Jump to Content Jump to Main Navigation
Corruption in International Investment Arbitration by Llamzon, Aloysius P (1st September 2014)

Part III Towards a Jurisprudence Constante in Investment Arbitration Decision-Making on Corruption, 9 Proving Corruption

From: Corruption in International Investment Arbitration

Aloysius P Llamzon

Corruption — Burden of proof — Standard of proof — International Court of Arbitration (ICC) — Law of treaties

(p. 225) Proving Corruption

A. Introduction: The Difficulty of Proving Corruption

9.01  When faced with allegations of corruption by parties, arbitrators are no longer concerned with questions of first principle—there is wide consensus that transnational corruption is unlawful as well as immoral, and should be sanctioned. The primary difficulty that most tribunals profess to encounter when faced with corruption allegations is how they are to be tested or proven. The most common modality in transnational investment—bribery—is perpetrated by parties who have power and resources, and collude precisely to avoid any trace of corrupt conduct.1 This raises hard questions about the proper evidentiary standard that must be applied; how parties are charged with burdens of proving corruption allegations; and under what circumstances, if any, such burdens of proof should shift. These evidentiary issues also lead to broader questions about the identity and role of arbitral tribunals as decision-makers of fact: how should arbitrators approach issues of bribery? Should they be more pro-active in trying to ferret out corruption? And if allegations are made, what are the evidentiary rules applicable, and do these rules implicitly aid or actually deter arbitrators from undertaking this sensitive task?

9.02  There are rare cases such as World Duty Free, Azpetrol, and Metal-Tech where the investor freely admitted, in varying degrees, to the facts that establish corruption during the proceedings. It is also likely that in the future, situations where investor corruption is admitted before national prosecutors and courts in their home States, such as in Siemens and Niko, will continue to grow, as capital-exporting States continue to take steps to police transnational corruption effectively. Those cases point to the potential impact national prosecutorial agencies have in affecting arbitral outcomes where corruption is at issue by aiding in the establishment of the underlying facts,2 a development that is to be encouraged, although not (p. 226) unqualifiedly.3 Ultimately, international tribunals must make independent assessments of key facts irrespective of the findings of national authorities.4

9.03  In the majority of cases where corruption may attend the investment, however, arbitrators will be faced with the question of how to ascertain whether or not corruption did occur on the basis of imperfect, episodic, and often contradictory evidence, with little aid from national authorities. The absence of an unbroken chain of facts leading to a morally certain conclusion that corruption did occur necessitates reliance on longstanding legal tools familiar to national and international courts and tribunals concerning rules of evidence on standards and burdens of proof, presumptions, and inferences that lead to a proper ‘connecting of the dots’.5 Rather than leaving parties to guess at the appropriate standards and burdens of proof to apply in cases of corruption and beyond, arbitrators ‘should give the parties some guidance, preferably prior to the taking of evidence, as to the burden of proof and the standard of proof, so that the parties can adapt themselves to those requirements and standards’6 thereby strengthening due process in arbitration.

B. Tribunals as Inquisitors? The Duty to Pursue Corruption Allegations

9.04  Summarizing the obligations of arbitrators faced with resolving issues concerning corruption, Bernardo Cremades stated: ‘[t]he position today is that the international arbitrator (p. 227) has a clear duty to address issues of bribery, money laundering or serious fraud whenever they arise in the arbitration and whatever the wishes of the parties and to record its legal and factual conclusions in its award. This is the only course available to protect the enforceability of the award and the integrity of the institution of international commercial arbitration.’7 While there has been some doubt about the precise scope and content of the duties of arbitral tribunals towards national authorities when dealing with corruption issues,8 there is no doubt that tribunals are obliged to the parties to address issues of corruption that are directly raised before them as part of their decision-making duty, particularly when corruption can potentially affect the outcome of the case.9 However, if tribunals are faced with only insinuations of corruption, or certain facts arise that lead to a suspicion of corruption, or tribunals suspect that a party is not forthcoming with all evidence that can be given to prove corruption, would tribunals nonetheless be under a duty to pursue and investigate?

9.05  For a number of reasons, it seems that tribunals are so obliged. As discussed previously, investment arbitrators bear obligations not only to the immediate parties to a dispute, but to the public of the host State and indeed the international community as a whole;10 it is no accident that public international law, including international anti-corruption law, forms part of the law that must be applied in investment arbitration. Thus, to the extent that the application of contemporary anti-corruption law in good faith requires the enforcement of those obligations, tribunals should deal with the issue propio motu as soon as certain facts emerge that raise concerns as to whether corruption attended the investment at some stage. Indeed, even apart from international anti-corruption norms per se, some commentators have argued that an arbitral tribunal has a duty to investigate a potential issue of corruption sua sponte, not least to ensure the enforceability of its award.11

(p. 228) 9.06  The practice of tribunals has been inconsistent, with some tribunals preferring to allow the adversarial process to take its course without intervening directly absent compelling indications.12 However, other cases exhibit more inquisitorial leanings. Metal-Tech v. Uzbekistan is instructive in this regard. The testimony of the claimant’s principal witness led to the admission of facts concerning the nature of the relationship between the investor and its contracted consultants, including the fact that the services provided were ‘lobbyist activity’ and not assistance with the operation, production, and delivery of the joint venture’s products as originally stated, and that US$4 million was paid).13 From these facts, the tribunal utilized its ex officio powers under Article 43 of the ICSID Convention, and by a series of procedural orders propio motu sought additional information from the parties. The tribunal subsequently called for additional testimony and evidence via another procedural order, and was able to obtain sufficient information to satisfy themselves of the existence of corruption in that case.14

9.07  There are practical limits of course to any tribunal’s power to investigate corruption—forcing discovery or testimony by compelling witnesses is difficult, as acknowledged in the case law and by scholars alike.15 However, Metal-Tech demonstrates that arbitral tribunals determined to ferret out the truth are not without tools to pursue their suspicions.

C. Burdens of Proof, Presumptions, Inferences, and ‘Red Flags’

9.08  The question of which party faces the burden of proving corruption has been relatively straightforward, despite some theoretical uncertainty as to the conflict of law rules that might apply.16 Following the maxim onus probandi incumbit actori (or actori incumbit probation), (p. 229) the prevailing principle is that each party has the burden of proving the facts on which it relies.17 When corruption is pled as a defence, the burden would fall on the host State; for investors seeking to prove corrupt solicitation or extortion by the host State’s public officials, the investor would bear the burden. The procedural rules applied by international courts and tribunals recognize this rule widely, and tribunals consistently confirm this in the case law.18

9.09  The more controversial question lies in what instances it would be appropriate, if any, to shift the burden of proof from the party asserting corruption to the party acting in defence. Constantine Partasides has noted that ‘a simple shifting of the burden of proof, all in one go, is rightly difficult for any lawyer to accept’.19 There are few endorsements of direct burden-shifting for corruption allegations in international arbitration20 and none in international investment arbitration.21 Burden-shifting appears to be more widely used in WTO adjudication, which allows for the burden of proof to be shifted once a prima facie violation has been established.22 However, the WTO’s practice is generally not found in other international courts and tribunals.23

(p. 230) 9.10  Is there something specific about the nature of corruption that justifies a re-allocation of the actori incumbit probation principle? Scholars, particularly those focused on improving the effectiveness of arbitration in combating corruption, have argued the necessity of burden-shifting.24 Institutions such as the World Bank employ burden-shifting as well in their corruption investigations, partly in appreciation of the fact that (as with arbitral tribunals) they have no subpoena or contempt powers.25 And as discussed in Chapter 7, the absence of any case where corruption was found to have existed (except for those few instances where admissions of the underlying facts supporting corruption were made by the investor’s primary witnesses) certainly buttresses the idea that the status quo is ineffective. The clandestine and highly complex nature of transnational corruption requires a candid admission that unless the evidentiary principles applied by the tribunal matches the ingenuity of those who are engaged in corruption, it will be difficult to find corruption in any arbitration.26 Indeed, there is some national legislation that shifts the burden on a party to explain why corruption should not be made out, once certain facts have been established; these methods have begun to take hold in the jurisprudence. In Metal-Tech, for example, the tribunal recognized the international community’s establishment of lists of indicators of corruption (often called ‘red flag’ principles) and considered the red flag principles drafted by a former Chief Justice of England and Wales, including an advisor’s lack of experience in the sector involved and any close personal relationship the advisor may have with the government that could improperly influence the latter’s decision.27

9.11  These ‘red flags’ and similar indicia of corruption can be conceived as potential forms of circumstantial evidence that, once established, can lead to a shifting of the burden of proof, requiring the rebuttal of allegations by evidence to the contrary, failing which certain inferences and conclusions might be drawn. Indeed, circumstantial evidence, particularly when direct evidence of corruption is unavailable, is widely, albeit cautiously, accepted as a tool to evaluate allegations of corruption by international tribunals.28 Methanex v. United States is instructive in this regard, as its endorsement of the ‘connect the dots’ methodology to reach a conclusion based on proven facts provides a detailed discussion of the applicability of circumstantial evidence in investment arbitration case law.29

(p. 231) 9.12  As for the use of adverse inferences, as early as 1939 Sandifer already noted that ‘the most effective sanction [international tribunals] have to impose upon parties negligent or recalcitrant in the production of evidence is the threat to draw an adverse inference against the party in the event of the continued refusal to bring forward the needed documents’.30 Similarly, the International Bar Association’s Rules on the Taking of Evidence in International Arbitration provide that ‘[i]f a Party fails without satisfactory explanation to produce any Document... ordered to be produced by the Arbitral Tribunal, the Arbitral Tribunal may infer that such document would be adverse to the interests of that Party’.31 The use of adverse inferences in appropriate situations is thus well accepted in principle.

9.13  While cases like Hamester v. Ghana emphasized that a tribunal would ‘only decide on substantiated facts, and cannot base itself on inferences’,32 more investment arbitration decisions concerning investor wrongdoing seem to draw conclusions on the basis of inference, either because a party did not produce evidence when asked to do so by the tribunal or because that party should have had within its possession exonerative evidence but did not produce it. The Metal-Tech tribunal considered that adverse inferences can indeed be drawn in appropriate instances to prove corruption: ‘the Tribunal may draw appropriate inferences from a party’s non-production of evidence ordered to be provided. In a number of cases, tribunals have indeed stated that they would draw inferences from non-production.’33 Indeed, Metal-Tech is a particularly instructive precedent for future tribunals, as the utilization of procedural orders to seek further information, coupled with the (sometimes subtle) reminder that failure to cooperate fully in satisfying the tribunal that all available evidence and testimony has been provided on a given corruption allegation may lead to adverse inferences, is not only sanctioned by international procedure,34 but can in the hands of a tribunal (p. 232) committed to unearthing the truth lead to strong findings of fact, no matter what standard of proof is advanced as relevant.35

9.14  In the final analysis, although presumptions and inferences can aid in establishing sufficient direct or circumstantial evidence that can discharge a party’s burden of proving its case or a particular defence asserted, questions on burdens of proof should not be confused with the establishment of sufficient proof. These are related but different questions, and in the area of investor wrongdoing, it is not so much the burden of proof as the standard of proof where much uncertainty remains.

D. Standards of Proof

9.15  The burden of proof identifies which party bears the obligation to prove a given allegation. The standard of proof defines the amount (or perhaps more accurately, the threshold) of evidence necessary to establish either an individual fact or contention or the party’s case as a whole. This distinction makes apparent, as well-expressed by the Rompetrol tribunal, ‘that the burden of proof is absolute, whereas the standard of proof is relative. [...] [I]f, according to basic principle, it is for the one party, or for the other, to establish a particular factual assertion, that will remain the position throughout the forensic process, starting from when the assertion is first put forward and all the way through to the end.’36 Standards of proof are ‘relative’ in the sense that ‘[w]hether a proposition has in fact been proved by the party which bears the burden of proving it depends not just on its own evidence but on the overall assessment of the accumulated evidence put forward by one or both parties, for the proposition or against it’.37 In essence, the burden of proof refers to who must prove a particular assertion, while the standard of proof refers to whether sufficient evidence has been provided to prove that assertion.

9.16  In national law, the applicable standard of proof varies between systems of common law and civil law. Broadly speaking, the general standard of proof for civil actions in common law is the ‘balance of probabilities’, while in civil law systems it is typically the ‘inner conviction’ of the adjudicator.38 International arbitration has often adopted the ‘balance of probabilities’ (p. 233) approach when setting a standard by which facts can be deemed proven.39 When serious allegations of wrongdoing are involved in civil proceedings, however, both systems generally demand a heightened standard of proof;40 the ‘clear and convincing’ proof standard, derived from U.S. law, is often employed.41 More recent cases in English law have cast doubt on whether it is proper to link the gravity of the misconduct alleged with the standard of proof that should apply; in Re B (Children), the House of Lords came to the conclusion that ‘the time has come to say, once and for all, that there is only one civil standard of proof and that is proof that the fact in issue more probably occurred than not’.42 Both the heightened standard and the ordinary standard have found their way into investment arbitration case law on investor wrongdoing,43 a reflection of the different legal traditions that inform arbitrators of varied backgrounds to issues of proof.

9.17  International commercial arbitration tribunals have, by and large, adopted high standards of proof: in one of the rare publicly-known ICC arbitrations that dealt with corruption issues concerning the validity of the main agreement (not merely a supporting consultancy arrangement as found for example in Judge Lagergren’s leading ICC Case No. 1110), Westinghouse v. National Power Corporation, Republic of the Philippines,44 the tribunal viewed bribery as a species of fraud in civil cases, thus demanding ‘clear and convincing evidence’ under the applicable U.S. and Philippine law:45 ‘fraud in civil cases must be proved to exist by clear and (p. 234) convincing evidence amounting to more than mere preponderance, and cannot be justified by a mere speculation. This is because fraud is never lightly to be presumed...’.46 Other international commercial arbitration decisions have adopted the same ‘clear and convincing’ standard of proof, or similarly high standards.47

9.18  Similarly, as seen from the trend analysis in Chapter 7, investment tribunals which are not similarly constrained by national standards of proof have largely adopted high standards of proof. In EDF v. Romania, for example, the tribunal placed, without any real attempt at harmonization, the fact that corruption is ‘notoriously difficult to prove’ against the ‘demand’ for clear and convincing evidence:

corruption must be proven and is notoriously difficult to prove since, typically, there is little or no physical evidence. The seriousness of the accusation of corruption in the present case, considering that it involves officials at the highest level of the Romanian Government at the time, demands clear and convincing evidence. There is general consensus among international tribunals and commentators regarding the need for a high standard of proof of corruption.48

9.19  This is not to say, however, that arbitrators have never considered less rigid evidentiary methodologies that better comport with the clandestine nature of transnational corruption. Siag v. Egypt bears discussion in this regard, as it places in clear relief how the choice of the applicable standard of proof directly affects whether a set of facts is considered to have crystallized into corruption. In that case, a divided tribunal equated the corrupt act complained of by Egypt to fraud, which in turn was to be proven through the ‘heavy’ standard of ‘clear and convincing evidence’, located somewhere between the ‘preponderance of evidence’/‘balance of probabilities’ standard for civil cases and the ‘beyond reasonable doubt’ standard in criminal cases under American law.49 However, the dissenting arbitrator, Professor Francisco Orrego Vicuña, believed that there was sufficient evidence to establish that corruption did occur in the procurement of the investor’s Lebanese citizenship, which would bear the legal repercussion that the investor was in actuality an Egyptian citizen and thus not permitted ratione personae to bring a BIT claim against Egypt. Professor Orrego began his analysis by emphasizing that his divergence from the majority did not in any way mean that the majority was condoning corruption.50 It was not the principles involved that were the source of difference; rather, it was a ‘different assessment of the evidence and whether it is sufficient to establish such impropriety’.51 The standard of proof the majority chose—that of clear and convincing evidence—was for Professor Orrego not appropriate. ‘It is my view that (p. 235) arbitration tribunals, particularly those deciding under international law, are free to choose the most relevant rules in accordance with the circumstances of the case and the nature of the facts involved, as it has been increasingly recognized.’52 In this case, the facts could be best judged under a standard of proof allowing the tribunal ‘discretion in inferring from a collection of concordant circumstantial evidence (faisceau d’indices) the facts at which the various indices are directed’.53

9.20  Professor Orrego’s view that arbitrators ‘are free to choose the most relevant rules’ based on the ‘circumstances of the case and the nature of the facts involved’ has been affirmed by the most recent cases in investment arbitration dealing with corruption and other forms of fraud and illegality. Metal-Tech is the most recent statement that evidentiary questions on corruption, including standards of proof, are based on international law due to the BIT being the lex causae and are therefore open questions, tribunals having ‘relative freedom in determining the standard necessary to sustain a determination of corruption’.54 This is so because international investment arbitration, and indeed the broader system of international dispute settlement, is characterized largely by principle rather than procedural formality, and the rules of evidence are neither rigid nor technical.55

9.21  Notwithstanding this freedom of action, however, the case law can still yield useful doctrine. The guidance drawn by the Rompetrol tribunal from arbitral practice on evidence is highly instructive: the tribunal considered ‘that there may well be situations in which, given the nature of an allegation of wrongful (in the widest sense) conduct, and in the light of the position of the person concerned, an adjudicator would be reluctant to find the allegation proved in the absence of a sufficient weight of positive evidence—as opposed to pure probabilities or circumstantial inferences. But the particular circumstances would be determinative, and in the Tribunal’s view defy codification.’ The Rompetrol tribunal then referred to Libananco v. Turkey:

In relation to the Claimant’s contention that there should be a heightened standard of proof for allegations of ‘fraud or other serious wrongdoing,’ the Tribunal accepts that fraud is a serious allegation, but it does not consider that this (without more) requires it to apply a heightened standard of proof. While agreeing with the general proposition that—the graver the charge, the more confidence there must be in the evidence relied on..., this does not necessarily entail a higher standard of proof. It may simply require more persuasive evidence, in the case of a fact that is inherently improbable, in order for the Tribunal to be satisfied that the burden of proof has been discharged.

9.22  While it would apply the ‘normal’ balance of probabilities standard to the factual issues before it, the Rompetrol tribunal held that it ‘will where necessary adopt a more nuanced approach and will decide in each discrete instance whether an allegation of seriously wrongful conduct (p. 236) by a [host] state official at either the administrative or policymaking level has been proved on the basis of the entire body of direct and indirect evidence before it’.56 Similarly, commentators have concluded, based on surveys of commercial arbitral awards involving bribery and corruption charges, that arbitrators frequently rely on indirect evidence of corruption in those cases where credible allegations of corruption have been made.57

9.23  Thus, the current state of evidentiary principles on corruption seems to be at the beginning of a move away from the uniformity and rigidity of high standards of proof, with tribunals refusing to be pinned down a priori either by particular standards or by formal rules on burden-shifting or presumptions. Given the inability of the prevailing doctrine to generate positive findings of corruption despite its anecdotal frequency and repeated invocation, this is surely a positive development. Without such room for the exercise of discretion, corruption would likely continue to reside in the margins of arbitral decision-making, and would have no practical impact on the outcomes of a vast majority of cases. That unsatisfactory status quowhere not a single investment arbitration case has resulted in a finding of corruption on the basis of contested evidence—would expose international arbitration to further criticism for being a ‘soft touch’ on corruption.58

9.24  Why arbitrators have hesitated for so long to adopt more flexible evidentiary processes that are equal to the ingenuity of those that conceal corruption59 is open to debate. The reason for choosing a high standard of proof seems to make good sense: higher standards of proof have been subscribed to by tribunals because of the serious legal consequences a finding of corruption would have.60 Reading the cases closely, it seems that adopting high standards of proof was motivated in part by the need to ensure that the serious consequences most commonly associated with corruption in international arbitration—contract invalidation, the unenforceability of the contract, the lack of jurisdiction, all acting to preclude any assessment of host State wrongdoing—would apply sparingly. And indeed, adopting a high standard of proof has the practical value of helping flush out the increasingly tactical or cynical use of corruption to derail the arbitral process.

(p. 237) 9.25  Thus, it would be too reductive to view the frequent employment of the ‘clear and convincing proof’ standard as done only because formal law demands that outcome: rather than simply identifying a standard they are pre-ordained to take, arbitrators are in reality choosing one standard over another in a situation where either (or all) standards proffered are squarely within that arbitrator’s scope of choice. In so doing, evidentiary principles, particularly the applicable standard of proof, are used as valves from which the pressure of applying anti-corruption norms can be regulated. Arbitrators choose higher or lower standards of proof based at least in part on their appreciation of the equities of the case and the role they believe corruption should play in the ultimate outcome of the arbitration.

9.26  With a nod to this reality, the less formalistic sensibility of Rompetrol and Metal-Tech towards the evidentiary rules to be applied to corruption issues is helpful. Because corruption is a serious charge with serious consequences attached, the degree of confidence a tribunal should have in the evidence of that corruption must be high. However, this does not mean that the standard of proof itself should necessarily be higher, or that circumstantial evidence, inferences, or presumptions and indicators of possible corruption (such as ‘red flags’) cannot come to the aid of the fact-finder. Tribunals are given the freedom and burden of choice, which they should not abdicate by rote reference to an abstract ‘heightened’ standard of proof.


1  In a similar fashion, a commentator has noted that ‘like most crimes and intentional misconduct, and perhaps more so, acts of corruption and collusion are specifically designed not to be able to be identified or detected.’ Karen Mills, Corruption and Other Illegality in the Formation and Performance of Contracts and in the Conduct of Arbitrations Relating Thereto, ICCA Congress Series No. 11 (A.J. van den Berg (ed.), Kluwer, 2003), 295.

2  In Niko, the tribunal acknowledged that national authorities were best placed to gather evidence and establish the necessary proof of corruption. See Niko Resources (Bangladesh) Ltd v. People’s Republic of Bangladesh, BAPEX, and PETROBANGLA, ICSID Case Nos. ARB/10/11 amd ARB 10/18, Decision on Jurisdiction dated 19 August 2013. para. 425 (‘In the present case, the acts of corruption of which the Claimant was convicted were committed in Bangladesh. If there were any other such acts committed they must have concerned persons making decisions in Bangladesh. Therefore, the authorities of Bangladesh were best placed to investigate and collect proof of corruption relevant for the present case.’) Please refer to Chapter 6 for full case references of the cases referred to in this chapter.

3  In Niko, the tribunal acknowledged that national authorities were better placed to gather evidence and establish the necessary proof of corruption. See Niko (n 2), para. 425 (‘In the present case, the acts of corruption of which the Claimant was convicted were committed in Bangladesh. If there were any other such acts committed they must have concerned persons making decisions in Bangladesh. Therefore, the authorities of Bangladesh were best placed to investigate and collect proof of corruption relevant for the present case.’)

There is always a risk, however, that domestic investigations are used either for political ends or even tactically to buttress the strength of a host State’s position in arbitration. In EDF (Services) v. Romania, the investor criticized the domestic investigation and exoneration of public officials; this stands in contrast to Fraport AG v. Philippines, where national prosecutorial authorities’ findings were that there was no violation of national ‘Anti-Dummy Law’ legislation, which ran against the host State’s position in the arbitration.

4  In Inceysa, the tribunal held that State Parties’ determinations as to the legality or illegality of investment are not determinative of the issue for the purposes of establishing the tribunal’s jurisdiction under a BIT (Inceysa Vallisoletana v. El Salvador, paras. 209–213). It is thus for the tribunal to determine if actions are fraudulent and domestic court findings regarding such actions are not res judicata for the purposes of determining jurisdiction under a BIT.

5  See Thunderbird v. Mexico, Separate Opinion of Thomas Wälde, para. 3 (citations omitted):

Since the arbitration has, as often or always, not elucidated all relevant facts, one needs to rely on standard practice of rules of evidence, burden of proof and presumptions to determine when the claimant and when the respondent has to bear the respective burden. In addition, as recently explained again by the Methanex v US award, what is unknown but relevant has to be dealt with by inference, i.e. by taking the ‘dots’ that are available, drawing explanatory lines between them (and then determine what explanation can be inferred by relying on burden of proof allocation, prima facie evidence and arbitral determination of the evidence). They need to be assessed not only from the lofty spheres of commercial arbitration law, but also with a real-life understanding of the ‘coal-face’ of foreign investment practices.

6  Jose Rosell and Harvey Prager, ‘Illicit Commissions and International Arbitration: The Question of Proof’ (1999) 15 Arb. Int’l 329, 348.

7  Bernardo Cremades and D.J.A. Cairns, ‘Trans-National Public Policy in International Arbitral Decision-Making: the Cases of Bribery, Money-Laundering and Fraud’, in Arbitration—Money Laundering, Corruption and Fraud (K. Karsten and A. Berkeley (eds.), ICC Publishing, 2003), 65, 85.

8  As recently as 2003, there seemed to be little by way of mandatory requirements of disclosure to national authorities. Cremades and Cairnes were ‘aware of no express legislation requiring a tribunal to disclose to regulatory authorities that criminal activity might have occurred in connection with a transaction before it’.Bernardo Cremades and D.J.A. Cairns, ‘Corruption, International Public Policy and the Duties of Arbitrators’ (2003) 58 Dispute Resolution Journal 83. More recently, the rules of some arbitration institutions specifically allow disclosure made in accordance with national law requirements. See Michael Hwang and Kevin Lim, ‘Corruption in Arbitration—Law and Reality’ (2012) 8 Asian Int’l Arb. J. 14–22, 48(referring to the 2010 SIAC Rules which provide that parties or arbitrators may disclose matters relating to an arbitration ‘in compliance with the provisions of the laws of any State which are binding on the party making the disclosure’).

9  That said, tribunals faced with direct corruption allegations have often sidestepped corruption allegations by host States when other issues have yielded the same outcome, e.g. a finding of no jurisdiction. The limited extent to which tribunals directly address corruption in fact has been discussed at length in Chapters 7 and 8.

10  The contrast with commercial arbitration, with its emphasis on party autonomy, in the treatment of corruption bears mention. In Dr Sayed’s extensive study, he found that ‘[t]he drive towards greater autonomy renders contracts of international trade apt to accommodate any content, including relations of corruption. Autonomy creates a space for the human will to express itself without restriction. Merchants may therefore agree on corruption, and the contract structures the method of its realization. The parties to such contracts rely on a certain inner solidarity as well as reciprocity and correlation of obligations, to preserve their respective rights. They also could perceive arbitration as a means to avoid that their contracts succumb to the grips of an official world where they would be perceived with disapproving eyes out of missionary moralism or hypocrisy.’ Abdulhay Sayed, Corruption in International Trade and Commercial Arbitration (Kluwer, 2004), 8, citing inter alia Yves Dezalay and Bryant Garth, Dealing in Virtue (University of Chicago Press, 1996), 245–6.

11  See Michael Hwang and Kevin Lim, ‘Corruption in Arbitration—Law and Reality’ (2012) 8 Asian Int’l Arb. J. 14–22.

12  For example, in Micula v. Romania, ICSID Case No. ARB/05/20, Decision on Jurisdiction and Admissibility, 24 September 2008, a tribunal faced an allegation that Mr Viorel Micula had obtained his Swedish nationality by fraud or material mistake. The tribunal noted:

Respondent has presented only limited evidence, none of which is sufficient to make the necessary showing’; it is also observed ‘The record does not include any elements which should lead the Tribunal to investigate facts that are not before it. Nor do Respondent’s allegations of facts lead to the need for opening a fact-finding procedure. Given the factual evidence presented by Respondent, the Tribunal, in its letter of 11 July 2008, directed Respondent how to proceed in the event that Respondent believed that it needed additional supplemental documentary production—an option that Respondent chose not to pursue. (see para. 95)

13  Metal-Tech Ltd v. Republic of Uzbekistan, ICSID Case No. ARB/10/3, Award of 4 October 2013, paras. 86, 274.

14  Metal-Tech (n 13), paras. 247–256.

15  Commenting on its treatment of the evidence on corruption provided, the F-W Oil v. Trinidad and Tobago tribunal recounted how F-W withdrew its corruption allegations at the very end of the hearing ‘under pointed questioning from the Tribunal itself as to whether there was any real evidence to sustain allegations of that breadth and gravity’. The tribunal was ‘naturally much concerned from the outset’ about the corruption allegations, not only because of their serious nature, but also ‘because it was faced with the problems inherent in investment arbitrations (by contrast with proceedings in a court of law): no evidence on oath, and no compellability of witnesses’. F-W Oil Interests, Inc. v. The Republic of Trinidad and Tobago, ICSID Case No. ARB/01/14, Award dated 3 March 2006, para. 211. See discussion in Chapter 6(B)(7). See also Matthias Scherer, ‘Circumstantial Evidence in Corruption Cases Before International Arbitral Tribunals’ (2002) 5(2) Int’l Arb. L. Rev. 29(‘[e]veryone familiar with international commercial arbitration concedes that the tools available for pre-hearing discovery are sharply limited’).

16  See Gary Born, II International Commercial Arbitration (Kluwer, 2009), 1858 (‘Allocating the burden of proof arguably presents choice-of-law questions. In particular, tribunals must decide whether to apply the law of the arbitral seat (on the theory that the burden of proof is “procedural”), the law governing the underlying substantive issues, or some international standard. [...] The better view is that the tribunals should allocate the burden of proof in light of its assessment of the applicable substantive law and procedures adopted in the arbitration. In so doing, the tribunal need not apply the burden of proof rules of any specific jurisdiction, but can instead fashion specialized rules in light of the particular substantive issues and procedures at issue in a specific instance’). An investment tribunal’s freedom to adopt evidentiary rules it deems more appropriate is greater given its international law remit.

17  Durward V. Sandifer, Evidence before International Tribunals (University Press of Virginia, 1975), 127 (‘[t]he burden of proof rests upon him who asserts the affirmative of a proposition that if not substantiated will result in a decision adverse to his contention’).

18  See e.g. Article 27(1), UNCITRAL Arbitration Rules; Case Concerning Military and Paramilitary Activities in and Against Nicaragua [1984] ICJ Rep. 392, 437 (‘it is the litigant seeking to establish a fact who bears the burden of proving it’); Metal-Tech (n 13), para. 237 (‘The principle that each party has the burden of proving the facts on which it relies is widely recognised and applied by international courts and tribunals.’)

19  Constantine Partasides, ‘Proving Corruption in International Arbitration: A Balanced Standard for the Real World’ (2010) 25(1) ICSID Review 47, 53.

20  In international commercial arbitration, ICC Case No. 6497 is one of the few examples where burden-shifting was endorsed by a tribunal. Even here, however, the tribunal cautioned that ‘such change to the burden of proof is only to be made in special circumstances and for very good reasons’.

21  Suggestions that the standard of proof should be reversed have been rejected as inconsistent with due process: see e.g.M. Hwang and K. Lim, ‘Corruption in Arbitration—Law and Reality’ (2012) 8 Asian Int’l Arb. J. 14;A. Mourre, ‘Arbitration and Criminal Law: Reflections on the Duties of the Arbitrator’ (2006) 22(1) Arb. Int’l 95, 102(arguing that a reversal of the burden of proof ‘does not seem to be acceptable or compatible with the right to a fair trial’). However, some commentators have advocated shifting the burden of proof to the allegedly corrupt party upon a prima facie showing of corruption: see K. Mills, ‘Corruption and Other Illegality in the Formation and Performance of Contracts and in the Conduct of Arbitration Relating Thereto’, in ICCA Congress Series No. 11—International Commercial Arbitration: Important Contemporary Questions (A. van den Berg (ed.), Kluwer, 2003), 288, 295 (‘Because of the near impossibility to “prove” corruption, where there is a reasonable indication of corruption, an appropriate way to make a determination may be to shift the burden of proof to the allegedly corrupt party to establish that the legal and good faith requirements were in fact duly met’);C. Lamm et al., ‘Fraud and Corruption in International Arbitration’ in Liber Amicorum: Bernardo Cremades (M.A. Fernandez-Ballesteros and D. Arias (eds.), Aspen, 2010), 699, 701(citing Mills, with tacit approval, and stating: ‘Another argument in favour of burden shifting is that the party accused of corruption is typically easily capable, if it is actually innocent of the allegations, of producing countervailing evidence’).

22  In EC—Hormones, for example, the WTO Appellate Body explained that ‘the initial burden lies on the complaining party, which must establish a prima facie case of inconsistency with a particular provision of the SPS Agreement on the part of the defending party, or more precisely, of its SPS measure or measures complained about. When that prima facie case is made, the burden of proof moves to the defending party, which must in turn counter or refute the claimed inconsistency.’

23  The prima facie rule has been applied consistently by WTO panels and the Appellate Body but is ‘at variance with the manner in which most other international courts [and tribunals] deal with the allocation of the burden of proof’. Chester Brown, A Common Law of International Adjudication (Oxford University Press, 2007), 93.

24  See e.g. Cecily Rose, ‘Questioning the Role of International Arbitration in the Fight Against Corruption’ (2014) 31 J. Int’l Arb. 183 (advocating the greater utilization of a number of solutions to perennial evidentiary problems in corruption; apart from identifying appropriate instances where the burden of proof should be shifted, drawing adverse inferences, placing greater reliance on circumstantial evidence, lowering the standard of proof, and drawing on factual findings in domestic proceedings are suggested).

25  See World Bank, Guidelines on Preventing and Combating Fraud and Corruption in Projects Financed by IBRD Loans and IDA Credits and Grants, 15 October 2006, <http://siteresources.worldbank.org/INTOFFEVASUS/Resources/WB_Anti_Corruption_Guidelines_10_2006.pdf>.

26  See Scherer, Circumstantial Evidence (n 15) (‘it is very rare that direct proof of corruption is available [...] [m]ost arbitral tribunals have to content themselves with circumstantial evidence’).

27  Metal-Tech (n 13), para. 293, citing Woolf Committee Report, Business Ethics, Global Companies and the Defence Industry: Ethical Business Conduct in BAE Systems Plc (2008), 25–6 and fn. 340 (‘...the red flag lists merely assemble a number of factors which any adjudicator with good common sense would consider when assessing facts in relation with a corruption issue...’). While Metal-Tech ultimately made no findings on burdens and standards of proof, the acknowledgment of ‘red flags’ is significant and opens the possibility of future tribunals adopting these principles when clear admissions from witnesses are not available to the tribunal.

28  See ICC Case No. 4145 (a fact can be proven by circumstantial evidence if its leads to ‘very high probability’); ICC Case No. 8891 (‘indices must be serious’).

29  In Methanex Corporation v. United States of America, NAFTA/UNCITRAL Arbitration Rules, Final Award on Jurisdiction and Merits dated 3 August 2005, the corruption allegation came in the form of an allegation by the investor that the decision of the Governor of the State of California to impose a ban that severely impacted the investment was secured by a competitor through corruption, i.e. through an attempt to ‘improperly influence the State of California’ by providing US$260,000 in political campaign contributions to Governor Davis. See Methanex, Part I, para. 5.

In a detailed analysis of the investor’s corruption-related allegations (see Methanex, Part III, Ch. B), the tribunal’s approach towards evaluating the evidence of corruption was particularly noteworthy. It adopted a ‘connect the dots’ methodology (Methanex, Part III, Ch. B, para. 2. (‘while individual pieces of evidence when viewed in isolation may appear to have no significance, when seen together, they provide the most compelling of possible explanations of events...’)), with the caveat that each ‘dot’ (i.e. key event) be examined in its own context as well as within the overall pattern, and that the tribunal must ultimately determine what the ‘dots’ are from the evidence presented by both parties:

Connecting the dots is hardly a unique methodology; but when it is applied, it is critical, first, that all the relevant dots be assembled; and, second, that each be examined, in its own context, for its own significance, before a possible pattern is essayed. Plainly, a self-serving selection of events and a self-serving interpretation of each of those selected, may produce an account approximating verisimilitude, but it will not reflect what actually happened. Accordingly, the tribunal will consider the various ‘dots’ which Methanex has adduced—one-by-one and then together with certain key events (essentially additional, noteworthy dots) which Methanex does not adduce—in order to reach a conclusion about the factual assertions which Methanex has made.

30  Sandifer, Evidence before International Tribunals (n 17), 101.

31  IBA Rule 9(5)–(6).

32  Hamester v. Ghana, para. 134, cited in Niko Resources v. Bangladesh (n 2), para. 424.

33  Metal-Tech (n 13), para. 245. See also: in commercial arbitration, ICC Case No. 5622 (‘It is true that it is possible to prove something through indirect evidence and that Art. 8 of the Swiss CC (3) does not exclude indirect evidence’); in inter-State adjudication, the ICJ in the Corfu Channel Case (UK v. Albania), Merits, Judgment of 9 April 1949, 1949 ICJ Rep. 4, 18 (States ‘should be allowed a more liberal recourse to inferences of fact and circumstantial evidence’).

34  See Articles 43–45, ICSID Convention; Rule 34, ICSID Arbitration Rules.

35  In this regard, see also Rompetrol v. Romania, ICSID Case No. ARB/06/3, Award of 6 May 2013, para. 181 (‘an ICSID tribunal is endowed with the independent power to determine, within the context provided by the circumstances of the dispute before it, whether particular evidence or kinds of evidence should be admitted or excluded, what weight (if any) should be given to particular items of evidence so admitted, whether it would like to see further evidence of any particular kind on any issue arising in the case, and so on and so forth. The tribunal is entitled to the cooperation of the parties in that regard, and is likewise entitled to take account of the quality of their cooperation. When paragraph (2) of Rule 34 lays down that “[t]he tribunal shall take formal note of the failure of a party to comply with its obligations under [that] paragraph and of any reasons given for such failure,” it no doubt intends, among other things, that a given tribunal is specifically authorized to draw whatever inferences it deems appropriate from the failure of either party to produce evidence which that party might otherwise have been expected to produce.’)

36  Rompetrol (n 35), para. 178.

37  Rompetrol (n 35), para. 178.

38  See Andreas Reiner, ‘Burden and General Standards of Proof’, (1994) 10(3) Arb. Int’l 325, 335. It might be assumed that it was to this ‘inner conviction’ standard referred to in a number of ICC arbitrations concerning corruption. See that ICC Case No. 7047 (a tribunal ‘must be convinced that there is indeed a case of bribery’); ICC Case No. 4145 (‘the Defendant’s accusation is not supported by direct evidence or even circumstantial evidence to be retained as convincing’).

39  Nigel Blackaby et al, Redfern and Hunter on International Arbitration (Oxford University Press, 2009), 387 (‘[t]he degree of proof that must be achieved in practice before an international arbitral tribunal is not capable of precise definition, but it may be safely assumed that it is close to the “balance of probability”’). The balance of probabilities standard is ‘a reasonable degree of probability, but not so high as is required in a criminal case. If the evidence is such that the tribunal can say: “We think it more probable than not,” the burden is discharged but, if the probabilities are equal, it is not.’ See at 387, citing Miller v. Minister of Pensions (1947) 2 All ER 372 (Denning J).

40  Reiner, Burden (n 38), 336 (‘the Anglo-Saxon and the continental systems require higher standards of proof for particularly important or delicate questions such as bribery or other types of fraud’).

41  Professor Wigmore provides an authoritative restatement of the standard: ‘[i]n several civil actions a different phrase has been adopted to signify a stronger persuasion [...] the phrase, viz., “clear and convincing proof” [...] [s]uch a test is in many or most States used for an issue involving fraud.’ John Henry Wigmore, A Students’ Textbook of the Law of Evidence (Foundation Press, 1935), 446.

42  In re B (Children)(FC), UKHL 35 (2008), para. 13. That said, ‘[t]here are some proceedings, though civil in form, whose nature is such that it is appropriate to apply the criminal standard of proof’, such as when the consequences of a finding are penal in nature. See para. 69. The need for a higher degree of probability that serious allegations are indeed true may yet be valid: as earlier stated by Lord Denning, ‘in civil cases the case must be proved by a preponderance of probability, but there may be degrees of probability within that standard. The degree depends on the subject-matter. A civil court, when considering a charge of fraud, will naturally require for itself a higher degree of probability than that which it would require when asking if negligence is established’: Bater v. Bater (1951) (Denning LJ).

43  Compare Siag v. Egypt, paras. 325–326 (‘It is common in most legal systems for serious allegations such as fraud to be held to a high standard of proof. The same can be said in international proceedings... The Tribunal accepts that the applicable standard of proof is greater than the balance of probabilities but less than beyond reasonable doubt. The term favoured by the Claimants is “clear and convincing evidence.” The Tribunal agrees with that test.’) with Rompetrol v. Romania (n 35), paras. 180–183 (applying the ‘normal rule of “balance of probabilities” as the standard appropriate to the generality of the factual issues before it’, while also ‘where necessary adopt[ing] a more nuanced approach’ when deciding ‘whether an allegation of seriously wrongful conduct...has been proved on the basis of the entire body of direct and indirect evidence before it.’)

44  Westinghouse Projects Company, et al. v. National Power Corporation, The Republic of the Philippines, Preliminary Award, 19 December 1991, reprinted in 7 Mealey’s Int’l Arb. Rep., No. 1, Sec. B (January 1992).

45  The tribunal determined the standard of proof by electing to apply the standards found in both U.S. and Philippine law; these national laws were found to be applicable because of the identities of the parties and the link between the facts to be determined and the rules being applied. Westinghouse Preliminary Award (n 44), 31.

46  Westinghouse Preliminary Award (n 44), 34.

47  As with Westinghouse, the Himpurna tribunal set the standard of proof at ‘clear and convincing proof.’ See Himpurna California Energy Ltd. (Bermuda) v. P.T. (Persero) Perusahaan Listruk Negara (Indonesia), Final Award of 4 May 1999, (2000) 14 Mealy’s Int’l Arb. Rep. 14, 43. Similarly, in ICC Case No. 6401, the tribunal stated: ‘bribery, a form of fraud, must be proved by “clear and convincing evidence”.’ ICC Case No. 5622 went even closer to an overtly criminal standard of proof: ‘bribery has not been proved beyond doubt’.

Indeed, some cases seem to adopt even stricter standards, employing language more commonly associated with criminal proceedings—see ICC Case No. 5622 (Hilmarton) (‘In the present case, bribery has not been proved beyond doubt’).

48  EDF v. Romania (Chapter 6, section B(14)), para. 221.

49  Siag (Chapter 6, section B(13)), para. 325.

50  Siag, Dissenting Opinion of Professor Francisco Orrego Vicuña (n 42), p. 1 (the difference in views ‘in no way reflects adversely on any of the distinguished counsel that have represented the Claimants in these proceedings, nor on my colleagues in the Tribunal, as none would be willing to condone [fraud or corruption]’).

51  Siag, Dissenting Opinion Orrego (n 50) p. 1.

52  Siag, Dissenting Opinion Orrego (n 50), p. 1. at 4, citing Abdulhay Sayed, Corruption in International Trade and Commercial Arbitration (Kluwer, 2004), 89–92.

53  Siag, Dissenting Opinion Orrego (n 50), p. 1. at 4, citing Sayed, Corruption in International Trade (n 52), 89–92.

54  Metal-Tech (n 13), para. 238. In a mild conflation of different evidentiary concepts, the tribunal stated: ‘Here, the question is whether for allegations of corruption, the burden should be shifted to the Claimant to establish that there was no corruption. Rules establishing presumptions or shifting the burden of proof under certain circumstances, or drawing inferences from a lack of proof are generally deemed to be part of the lex causae. In the present case, the lex causae is essentially the BIT, which provides no rules for shifting the burden of proof or establishing presumptions. Therefore, the Tribunal has relative freedom in determining the standard necessary to sustain a determination of corruption.’

55  See Rompetrol (n 35), paras. 178, 181.

56  Rompetrol (n 35), para. 183.

57  C. Lammet et al., ‘Fraud and Corruption in International Arbitration’ in M.A. Fernandez-Ballesteros and D. Arias, Liber Amicorum: Bernardo Cremades (Wolters Kluwer, 2010), 699, 703 (citing A. Crivellaro, ‘Arbitration Case Law on Bribery: Issues of arbitrability, contract validity, merits and evidence’ in K. Karsten and A. Berkeley (eds.), Arbitration: Money Laundering, Corruption and Fraud (ICC, 2003), 109).

58  See Former Kenyan Attorney-General Amos Wako, quoted in Alison Ross, ‘The Man Behind Kenyan Arbitration’, Global Arbitration Review, 20 January 2012 (‘Unfortunately, in the case of international contracts it can be hard to prove corruption and to point the finger at a particular player. We must ensure that international arbitration tribunals do not develop an unacceptable reputation for being “a soft touch” on corruption and other forms of illegality. Arbitrators need to emerge from their ivory towers, recognise how difficult an allegation of corruption can be to substantiate and show procedural flexibility to take these difficulties in proving corruption into account.’)

59  See Partasides, Proving Corruption (n 19), 51, 56. (‘How can our process ensure that it is equal to [the ingenuity of those that conceal corruption]?’ To which he answers: ‘Our process will not ensure that it is equal to the ingenuity of corruptors, and/or those that they corrupt, if arbitrators simply fail to take account of that ingenuity, and where necessary adapt for it, in the conduct of their proceedings.’)

60  Indeed, a number of tribunals in fraud and corruption have adopted Judge Higgins’ view of there seeming to be ‘general agreement’ that ‘the graver the charge the more confidence there must be in the evidence relied on’. Oil Platforms (Islamic Republic of Iran v. United States of America) (2003) 42 I.L.M. 1334,1384–6 (Higgins J, Separate Opinion), cited with approval in Rompetrol (n 35), para. 182 and Metal-Tech (n 13).