This paper was first published in (2026), 22, Asian International Arbitration Journal, Issue 1, pp. 15-30, Excluding Illegally Obtained Evidence in International Arbitration: A Consent-Based Approach to (In)Admissibility – Kluwer Law Online. 

SECTION I: INTRODUCTION

Almost ninety years ago, the US Supreme Court in Nardone v. United States (1939) was called to rule upon the admissibility of evidence procured through warrantless wiretapping.¹ The evidence in question, obtained in breach of section 605 of the extant Communications Act, 1934, had been used by the trial court to convict Nardone and the other defendants of smuggling alcohol. The conviction was subsequently upheld by the Second Circuit Court of Appeals, which held that excluding the evidence would be a “reversible error,” or an error of sufficient seriousness to warrant the judgment being reversed on appeal.²

When the case reached the US Supreme Court, however, the Court did exactly that. Justice Felix Frankfurter, writing for the Court, held that Nardone had satisfied the burden of proving that the wiretapping had been unlawfully employed, thus demonstrating that “a substantial portion of the case against him was a fruit of the poisonous tree.”³ The substance of the principle originated in the judgment passed in Silverthorne Lumber Co. v. United States (1920). However, it was Justice Frankfurter’s pithy metaphor which came to widely describe this position of law, i.e., that illegally obtained evidence (“IOE”) is like “fruit” from a “poisonous tree” — tainted, and inadmissible due to the illegal way it had been procured.

This doctrine continues to enjoy considerable influence in the US, where cases like Plymouth Sedan v. Pennsylvania (1965) have affirmed its application in civil proceedings before national courts. ⁴

In contrast, in other jurisdictions, IOE is assessed not only with reference to the manner in which it was secured, but also — and primarily — on other factors, such as its relevance, reliability, and materiality. Interestingly, this is true of both civil and common law jurisdictions.

For instance, in the UK and most Commonwealth countries, judges will weigh competing considerations to decide whether to admit IOE, particularly when the judgment has important public policy implications. In the seminal case of R v. Leathem (1861), the prosecution sought to admit a letter written by a person accused of bribery under the Corrupt Practices Prevention Act, 1854.⁵ This letter had been discovered because of an inadmissible statement made by the accused.

Nonetheless, it was admitted into evidence, with Justice Crompton famously holding, “It matters not how you get it; if you steal it even, it would be admissible in evidence.” ⁶ More recently, in Wong Keng Leong Rayney v. Law Society of Singapore (2007), ⁷ a sting operation caught a solicitor offering referral fees in exchange for conveyancing work. Disciplinary charges were brought against him. The solicitor objected to the use of the recording, contending that the evidence against him was illegally obtained. The disciplinary committee rejected this argument and held that the evidence was admissible. The solicitor challenged this decision before the Singapore High Court. The High Court, while dismissing the challenge, clarified that the general principle under the Evidence Act is that all relevant evidence is admissible. This also applied to improperly obtained evidence unless it operated unfairly at trial. The Court further highlighted that the question of unfairness is not concerned with how the evidence was obtained, but with whether its prejudicial effect at trial might exceed its probative value. ⁸

Many civil law countries take a similar view — in Sweden, judges enjoy a wide degree of latitude under the principle of “fri bevisprövning,” or “free evaluation of evidence.” This allows both sides to adduce and use any evidence, with the court ultimately evaluating its value and impact.⁹ Other civil-law countries that utilize and apply such standards are France, Germany, Italy, the Netherlands, and Spain. Courts in these countries may choose to exclude such IOE on a balance of considerations.

It perhaps points to the practical value of such a flexible standard that both civil and common law systems have converged on a shared understanding, i.e., the exclusion of IOE is flexible depending on the facts and circumstances in each case, not automatic.

Unfortunately, in contrast to the state of the law in relation to IOE in civil proceedings in national courts, the rules governing the admissibility of IOE in arbitration are less clear-cut. This paper argues that parties’ consent to arbitrate cannot reasonably be construed to encompass the use of IOE by parties in the course of proceedings. In fact, its admission fundamentally vitiates the procedural bargain underlying the arbitration agreement. An illegality of this sort would taint the integrity of the arbitral process and thatof any resulting award.

In international arbitration, matters of procedure are nominally governed by the “lex arbitri,” the law of the seat. In practice, however, procedural aspects of arbitral proceedings are dependent on rules of arbitral institutions or ad hoc arrangements agreed on between an arbitrator and parties in ad hoc arbitrations. By and large, these rules and arrangements provide minimal guidance on procedural particulars, instead conferring an expansive discretion upon the arbitral tribunal. This flexibility, long hailed as arbitration’s strength, becomes its vulnerability when confronted with problems such as the treatment of IOE.

When faced with issues without clear doctrinal moorings, arbitrators may borrow from domestic court practices as well as any standards prescribed under the lex arbitri. While this approach of reliance on the practice and procedures of courts may be acceptable on questions of document production or expectations of reasoning in a written award, it is less appropriate when it comes to the treatment of IOE. When national courts admit IOE, they do so as organs of a sovereign state with inherent jurisdictional authority granted by legislation. While arbitral tribunals have been granted wide discretionary powers that are subject to the consent of the parties and applicable laws, they also have the flexibility to conduct the arbitration in a manner the tribunal considers appropriate. However, their authority is derivative, contingent, and bound by the parties’ consent. Put simply, arbitration and consequently, the arbitral tribunal, is a creature of contract.

Consistent with this logic, this paper argues that the solution lies not in refining the existing balancing tests, but in reconceptualizing the problem. Particularly, rather than asking how IOE should be treated based on conflicting policy considerations, arbitral tribunals should ask a more fundamental question: does the admission of such evidence even fall within the scope of the parties’ consent?

SECTION II: COURTS v. TRIBUNALS

While courts and arbitral tribunals discharge an adjudicatory function, as fora for the resolution of disputes, the two are fundamentally dissimilar. Courts are organs of the state, established by statute or even directly by the Constitution. They derive their authority from the sovereign power of the state and exercise broad, predetermined powers over a wide range of disputes. Their jurisdiction is inherent and cannot be contracted “out of” or overridden by private agreements except in well-defined circumstances, (e.g., waiver or limitation of liability clauses).

While tribunals also enjoy wide discretionary powers vested by the lex arbitri, arbitral tribunals ultimately derive their jurisdiction from the contractual agreement of the parties: their jurisdiction is both created and constrained by the terms of the parties’ consent.

Yves Derains puts it like this: state courts have “two sorts of powers: the power to pronounce the law, or jurisdictional powers, derived from the Roman jurisdictio, and the power to command, derived from the Roman imperium.”¹⁰ Of these two, arbitral tribunals possess the former, but not the latter — which is why they require national courts to intervene and supply their “powers of coercion” to grant exequatur and make an arbitral award enforceable.¹¹ The Indian Supreme Court, in MD Army Welfare Housing Organization v. Sumangal Services Pvt. Ltd. (2003), summed it up in these terms:

An arbitrator cannot be equated with a court of law. Whereas court has an inherent power; an arbitrator does not have. It is a tribunal of limited jurisdiction. Its jurisdiction is circumscribed by the terms and reference. An arbitrator can act only within the four corners of the agreement and not beyond. ¹²

Of course, these structural differences have further consequences. Court proceedings are generally, with narrowly defined exceptions, conducted in public as the principle of “open justice” requires transparency, accountability, and public confidence in the judicial system. It allows scrutiny of judicial decisions, discourages bias or unfairness, and upholds the rule of law. Even in purely commercial cases between private parties, the court’s role in enforcing legal rights and shaping precedent justifies public access.

Arbitration, by contrast, is “private” by default, being based on party autonomy. The tribunal exists only because the parties consented to create it, and issues awards that are either published on a redacted basis or kept confidential, both at the parties’ discretion. Confidentiality is one of arbitration’s main attractions — not only for protecting sensitive business information, but also for preventing the sort of public exposure concomitant with traditional litigation. Institutional rules mostly reinforce this confidentiality, though laws in some jurisdictions, (e.g., Australia¹³) impose limits on absolute secrecy in arbitration.

Additionally, courts operate within a tightly structured system of appellate review, allowing for oversight and correction of legal or procedural errors, while arbitration is designed to be speedy, final, and binding, with deliberately few avenues for appeal or even other judicial intervention.

This combination of confidentiality, case-by-case decision making, and lack of formal evidentiary rules means that arbitration lacks the structured safeguards of judicial proceedings. In fact, in the words of Catherine Rogers, “[i]nternational arbitration dwells in an ethical no-man’s land.”¹⁴

This is particularly true for the admission of IOE. Though tribunals operate within a narrower scope of authority, grounded in party consent, that very limitation allows them to avoid rigid rule-following and adopt a comparatively flexible approach to matters such as evidence-taking. In fact, the rules of virtually all major arbitral institutions offer little guidance on the admissibility, legitimacy, or value of IOE.

For example, Rule 36(1) of the International Centre for Settlement of Investment Disputes (“ICSID”) Arbitration Rules, 2022 states that ‘[t]he Tribunal shall determine the admissibility and probative value of the evidence adduced.’ Similarly, Rule 32.3 of the Arbitration Rules of the Singapore International Arbitration Centre (“SIAC”), 2025 states that a tribunal shall determine admissibility, relevance, materiality, and weight of all evidence, further clarifying that it ‘not required to apply the rules of evidence of any applicable law in making such a determination.’ Article 22.1(vi) of the London Court of International Arbitration (“LCIA”) Arbitration Rules, 2020 and Article 22.2 of the Hong Kong International Arbitration Centre (“HKIAC”) Administered Arbitration Rules, 2024 similarly leave it to the tribunal to decide whether to apply ‘strict rules of evidence.’

Given the above framework, parties often turn to gap-filling ‘soft-law’ instruments, with one of the more prevalently used instruments being the International Bar Association (‘IBA’) Rules on the Taking of Evidence in International Arbitration, 2020 (‘2020 IBA Rules’). However, the 2020 IBA Rules are supplementary in nature, designed not to override existing arbitration rules or matters of procedure agreed on between the parties and/or set by the tribunal, but to fill the lacunae left by them. Therefore, they still allow tribunals a great deal of discretion in their application.

The Preamble to the 2020 IBA Rules itself makes it clear that they are ‘not intended to limit the flexibility that is inherent in, and an advantage of, international arbitration.’ This has been reinforced by the amendment to Article 9.3 of the 2020 IBA Rules. This revised provision expressly states that a tribunal ‘may’ exclude IOE, either suo moto or on the request of a party. However, this change is largely accepted as a codification of existing arbitral practice versus a substantive shift. The IBA Commentary on the 2020 Rules further confirms that Article 9.3 was intended to be permissive because of the lack of a universal approach to the admissibility of IOE.¹⁵

Past decisions, some of which are discussed below, have turned on various factors, including: (1) whether the party seeking to introduce the evidence was involved in the illegality; (2) the materiality of the evidence to the outcome of the case; (3) whether the information was already in the public domain; and (4) the severity of the illegality involved. Rather than establishing a firm prohibition, the amendment elected to preserve arbitral discretion.¹⁶ Therefore, unlike Article 9.2 of the 2020 IBA Rules, which requires the exclusion of evidence on specific grounds, such as legal privilege or unreasonable burden of proof, Article 9.3 does not obligate the tribunal to exclude IOE, unless they deem it suitable or necessary.

SECTION III: IOE EXCEEDS THE SCOPE OF THE CONSENT TO ARBITRATE

3.1 IOE OUGHT TO BE EXCLUDED

This paper proposes that IOE should be excluded, as the acceptance of IOE undermines the scope of consent to arbitrate disputes that is embodied in an arbitration agreement. First, because the principles of contractual interpretation directly contradict its implied admissibility; and second, because implied duty of bona fides — recognized and upheld as an implied contractual term in most legal systems under discussion in this paper — requires that a party not be required to expend time and costs refuting IOE.

This consent-based issue with IOE is true even for jurisdictions where there is a distinction drawn between questions of admissibility and weight of IOE, i.e., in systems where a tribunal may be more inclined to receive IOE into the record and subsequently rationalize it by discounting its weight. This paper argues that such a practice still risks eroding the parties’ consent, regardless of the evidence’s perceived value, as its very presence alters the procedural equilibrium of the arbitration.

Suppose that the phrase ‘irrespective of whether the evidence considered in the arbitration is illegally obtained’ is inserted into an arbitration clause. It is evident that no rational party would sign such an agreement. If a party would not agree to such a provision in express terms, it ought not be implied into the contract either: what cannot be done directly cannot be done indirectly. Admissibility of IOE v. Duty to act in good faith.

In addition to undermining the express consent of the parties to arbitrate, admissibility of IOE might undermine the duty of good faith implicit in contract as recognized by the majority of the legal systems under discussion in this paper. Often defined in opposition to ‘bad faith’, which involves ‘a design to mislead or deceive another’¹⁷ or dishonesty of belief, purpose or motive¹⁸ (i.e., sustained deception), ‘good faith’ imposes on parties the obligation to be sincere in their conduct. As Kotzur writes for the Max Planck Encyclopedia of Public International Law, ‘[b]ona fides is the most fundamental principle of substantive law also applicable to the proceedings before international courts and tribunals’, of which there are a ‘series of “concretizations” in the field of procedural law.’¹⁹ Schreuer, in his legal opinion in Stati v. Kazakhstan, described the doctrine of “clean hands” as one such concretization,²⁰ citing cases like Diversion of Water from the River Meuse (1937),²¹ the Namibia Advisory Opinion (1970), and more recently, the Hungary-Slovakia Gabčikovo-Nagymaros Project dispute (1997),²² to demonstrate how approaching a forum with “unclean hands” could even strip a party of locus standi.

These principles operate powerfully in favor of excluding IOE — not merely as ubiquitous legal ideals, but as an implied obligation in agreements. It is settled law that the duty of good faith, or “fair dealing,” is a pervasive contractual duty — as the New York Court of Appeals had held in the early case of Kirke La Shelle Co. v. Paul Armstrong Co. (1933):

In every contract, there is an implied covenant that neither party shall do anything which will have the effect of destroying or injuring the right of the other party to receive the fruits of the contract.²³

(emphasis supplied)

This sort of reasoning has found support across jurisdictions, and even before international courts and tribunals. For example, the Canadian Supreme Court in Bhasin v. Hrynew (2014) cited a range of cases to conclusively lay down that no matter its source, good faith “imposes as a contractual duty a minimum standard of honest contractual performance.”²⁴ On the other hand, for civil-law systems (such as those in Continental Europe) the duty of good faith is many a times codified into the law, such as “Treu und Glauben” (“trust and faith”) under Article 5(3) of the Swiss Constitution, or “redelijkheid en billijkheid” (“reasonableness and fairness”) under Article 6:248 of the Dutch Civil Code.

Internationally, too, the concept has been entrenched in Articles 26 and 33(1) of the Vienna Convention on the Law of Treaties, 1969 (“VCLT”), relied upon by the ICJ in Djibouti v. France (2008). In fact, much like the Canadian Supreme Court, the ICJ in Nuclear Tests (1974) had held that “one of the basic principles governing the creation and performance of legal obligations, whatever their source, is the principle of good faith,”²⁵ speaking to the normative force possessed by the idea.

More pertinently, even international arbitral tribunals have repeatedly invoked and applied the doctrine of good faith. In Inceysa v. El Salvador²⁶ and Plama v. Bulgaria,²⁷ the tribunals dismissed the investors’ claims because their conduct (deception, misrepresentation, concealment of material facts) betrayed the fundamental duty of good faith. Similarly, in Phoenix v. Czech Republic, the tribunal rejected a claim where the investor sought treaty protection after the dispute had already arisen, emphasizing that its conduct amounted to an attempt to abuse the arbitration process.²⁸ Most notably, in Sanum v. Laos, the tribunal held:

The principle of good faith arises in investment treaty arbitrations in various contexts. Tribunals, of course, regularly refer to Article 31(1) of the VCLT for the rule that treaties shall be interpreted in good faith. The obligation extends to a duty of parties to arbitrate in good faith.²⁹

(emphasis supplied)

If tribunals can bar entire claims on the basis of “good faith,” it would be an extension from that to argue that they can and ought to use the same reasoning to exclude IOE. In the vast majority of cases, such evidence cannot be admitted without “tainting” or compromising the integrity of the arbitral process.

3.2 THE BALANCING TEST

Consider another iteration of this argument. As mentioned above, Article 9.2(b) of the 2020 IBA Rules compel the exclusion of documents protected by legal privilege. Likewise, in cases like Libananco v. Turkey,³⁰ tribunals have summarily refused to admit privileged documents, irrespective of their substance and relevance. In Libananco, the Tribunal stated that it attributed “great importance to privilege and confidentiality, and if instructions [had] been given with the benefit of improperly obtained privileged or confidential information, severe prejudice [could] result.”³¹ In Wee Shuo Woon v. HT S.R.L,³² the Singapore Court of Appeal held that privileged documents did not lose this privilege after the WikiLeaks hack placed them in the public domain — if anything, “the equity in favour of restraining the use of privileged documents [was] even stronger in the case of a party who had its privileged documents accessed and taken through stealth and unlawful means.”³³

Essentially, it is implicit in the above provisions of the 2020 IBA Rules that some protections are so fundamental that they override considerations of relevance or materiality. To make parties fear that their private communications with counsel may be exposed and admitted fundamentally undermines the right to a fair defence. This is precisely the reason for tribunals to strictly protect legal privilege. Without it, parties cannot seek legal advice freely, fearing that their own words may later be used against them.

But does the same logic not extend beyond privileged communications to IOE? Why do arbitral tribunals administer a “balancing” test when confronted with other forms of evidence, outside privileged communication, to, for example, IOE? Such a balancing test usually weighs the probative value of the evidence against other considerations, such as due process, party conduct, and the severity of the illegality involved. Consequently, even when tribunals reject IOE, they do so not merely because it was unlawfully obtained, but because other disqualifying factors — be it unreliability, immateriality, even privilege — have cumulatively tipped the scale.

In Methanex v. USA, the documents the Claimant procured by “diving” through Respondent’s dumpsters were rejected not only because the Claimant had committed trespass and theft, but also because the documents themselves had only “marginal evidential significance.”³⁴ Similarly, in EDF v. Romania, the exclusion of audio recordings made in breach of Romanian law was done on the basis of the illegal procurement of the recordings but also their doubtful authenticity, and lack of clear probative value.³⁵

In cases where IOE was admitted — for instance in Caratube v. Kazakhstan — the tribunal held that the non-culpable involvement of Caratube in the KazakhLeaks and the significance of the leaked documents together outweighed Kazakhstan’s objections to their admissibility.³⁶

Simply put, if tribunals remain willing to admit IOE under a balancing test, they create an environment where parties face the risk that IOE will be admitted if it has sufficient probative value. This erodes trust in the arbitral process, shifting it from a system based on mutual consent and fairness to one where parties must assume that wrongdoing may be rewarded if the resulting evidence is relevant.

SECTION IV: DOES ADMITTING IOE AMOUNT TO PROCEDURAL FRAUD ON THE RESISTING PARTY?

“Evidentiary fraud” as a species of procedural fraud is becoming increasingly influential across jurisdictions. In FIC Properties Sdn Bhd v. PT Rajawali Capital International (2024),³⁷ the Singapore International Commercial Court (“SICC”) considered this issue. The SICC held that the definition of “fraud” under section 24(a) of the International Arbitration Act, 1994 included “procedural fraud,” defined as “when a party commits perjury, conceals material information and/or suppresses evidence that would have substantial effect on the making of the award.” The SICC, relying on the decision of the Singapore Court of Appeal in Bloomberry Resorts and Hotels Inc v. Global Gaming Philippines LLC (2021),³⁸ further clarified that only procedural misconduct aimed at deceiving the arbitral tribunal would amount to procedural fraud. Though this test was ultimately not satisfied in FIC Properties (and the impugned award not annulled) it provided valuable guidance on the notion of “evidentiary fraud” in the context of enforcing awards.

Similarly, in January 2023, the Amsterdam District Court denied leave to recognize and enforce the USD 350M arbitral award obtained by the Claimants in Statis v. Kazakhstan, holding that they had submitted false and incomplete information, misled the tribunal, and actively concealed relevant facts.³⁹ This was the fourth time (after Belgium, Luxembourg, and England & Wales) that a court of first instance had declined to enforce the Statis award on the basis of the Claimant’s evidentiary misconduct.

This idea is not contemporary — English law, for example, has long upheld this principle. As far back as Jet Holdings Inc v. Patel (1990),⁴⁰ courts have affirmed that, in the context of resisting award-enforcement, fraud had to be understood in broad terms, encompassing “every variety of mala fides and mala praxis whereby one of the parties misleads and deceives the judicial tribunal.”⁴¹ This broad interpretation was reflected in subsequent decisions, such as Profilati Italia SrL v. PaineWebber Inc (2001)⁴² and Elektrim SA v. Vivendi Universal SA (2007),⁴³ where deliberate suppression of evidence was deemed tantamount to fraud. An award made on that strength, therefore, was an award “procured” in a manner contrary to public policy. In Thyssen Canada Limited v. Mariana Maritime SA (2005), Justice Cooke specifically emphasized that evidentiary misconduct, whether through perjury, concealment, or manipulation, would result in the setting-aside of an award if it directly caused substantial injustice.⁴⁴

This evolving jurisprudence provides a compelling argument that IOE can be located within this same category of procedural fraud. Evidently, the evolving definition of “procedural fraud” has subsumed within itself the gamut of evidentiary misconduct, ranging from perjury to suppression of evidence, and even concealment. This paper argues that introduction of IOE, too, falls under this category, as it similarly distorts the arbitral process in an inherently deceptive and coercive manner.

While describing a complete, step-by-step practical approach to the question of the admissibility of IOE is beyond the scope of this paper, tribunals may consider adopting a two-step assessment.

The first, or the threshold question which the tribunal must consider, is whether the allegation of illegality implicates the integrity of the arbitral process itself. An apt example would be evidence obtained through hacking, bribery, or breach of confidentiality. If so, an early ruling on admissibility is warranted, as allowing such material into the record could risk tainting the entire proceedings. Here, the tribunal does not act beyond its mandate but in fulfilment of it, since international arbitral practice gives tribunals wide discretion to decide on questions of admissibility under most procedural rules.

If this threshold is not met, then the tribunal can admit the evidence provisionally, note the objection, and reserve its decision on the final admissibility of the evidence once it assesses weight and relevance in the final award. This avoids procedural gridlock and limits the risk of frivolous objections being used strategically to delay the proceedings.

The risk of abusive objections can also be managed by requiring the objecting party to substantiate its claim of illegality promptly and with specificity. On the point of the IOE itself, tribunals may apply a filter framed around five questions: is the evidence authentic? And if yes, then examining who committed the alleged wrong in its procurement; the nature of the wrong; the nature of the evidence; and lastly, who stands to benefit from its admission.⁴⁵ Perhaps such an approach ensures procedural efficiency while preserving the tribunal’s authority to reject evidence that truly compromises the integrity of the process.

SECTION V: ACCEPTANCE OF IOE: A READY CHALLENGE TO THE RESULTING AWARD?

Further, situating IOE in the category of ‘procedural fraud’ has significant practical implications. Across jurisdictions, an award may be annulled or its enforcement may be refused on the ground of ‘public policy.’ Article 34(2)(b)(ii) of the Model Law allows an award to be set aside if it is contrary to the public policy of that state. Similarly, Article 36(1)(b)(ii) of the Model Law allows a court to refuse recognition or enforcement of an award for the same reason. Article V(2)(b) of the New York Convention, too, gives enforcing courts the discretion to refuse enforcement if enforcing the award would be contrary to the public policy of that country.

On the question of what constitutes ‘public policy’, there has traditionally been a distinction between the approach of the courts in annulment proceedings on the one hand, and non-enforcement proceedings on the other. Courts at the seat, when dealing with an annulment or set-aside application, most often apply domestic policy standards to the application, reasoning that a country has the sovereign right to ensure that awards rendered under its law comply with its core legal values. On the other hand, enforcing courts often refer to ‘international’ or ‘transnational’ public policy decisions, which tends to be a stricter standard than domestic public policy. This is to prevent excessive interference with the finality of arbitral awards, while still protecting core legal principles. A case from Colombia illustrates this distinction. In Tampico Beverages Inc. v. Productos Naturales (2017), the Supreme Court of Colombia was confronted with an ICC award that allegedly violated Colombian domestic public policy due to an arbitrator’s conflict of interest.⁴⁶ Interestingly, while the Court acknowledged the potential violation under domestic standards, it held that international public policy was still distinct and required reference to international norms. Therefore, to assess whether non-enforcement was appropriate, the court looked beyond national law and considered the 2014 IBA Guidelines on Conflicts of Interest, treating these principles as representative of global best practices.

However, first, over time, the standards of ‘domestic’ and ‘transnational’ public policy are converging, with jurisdictions like France including an ‘international public policy’ test as a ground for annulment, as reflected in Article 1520(5) of the French Code of Civil Procedure. This reflects a broader shift toward a more globally consistent standard of public policy. Second, ‘fraud’ is indubitably a component of this transnational public policy — whether due to its ubiquity across legal systems, or because it autonomously meets the definition of a transnational norm. Some guidance can be found in the preparatory materials of the Model Law. For instance, the UNCITRAL Report on the work of its eighteenth session stated:

297. In discussing the term ‘public policy’, it was understood that it was not equivalent to the political stance or international policies of a State but comprised the fundamental notions and principles of justice … It was understood that the term “public policy,” which was used in the 1958 New York Convention and many other treaties, covered fundamental principles of law and justice in substantive as well as procedural respects. Thus, instances such as corruption, bribery or fraud and similar serious cases would constitute a ground for setting aside.⁴⁷

Given that virtually every jurisdiction treats fraud as contrary to domestic public policy, fraud arguably lies within the scope of generally accepted definitions of public policy that would be accepted in a cross-border transnational context. This applies not only to substantive fraud, such as fraud in the underlying contract or transaction, but also to procedural fraud, which taints the arbitration process itself.

In the words of Lord Denning in Lazarus Estates v. Beasley (1956): “fraud unravels everything.”⁴⁸ This enduring legal maxim reflects the gravity with which fraud is treated across the globe. More recently, the Delhi High Court concurred with the Indian Supreme Court’s finding on fraud in the long-running Devas v. Antrix saga, holding that:

After affirming the concurrent finding of fraud the Supreme Court has held that if the seeds of the commercial relationship between Antrix and Devas were a product of fraud perpetrated by Devas, every part of the plant that grew out of those seeds, such as the Agreement, the disputes, arbitral awards etc., are all infected with the poison of fraud. A product of fraud is in conflict with the public policy of any country including India.⁴⁹

(emphasis supplied)

Arbitral tribunals do not operate in a vacuum. They have a duty to ensure that the awards they render are not just comprehensive or well-reasoned, but also capable of enforcement. As Julian Lew puts it, “[t]he ultimate purpose of an arbitration tribunal is to render an enforceable award.”⁵⁰ An award that risks annulment at the seat or refusal of enforcement in key jurisdictions is, in many ways, a failed award. This principle is not merely intuitive or theoretical, but is embedded in institutional rules. For instance, Rule 3.5 of the SIAC Rules, 2025, requires the SIAC Court, President, Vice President, Registrar, SIAC Secretariat as well as the Emergency Arbitrator and Tribunal to endeavor to ensure the enforceability of an award. Similarly, Article 42 of the ICC Rules, 2021 explicitly directs arbitral tribunals to “make every effort to make sure that the award is enforceable at law.” Likewise, Article 32.2 of the LCIA Rules, 2020 mirrors this expectation, emphasizing that tribunals must use “reasonable efforts” to produce awards that are “legally recognised” and “enforceable at the arbitral seat.” An award that cannot withstand scrutiny is an award at risk, and one that arbitrators have a codified duty to avoid.

In fact, courts have increasingly emphasized that it is not just the correctness of an arbitral tribunal’s decision that matters, but also the integrity of the process leading to the award. Consider the landmark decision of the High Court of England & Wales in Nigeria v. P&ID (2023), in which the High Court set aside an award for USD eleven billion in damages including interest, on the basis that it had been procured through fraud and serious procedural irregularities.⁵¹ The Court had emphasized that even if an arbitral tribunal reaches a substantively “correct” decision, the legitimacy of the award was fundamentally compromised if the arbitral process itself was tainted. In this case, P&ID had engaged in multiple forms of misconduct, including bribing Nigerian officials to secure the underlying contract and improperly accessing Nigeria’s privileged legal documents during the arbitration. The Court found that these actions so seriously violated principles of due process and fairness that it was impossible to uphold the award.

Even if an award survives annulment at the seat, it remains vulnerable at the enforcement stage. A party resisting enforcement could argue that reliance on IOE violates fundamental principles of due process and fairness, triggering public policy objections under the New York Convention and national laws. This means that tribunals admitting IOE may inadvertently jeopardize the finality and enforceability of their awards. Given the increasing judicial scrutiny of procedural fairness in arbitration, reliance on IOE is not just an evidentiary issue; it is a strategic liability. If an award is ultimately unenforceable, then the entire arbitration process is rendered futile.

SECTION VI: CONCLUSION

At its core, arbitration is a system built on trust: trust placed by the parties in the tribunal, and trust that fairness will prevail over procedural manoeuvring. But this trust is fragile and erodes each time misconduct is allowed to go unchecked. IOE presents an interesting challenge to the system’s integrity. If tribunals permit parties to rely on IOE, it sends a message that procedural fairness is negotiable, and that arbitration is not governed by legal principles, but by opportunism.

This is particularly apposite for a consent-based system like arbitration, which has been conceptualized with the sole purpose of resolving disputes in a way that parties have agreed upon, including an implied agreement that parties will conduct themselves in a bona fide manner.

In this regard, a parallel may be drawn to claims raised by parties which are tainted by deceit, fraud, jurisdictional manipulation, or fraudulent misrepresentation: in the past, tribunals themselves have refused to recognize such claims. This reinforces a simple truth: what is gained through bad faith cannot be legitimized by the arbitration process. If tribunals have barred entire claims on this basis, then the same logic must apply to evidence as well. To do otherwise would be to create an untenable contradiction; where a claim can be dismissed for deception, yet evidence obtained through deception is welcomed into the hearing room.

For the pragmatic, the practical consequences are just as troubling. Acceptance of IOE incentivizes unlawful conduct, transforming arbitration from a principled adjudicatory process premised on fairness into an opportunistic battle where the most unscrupulous party wins. If this is permitted, it may result in a situation where tribunals legitimize conduct that would perhaps be condemned by a national court, compromising arbitration’s longstanding reputation as a fair and neutral forum.

Ultimately, the question of excluding IOE is not just about which legal doctrine prevails, but also the foundational pillars we want arbitration as a dispute resolution mechanism to stand on. If arbitration is to remain a process built on trust, fairness, and procedural integrity, then inclusion of IOE is a compromise we cannot afford to make. In the past decades, dedicated efforts have been made worldwide to increase the trust of parties in arbitration as an effective alternative to a court-based adjudication process. In order to maintain this trust, which has been painstakingly built over the years, IOE must be excluded, not just as a matter of principle, but as a necessary safeguard against the erosion of arbitration itself.

Footnotes

1. 308 U.S. 338.
2. 90 F.2d 630 (2d Cir. 1937).
3. Supra n. 1, at 308.
4. 380 U.S. 693.
5. 1861, 8 Cox C.C.
6. Ibid., para. 501.
7. (2007) 4 SLR(R) 377.
8. Ibid., paras 27 and 40.
9. Chapter 35, s. 1, Rättegångsbalken (Swedish Code of Judicial Procedure): rätten skall efter samvetsgrann prövning av allt, som förekommit, avgöra, vad i målet är bevisat (“The court shall, after a conscientious examination of everything that has occurred, decide what is proven in the case.”).
10. Yves Derains, State Courts and Arbitrators, in Special Supplement 1999: Arbitration in the Next Decade: Proceedings of the International Court of Arbitration’s 75th Anniversary Conference, International Chamber of Commerce, Eric A. Schwartz, Francis P. Donovan and Fali S. Nariman.
11. Ibid.
12. 2003 INSC 528.
13. Section 22 read with s. 23C-G, International Arbitration Act 1974.
14. Catherine A. Rogers, Fit and Function in Legal Ethics: Developing a Code of Conduct for International Arbitration, 23 Mich. Int’l L.J. 341–423, at 342 (2002).
15. Commentary on the Revised Text of the 2020 IBA Rules on the Taking of Evidence in International Arbitration, International Bar Association (30 Jan. 2021), at 30, https://www.ibanet.org/MediaHandler?id=4F797338-693E-47C7-A92A-1509790ECC9D (accessed 30 Mar. 2025).
16. Ibid.
17. Stath v. Williams, Ind. App., 367 N.E.2d 1120, 1124.
18. Black’s Law Dictionary, 10th Edition.
19. Markus Kotzur, Good Faith (Bona Fides), Max Planck Encyclopedia Public International Law, Vol. IV, p. 514 (2012).
20. Legal Opinion by Christoph Schreuer, Ascom Group S.A., Anatolie Stati, Gabriel Stati and Terra Raf Trans Traiding Ltd. v. Republic of Kazakhstan (I), SCC Case No. 116/2010 (21 Jan. 2020).
21. 1937 PCIJ, Series A/B (No. 70) (Judgment of 28 Jun.) at 25.
22. Gabčikovo-Nagymaros Project (Hungary v. Slovakia), 1997 ICJ 7, 76, Judgment of 25 Sep., para. 133.
23. 263 NY 79 (NY 1933), para. 87.
24. 2014 SCC 71.
25. Nuclear Tests (Australia v. France), Judgment, ICJ Reports 197, para. 46.
26. Inceysa Vallisoletana, S.L. v. Republic of El Salvador, ICSID Case No. ARB/03/26, Award (2 Aug. 2006).
27. Plama Consortium Limited v. Republic of Bulgaria, ICSID Case No. ARB/03/24, Award (27 Aug. 2008).
28. Phoenix Action, Ltd. v. The Czech Republic, ICSID Case No. ARB/06/5, Award (15 Apr. 2009).
29. Sanum Investments Limited v. Lao People’s Democratic Republic (I), PCA Case No. 2013–13, Award (6 Aug. 2019), para. 171.
30. Libananco Holdings Co. Limited v. Republic of Turkey, ICSID Case No. ARB/06/8, Decision on Preliminary Issues (23 Jun. 2008).
31. Ibid., para. 80.
32. [2017] SGCA 23.
33. [2017] SGCA 23, para. 52.
34. Methanex Corporation v. United States of America, UNCITRAL, Final Award of the Tribunal on Jurisdiction and Merits (3 Aug. 2005), para. 56.
35. EDF (Services) Limited v. Romania, ICSID Case No. ARB/05/13, Award (8 Oct. 2009).
36. Caratube International Oil Company LLP and Devincci Salah Hourani v. Republic of Kazakhstan, ICSID Case No. ARB/13/13, Award of the Tribunal (27 Sep. 2017).
37. [2024] SGHC(I) 33.
38. [2021] SGCA 9.
39. C/13/712678/KG RK 22–65/MDvH/MV.
40. [1990] 1 Q.B. 335.
41. Ibid., para. 346.
42. [2001] 1 Lloyd’s Rep 715.
43. [2007] EWHC 571 (Comm).
44. [2005] APP.L.R. 02/23.
45. N. Singh, A Fourfold Test for Evaluating the Admissibility of Illegally Obtained Evidence in International Arbitration, Revista Romana de Arbitraj 85 (2022), Vol. 16, Issue 3.
46. 11001-02-03-000-2014-01927-00.
47. Report on the work of its eighteenth session, UNCITRAL (A/40/17) (1985).
48. 1956(1) QB 702.
49. 2023:DHC:1933-DB.
50. Julian Lew, The Law Applicable to the Form and Substance of the Arbitration Clause, in Improving the Efficiency of Arbitration Agreements and Awards: 40 Years of Application of the New York Convention, 114–145, Vol. 9, Albert Jan van den Berg ed., ICCA Congress Series, 1998 Paris, Kluwer Law International 1999.
51. [2023] EWHC 2638 (Comm).