Evaluating (In)Consistency In Investor-State Arbitration: A Roadmap

Clinic: Queens University and UOttawa, Spring 2018

Beneficiary: Global Affairs Canada

Students making a presentation on the project (above).


Read the full report here.

The uOttawa students of the uOttawa-Queen's Joint Clinic carried out two projects for Global Affairs Canada related to the UNCITRAL Working Group III on Investor-State Dispute Settlement Reform.


The United Nations Commission on International Trade Law (UNCITRAL) gave its Working Group III the responsibility to work on the reform of investor-state dispute settlement (ISDS). The mandate of the working group is to identify the growing concerns surrounding ISDS, to determine whether reforming the ISDS system is advisable, and if so, to make recommendations to the UNCITRAL Commission.

Both projects undertaken by the uOttawa-Queen's Joint Clinic are focused on the threshold question under discussion in UNCITRAL Working Group III: whether investor-state awards have been inconsistent, and whether the ISDS system is inconsistent and incoherent, as alleged by critics.

Focus of the Projects:

Project 1: Evaluating (In) Consistency in Investor-State Arbitration: A Roadmap - The project provides a roadmap for examination of whether there is inconsistency of investor-state awards and if so, why?

Project 2: Étude de Cas: L'Incohérence dans La Jurisprudence sur La Norme Detraitement National en Driot International des Investissements Étrangers - The French memorandum provides a case study of the investor-state awards on the National Treatment principle.

The summary of Project 1 is provided below. To read and download the full report, please visit here. To read more about Project 2, see here.

Executive Summary

Conflicting decisions by tribunals arbitrating investor-State disputes has led to critics proposing overhauling the dispute settlement system. However, the investor-State dispute settlement system is responsible for applying roughly 3,000 treaties, each possessing their own characteristics that could justify apparently inconsistent decisions.

Substantial Inconsistency Attributable to Treaty: Our study of identified areas of controversy and alleged inconsistency show that different outcomes in three areas can be explained by the different treaties applicable in each case. However, even in those cases, the tribunals do not decide with perfect consistency or, in one instance, relies on a detail in a treaty that may not justify their conclusion differing from previous decisions. Specifically:

  • Most Favoured Nation: These clauses can be subdivided into six different types of clauses and the vast majority of tribunals determine whether procedural provisions can be incorporated based on the type of clause at issue.

  • Denial of Benefits: These clauses can be subdivided into three different types, and, with a few exceptions, tribunals consistently determine that benefits can be denied retrospectively based on the type of clause at issue.

  • Sovereign Bonds as Investments: These bonds are consistently found to be investments where the treaty’s definition of “investment” refers to public sources of obligations, but one tribunal found sovereign bonds were not investments where the treaty did not specify public obligations.

Most Inconsistency is Unjustified: In eight areas, we find that the inconsistent decisions cannot be explained by differences in treaty language or facts. Instead, these tribunals applied different interpretative approaches or analysis to arrive at widely different conclusion despite similar laws and facts in each.

  • CME and Lauder v the Czech Republic: Each tribunal applied a different analysis of causation and only one found that the State action, identical in each case, caused a compensable loss to the investor.

  • Countermeasures: The inconsistent decisions arose from the characterization of the investors rights under the treaty: If the investor has independent rights from the home State, then those rights cannot be denied as a countermeasure targeting the home State. Otherwise, countermeasures can be validly enacted.

  • Fair and Equitable Treatment: These clauses vary widely but even among tribunals applying the same treaty, there exists inconsistency in the content and level of the clause’s obligations.

  • ICSID’s Definition of Investment:” Some tribunals apply ICSID Article 25(1) as a limit on the “investments” within the tribunal’s jurisdiction and apply one of several variations of the Salini test, and apply varying tests to individual elements of that test. Other tribunals only limit their jurisdiction according to their empowering investment treaty.

  • Sales Contracts as Investments: Some tribunals apply the Salini test and come to varying determinations. Other tribunals, despite broad treaty language, categorically presume that sales and services contracts are not investments.

  • Awards as Investments: Despite similar treaty language, tribunals arrive at different conclusions based on whether an arbitral award should be treated as distinct from the contract from which it arose and whether the Salini test applies.

  • Umbrella Clauses: Tribunals applying similar provisions differ on their applicable to commercial contracts and the impact of exclusive jurisdiction clauses within those contracts.

Improving with Time and Continued Arbitration: Some of these areas of inconsistency may, over time, warrant less concern. There are three areas where the decisions would be consistent but for a single tribunal, or even a single tribunal member. And there is one area where an inconsistent decision has already lost much of its relevance based on the growing body of tribunals who have rejected its reasoning.

  • Awards as Investments: GEA v Ukraine is the only tribunal to find that an award is analytically distinct from the contract it arose from and is not an investment. Later tribunals have not followed their analysis.

  • Countermeasures: Archer Daniel Midlands Co and Tate & Lyle Ingredients Americas, Inc v Mexico was the only tribunal to find, on the key issue, that investors did not have individual rights included a dissent on that issue.

  • Sovereign Bonds: Only the Postova Banka v Greece tribunal found sovereign bonds were not investments and there was an unpublished or unwritten dissent, potentially on that issue.

  • Umbrella Clauses: The first tribunal to address the issue, SGS v Pakistan, has been frequently criticized, rarely followed, and commentators suspect the position may have been abandoned within investor-State arbitration.

Vague Language Leads to Inconsistency: Other areas of inconsistency can potentially be attributed to vague clauses in the treaties, where a lack of guidance for tribunals unsurprisingly results in varying outcomes.

  • Procurement: In NAFTA, procurement is defined in some chapters but undefined in Chapter 11. As a result, tribunals either apply the definition from another chapter to Chapter 11, or interpret the absence of a definition as meaning that the ordinary meaning of the word is intended to apply.

  • Fair and Equitable Treatment: The lack of clearly defined legal obligations contained in treaties, amplified by the variations in language among these clauses, leave tribunals to determine the meaning and scope of the clause.

Newer Treaties Provide Clarification: Many modern treaties now include language that helps to address these areas of inconsistency however our review of tribunals’ analysis shows that tribunals will depart from applying a treaty provision’s ordinary meaning. This tendency to depart has been, and will likely continue to be, a significant source of inconsistency.

Rebuttable Presumptions: Our review of tribunals’ analysis suggests that some tribunals appear to operate on a presumption that can be difficult to displace. Some tribunals, and even some commentators, hold presumptions about the distinction between contract and treaty law, the effect of various types of clauses, and what assets are presumably excluded from investment-treaty protection.

  • Umbrella Clauses: The SGS v Pakistan tribunal held a presumptive distinction between domestic and international law and so was reluctant to elevate contractual breaches to treaty breaches.

  • Most Favoured Nation: These clauses are often seen as being homogenous, rather than as a general descriptor for six different types of clauses which should lead to different results. Where a treaty uses language that expressly expands the clause, then tribunals tend to find the clause can incorporate procedural provisions from other treaties.

  • Sales Contracts as Investments: The Romak v Uzbekistan tribunal indicated that a treaty must “leave no room for doubt” in its language in order to include sales contracts as investments, indicating that including such contracts would be counter-intuitive.

A Higher Threshold for Interpretation: The analysis also indicates how these presumptions can be displaced. Wording that is “clear and convincing” or leaves “no room for doubt” may be sufficient to ensure that tribunals apply their empowering instruments according to the treaties’ ordinary meaning and address the areas of unjustified inconsistency in investor-State arbitration.

We have found significant areas of inconsistency. Some inconsistency is largely attributable to differences in law. There are promising signs that some decisions may become outliers and the greater body of decisions will be consistent. The clarified language in recent treaties should increase consistency in the future. However, current clarified language may not be enough. In order to ensure consistent decisions, treaty drafters should use not just clear, but clear and convincing language expressing their intent in order to ensure tribunals comply with the letter of the law.

To read and download the full report, please visit here.

#Investment #Arbitration #ISDS #UNCITRAL #uOttawaQueens #uOttawa #GAC #Negotiation