THE CHOICE OF SEAT IN INVESTMENT ARBITRATION

Author: Enyo Law LLP

Introduction

It is well recognised that the choice of the place, or seat, of arbitration is a matter of major, if not critical, importance. However, although the choice of seat and its implications have received much attention in the context of international commercial arbitration, there has been little, if any, discussion of the implications of this choice on investment treaty arbitration. This may be in part because of the fact that the majority of investment treaty arbitrations have been brought under the ICSID Convention (approximately 53 per cent), with seat not being a concern in the context of such delocalised arbitrations. However, for the remaining 47 per cent of non-ICSID investment treaty arbitrations, the choice of seat will have as much significance as for commercial arbitrations.

The Importance of The Seat in International Arbitration

The choice of the legal seat has important legal and practical implications for both commercial and investment treaty arbitrations. Most significantly, the choice of seat would subject the arbitration to the supervisory jurisdiction of the court of the seat, which would have the power to set aside the award. They would also exercise other powers in support of arbitration (e.g., potentially consider challenges to arbitrators, make default arbitrator appointments, assist the tribunal with the taking of evidence or order provisional measures in support of arbitration).

The seat of arbitration is also significant from the point of view of the enforcement of a resulting arbitral award. The award will be deemed to have been legally rendered at the seat. Although the 1958 New York Arbitration Convention on the Recognition and Enforcement of Foreign Arbitral Awards (the New York Convention) applies to all ‘arbitral awards made in the territory of a State other than the State where the recognition and enforcement of such awards are sought’, a number of contracting states have made reservations under Article 1(3) of the Convention, stipulating that they will only recognise and enforce foreign arbitral awards on the basis of reciprocity (i.e., where the award was rendered in the territory of another contracting state). Thus, an award rendered in a state that is not a party to the New York Convention may be more difficult to enforce worldwide. However, because the most popular seats for investment treaty arbitration are state parties to the New York Convention, this is less likely to be a significant concern in the investment treaty context.

Finally, the choice of the legal seat of arbitration will involve a number of practical and logistical implications. While the choice of a seat of arbitration does not necessarily mean in all cases that the hearing venue will be at the seat, this will more often than not be the default position, particularly when the parties are unable to reach an agreement on another venue that would be convenient for all hearing participants. Assuming the hearing takes place at the seat of arbitration, its choice can thus significantly affect the costs incurred by the parties, and potentially inconvenience one side more than the other (owing to, inter alia, the duration of travel, jet lag and availability of office premises that the parties can use during the hearing). In case a hearing may need to be conducted virtually, the law of the seat may affect whether this is possible. Finally, the law of the seat of some jurisdictions may not allow the remote signing of the award and would require it to be signed in person. All these considerations should be taken into account in the choice of seat.

The Framework for Choosing a Seat in Investment Treaty Arbitration

It is uncommon for investment treaties to identify specifically the seat of arbitration under that treaty. Many investment treaties specify that the arbitration shall take place in a state party to the 1958 New York Convention, with a view to maximising the enforceability of the resulting award. Some bilateral investment treaties (BITs) also specify that the arbitration cannot be held in the territory of a contracting party. One exception is the now-defunct North American Free Trade Agreement (NAFTA), which provided that the place of arbitration had to be chosen in one of the NAFTA states.

The lack of attention to the seat of arbitration in investment treaties can be perhaps because of the fact that, to date, the majority of investment treaty arbitrations have been conducted under the International Centre for Settlement of Investment Disputes (ICSID) Convention. ICSID arbitrations are not subject to the supervisory jurisdiction of state courts and as such are deemed ‘delocalised’ and do not have a legal seat.

Nevertheless, the choice of seat will still be a significant issue in non-ICSID Convention investment treaty arbitration, such as arbitration under the ICSID Additional Facility Rules, other institutional rules (such as the rules of the Stockholm Chamber of Commerce (SCC), International Chamber of Commerce (ICC) or London Court of International Arbitration (LCIA)) and ad hoc arbitration (whether under the UNCITRAL Rules or not). Most arbitration rules contain provisions dealing with the choice of seat in the event that the arbitration agreement does not specifically select the seat.

In particular, the ICSID Additional Facility Rules stipulate that arbitration proceedings shall only be held in states that are parties to the 1958 New York Convention. Subject to this requirement, the place of arbitration shall be determined by the tribunal after consultation with the parties and the Secretariat. Similarly, under the UNCITRAL Rules, in the absence of agreement by the parties, the place of arbitration shall be determined by the arbitral tribunal ‘having regard to the circumstances of the case’.

Under the SCC and ICC rules, when not agreed by the parties, the seat is determined by the institution. Under the LCIA Rules, in default of agreement by the parties, the seat of the arbitration shall be London (United Kingdom), ‘unless and until the Arbitral Tribunal orders, in view of the circumstances and after having given the parties a reasonable opportunity to make written comments to the Arbitral Tribunal, that another arbitral seat is more appropriate’.

How Seats Have Been Chosen to Date in Investment Treaty Arbitrations

There is a relatively limited number of decisions by arbitral tribunals in the public domain determining the place of arbitration. This may be because, in many cases, the parties are able to reach an agreement on the seat, and also because many procedural orders on this issue are not made public. Institutions’ decisions on these matters are likewise not published.

Tribunals determining where an investment treaty arbitration shall be seated have generally taken into account the criteria set out in the UNCITRAL Notes on Organizing Arbitral Proceedings. These are as follows:

1. Suitability of the law on arbitral procedure of the place of arbitration. This is by far the most important consideration to consider when determining the place of arbitration. In practice, most tribunals have been unwilling to undertake a detailed comparison of the arbitration laws of the proposed seats and have generally found that the laws of the most commonly chosen seats of arbitration fulfil this requirement. For instance, in Philip Morris v. Australia, the tribunal, facing a choice between London and Singapore, noted that ‘both Singapore and London fulfil the requirements of a place for a BIT arbitration. The arbitration law as well as the judiciary, should it become involved, are well equipped in both States to deal with the present dispute’. This is despite the fact that, at the time, there had been only one investment treaty case with a seat in Singapore, while English courts had reviewed a number of investment treaty awards. However, in UPS v. Canada, in preferring Washington, DC, over Ottawa, the tribunal gave some weight to the position of the Canadian government as to the standard of review of NAFTA awards in Canadian courts, as expressed in the challenge of the earlier Metalclad award.
2. Whether there is a multilateral or bilateral treaty on enforcement of arbitral awards between the state where the arbitration takes place and the state or states where the award may have to be enforced. Because all commonly used seats of investment arbitration are in state parties to the New York Convention, this has not been a significant distinguishing factor in the choice of one seat over another.
3. Convenience of the parties and the arbitrators, including the travel distances. This was the deciding factor considered by the tribunal in Philip Morris when preferring Singapore over London, noting that factors of convenience and travel distances ‘may not only be relevant for hearings of this Tribunal, but also for subsequent disputes before the domestic courts at the place of arbitration should such disputes arise’. In some arbitrations it has been more of a matter of a ‘balance of inconvenience’ rather than a balance of convenience, with the tribunal preferring a location that would be least inconvenient for the parties, their counsel and the arbitrators.
4. Availability and cost of support services needed. Although hearing facilities and support services should be available in most leading international arbitration seats, the cost of these facilities and services may vary significantly. One factor to consider in this regard is that, in cases administered by ICSID (both under the ICSID Additional Facility Rules and ad hoc arbitrations), parties can benefit from hearing facilities in Washington, DC, and Paris at no additional fee. In cases administered by the Permanent Court of Arbitration (PCA), parties can likewise use the facilities of the Peace Palace in The Hague, as well as hearing facilities in countries with which the PCA has concluded host country agreements, at no charge.
5. Location of the subject matter in dispute and proximity of evidence. While taken into account by some tribunals, this has generally been a less important consideration. In the context of an investment treaty dispute, the subject matter of the dispute (i.e., the investment), as well as the majority of the connected evidence will be located in the host state. Holding the arbitration in the host state itself, however, could be potentially at odds with the principle of neutrality, as discussed below.
6. Any qualification restrictions with respect to counsel representation. Such restrictions are rare but can be a very important consideration where present (e.g., in Nigeria). The most commonly used seats of investment treaty arbitration generally do not restrict counsel representation to lawyers qualified in those jurisdictions, so this is less likely to be a concern.

In addition to the criteria set out in the UNCITRAL Notes on Organizing Arbitral Proceedings, tribunals have also considered certain additional criteria, the most important of which is neutrality. As noted by the Ethyl tribunal:

the fact that the UNCITRAL Notes omitted ‘perception of a place as being neutral’ from its list of criteria for selection of a place of arbitration because it was ‘unclear, potentially confusing’ does not mean that such criterion cannot be considered. UNCITRAL, in taking this step, itself indicated ‘that the arbitral tribunal, before deciding on the place of arbitration, might wish to discuss that with the parties’.

Given the highly political nature of many investment treaty cases and the fact that, by definition, they are adverse to the host state (including, occasionally, allegations of breaches of international law by that state’s judiciary), it is of particular importance to insulate investment treaty proceedings from any interference of the courts of the host state. It is therefore extremely uncommon for investment treaty proceedings to be seated in the host state. Although a seat of arbitration in the home state of the investor is less likely to give rise to concerns about potential interference with the arbitration, considerations of neutrality are likely to favour choosing a seat of arbitration in a third state.

In light of the recent trend whereby states have been seeking to set aside awards out of intra-EU BITs further to the Court of Justice of the European Union (CJEU)’s Achmea decision (see Section V. v), considerations of neutrality would dictate that intra-EU arbitrations have a seat outside the European Union, thus insulating them from review by European courts.

Another criterion which, in practice, appears to influence the selection of the seat of investment treaty arbitration, despite its lack of formal legal significance, is the location of the administering institution. For example, of the 61 non-ICSID cases administered by ICSID, the seat of which is known, 23 arbitrations had their seat in Washington, DC, which was the single most commonly chosen seat for such arbitrations. In Merrill v. Canada, the tribunal considered this fact, noting that ‘Washington, D.C. is the seat of ICSID, the administering institution of this case, that it has been accepted on various occasions as the place of arbitration and that it has developed the reputation of being an independent venue for many international organizations’. The Hague is commonly chosen as the seat of arbitration in ad hoc arbitrations administered by the PCA. Most SCC investment treaty arbitrations have their seat in Stockholm. Proximity to the administering institution does indeed offer certain logistical advantages and can make organising a merits hearing more expedient. The choice of particular institutional rules, or the agreement on a particular administering institution in the context of an Ad Hoc Arbitration, thus serves as a strong predictor of the eventually chosen seat.

Standards of Review of Investment Awards by Courts at The Seat

Of the criteria listed above, by far the most significant consideration in the selection of an appropriate seat of an investment treaty arbitration is the degree of deference accorded by local courts to awards of investment tribunals, and particular to their findings on their jurisdiction, which is the most frequent basis for challenges of investment awards before state courts. This section provides a high-level overview of the standard of review of arbitral awards in the most commonly chosen seats of arbitration and surveys the most important and recent jurisprudence.

England

The 1996 Arbitration Act sets out two main grounds for the set-aside of an investment treaty award rendered in England – lack of substantive jurisdiction (Section 67) and serious irregularity (Section 68). To date, there have been seven applications to set aside an investment arbitration award in England, three of which have been successful.

The first investment arbitration award to be challenged before English courts was the 2004 Occidental v. Ecuador award. Ecuador applied to set aside the award based on both Section 67 and Section 68 of the Arbitration Act. Occidental raised an objection that Ecuador’s challenge required the English court to interpret provisions of the BIT between the United States and Ecuador (to which the United Kingdom was not party), in contravention of a rule of English law making such an issue ‘non-justiciable’. This objection was denied by the High Court, and the Court of Appeal confirmed that an award rendered in England under an investment treaty was subject to the supervisory powers of English courts (even if the United Kingdom was not party to the relevant treaty and it did not form part of English law). The Court of Appeal also did not consider that courts should exercise any sort of ‘judicial restraint’ in reviewing the award.

Although it was allowed to proceed, Ecuador’s challenge to the award before English courts, which concerned the scope of the US–Ecuador BIT’s tax carve-out, was ultimately unsuccessful. The English courts found that Occidental’s claims did fall within the BIT’s scope (as it fell within the scope of an exception to the treaty’s tax carve-out). In Occidental, the Court of Appeal reaffirmed that the test for review of the tribunal’s jurisdictional decision under Section 67 of the Arbitration Act was whether the tribunal was ‘correct’ in its decision on jurisdiction, and that this is achieved by way of rehearing the matters before the arbitrators. The standard of review of jurisdictional decisions by English courts is thus that of a de novo review, which in practice can mean a fairly costly trial before English courts on jurisdictional matters (potentially involving fact and expert witness evidence).

Despite the wide-ranging powers of review of investment treaty tribunals, English courts have, until recently, mostly agreed with tribunals’ jurisdictional findings and have favoured a relatively expansive interpretation of jurisdictional provisions in investment treaties (in line with English courts’ pro-arbitration approach generally). In Czech Republic v. European Media Ventures SA, Mr Justice Simon rejected the Czech Republic’s argument that the arbitration clause under the Belgium and Luxembourg Economic Union–Czech Republic BIT was only limited to the quantification of damages incurred and did not extend to establishing whether the breach occurred. In The Kyrgyz Republic v. Stans Energy Corporation and Another, the court agreed with the tribunal’s interpretation of the scope of application of Kyrgyzstan’s Foreign Investment Law.

Several English courts have disagreed with investment tribunals’ findings. In GPF GP SÀRL v. The Republic of Poland, the investor succeeded in setting aside a jurisdictional award that had limited the scope of the arbitration clause in the Belgium and Luxembourg Economic Union–Poland BIT to expropriation claims only. Mr Justice Bryan considered that the arbitration agreement under the BIT was broader in scope and included claims of breach of fair and equitable treatment as well, and substituted the tribunal’s jurisdictional finding with his own. Similarly, in Gold Pool JV Limited v. The Republic of Kazakhstan, Mr Justice Baker set aside an award that had declined jurisdiction over a claim against Kazakhstan under the Canada–USSR BIT on the grounds that Kazakhstan was not a successor to this treaty; Baker J found that the 1992 Declaration on Economic Co-operation between Canada and Kazakhstan evidenced the parties’ agreement to be bound by this BIT and that, accordingly, the tribunal was wrong to decline jurisdiction.

In Republic of Kazakhstan v. World Wide Minerals Ltd & Anor, an English court for the first time (partially) set aside an investment award under Section 68 of the Arbitration Act. In that case, Judge Pelling QC found that the tribunal awarding damages to the defendants by reference to an argument that the defendants had not advanced during the hearing or prior written pleadings amounted to a serious procedural irregularity, set aside the damages part of the award and remitted it to the tribunal for further consideration.

Overall, while not overly deferential to investment tribunals, English courts have applied their pro-arbitration stance to the review of investment treaty awards, and England continues to be a safe seat for investment treaty arbitration.

France

France has been one of the most popular seats of investment arbitration, and the relatively large number of set-aside applications in relation to resulting awards (eight of which were wholly or partially successful) relates to that fact. French courts may set aside an award on one of the following grounds:

1. the arbitral tribunal wrongly upheld or declined jurisdiction;
2. the tribunal was irregularly constituted;
3. the arbitral tribunal ruled without complying with the mandate conferred on it;
4. the due process requirement was violated; or
5. recognition or enforcement of the award would violate international public policy.

In reviewing investment awards, French courts have taken the same approach as in relation to commercial arbitration awards and have considered all elements of law and fact permitting to establish the scope of the arbitration agreement.

Of the set-aside applications that have recently been successful, three concern the awards being tainted by evidence of bribery or money-laundering, reflecting a particularly robust approach adopted by French courts in relation to these matters. In particular, in Ghenia and Sorelec, consent awards embodying settlement agreements were set aside because of the fact that the underlying settlements were allegedly tainted by fraud. In Belokon, allegations were made by Kyrgyzstan that the Kyrgyz bank that belonged to the investor was involved in money-laundering activities; although the arbitral tribunal did not find that sufficient evidence was present to make a finding of money laundering, the French court considered that ‘grave indicia’ of money laundering were present (including the fact that the profits of the bank were disproportionate to the state of the Kyrgyz economy; in the view of the French court, ‘such an astonishing success, in such a brief period of time, in such a poor country, was not explainable by orthodox banking practices’). As a result, the court considered that the award would allow Mr Belokon to ‘benefit’ from illegal activities, which would be contrary to international public policy, and set aside the award on that basis.

However, some more recent decisions suggest that French courts may be now taking a more moderate approach in respect of awards allegedly tainted by corruption. For instance, in Nurol Insaat Ve Ticaret Anonim Sirketi v. Libya, the French court refused to set aside a jurisdictional award under the Turkey–Libya BIT, where the investor had allegedly engaged in bribery; the court considered that such allegations of corruption pertain to the merits of the case rather than to the jurisdiction of the tribunal and thus should be decided by the tribunal itself during the merits phase.

Several other investment awards have been set aside on jurisdictional grounds.

In Rusoro, the French court partially set aside the damages part of the award insofar as the damages concerned a time prior to the relevant BIT’s entry into force. This decision was subsequently overturned by the Court of Cassation, which found that the matter pertained to admissibility and not jurisdiction, and the relevant part of the award could thus not be set aside on the basis that the tribunal had exceeded its jurisdiction.

In García Armas, the French court found that a jurisdictional award in favour of two dual nationals in an investment treaty claim against Venezuela must be set aside in its entirety because the claimants were not Spanish nationals at the time when their initial investments were made. Despite the fact that the claimants had already withdrawn claims in the arbitration in respect of initial investments made at a time when they were not Spanish nationals (and were awarded damages in relation to subsequent investments), the court declined to set aside the award only partially and chose to annul the entire award. However, subsequently the Court of Cassation overturned the decision of the lower court, finding that it improperly added the condition that the investors hold Spanish nationality at the time of the making of the investment, and remanded the decision back to the Court of Appeal.

In Oschadbank, the French court set aside a US$1 billion award against the Russian Federation in relation to the expropriation of Ukrainian nationals’ assets in Crimea, finding that the Ukraine–Russia BIT did not protect investments made before 1992 and that the investments in question were indeed made before this date.

Additionally, in DS Construction FZCO, the French court set aside an award under the Organisation of Islamic Cooperation Investment Agreement (the OIC Agreement) on the basis that the tribunal had been irregularly constituted. The tribunal had been appointed by the Permanent Court of Arbitration as an appointing authority further to the ‘import’ of the UNCITRAL ad hoc arbitration rules via the most-favoured nation (MFN) clause in the OIC Agreement.

The relatively large number of successful set-aside applications of investment awards before French courts could potentially call into question the finality of investment arbitration and the attractiveness of Paris as a seat of investment arbitration.

Switzerland

Being a popular seat of investment arbitration, Switzerland has likewise seen a number of challenges to investment awards, all but one of which have so far been unsuccessful.

The grounds for setting aside arbitral awards are stated in Article 190 of Chapter 12 of the Federal Statute on Private International Law and are as follows:

1. if the sole arbitrator was not properly appointed or if the arbitral tribunal was not properly constituted;
2. if the arbitral tribunal wrongly accepted or declined jurisdiction;
3. if the arbitral tribunal’s decision went beyond the claims submitted to it, or failed to decide one of the items of the claim;
4. if the principle of equal treatment of the parties or the right of the parties to be heard was violated; and
5. if the award is incompatible with public policy.

The low number of investment awards set aside to date may be at least in part because of the comparatively restrained and deferential approach to the review of decisions of investment tribunals adopted by Swiss courts. For instance, in the RECOFI case, the claimant was unsuccessful in its arbitration against Vietnam and was seeking to set aside the award before the Swiss Federal Tribunal on the basis that the tribunal allegedly misinterpreted the definition of investment in the France–Vietnam BIT. When interpreting the scope of protected investment as part of the set-aside proceedings, the Swiss Federal Tribunal did not approach it de novo, but rather, noting that the definition of investment under the BIT was interpreted by ‘three arbitrators whose experience in the matter and international reputation are recognised by both parties’, it would not ‘needlessly depart from the unanimous opinion expressed by specialists’.

In Clorox, Switzerland’s only set-aside decision to date, the Swiss Federal Tribunal set aside an award in favour of a state that had declined jurisdiction over a dispute, finding that the arbitral tribunal had inappropriately implied criteria for protected investments that were not found in the text of the BIT in question (namely, that the investment be ‘actively’ invested in the host state) and remitted the case back to the tribunal.

Some of the most recent set-aside applications before Swiss courts were based on allegations that the underlying investment was tainted by corruption or fraud and that the resulting awards were thus incompatible with public policy. In Ukrnafta, the Russian Federation sought to set aside an award arising out of the expropriation of assets of Ukrainian nationals following the annexation of Crimea; it argued that one of the investors’ shareholders ‘came into his fortune through fraudulent schemes and corruption’ and that, therefore, an award in favour of the investor was contrary to public policy. The Federal Tribunal rejected this argument, noting that the Russian Federation failed to bring such evidence before the arbitral tribunal, and it was not open to it to bring it before the Swiss courts.

Most recently, Swiss courts refused to set aside an award against Spain in favour of a group of solar investors. Spain had argued that the tribunal’s refusal to take into account the Achmea decision of the CJEU and re-examine its jurisdiction over the dispute (after having already rendered a preliminary award rejecting Spain’s objections to the tribunal’s jurisdiction concerning the invalidity of intra-EU BITs) breached Spain’s right to be heard and that the award was as such contrary to public policy. The Swiss Federal Tribunal rejected the application, noting that the arbitral tribunal’s 2018 procedural order declining to re-examine its jurisdiction should have been appealed before Swiss court within 30 days and that Spain had therefore filed its set-aside application too late, and that in any event the tribunal gave reasons for dismissing the request to reconsider its jurisdictional decision and, therefore, could not be accused of violating Spain’s right to be heard or causing a denial of justice.

Swiss courts have therefore predominantly adopted a rather light-touch approach to the setting aside of investment treaty awards, largely deferring to tribunals in their determination of both jurisdictional and procedural matters and limiting the scope of matters that can be relitigated after the tribunal rendered its award.

Netherlands

The Dutch Arbitration Act provides for two forms of recourse against arbitral awards: setting aside and revocation.

An award can only be set aside on the following grounds:

1. absence of a valid arbitration agreement;
2. the arbitral tribunal was composed in violation of the applicable rules;
3. the arbitral tribunal has manifestly not complied with its mandate;
4. the award was not signed or did not contain reasons in accordance with the Dutch Code of Civil Procedure; or
5. the award, or the manner in which it was made, violates public policy.

An award may be revoked if it is wholly or partially based on fraud committed in the arbitration, on forged records that turn out to have been forged after the award was made, or if relevant documents have been withheld by the other party.

A relatively limited number of set-aside applications have been brought in relation to investment treaty awards in the Netherlands. These have generally been ultimately unsuccessful, although some were successful in the court of first instance, only to be overturned by higher-level courts.

The most high-profile case is the one in relation to the Yukos award. The first instance court set aside the award for lack of jurisdiction under the Energy Charter Treaty (ECT), disagreeing with the tribunal’s finding that the ECT provisionally applied to Russia. In clarifying the standard of review of investment treaty awards by Dutch courts, the Hague District Court noted that although the arbitral tribunal ‘was qualified to assess its jurisdiction’, ultimately the Dutch court had to establish whether or not there was a valid arbitration agreement. Given the fundamental importance of the right of access to courts, such an assessment is not done ‘restrictively’, and the onus is on the defendants, as the party seeking to establish the tribunal’s jurisdiction, to prove that the tribunal is competent. This standard was confirmed by the Hague Court of Appeal (which noted, however, that it did not need to decide which party bears the burden of proof as to the existence of the arbitration agreement). Dutch courts thus effectively adopt a standard of de novo review, with no particular deference accorded to the tribunal’s determination of the relevant matters.

The decision of the Hague District Court was subsequently overturned by the Hague Court of Appeal, which reinstated the award. More recently, the Dutch Supreme Court rejected all of the Russian Federation’s grounds for appeal except for one: it considered that the Court of Appeal had erred when it prevented the Russian Federation from raising a fraud allegation on procedural grounds. This fraud allegation is now before the Amsterdam Court of Appeal.

In 2021, in the García Armas case, the Hague Court of Appeal confirmed that, in set-aside applications concerning the existence of a valid arbitral agreement as well as compliance with due process, Dutch courts will depart from the general principle that ‘the State court must exercise its power to set aside an arbitral award with restraint’ and that, with respect to these questions, ‘the arbitral award is subject to a full review’. However, this standard only appears to apply in a one-way sense, when an arbitral tribunal has made a finding that it has jurisdiction. When a tribunal has reached a conclusion that it does not have jurisdiction, Dutch law does not give courts a power of de novo review of such a finding, because ‘right of access to justice is not at stake’. In these circumstances, the court may not engage in a ‘substantive assessment of the arbitral tribunal’s judgment that it lacks jurisdiction’ and has no right to set aside the award ‘if it arrives at a different conclusion than the arbitral tribunal’. The Hague Court of Appeal thus refused to set aside a jurisdictional award where the tribunal found that it had no jurisdiction over claims brought by dual Spanish–Venezuelan nationals against Venezuela.

Dutch courts have thus adopted a fairly invasive standard of review when it comes to awards establishing jurisdiction, but a hands-off approach to the review of awards declining jurisdiction, arguably tilting the playing field in an anti-arbitration direction.

Sweden

The grounds for setting aside an award rendered in Sweden are listed in Section 34 of the Swedish Arbitration Act and are as follows:

1. the award is not covered by a valid arbitration agreement between the parties;
2. the award has been made by the arbitrators after the period set by the parties has expired or the arbitrators have otherwise exceeded their mandate;
3. the arbitral proceedings should not have taken place in Sweden;
4. an arbitrator has been appointed contrary to the parties’ agreement or the Swedish Arbitration Act;
5. an arbitrator was unauthorised because he or she did not possess full legal capacity in relation to his or her actions and his or her property, or because he or she was not impartial; or
6. another irregularity had occurred during the course of the proceedings, without fault of any party, and that irregularity may have influenced the outcome of the matter.

Additionally, under Section 33 of the Swedish Arbitration Act, an award is invalid if (1) it includes a question that, according to Swedish law, may not be decided by arbitrators, (2) the award, or the way in which it was made, is clearly incompatible with the basic principles of the Swedish legal system, or (3) it does not fulfil the requirement in relation to its written form and signature.

Although more than a dozen set-aside applications in relation to investment treaty awards have been brought in Sweden, only three have been ultimately set aside. This may be in part because of the application of the ‘doctrine of assertion’ to the question of existence of an arbitration agreement by Swedish courts in some cases, whereby:

arbitrators, when ruling on their jurisdiction, should not examine the existence of any facts which the claimant alleges establish a legal relationship that is covered by the arbitration agreement. To establish jurisdiction for the arbitral tribunal, it is thus sufficient that the claimant asserts that the facts that allegedly create jurisdiction exist.

The assertion doctrine was endorsed by the Swedish Supreme Court in Petrobart (I), in which the claimant sought to set aside the award of the tribunal declining jurisdiction under Kyrgyzstan’s Foreign Investment Law. Although the Svea Court of Appeal, applying a standard of de novo review, upheld the tribunal’s jurisdictional finding, the Supreme Court reversed its decision, applying the doctrine of assertion, and found that the tribunal should have established jurisdiction.

However, some Swedish court decisions have cast doubt on the applicability of the doctrine of assertion, adopting an approach closer to de novo review. In Petrobart (II), the same claimant commenced proceedings against Kyrgyzstan under the ECT. Kyrgyzstan sought to set aside the award in favour of the Gibraltar-incorporated investor primarily on the basis that the ECT does not apply to Gibraltar, because the United Kingdom (which is in charge of Gibraltar’s external relations) had not ratified the ECT on behalf of Gibraltar, even though the United Kingdom had previously provisionally accepted that the treaty applied to Gibraltar. The Svea Court of Appeal considered de novo whether there was a valid arbitration agreement between the parties, hearing witness and expert evidence on the matter. It reached the same conclusions as the tribunal and refused to set aside the award.

In Quasar de Valores, the Svea Court of Appeal overturned a decision of the district court (which had applied the doctrine of assertion and concluded, based on a prima facie review, that a valid arbitration agreement was in place) and conducted a de novo review of the matter. The arbitration in question had been commenced under the Spain–Russia BIT, which contained an arbitration clause limited in scope to the determination of the amount or method of payment of compensation for expropriation. The tribunal interpreted the clause as also extending to the question of whether compensation is ‘due’. The tribunal rejected the investor’s argument that the BIT’s MFN clause could be used to ‘import’ a broader dispute resolution provision from another treaty. The Svea Court of Appeal reached the same conclusion as the tribunal on the MFN issue, but disagreed with its finding on the scope of the dispute resolution clause and set aside the award.

In some of the more recent set-aside applications, Swedish courts have been grappling with the issue of compatibility of investment treaty awards under intra-EU BITs with Swedish law. In PL Holdings, Poland sought to set aside an award under the Luxembourg–Poland BIT, citing the Achmea decision of the CJEU, on the basis that resolving such a dispute in arbitration was contrary to Swedish ordre public and not being capable of being resolved in arbitration. The Svea Court of Appeal rejected Poland’s objections and also found that they had not been raised in a timely manner. However, the Swedish Supreme Court considered that it was unclear how EU law should be interpreted with regard to the issues that had arisen in this case. On that basis, it referred to the CJEU the question of whether, in light of the Achmea case, where there was an intra-EU BIT but the state had failed to make a valid objection to the tribunal’s jurisdiction, there could still be a valid arbitration agreement.

On 11 February 2021, the Svea Court of Appeal granted a request by Italy to seek a preliminary ruling from the CJEU before ruling on the state’s application to set aside an ECT award in favour of Danish and Luxembourgish investors. This was a departure from the position previously adopted by the Svea Court of Appeal, which refused to make a request for preliminary ruling by the CJEU in three earlier set-aside applications made by Spain. However, the Svea Court of Appeal later considered that it was no longer necessary to maintain this request for a preliminary ruling, in light of the CJEU’s decisions in Komstroy v. Moldova and PL Holdings v. Poland.

These decisions highlight the risks of choosing a seat in the European Union post-Achmea in intra-EU arbitrations.

Overall, although the standard of review of investment awards has not been applied consistently by Swedish courts in all cases, they appear to be leaning towards de novo review. Nevertheless, Swedish courts have exercised these powers relatively sparingly. It remains to be seen what effect the recent trend towards referring questions for preliminary determination to the CJEU would have on Sweden’s role as a popular investment arbitration seat, particularly in intra-EU disputes.

United States

The grounds for setting aside arbitration awards in the United States are set out in Chapter 1, Section 10 of the Federal Arbitration Act (FAA) and are as follows:

1. the award is a result of corruption or fraud;
2. evident partiality or corruption of an arbitrator;
3. arbitrator misconduct, such as refusing to hear pertinent and material evidence; or
4. the arbitrators exceeded their powers, or so imperfectly executed them that a mutual, final and definite award was not made.

US federal courts are split as to whether, in addition to these four statutory grounds, a ‘manifest disregard of the law’ doctrine remains a separate basis for setting aside an award under the FAA. The DC Circuit (where most set-aside applications in relation to investment treaty awards are heard) has expressed doubt about the application of this doctrine, while the Second Circuit (which encompasses New York) has found that ‘manifest disregard’ remains a valid ground for vacating arbitration awards.

To date, all set-aside applications in relation to investment treaty awards brought before US courts have been unsuccessful.

The leading authority on the standard of review of investment treaty awards by US courts is the Supreme Court decision in BG Group Plc v. Republic of Argentina, which held that reviewing courts cannot review arbitral awards de novo but, rather, they must do so with ‘considerable deference’. In the underlying arbitration, the arbitrators found that they had jurisdiction over the dispute despite the fact that BG had not complied with the 18-month local court litigation requirement under the UK–Argentina BIT before commencing the arbitration. The Supreme Court overturned the decision of the Court of Appeals of the District of Columbia, which had considered that the local litigation requirement was a matter for the court to decide de novo (i.e., without deference to the arbitrators’ views).

Subsequent US courts considering set-aside applications in relation to investment treaty awards all applied a similarly deferential standard of review. For instance, in AWG v. Argentina, the US District Court for the District of Columbia rejected Argentina’s challenge to the award based on the tribunal allegedly exceeding its powers by failing to apply applicable law in its computation of damages and its evaluation of the necessity defence, noting that Argentina could not overcome the ‘high hurdle requiring deference to the Tribunal’s determinations’. In China Heilongjiang v. Mongolia, the US District Court for the Southern District of New York adopted a deferential standard of review of the award, in circumstances where the question as to the scope of issues covered by the China–Mongolia BIT’s arbitration clause had been clearly put before the arbitral tribunal.

In light of the deferential standard of review of investment treaty awards adopted by US courts (particularly post-BG), US seats, notably Washington, DC, and New York, should be considered safe choices for investment arbitration.

Canada

A number of investment treaty awards (mostly, but not exclusively NAFTA awards) have been challenged before Canadian courts (notably, the courts of Ontario but also of British Columbia).

The International Commercial Arbitration Act in both Ontario and British Columbia has implemented the UNCITRAL Model Law on International Commercial Arbitration (the UNCITRAL Model Law). Thus, the grounds for setting aside an arbitral award in these jurisdictions correspond to the grounds set out in Article 34 of the UNCITRAL Model Law, namely:

1. if the party making the application furnishes proof that:
2. a party to the arbitration agreement was under some incapacity, or the agreement is not valid;
3. the party making the application was not given proper notice of the appointment of an arbitrator or of the arbitral proceedings or was otherwise unable to present his or her case;
4. the award deals with a dispute not contemplated by or not falling within the terms of the submission to arbitration, or contains decisions on matters beyond the scope of the submission to arbitration, provided that, if the decisions on matters submitted to arbitration can be separated from those not submitted, only that part of the award that contains decisions on matters not submitted to arbitration may be set aside; or
5. the composition of the arbitral tribunal or the arbitral procedure was not in accordance with the agreement of the parties; or
6. the court finds that:
7. the subject matter of the dispute is not capable of settlement by arbitration under the law of the seat of arbitration; or
8. the award is in conflict with the public policy of the state of the seat of arbitration.

Apart from an early partial set-aside of the award in Metalclad v. Mexico by the courts of British Columbia, all other applications to date to set aside treaty awards before Canadian courts have been unsuccessful. In particular, Ontario courts have dismissed applications to set aside the award by the investors in Bayview v. Mexico and Joshua Dean Nelson v. Mexico, as well as the respondent’s application to set aside in Cargill v. Mexico, Feldman v. Mexico, Luxtona v. Russia and B-Mex v. Mexico They have also refused to set aside awards when the application for review was made by the Canadian government in SD Myers Inc and Bilcon.

The leading Canadian case setting out the standard of review of investment treaty awards is the decision of the Ontario Court of Appeal in Mexico v. Cargill. Overturning a more deferential test of ‘reasonableness’ adopted by a lower level court, the Ontario Court of Appeal concluded that the standard of review of the award is ‘correctness’, ‘in the sense that the tribunal had to be correct in its determination that it had the ability to make the decision it made’. At the same time, the Ontario Court of Appeal noted that:

the fact that the standard of review on jurisdictional questions is correctness does not give the courts a broad scope for intervention in the decisions of international arbitral tribunals. To the contrary, courts are expected to intervene only in rare circumstances where there is a true question of jurisdiction.

The jurisdictional challenge in that case involved an argument by Mexico that the tribunal had no jurisdiction to award compensation for upstream losses as the losses were suffered by Cargill as a producer in the United States rather than as an investor in Mexico, which was dismissed by Canadian courts.

The correctness test has been applied by subsequent Canadian courts reviewing investment treaty awards, including in the Luxtona decision. Although the set-aside application by the Russian Federation against the Luxtona award is still pending, Ontario courts have considered the standard of review and the scope of evidence that can be considered by the court as part of the set-aside proceedings in a number of interlocutory decisions. In a 2018 decision, Mr Justice Dunphy noted that, ‘whether or not described as a hearing “de novo”’, the record before the court cannot be limited to the evidence that had been put forward before the arbitral tribunal. In a 2019 decision, Mr Justice Penny cited Cargill, noting that ‘in a jurisdictional challenge, the court must exercise care to ensure that the issue in the challenge characterized as “jurisdictional” is, in fact, an issue of “true” jurisdiction’. The judge described it as ‘a review on correctness, without any deference, in which the court must come to its own conclusion on whether the tribunal had jurisdiction, but a “review” nevertheless’ and distinguished it from a de novo standard of review by English courts. Penny J also limited the scope of the new evidence that could be presented in the set-aside proceedings.

More recently, in B-Mex, the Superior Court of Justice of Ontario applied the Cargill correctness standard and refused to set aside a jurisdictional award that had been challenged by Mexico on the basis that allegedly not all the investors who had brought the claim were named in the notice of dispute.

To conclude, Canadian courts, following Cargill, have adopted a ‘correctness’ standard in reviewing whether or not an investment treaty tribunal had jurisdiction. So far, Canadian courts have exercised their power of review sparingly, while at the same time not according deference to tribunals’ findings. Their approach as to the extent of new evidence that can be adduced as part of the set-aside proceedings appears to still be evolving.

Singapore

A jurisdiction that has been growing in popularity as a seat of arbitration is Singapore, which is an UNCITRAL Model Law jurisdiction. Investment treaty awards rendered there can thus be set aside based on the grounds listed in Article 34 of the UNCITRAL Model Law (as set out in Section V.(vii)) and on two additional grounds: (1) the making of the award was induced or affected by fraud or corruption; or (2) a breach of the rules of natural justice occurred in connection with the making of the award by which the rights of any party have been prejudiced.

To date, there have been three applications to set aside investment treaty awards in Singapore, only one of which has been successful.

In Sanum, Laos sought to set aside the jurisdictional award, arguing that the tribunal lacked jurisdiction because Sanum, an entity registered in Macao, was not a protected investor under the China–Laos BIT. Laos also asserted that the tribunal’s jurisdiction under the BIT was limited to disputes concerning the amount of compensation for expropriation. Although the Singapore High Court granted Laos’ set-aside application, the Singapore Court of Appeal reversed this decision and reinstated the award. On the BIT’s applicability to Macao, the Court of Appeal found, among other things, that China and Macao elected not to deviate from the Moving Treaty Frontiers Rules, under which China’s treaties would automatically apply to Macao as soon as it became part of China. On the scope of the tribunal’s jurisdiction, the Court of Appeal found that the BIT applied not only to disputes relating to the quantum of an expropriation claim but also to disputes on whether an expropriation had taken place. Similar to the position in England, the Singapore courts apply a de novo standard of review. In Sanum, the Court of Appeal further clarified that in conducting such a review, it will consider the tribunal’s findings, which may be persuasive, but it is not bound to accept its findings.

In the latest chapter of the Sanum saga, the claimants challenged the resulting merits awards in the arbitrations, where their claims were rejected by the tribunals on the grounds that the claimants had engagement in corruption. The claimants argued that the corruption defences had been introduced in breach of a 2014 settlement deed with Laos that had allegedly ‘frozen’ the claims and evidence in the arbitration. The Singapore International Commercial Court rejected the set-aside bid, finding that the settlement deed did not bar the tribunal from considering these allegations, or from accepting new evidence introduced by Laos. In any event, the Court found that the tribunal would have likely found illegality even without such evidence. Of particular note is the Court’s pronouncement that arbitration tribunals have a ‘public duty’ to hear and consider allegations of corruption, and that parties cannot escape that duty through procedural agreements.

In Swissbourgh and others v. Lesotho, Lesotho sought to set aside the award issued by the tribunal under the Southern African Development Community Protocol on Finance and Investment (the SADC Protocol) in favour of the investors. In its application before the High Court of Singapore, Lesotho argued that the claimant’s investment was not protected under the SADC Protocol. The Singapore High Court accepted Lesotho’s submissions and set aside the award.¹³⁷ The decision was later confirmed by the Singapore Court of Appeal.

While Singapore takes pride in being an arbitration-friendly jurisdiction, parties contemplating the choice of Singapore as a seat of investment treaty arbitration will need to bear in mind the de novo standard of review applied by Singaporean courts to the review of investment awards and the not-insignificant risk that such awards may be set aside.

Conclusions and Best Practice Recommendations

The importance of the choice of seat in non-ICSID investment arbitrations cannot be overstated. Unlike ICSID arbitration with its autonomous annulment mechanism and clearly specified grounds and standard for annulment, standards of review vary significantly in non-ICSID set-aside proceedings, depending on the law of the seat of arbitration. They range from complete de novo review (such as in England and Singapore) to a high standard of deference to tribunals’ findings (such as in Switzerland), with other jurisdictions falling somewhere between (with France, Netherlands and Sweden more inclined towards de novo review, whereas Canada and the United States take a more deferential approach).

In addition to significant variations in the standard of review, the approach taken by courts in practice also varies significantly. For instance, Dutch courts, while espousing a de novo standard of review, have in practice been relatively restrained in setting aside awards, whereas French courts have been much less so. As the recent wave of decisions setting aside investment awards in France demonstrates, the most popular seats of arbitration are not always the safest ones. The choice of a seat of investment arbitration can thus make a very significant difference to the outcome of the dispute and can turn a victory in the arbitration into a pyrrhic one.

A de novo standard of review, with potentially new evidence and lengthy hearings, would entail significant costs for both parties and delay the definitive resolution of the dispute. Additionally, not all judges in all jurisdictions possess the requisite knowledge and experience to substantively review the decisions of arbitral tribunals composed of experts in international law. Thus, in the interests of finality and saving costs, parties and tribunals would be well served to choose seats with a more deferential standard of review. Additionally, to ensure neutrality and a level playing field for the parties in intra-EU arbitrations, seats outside the European Union should be chosen for such arbitrations.

While logistical considerations are also important, parties and tribunals should not let them affect the choice of the best seat from a legal perspective. To the extent expedient, the hearing venue can and should be decoupled from the legal seat of the arbitration. This will allow the parties to have the best of both worlds and combine a hearing at a convenient venue with a safe and predictable legal environment.