under Schedule VII. This will also significantly affect disputes in specialized areas such as maritime, sports, construction, etc., which call for professionals who have experience not only as arbitrators, but also in these respective specific fields, given the limited number of people who possess even the minimum qualifications required to arbitrate over such matters. 

After the constitution of the tribunal, either party may challenge an arbitrator’s appointment before the Tribunal, if there are “justifiable doubts as to their independence or impartiality”, determined based on grounds stated in Schedule V, and such doubts as to independence or impartiality have to be determined as a matter of fact by the Arbitral Tribunal under Sec. 13 A&C Act. If the tribunal rejects such a plea, and an award is rendered, either party may then challenge an award so rendered before a competent court.[6] Schedule V is noticeably wider than Schedule VII. It is interesting to note that sub-clause (24) includes current and past involvement of the arbitrator in another arbitration on a related issue. However, it imposes a caveat that this only becomes a ground to challenge when such a related-issue arbitration involves one of the parties or an affiliate of one of the parties. Thus, it fails to account for scenarios where an arbitrator has acted as counsel in a substantively similar matter that did not involve any of the parties. This points towards the standard of impartiality under the A&C Act being limited to ‘specific and direct’ proximity, discussed in greater detail in Part III of this article. Thus, from a bare reading of the Schedules, the A&C Act effectively fails to account for bias arising as a result of double-hatting.

As to the question of judicial interpretation, it is clear that the threshold for de jure ineligibility is very high, and does not include double-hatting. In Quess Corp Ltd vs Nectore Cloud Pvt Ltd (2023), even when the arbitrator had acted as a counsel for the other party’s advocate, the Bombay High Court held that there was no clash of interest. Similarly, in Sheetal Maruti Kurundwade v Metal Power Analytical Pvt Ltd & Ors (2017), where the respondent’s advocate had engaged one of the arbitrators as a counsel, the Bombay High Court dismissed the plea for disqualification, while highlighting how this is a common practice.

“…Courts appoint practising Advocates in ad hoc arbitrations. Some practising Advocates are also empanelled with arbitration institutes and forums, and serve in institutional arbitrations. The point that Mr Lulia attempts to make seems to me to strike at the very root of professional independence. It questions the very basis on which our system functions.”

In  HRD Corpn. v. GAIL (India) Ltd (2018), the arbitrator had previously provided a legal opinion to a party. While this case acutely falls under sub-clause (24), the SC held in obiter dicta that item 16, i.e. “The arbitrator has previous involvement in the case”, cannot be read as including previous involvements in another arbitration on a related issue. It was held that such an interpretation would render sub-clause (24), which imposes the additional caveat, i.e. “involving one of the parties or an affiliate of one of the parties”, largely ineffective. Thus, judicial interpretation also appears to limit bias to instances where the arbitrator previously acted as legal counsel or advisor, rather than merely having served as an arbitrator in related matters. Further, sub-clause (24) of Schedule V is the guiding provision in such cases, and effectively dictates the extent to which the A&C Act seeks to regulate double-hatting, or where they served as arbitrator in a related case involving one of the same parties within the past three years. 

III. ADDRESSING BIAS IN ARBITRATION: SELF-PRESERVATION VS. PRE-DISPOSITIONAL BIAS

The A&C Act uses a proximity-based standard of bias, where prior direct relationships with a party within a recent, defined timeframe are presumed to give rise to justifiable doubts about impartiality. The A&C Act only takes ‘direct and specific proximity’ to the dispute into consideration, and not ‘general proximity’. This standard, however, does not cover a wide variety of conflict of interest situations, including double-hatting. Patel J. highlights that impartiality needs to be understood in the context of remoteness of interest in question. Paul Craig draws a distinction between pecuniary and personal interest. A pecuniary interest, however small, is usually a valid ground for disqualification, with two notable exceptions. Firstly, when it is not personal to the decision maker and secondly, when it is too remote. In such situations, these matters are to be treated as personal interests as per Craig. While Patel J. only contextualizes bias as arising from a vested ‘interest’, it is argued that there is a class of biases that do not arise from vested interest, i.e. subconscious biases, or pre-dispositional biases.

Broadly, bias arising from pecuniary or personal interests is self-preservation bias, arising when the arbitrator’s decision has the potential to directly or indirectly impact their personal or professional interests. For instance, concurrent roles as both arbitrator and counsel might tempt one to add a line in the award that bolsters their position as counsel in another case. Similarly, if the arbitrator’s spouse/relative works for any of the parties, one might subconsciously aim to avoid adverse rulings against the said party, driven by a sense of self-preservation. Even in scenarios where the arbitrator has previously provided legal advice on the same dispute, their decisions might be influenced by the need to uphold the credibility of their past advice. In all such cases, the locus of the bias is tied to safeguarding the arbitrator’s own interests. The author argues that, unlike self-preservation bias, predisposition-bias arises from pre-existing factors that make the arbitrator biased either in favour of, or against a party or issue based on their predispositions. These factors may include shared affiliations, ideological alignments, or other subjective influences. For instance, an arbitrator might be predisposed to favour someone from the same caste or community, or one may be predisposed to side with an athlete in a sports arbitration, if the arbitrator themself has been an athlete in past. Unlike self-preservation bias, pre-dispositional bias is not about protecting oneself but stems from past experiences, broader influences or associations that shape the arbitrator’s perspective/outlook. In the context of double-hatting and the RSE Holdings case discussed before, Ms. Frey, being a counsel in multiple ECT disputes, could have pre-disposed her to unconsciously favour certain legal interpretations or argumentative frameworks she had previously advanced as counsel. Rather than consciously seeking to benefit herself, her repeated engagement with similar substantive issues could have subtly shaped her approach, thereby compromising her ability to approach the dispute with a fresh and neutral perspective.

Daniel Kahneman, in his book Thinking, Fast and Slow, argues that one often associates new information with existing patterns or thoughts, rather than creating new patterns for each new experience. In this context, an arbitrator may only be able to conceptualize a case within the framework within which he has previously argued, given the similarity of substantive issues. As Greenwald and Banaji note, “Implicit attitudes are introspectively unidentified (or inaccurately identified) traces of past experience that mediate favorable or unfavorable feeling, thought, or action toward social objects.” These biases operate outside conscious awareness, meaning an arbitrator may unknowingly carry predispositions, rooted in personal background or social identity, that subtly affect decision-making.

The question that arises at this juncture is, to what extent can a bias such as this be regulated, or rather, is it even desirable to completely regulate it? Harding’s famously known ‘standpoint theory’ is essentially premised on the fact that knowledge is shaped by one’s experiences. Thus, it is practically impossible to find a completely neutral third person/judge. However,  parties have the right to appoint a ‘neutral’ arbitrator. The standard of neutrality, however, has to be reasonable and practical. For instance, allowing parties to challenge an appointment and allege bias on the grounds of the arbitrator belonging to the same caste or nationality is impractical. Further, such a claim is not even remotely related to the substantive issue of the case. Here, the test that the author proposes is to check if the claim finds ‘general proximity to the dispute’, as opposed to the direct and specific proximity standard. Thus, the ability and potential to affect the ‘dispute’, as opposed to the parties. Such an assessment essentially eliminates challenges based on shared affiliations, ideological alignments, like the arbitrator being from the same caste or nationality, as such grounds are not ‘proximate to the dispute’, while also allowing assessment of other factors that might not be proximate to the parties, but are relevant to the dispute at hand.

IV. PROPOSED FRAMEWORK FOR TACKLING DOUBLE-HATTING

Having established the difference between pre-dispositional and self-preservation bias, and the need to include double hatting within the current framework, here the author suggests a framework to mechanise the distinction.

Scholars have suggested imposing a quantitative limit on the number of times an individual can switch between roles. Notably, Devarkonda argues that double-hatting raises both real and perceived risks of bias, and imposing a definitive limit on the number of times an individual can alternate between serving as arbitrator and counsel helps enhance public confidence and ensures consistency. While this suggestion seems effective on paper, Devarkonda doesn’t talk about how this would actually take effect. Further, the consequences of non-compliance, i.e. if such a limit is crossed by an individual, are unclear. Moreover, if such a number is exhausted, it would preclude the individual from practising as either an arbitrator or a counsel in a specific area. Not only is this ‘area’ of substantive law difficult to define, but this makes the solution proposed largely impractical. Further, the existing scholarship also often recommends introducing a cooling-off period, but even this does not guarantee the elimination of bias. This is because it’s next to impossible to accurately determine a duration that would rid one of pre-dispositional biases developed as a result of double-hatting.  

Notably, in cases such as St. Gobain Performance Plastics v. Venezuela and Blue Bank International & Trust (Barbados) v. Bolivarian Republic of Venezuela, a challenge based on the arbitrator’s prior professional advocacy was outright rejected. Whereas in cases such as Grand River Enterprises v. United States, a challenge based on concurrent appointment was upheld. It is thus clear that most investor arbitration panels don’t view past engagement as much of a problem, as concurrent engagement as arbitrator and counsel in cases with substantive similarities. In line with this, the author proposes imposition of a dual duty on an arbitrator.

Firstly, the duty to refrain from engaging in substantively similar matters in any capacity concurrently, which is a negative post-constitution (of the arbitration tribunal) duty. The UNCITRAL Draft Code of Conduct for Arbitrators in International Investment Dispute Resolution prohibits an arbitrator from acting in the capacity of a legal representative or an expert witness in a case dealing with the same measure, related party, or provision of the same instrument of consent. This burden is a negative post-constitution duty. Adoption of such a provision effectively deals with the issue of concurrent appointments, i.e. if the proposed arbitrator is concurrently arguing a substantively similar matter as counsel in another case.  

Secondly, an enhanced disclosure duty u/s 11 A&C Act, which is a positive pre-constitution of the arbitration tribunal) duty. Attention must be paid to the Draft ICSID Arbitration Rules, which, through Rule 13(4), allows a person who has been involved in an issue of a similar nature in any capacity ONLY by agreement of the Parties. ICSID Regulations, however only superficially deal with the issue, leaving many operational questions unanswered; Cardoso and Cuadrado critique Rule 13(4), identifying two major issues in its implementation - difficulty in identifying overlapping cases at appointment, and uncertainty over start and duration of three-year ban obligations. The author recommends the adoption of a similar provision with the following changes. Art. 11(2)(iii) already expects arbitrators to disclose “Any other proceeding involving a disputing party or an entity identified by a disputing party”; this may be amended to “Any other proceeding involving a disputing party or an entity identified by a disputing party or involving a substantively-similar matter”, in line with the general proximity approach suggested earlier. Further, to complement this, the author also suggests an amendment to Schedule V, specifically sub-clause (24), to allow for parties to challenge an appointment on grounds of double-hatting. Sub-clause(24) effectively addresses past engagements, but imposes the additional caveat of direct and specific proximity, as discussed before. If this additional caveat is removed, sub-clause (24) would effectively bridge the gap that the ICSID Regulations fail to address. This effectively allows a party to raise double-hatting as an issue before the tribunal before an award is rendered, and also allows it to be used as a ground to set aside an award u/s 34 A&C Act. However, because it’s not a criterion under Schedule VII, it cannot be used as grounds to allege de jure pre-constitution ineligibility.

V. CONCLUSION

Arbitration offers parties a unique opportunity to have considerable control over the adjudication process, unlike state-affiliated adjudication mechanisms. Ensuring that parties are convinced that the adjudicator, a private professional with no public duties vested by state, is unbiased, independent and impartial is a key component of this process. In such a scenario, it is important to uphold the integrity of the process and party autonomy effectively and holistically. The biggest concern with prohibiting double-hatting remains its impact on the pool of potential arbitrators that parties can select from. In fact, the ICSID in its Code of Conduct Background Papers on Double Hatting noted that in  47% of ISDS cases, at least one arbitrator also serves as counsel. Further, the current structure mirrors the system followed by most judiciaries, where acquiring experience as counsel remains the normal path leading to appointments as a judge. A blanket ban would thus be virtually impossible, and could lead to the creation of a class of professionals who have no experience in contentious practice, defeating the purpose of arbitration.

---

*Siddhanth Shinde is a 4th Year B.A. LLB (hons.) student at MNLU Mumbai.

[1] See https://repository.nls.ac.in/cgi/viewcontent.cgi?article=1097&context=nlsblr, https://dailyjus.com/world/2022/01/regulating-double-hatting-in-investment-treaty-arbitration-latest-icsid-uncitral-proposals, 

[2] Rule 20, SIAC Rules 2025; Sec. 2, SIAC Code of Ethics for an Arbitrator; Art. 11.2, HKIAC Rules 2024; Art. 11(2), ICC Rules 2021; Art. 5.4, LCIA Rules 2020

[3] Sec. 14(1)(a) A&C Act; See Manikankana Saha v. Leela Devi Kasera, 2023 SCC OnLine Cal 438; Steel Authority of India Ltd. v. British Marine PLC, 2016 SCC OnLine Del 5599.

[4] Bharat Broadband Network Ltd. v. United Telecoms Ltd., (2019) 5 SCC 755.

[5] See Paras 1,8 & 15, Schedule VII, A&C Act, 1996.

[6] Sec. 34 A&C Act, 1996