Criticising the Jurisprudence of ‘Risk’ in Investor-State Arbitrations
Introduction
To cross the proverbial ‘fortress moat’ to the International Centre for Settlement of Investment Disputes (ICSID), a claimant must first lower the bridge by persuading the tribunal that their transaction constitutes an investment under Article 25 of the ICSID Convention. This question partially turns on whether the investor has assumed a ‘risk’, an inherent characteristic of an investment undertaking. This essay shall criticise the jurisprudence concerning the meaning of investment ‘risk.’ To accomplish this, this essay shall: (a) summarise why the definition of risk is significant by contextualising it within the Salini Test; (b) explain how arbitral tribunals have interpreted risk by referring to the jurisprudence established in Joy Mining v. Arab Republic of Egypt and Romak v. Republic of Uzbekistan; (c) argue that this current interpretation is overly restrictive to the detriment of both investors issuing debt and investors subjected to particular sovereign acts; (d) illustrate that the current interpretation is also overly inclusive concerning the requirement that an investor remains uncertain as to the total amount he or she will spend, and; (e) briefly outline and evaluate proposals to remedy these problems. Having done so, this essay will effectively conclude that the current jurisprudence of risk in investor-state arbitrations is unsatisfactory, arguing that a shift from the current ‘investment risk’ requirement to an objective ‘intention-reliant’ definition of investment for ICSID jurisdiction to arise may be preferable.
The Significance of ‘Risk’ to Investor-State Disputes
The task of criticising the existing arbitral jurisprudence on risk is crucial, given the importance of investment risk within the broader ICSID rules. Article 25 of the ICSID Convention limits the centre's jurisdiction to legal disputes “arising directly out of an investment” between a contracting state (or its agencies or subdivisions) and nationals of another contracting state (the investing party). Consequently, any investor-state tribunal must first determine whether a particular undertaking qualifies as an ‘investment’ to assume jurisdiction over the dispute (and, by extension) ultimately adjudicate on it.
To satisfy the meaning of ‘investment’ under Article 25, tribunals have adopted the following ‘double-keyhole’ approach. Firstly, it must be ascertained whether the concerned undertaking satisfies the meaning of ‘investment’ provided by the operative Bilateral Investment Treaty (BIT) between the plaintiff state and the state of which the claimant is a national. Secondly, the transaction must satisfy the following three criteria: (a) it must involve a contribution by the Investor, (b) it must exist for a particular duration, and (c) it must involve some assumption of risk by the Investor. These latter criteria were first established in the 2001 case of Salini & Italstrade v. Kingdom of Morocco and have since crystallised into the ‘Salini Test’, which (as observed by the Electrabel v. Republic of Hungary tribunal) has become the predominant methodology used by arbitrators to establish the existence of an investment.
Thus, the exclusion or inclusion of particular factors as consecutive of this element of risk effectively determines when the Salini Test, and by extension, Article 25, is satisfied to enable ICSID jurisdiction. In this context, states are interested in a definition of risk which is not too broad or easily fulfilled to allow an unmanageable flood of claims from projects that cannot be reasonably classified as investments (the ‘Stable Legal Order Requirement’). Simultaneously, any definition of risk must still be inclusive enough to apply to the various types of investments that may arise to ensure that a particular claim is not unfairly excluded from the ICSID system, consequently leaving the investor without an effective remedy (the ‘Flexibility Requirement’). This latter prospect is especially threatening to the entire legitimacy of the investor-state dispute resolution system, given that ICSID was founded primarily to enable foreign nationals to enforce their legal rights under international investment treaties. It follows that the meaning of risk is crucial in determining the material scope of the ICSID system.
Joy Mining and Romak: Outlining the Current Legal Standard
Investor-state arbitral tribunals have sought to portray investment risk for the purposes of Article 25 as more substantial and severe than common commercial risk. This standard was first reflected in Joy Mining v. Arab Republic of Egypt, which concerned the supply of specialised phosphate extraction equipment by Joy Mining to an Egyptian state mining body. The dispute arose when the Egyptian government (claiming the equipment was faulty) withheld bank guarantees owed to Joy Mining, alleging a failure of contractual performance. Joy Mining subsequently brought an ICSID claim, arguing that the Egyptian government’s failure to release the bank guarantees was equivalent to expropriation in breach of the United Kingdom-Egypt BIT, resulting in the wrongful deprivation of investment returns.
Egypt disputed that the Salini Test was satisfied to trigger Article 25 jurisdiction, arguing that Joy Mining bore no genuine risk in the transaction, given that the contract was merely a standard recurrent supply agreement through which payment was issued using an irrevocable letter of credit. This view was accepted by the ICSID tribunal, which reasoned that the risks faced by Joy Mining were no different to those present in any other commercial transaction, such as the possibility of contract termination or contractual non-performance. The tribunal’s conclusion was supported by the language of the terms governing the bank guarantees, which were entirely standard and excluded any reference to an investment made under Egyptian law. Furthermore, the claimant took no steps to take advantage of the special incentives offered by the Egyptian government to international investors. Hence, the tribunal concluded that Joy Mining’s undertaking could not be distinguished from an ordinary commercial transaction, and therefore, no investment risk was readily apparent. It is, therefore, clear that the mere existence of a risk common to all transactions (otherwise called the ‘risk of doing business generally’) would not satisfy the Salini Test and that the existence of ‘investment risk’ requires much stricter factual circumstances.
Romak v. Republic of Uzbekistan outlined what specific conditions would establish a risk capable of fulfilling the Salini Test, being: (a) uncertainty as to whether the investor will receive a return on their investment; (b) uncertainty as to the amount the investor will spend (even where all involved parties fulfil their contractual obligations) such that the “investor simply cannot predict the outcome of the transaction”, and; (c) that the alleged risk is distinct from any ordinary commercial risks. In 2015, a fourth element was added to the Romak Criteria by Postova Bank v. Hellenic Republic, which provided that the alleged risk must be distinct from ‘sovereign risk’, defined as government interference and non-performance in a transaction or contractual relationship. In arriving at this conclusion, the Postova Bank tribunal (relying on Romak) distinguished sovereign risk from investment risk on the basis that the former was not a “risk inherent in the investment operation in its surrounding [specifically] the success or failure of the economic venture concerned.” It follows that under the present jurisprudence, investment risk is distinct from sovereign or commercial risk and purely concerns the inherent operation of the venture itself.
Having outlined and explained the interpretation of investment ‘risk’ assigned by Joy Mining and elaborated upon in Romak and Postova Bank, this essay shall proceed to critique the above jurisprudence.
Does the Existing Arbitral Jurisprudence Create An Overly Restrictive Standard for Determining ICSID Jurisdiction?
The Romak Criteria are arguably overly restrictive for the following reasons: (i) criterion (a) unjustifiably excludes particular accepted forms of investments, and (ii) the Postova Bank precedent fails to acknowledge that sovereign risk may be the maximum risk an investor may face, given the nature and status of sovereigns and the possibility that a state’s internal affairs may create severe risks. Concerning point (i), it is submitted that there are uncontroversial types of investments where the exact amount of return is known. For example, bonds that pay out a certain sum and which are redeemable at a specific stated value or any loan agreement where the rate of interest or sums payable is expressly highlighted and fixed. Such an investment would prima facie fail to satisfy the Romak Criteria for lacking any element of uncertainty vis-à-vis investment returns. Consequently, the current approach may enable a discriminatory system based on two classes of ‘equity’ and ‘debt’ investors, whereby only the former is perceived as facing any inherent investment-related risk and thus can benefit from the ICSID system.
It is difficult to justify this perception. Indeed, as acknowledged by the arbitral tribunal in Yukos Capital v. Russian Federation, a loan represents an investor’s continual stake in a project’s success or failure, whereby the operative risk vests in a failure by the debtor to repay. Importantly, this risk is distinct from that of an ordinary transaction for the sale of goods or raw materials, where the seller would not ordinarily acquire a longstanding proprietary or economic interest in the economic venture itself. It follows that the type of risk faced by ‘debt’ investors indeed exists and arguably relates to the inherent nature of a project (as opposed to being a general commercial risk). Hence, the effect of excluding investments where the amount of return is certain creates the absurd result of an investor being unable to prove the existence of genuine investment-related risk, thereby fatally undermining the ‘Flexibility Requirement.’ This is indicative of the overly restrictive nature of the precedent in Romak.
Concerning point (ii), the view established by Postova Bank may be challenged, given that sovereign interference in a contractual relationship may be a real and severe threat faced by international investors. Indeed, state actions (such as the nationalisation of private assets, sovereign defaults, military activities and political instability) can significantly undermine investors’ core legal rights and the integrity of their investment. This view is well supported by the facts of Gavrilovich v. Republic of Croatia, where the tribunal held that the “economic and political circumstances” initiated by the Croatian War of Independence were sufficient to create investment-related risks capable of satisfying the Salini Test. These risks specifically concerned the possibility that (a) Croatian or Serbian military operations would damage or destroy the claimant’s assets and (b) given that the investments were located in regions controlled by Serbian separatists, the claimant may be prevented from accessing or asserting legal title to them depending on the war’s territorial outcome. Suppose we accept the understanding adopted in Romak that investment risk must, by definition, reach a certain threshold of severity to undermine the inherent operation or success of an investor’s undertaking; it follows that the type of sovereign interference in Gavrilovich ought to be considered a form of investment risk. Indeed, the physical destruction of an investment or the arbitrary loss of an investor’s legal title to these assets are substantially more severe outcomes than those potentially inflicted by ordinary contractual risks.
This is not to say that every form of sovereign risk will meet this threshold but that those which do (as was arguably the case in Gavrilovich) share much in common with investment risks and should not, resultantly, be distinguished from it. It follows that the blanket exclusion of sovereign risk by Postova Bank is unjustified and incoherent because it fails to recognise this fact. It is also unfair because, due to such incoherence, this interpretation (like criterion (a) of the Romak Criteria) precludes investors from proving the existence of a genuine and significant investment risk. It is, thus, fortunate that more recent ICSID tribunals have departed from Postova Bank’s restrictive interpretation, such as in SCB (Hong Kong) v. United Republic of Tanzania, which expressly held that the risk of sovereign interferences by a home state would be constitutive of an investment risk.
Given the above, this essay shall now proceed to explain how criterion (b) of the Romak Criteria may be overly inclusive in a negative sense.
Does the Existing Arbitral Jurisprudence Create An Overly Inclusive Standard for Determining ICSID Jurisdiction?
Before introducing the explanation below, the Author must concede that an argument for the Romak Criteria’s over-inclusiveness may appear to conflict with earlier arguments concerning its overly restrictive nature. To preclude the possibility of any contradiction or confusion, it must be reiterated that both these arguments relate only to specific limbs of the test and not the test as a whole. In this context, criterion (a) of Romak is only overly restrictive and criterion (b) is only overly inclusive, such that the forthcoming examination is made without prejudice to any previous argument. Having clarified this, this essay can now begin its analysis of over-inclusiveness.
It should be recalled that criterion (b) of the Romak Criteria requires an investor to be uncertain about the amount he or she will spend throughout an investment’s lifetime, even where all other involved counterparties fulfil their contractual obligations. It is submitted that this criterion is too easily satisfied and can include virtually every long-term transaction, potentially contributing to an overly expansive application of Article 25 jurisdiction. It should be acknowledged that this requirement was likely established in response to the tribunal’s comments (at paragraph 58 of the Award on Jurisdiction) in Joy Mining, which emphasised the need to distinguish contracts for the sale of goods, where the total amount to be paid is usually firmly established, from investment undertakings. However, the Romak tribunal did not elaborate on the amount of uncertainty needed, meaning that this limb of the criteria may be satisfied if there is merely the slightest doubt regarding the complete costs an investor may incur.
To explain how this criterion could be satisfied by virtually any collection of factual circumstances, consider the following hypothetical cross-border transactional scenario whereby a Jordanian company (Y) must annually pay an American company (X) a $500 sum. In this context, Darius Chan has highlighted that it may be argued that the amount Y will spend is unknown due to any number of factors, such as (i) fluctuation in the exchange rate between the Jordanian Dinar and the US Dollar, and (ii) inflation, which may alter the actual value of the $500 as time passes. Crucially, these factors operate independently of whether the parties have satisfied their contractual obligations. Therefore, it is clear that this rule is ill-suited to accommodate long-term transactions. This is reinforced by the fact that the tribunal in Romak was merely concerned with a simple supply contract, which (due to its shorter lifespan) would have been less vulnerable to macroeconomic trends such as fluctuating currency exchange rates. Hence, while the present formulation would have been sufficient to deal with the short-term nature of a Romak-like transaction, it should not be considered a general standard applicable to all investment disputes.
The overly expansive application of criterion (b) enables investors to easily fulfil Article 25 jurisdiction by tying the existence of risk to the operation of generic trends such as monetary inflation, potentially leading to an unmanageable flood of ICSID claims against states. Such an outcome undermines the raison d'etre of the Romak Criteria itself, which is to ensure that only circumstances of sufficient severity and novelty may suffice to satisfy the Salini Test’s risk element. Furthermore, such an outcome may also explain the attempt by some tribunals to restrict what amounts to ‘investment risk’ as a means to reduce their workload and allow the ICSID to remain effective.
Additionally, it may also be argued that criterion (b) should be interpreted within the context of ‘operational risk’ to circumvent these shortcomings. Postova Bank defined ‘operational risk’ as a risk “inherent in the investment operation in its surroundings.” Seeking to elaborate on this statement, the SCB (Hong Kong) tribunal provided that an investor's liability for the maintenance, repair and operation of a piece of infrastructure would be an example of the assumption of ‘operational risk’. Hence, the continuing possibility of accidents or repairs would guarantee that an investor would be uncertain of the total amount he or she will spend. While limiting criterion (b) to more specific circumstances, this approach is potentially restrictive in its own right and should not be adopted. This is because the standard of ‘operational risk’ cannot be applicable vis-à-vis investments where there is nothing for the investor to operate or maintain. It cannot, consequently, serve as a means of general interpretation.
Potential Reform?
The issues outlined above are not easy to resolve. Indeed, the task of any reformist is made more complex by the competing demands of investors and states for a sufficiently adaptable (yet predictable) definition of investment risk and the need to ensure the existence of a clear boundary between investment risk and general commercial risk. In this context, two potential solutions are worth examining.
Firstly, it has been proposed that the Romak Criteria be abandoned in favour of earlier ICSID jurisprudence on investment risk. In particular, Darius Chan has adopted the view that the standard in Quiborax & Fosk v. Plurinational State of Bolivia is preferable to the approach taken in Romak. In Quiborax & Fosk, the tribunal established that the required risk need not be strictly investment-related and could include “market, financial and political risks.” This is an arguably tempting proposition for the following two reasons. Firstly, it resolves the aforementioned unfairness facing debt investors by neatly recognising the risk of a debtor’s default under the banner of ‘financial risk.’ The same can be said for investors facing sovereign interferences thanks to the tribunal’s inclusion of ‘political risk.’ Secondly, it does not require that an investor be uncertain about the amount he or she will ultimately spend. However, the broadness of this standard may be extended to deem almost any set of circumstances as including some risk. This is evidenced by the cases of FEDAX v. Republic of Venezuela and Deutsche Bank v. Democratic Socialist Republic of Sri Lanka, where Quiborax & Fosk was relied upon to conclude that the ‘very existence of a dispute’ would demonstrate the presence of a risk undertaken by an investor. It follows from such reasoning that the very act of an investor filing an ICSID claim would automatically satisfy the risk component of the Salini Test, which is clearly an unsatisfactory and extreme interpretation, as it renders this element of the test entirely meaningless. It is thus evident that the Quiborax & Fosk standard may create its own problems of over-inclusiveness and ought not to be adopted as a replacement for the Romak Criteria.
It may, perhaps, as advocated for by academics Rudolf Dolzer and Christoph Schreuer, be time to recognise that the concept of risk has only minimal value when characterising an investment, as almost every transaction “which extends beyond the day of its conclusion” will involve circumstances of uncertainty for those involved, and thus an element of risk. In this context, a second proposal by Mavluda Sattorova has involved abolishing the Salini Test (including its risk element) entirely in favour of an objective and technical definition of investment. In this regard, she refers specifically to two potential definitions. Firstly, the definition provided by Article 45 of the Free Trade Agreement between Mexico and the EFTA (European Free Trade Association), which provides that an investment must “be made in accordance with the laws and regulations of the parties […] for the purpose of establishing lasting economic relations [and] give the possibility of exercising an effective influence on the management thereof.” Secondly, the view taken by the European Court of Justice (ECJ) that foreign direct investment includes the following core variants: (i) the creation and extension of branches of new undertakings belonging solely to the person providing the capital or the complete acquisition of existing undertakings; (ii) the participation in new or established undertakings to create or maintain a lasting economic relationship; and (iii) the issuing of long-term loans to create or maintain a lasting economic relationship.
These definitions benefit from the following two features. Firstly, they focus on the reasonable intentions of the involved parties to create a long-term economic relationship and, therefore, may be empirically satisfied by reference to their actions, communications, or the particular language employed in their contractual arrangements. Importantly, such an ‘intention-reliant’ (as opposed to ‘risk-reliant’) approach enables the definitions to be neutral vis-à-vis the type of risks concerned (thereby ensuring their inclusiveness and adaptability to satisfy the ‘Flexibility Requirement’) while still establishing some hurdle of conduct or evidence to be satisfied (thus limiting the application of Article 25 to meet the ‘Stable Legal Order Requirement’). Secondly, neither appears to exclude any particular class of investor, namely debt investors, which are accommodated explicitly in part (iii) of the ECJ’s definition. Therefore, at least in this regard, they do not suffer from the aforementioned problems relating to an unfair or overly restrictive application. It follows that Sattorova’s solution allows for a more effective legal mechanism than the risk-reliant Salini Test. Thus, it should be seriously considered as a substitute for the Romak Criteria in future ICSID jurisprudence.
Conclusion
This essay sought to outline and evaluate ICSID jurisprudence on the meaning of risk. It opened this discussion by contextualising the importance of this definition in determining ICSID jurisdiction with reference to the Salini Test. It followed this by explaining the significance of Joy Mining and Romak and criticising particular elements of the Romak Criteria as overly restrictive and overly inclusive at the same time. Lastly, this essay outlined two proposals for resolving these problems to conclude that abolishing the requirement for risk entirely in favour of an objective ‘intention-reliant’ definition of investment may be preferable. While the above arguments are far from the last word on this subject, this essay hopes to have made a worthwhile contribution to the discussion on the challenges facing the investor-state dispute settlement process.
