Introduction

This contribution offers some thoughts, illustrated by a zig-zagging jurisprudence, on the question whether a host state can validly raise the defence of investor corruption to avoid liability for breach of investment protection standards, in circumstances where the state participated in or condoned an investor's corrupt actions, for example by soliciting and/or receiving bribes from the investor, or refusing to take action against the corrupt investor and complicit state officials.

The following issues are examined:

a) (Non-)Attribution to the state of the corrupt behaviour on the part of a state official: World Duty Free v Kenya (2006);

b) Causal link between bribe and benefit: Niko v Bangladesh (2013);

c) Use by the tribunal of ex officio investigative powers: Metal-Tech v Uzbekistan (2013).

1. Attribution - World Duty Free (WDF) v Kenya

It is worth noting that WDF v Kenya,1consistently hailed as poster child for the treatment of corruption in investment treaty arbitration, was neither a typical case on its facts, nor was it a treaty case.

This was a case conducted under the ICSID Rules pursuant to Article 9 of the contract between the claimant and the Republic of Kenya for the operation of duty-free complexes at the Nairobi and Mombassa international airports, stipulating that the tribunal was to apply English law. Article 10(A) provided that the Contract was governed by and construed in accordance with Kenyan law. The tribunal found that English law and Kenyan law presented no material differences on the points at issue. Therefore - despite lengthy digressions on international public policy in the Award - this is a domestic law case of contractual liability.

At the hearing, the claimant's Mr Ali took the stand to reveal that, in order to arrange the necessary authorisations for the implementation of the agreement, he had been required to make, and did make, a 'personal donation' of US$2 million the then-president, Daniel arap Moi. Kenya sought an order that the agreement was unenforceable because it had been procured by bribery. The tribunal agreed.

The WDF tribunal gave scant consideration throughout to the criminality of the actions of the taker of the bribe,2the President of Kenya, and how the conduct of the bribe-taker impacted, if at all, on the respondent state's liability as a matter of English and Kenyan law (as opposed to international law).

The tribunal made passing reference to this in a couple of sibylline sentences: "Mr. Ali's payment was received corruptly by the Kenyan head of state; it was a covert bribe; and accordingly its receipt is not legally to be imputed to Kenya itself. If it were otherwise, the payment would not be a bribe." [paragraph 169] (emphasis added). Similarly, at paragraph 178, when rejecting as incompatible with English law the claimant's plea for a 'rebalancing exercise' reflecting the relative misconduct of the claimant and the Kenyan President, the tribunal said: "Further, the tribunal does not identify the Kenyan President with Kenya; and in any balancing exercise between Kenya and the claimant, the balance against the claimant would remain one-sided."

The tribunal appeared to equate the covert nature of the bribe with the lack of attribution to the Kenyan state ("accordingly"). In other words, the President was not acting as President when he was taking the bribe, he was acting as Mr Daniel arap Moi. This schizophrenic approach rather begs the question on what basis, in such circumstances, the 'personal donation' was a bribe at all (namely a payment to an influential official in order to secure the implementation of a contract with the state), if the bribe taker (who happens to be the President) is purported not to accept the 'donation' in his capacity as President because the payment was covert. Yet the bribe taker is suddenly cloaked back into his official capacity as President for the purpose of condemning the bribe giver for paying the bribe. The tribunal recognised this uneasy state of affairs when it said [paragraph 173]: "The offence [of bribery at English common law] lies in bribing a person to exercise his public duty corruptly and not in accordance with what is right and proper for the state and its citizens" (emphasis added) but drew no legal consequence from it, except that each party was made to bear its own legal costs. Query whether this is sufficient, or all that a tribunal can do. We return to this below.

Whilst the WDF tribunal recited at length the position in international law and English and Kenyan law on corruption, it said nothing on the English law position of whether, and how, acts by the president of a state may be imputed to the state itself, in circumstances where the state is not acting de jure imperii, but in a commercial manner, such as in contractual relations and the dispensing of commercial contracts.

It would have been interesting, and arguably correct, to see the WDF tribunal approach the issue from the angle of the attribution of the President's knowledge and corrupt behaviour to the state on the basis of the principles that apply in the attribution of the knowledge and actions of a Company Director to the Company itself. It is suggested that the flaws in the WDF tribunal's rationale, drawing a watertight distinction between the state and President Moi, are that it ignored (a) the commercial context of the dispute; and (b) the basic fact that, like a Company, a state acts through its officials.

In Jafari-Fini v Skillglass Ltd. [2007] EWCA Civ 261 the issue was whether a Company had committed a "Major Default" as defined in a Facility Agreement by failing to disclose a bribe paid by the claimant to procure the company's consent. The first question was whether the judge at first instance had been right to find on the facts that a bribe had been paid at all, and the second question was whether, if that was so, he had also been right to find that the bribe had come to the attention of the Company, which involved determining whether the knowledge of a Director, Mr Webster, was to be attributed to the Company.

By a majority, the English Court of Appeal decided that the judge was right on both questions, and notably on the second point that Mr Webster's directorship and his knowledge of the bribe made the Company equally knowledgeable.

Assuming a tribunal attributes the official's bribe taking to the respondent state, what consequence does this entail? In WDF, had President Moi's actions been attributed to Kenya, as it is submitted they should have been, presumably the respondent could not have been heard to argue for the dismissal of the claim on the basis that the contract was void in its entirety. One possible outcome, given that both parties are equally at fault, would be to void the contract as tainted by fraud and with no innocent party, thereby in essence giving effect to the expropriation carried out by the state, and leaving the investor without compensation as a penalty for corruption. So the WDF result was right, albeit the means of getting there were not. What this illustrates is the helplessness of arbitration tribunals in deterring corruption, because in the instant case the state of Kenya (and President Moi, who was never prosecuted) not only got away with it, it based a successful defence on it.

2. A Results-Based Criterion : Niko v Bangladesh

The 2013 case of Niko v Bangladesh3is similar to WDF in context, in that (a) it is not a treaty case; and (b) the claimant's corrupt activity was common ground on the facts of the case.

The claimant, the Barbados subsidiary of a Canadian oil and gas exploration company, entered into a Joint Venture Agreement (JVA) with BAPEX, the second respondent and the state-owned Bangladeshi oil and gas exploration and production company. The JVA contained an arbitration clause providing for ICSID arbitration.

By the time the arbitration started, the claimant, on an Agreed Statement of Facts, had been convicted in the Canadian Courts on two counts of attempting to bribe the Bangladesh Energy Minister, one in connection with securing a sale and purchase agreement for the gas post the JVA, the other to ensure that the claimant would be adequately compensated in relation to two blowouts that occurred whilst drilling the fields. In the Canadian Courts specifically, the Crown was 'unable to prove' that the bribes had indeed secured the influence sought.

The respondents did not argue that the JVA was void or voidable. Rather, they raised the question of corruption as a bar to ICSID jurisdiction. Placing their argument on a higher level, they posited that ICSID jurisdiction should be denied to a claimant who had attempted to procure a contract by bribery, whether or not that attempt succeeds, on the basis that in so doing the claimant acted in a manner that contravened good faith and international public policy, and so should not be granted the benefit of the ICSID mechanism.

The claimant argued that this was not a treaty case, but a contractual case, and that there had to exist a nexus between the alleged corruption or bad faith, and the contract sought to be enforced.

The tribunal upheld its jurisdiction. It found that there was insufficient evidence to lead to the conclusion that other acts of corruption were perpetrated by Niko, aside from those acts the subject of the Canadian criminal proceedings.

The tribunal applied the following burden of proof: "The tribunal is aware that acts of corruption are often difficult to prove, and arbitral tribunals have only very limited means to reach their conclusions. While they must bear in mind these difficulties they must also be aware that findings of corruption are a serious matter which should not be reached lightly. As the tribunal put it in Hamester v Ghana, a tribunal would "only decide on substantiated facts, and cannot base itself on inferences"." [paragraph 424]. The tribunal's refusal to consider inferences is in contrast to the approach in Metal-Tech v Uzbekistan below.

The tribunal was clearly impressed by the Canadian prosecutors' statement that they were unable to prove a causal link between the bribery and any resulting benefits:

Finally, as concerns the effect of the acts of corruption, the tribunal notes that the Canadian authorities were "unable to prove that any influence was obtained as a result of providing the benefits to the Minister." No allegation to the contrary was made in this arbitration. Bearing in mind the quoted finding of the Supreme Court of Bangladesh concerning the absence of "fraudulent means" in the making of the JVA, the tribunal has no reason to believe that corruption had any influence in the conclusion or the content of the JVA or the GSPA. [paragraph 429]

The tribunal therefore found that since the agreements and arbitration clauses therein had not been avoided, they remained in force and binding [paragraph 465]. Therefore hearing and resolving claims under valid agreements did not, as argued by the respondents, jeopardize the integrity of the ICSID system [paragraph 475].

The tribunal deferred to the merits the question whether, as the respondents contended, the claimant's investment had not been made in good faith: "In a contractual dispute as the present one, alleged or established lack of good faith in the investment does not justify the denial of jurisdiction but must be considered as part of the merits of the dispute." [paragraph 471] (emphasis added).

Finally, the tribunal dismissed the respondents' contention that the doctrine of 'clean hands' operated to prevent jurisdiction over the claimant's claim. In so doing, the tribunal relied solely on principles of public international law, a surprising choice in a private law case based on contract. It concluded that the respondents' clean hands argument failed to meet the public international law criteria as set out by the UNCLO S Arbitral Tribunal in Guyana v Suriname,4to the following effect: (i) the breach must concern a continuing violation; (ii) the remedy sought must be "protection against continuance of that violation in the future"; not damages for past violations; and (iii) there must be a relationship of reciprocity between the obligations considered.

In the instant case, said the tribunal, "the violation on which the respondents rely is not continuing, but consisted in two acts that have been completed long ago; the remedy which the claimant seeks does not concern protection against this past violation; and there is no relation of reciprocity between the relief which the claimant now seeks in this arbitration and the acts in the past which the respondents characterise as involving unclean hands." [paragraph 483]

In any event, the tribunal found that the respondents were estopped from relying on allegedly unclean hands since the corrupt acts occurred before the GSA was concluded and the respondent relied on these events only now.

What lessons can be drawn from Niko?

Not only must the respondent prove corrupt behaviour, but Niko opens the door to the additional requirement of having to prove that the corrupt behaviour brought about the intended benefit - probably the only aspect of corruption more impossible to prove than the covert act itself - and presumably, according to the tribunal's words, on the basis of 'substantiated facts'. It may be ironic that the Canadian prosecutors' inability to prove that very causal link did not stop the Canadian Courts from imposing a fine of several million dollars to Niko, and a three-year probation period.

Another lesson may be that it is risky for a respondent to raise corruption as a bar to jurisdiction, and that it stands a better chance on the merits - but only if the respondent will not avoid the contract.

3. The Tribunal 's ex officio inquisitive Powers : Metal -Tech v Uzbekistan

Metal-Tech5is a treaty case pursuant to the Israeli-Uzbek BIT, and cut of a different cloth.

Here the tribunal declined jurisdiction on the claimants' claims (and the respondent's counterclaims) on what has been described in commentary as a 'technical' ground that the state had not given its consent to arbitration in the case of an investment 'implemented' (namely, argued the respondent, made and operated) in contravention of the legality requirement in the Israel-Uzbek BIT. The legal contravention was the payment of some US$4 million in bribes by the Israeli claimant to various Uzbek state officials under the cover of consultancy fees.

How the tribunal arrived at its finding of corruption is instructive and, it is submitted, right. In contrast to many historical pronouncements that corruption must be proven by no less than 'substantiated facts and conclusive evidence' (making it in practice near-impossible to prove), and in contrast to many tribunals' complete blank out of the state's culpable role in the corruption process, the tribunal in Metal-Tech embarked upon a meticulous analysis of the facts and proceeded to make findings of corrupt practices on the basis of 'red flags':

293. For the application of the prohibition of corruption, the international community has established lists of indicators, sometimes called "red flags". Several red flag lists exist, which, although worded differently, have essentially the same content. For instance, Lord Woolf, former Chief Justice of England and Wales, included on his list of "Key Red Flags" among other things" (1) "an Adviser has a lack of experience in the sector;" (2) "nonresidence of an Adviser in the country where the customer or the project is located;" (3) "no significant business presence of the Adviser within the country; (4) "an Adviser requests 'urgent' payments or unusually high commissions;" (5) "an Adviser requests payments be paid in cash, use of a corporate vehicle such as equity, or be paid in a third country, to a numbered bank account, or to some other person or entity;" (6) "an Adviser has a close personal/professional relationship to the government or customers that could improperly influence the customer's decision". As has been seen above in the section entitled "Key facts" and as will become further evident in the course of the analysis under Uzbek law, many of these red flags are present here.

It did so as a matter of its duty and exercising ex officio powers to ask probing questions of the parties and issuing Directions with specific requirements:

241. The payment of such substantial sums having been admitted, the tribunal considered it its duty to inquire about the reasons for such payment. First, at the January Hearing itself, the tribunal observed that, given the disclosure of facts unknown until then, it needed more information from the Parties. In the exercise of its ex officio powers under Article 43 of the ICSID Convention, the tribunal therefore invited the Parties in PO 7 to provide that information. In PO 10, the tribunal once again exercised its ex officio powers to call for additional testimony and evidence.

Its inquiry included the issuance of specific Directions:

92. [The tribunal] also noted that it would be assisted by receiving further information about the services provided by the Consultants. Specifically, it directed the claimant to submit information about the payments made to each individual Consultant, specifying (a) the specific services rendered, (b) which Consultant rendered the service, and (c) when the services were rendered. The tribunal also asked for responses to the following questions:

a) Where payments were made to companies (ie.., MPC Companies or Lacey International):

• For which individual Consultant were these payments meant?

• What was the amount paid to each individual Consultant?

• What service was the payment intended to remunerate?

• When was this service rendered?

b) Where payments were made to individual Consultants:

• What service was the payment intended to remunerate (beyond mere reference to the contracts under which the service was rendered)?

• When was this service rendered?

In drawing inferences, the tribunal looked at the following factors:

• The relative size of the consultancy payments overall (US$4 millon) in relation to the capital contribution from Metal Tech (US$500,000);

• The low cost of living in Uzbekistan (which made the size of the bonus payments disproportionate - 50 times their Uzbek salaries);

• The nature of the services rendered - Uzbek law does not recognize "lobbying", and the individuals were not certified as lobbyists, or qualified to provide the "technical expertise" in the molybdenum industry;

• The connections of two individuals with state officials responsible for the approval, establishment and operation of the claimant's investment;

• Secret contracts and payments made via offshore companies, rather than in a transparent manner;

• Payments made irrespective of services or evidence of services;

• The claimant's failure to provide answers and evidence (documentary or testimonial) in reply to the tribunal's directions above.

The tribunal came to the conclusion that, "[w]hile the tribunal does not believe that the claimant sought to conceal evidence, the inference that inexorably emerges from this dearth of evidence is that the claimant can provide no evidence of services, because no services, or at least no legitimate services at the time of the establishment of the claimant's investment, were in fact performed. The tribunal will bear this inference in mind when further assessing the facts." [paragraph 265]

In light of this finding the tribunal found it unnecessary to embark upon an analysis of the claimant's violation of transnational public policy, and indeed the requisite burden of proving allegations of corruption, finding the red flags sufficiently compelling.

The tribunal's approach, although undeniably right in principle and outcome, presents some risk in practical consequences, and requires a robust and experienced tribunal equipped with thorough preparation.

Some of the risky consequences are as follows.

One is the personal safety of witnesses. In asking its probing questions, the tribunal found that several individuals, who on paper clearly had the requisite knowledge to answer the tribunal's questions, refused to come forward and testify on behalf of the claimant on safety grounds. The tribunal dismissed their concerns as 'unconvincing' [paragraph 264], stating that the tribunal could have made necessary arrangements for their safety and protection. It is legitimate to question what these arrangements might have been and how an international tribunal, lacking the powers of a state court and unable to summon police protection, might effectively protect the individuals concerned from retaliation, personal or against their families, in their home country.

Another consequence is that the tribunal's approach might be open to accusation that it effectively assisted the respondent in proving its case, and of an attendant lack of due process.

That said, of the uneven investment arbitration jurisprudence on the treatment of corruption allegations in existence so far, Metal-Tech is undeniably the more authoritative and convincing instance. In staying away from lofty pronouncements and focusing on fact-finding, the tribunal ended up with a sufficient number of red flags, from which a compelling picture emerged allowing it to draw inferences that were more useful than the elusive 'conclusive evidence' required in other cases.

Another area where the Metal-Tech tribunal showed its perceptiveness and realism is in its recognition that, in cases of corruption, it often takes two to tango. Consequently, the tribunal found that the parties were to share in the costs:

389. (…) the tribunal is sensitive to the ongoing debate that findings on corruption often come down heavily on claimants, while possibly exonerating defendants that may have themselves been involved in the corrupt acts. It is true that the outcome in cases of corruption often appears unsatisfactory because, at first sight at least, it seems to give an unfair advantage to the defendant party. The idea, however, is not to punish one party at the cost of the other, but rather to ensure the promotion of the rule of law, which entails that a court or tribunal cannot grant assistance to a party that has engaged in a corrupt act."

422. More important, the tribunal's determination is linked to the ground for denial of jurisdiction. The tribunal found that the rights of the investor against the host state, including the right of access to arbitration, could not be protected because the investment was tainted by illegal activities, specifically corruption. The law is clear - and rightly so - that in such a situation the investor is deprived of protection and, consequently, the host state avoids any potential liability. That does not mean, however, that the state has not participated in creating the situation that leads to the dismissal of the claims. Because of this participation, which is implicit in the very nature of corruption, it appears fair that the Parties share in the costs. (emphasis added).

Conclusion

The uneven state of investment arbitration jurisprudence to date on allegations of corruption is a stark reminder of the lack of consensus on how to deal with corruption in international disputes, beyond the widely shared opprobrium on corruption and its effects generally. Allegations of corruption push the limits of the arbitral process, and their awkward treatment to date by international tribunals arguably did little to deter litigants from bringing disputes involving corrupt activity at one point in the background.

Metal-Tech may herald a new era of welcome initiative and horn-taking by investment tribunals.

What Metal-Tech demonstrates to the savvy investor is that it is taking an investment claim tainted by corruption before a pro-active tribunal at its peril. The difficulty remains that, upon a finding of corruption, states still get away with relative impunity aside from a costs order. There is little else a supra-national tribunal can do. More biting sanctions lie at a policy level.

1
World Duty Free Company Limited v The Republic of Kenya, ICSID Case ARB/00/7, Award, 4 October 2006.

2
Albeit the applicable Kenyan statute is clear in making the bribe-taker an offender: "Indeed, by Sections 3 and 4 of the Kenyan Prevention of Corruption Act 1956, (being in force at the time of the bribe in March 1989), it was a felony for any person or agent corruptly to give or to receive any gift, fee, reward, consideration or advantage whatever, as an inducement to, or reward for, any member, officer or servant of a public body doing, or forbearing to do, or having done or forborne to do, anything in respect of any matter or transaction whatsoever, actual or proposed or likely to take place, in which the public body is concerned…". (emphasis added) [para. 170]

3
Niko Resources (Bangladesh) Limited v People's Republic of Bangladesh et al, ICSID Cases No ARB/10/11 and ARB/10/18, Decision on Jurisdiction, 19 August 2013.

4
Guyana v Suriname, PCA, Award of 17 September 2007 (under UNCLO S Ch VII) [paragraphs 420-421.

5
Metal-Tech Limited v The Republic of Uzbekistan, ICSID Case No. ARB/10/3, Award, 4 October 2013.