In order to be truly effective, the reforms currently being pursued within the general investment regime must address both the content of investment treaty clauses and the way in which these clauses fit together. The actors undertaking these reforms should give particular attention to the relationship between these clauses for a variety of reasons, including those related to the characteristic features of “old-generation” treaties. In fact, these treaties are considered to be structurally unbalanced because they set out obligations that are burdensome for states alone while establishing rights in favour of investors only. Also, some of the more vaguely drafted clauses of these treaties have been extensively interpreted by arbitral tribunals to the benefit of investors but to the detriment of states. Logically, these two problems are compounded by the overlapping content of certain clauses. This overlap is actually favoured by the simultaneous presence in the same treaty of provisions of essentially similar object or purpose. Not only does this overlap disrupt the overall consistency of the treaty, but, more importantly, it extends the range of obligations supported by the host states and, by extension, the entitlements enjoyed by investors.
Unfortunately, these inconsistencies observed with standard treaties, which have caused countless practical problems for states, persist today with modern treaties. In this respect, a major uncertainty that must be attended to in these treaties relates to the wording of certain provisions such as the loss compensation clause or the full protection and security (FPS) clause. Indeed, as currently drafted, each of these two clauses, when taken in isolation first of all, presents practical difficulties in terms of interpretation and application. Such difficulties are then compounded when these clauses are applied in combination with others. Finally, the very question of the need to maintain these two clauses in the same treaty arises when one considers that they can sometimes be invoked at the same time (in times of armed conflict, domestic disturbances, or internal tensions). In general, these two clauses are a potential source of costly arbitration risks for states due to the diverging interpretations to which they are still subject. Today, investors are likely to invoke these clauses with considerably increased frequency due to the resurgence of crisis situations in several host states.
Investment treaty negotiators could therefore seize the opportunity offered by the various ongoing investment policy reform processes to give closer attention to the structural consistency of these treaties in general, and to these two provisions specifically. This brief reflection aims to draw the attention of those negotiating these types of treaties to the challenges presented by the linkages between the different clauses therein. The objective is to supply negotiators with useful information that specifically enables an improved drafting of these two types of clauses. The present note therefore provides a number of recommendations for rephrasing these clauses should the need arise and for minimizing the arbitration risks generated when the courts permit an uncontrolled extension of their scope.
Two questionably worded clauses that inject vagueness and confusion into treaties
Most investment treaties contain both a loss compensation clause and a protection and security (PS) clause. Each of these clauses can be drafted in a variety of ways.
First of all, in some treaties, the PS provision is dealt with in isolation from the other clauses, with a combination of the words “protection” and “security.” In other treaties, only one of these terms is used. Sometimes, this PS standard is expressed in association with other clauses in the same article. In all cases, the term PS is accompanied by a variety of qualifiers such as “constant,” “integral,” “total,” “full,”2 or the phrase “in accordance with the customary international law.” Usually, however, this standard is referred to as “full protection and security.” The writers generally settle for simply mentioning an FPS requirement without providing any further information, though others sometimes endeavour to more or less define its substance.
The loss compensation clause is also drafted in a variety of ways. In some treaties, the parties provide for an obligation to offer compensation to the investors, regardless of who caused the damage and regardless of the circumstances in which it occurred. Other treaties fail to specify the conditions under which the loss could be compensated. In yet others, the conditions for compensation are related to the origin, context, and nature of the injury suffered by the investors. Sometimes, the situations that can lead to the damage are listed in a restrictive fashion, and sometimes open-ended language is used to enumerate them. In practice, this enshrinement of broad obligations very often favours divergent, even dicey, interpretations.
These issues emanating from the use of open-ended language in the listing of the obligations owed by states toward investors become even more complex when one considers how these commitments can be combined with each other. Indeed, given that it is sometimes combined within the same article with other commitments such as those relating to fair and equitable treatment (FET), expropriation, or non-discrimination, the wording of the FPS clause can create additional confusion about the various obligations that these clauses impose upon the states. One such example is given by the wording of Article 5 of the Argentina–France BIT of July 3, 1991, which suggests that the FPS standard is a component of FET.
This confusion is also noteworthy when an FPS clause is laid down alongside a loss compensation clause in the same treaty. In such a situation, there exists a likelihood of confusion between the obligations set out by each. This confusion has, moreover, been put into perspective by two arbitrations. Indeed, in the case of AAPL v. Sri Lanka, the tribunal that interpreted Articles 2 and 4 of the Sri Lanka–United Kingdom BIT suggested that when an FPS standard and a loss compensation standard are set out within the same treaty, the latter must be considered as a special provision deviating from the former, which is assumed to be general, because the two standards relate to the same subject. This approach was also adopted by the tribunal that dealt with the case of L.E.S.I S.p.A. and Astaldi S.p.A. v. Algeria relative to the understanding of Article 4 of the Algeria–Italy BIT.
In summary, variations in wording coupled with sometimes confusing and ambiguous language in the loss compensation and FPS clauses are likely to create practical problems of application when host states are called upon to comply with the obligations that these clauses entail. These clauses have also undergone multiple interpretations by the courts, which sometimes tend to extend or restrict their scope.
Two clauses of uncertain scope: From extension to restriction by the courts
The FPS clause, as it continues to be drafted in investment treaties, remains problematic in its implementation and is a source of arbitration risk for the host state. Indeed, when drafting the clause, the writers tend to use wording such as “Each Party ensures, to the greatest extent possible,” “insofar as possible,” “reasonably,” “at all times . . . by reasonable means,” or “by adequate measures.” Sometimes it is the investment alone that appears to benefit from such an FPS provision, and sometimes this provision must be understood as benefiting both the investor and the investment. This indeterminacy of language raises the problem of the scope of the obligation incumbent on the host state. In this regard, most tribunals rightly consider that the obligation requires the host state to take a diligent course of action that is both preventive and specific in order to prevent damage that could be suffered by investments not only in normal situations but also in exceptional circumstances. But divergences persist among the tribunals regarding which harmful actions are covered by the clause. While for some courts the FPS clause must apply only to the actions of private persons, others consider on the contrary that its scope extends to those of state bodies. Yet others claim that the FPS clause covers both the actions of private persons and those of the state.
On a more concrete level, the divergences of interpretation concern the scope of the obligation, in particular the type of behaviour required of the state (precautionary efforts by state bodies, or vigilance with regard to third parties), the types of infringement prohibited (physical and/or non-physical), and the degree of diligence required in all cases (objective or subjective diligence). In addressing these issues, the courts have taken divergent approaches, thereby obscuring the scope of the obligation. This uncertainty therefore exposes the state to arbitration risks. This risk is on the increase for all host states that are currently dealing with armed conflicts, terrorist attacks, or phenomena such as coups, insurrections, riots, and various disturbances during which the defense and security forces of the state are at work to safeguard its essential interests. For example, in the cases of AMT v. Zaire and AAPL v. Sri Lanka, the investors went to arbitration to challenge the conduct of state bodies on the basis of the FPS clause. It is therefore difficult to predict when the clause will be invoked, unless it specifies, for example, which types of investment (tangible or intangible) and investor (natural person, legal person or employee) are covered. This vagueness of the material scope of the obligation is what has allowed certain tribunals to extend the scope of the clause, by considering that it is not limited to physical infringements but also extends to non-physical infringements.
As for the loss compensation clause, there is also considerable uncertainty regarding its legal scope. First of all, its different wordings in the various treaties make it impossible to know, a priori, whether or not it places a specific autonomous obligation upon the host state. The language used is simultaneously binding (“Each party shall grant”) and unconditional (“Each party will grant”). Through this type of language, the state places itself under an unconditional obligation. The risk presented by this manner of writing is that regardless of the circumstances in which the investments suffered losses and regardless of the conduct of the state at the time of the losses, an investor could arraign the state before an arbitration tribunal on the basis of this clause as soon as it claims to have suffered losses. Then, by accepting a non-discriminatory compensation commitment, states can be attacked in arbitration for violating the national treatment or most favoured nation treatment principle in the compensation. Still today, in some treaties, the loss compensation clause covers the measures enacted by the state for purposes of requisition. It requires the state to compensate investors for these requisitions, whatever the reasons behind these requisitions may be.
Thus, for either the FPS clause or the loss compensation clause, the absence of an exception in their wording leaves the state with excessively reduced room for manoeuvre because there is no longer any regard for considerations related to the situation of the state at the time of the events giving rise to the damage.
In this time of reform of the international investment regime, it is appropriate to take stock of these clauses in order to improve their wording should they continue to appear in new-generation treaties.
Two clauses worth reviewing: Removal versus reformulation
Despite these uncertainties and potential risks related to the implementation of these two clauses, the reforms currently under way pay them too little attention. Both the treaties in force and the recent model BITs unfortunately reuse conventional wording. This is the case, for example, for the Burkina Faso–Turkey BIT of 2019, the Canada–Moldova BIT of 2022, and the new model BITs of Canada (2021), Italy (2021), and Morocco (2019).
In this time of investment treaty reform, a pivotal question that must be answered by the relevant actors is whether it remains necessary for all of the conventional clauses of old-generation agreements to be systematically included in modern agreements. Especially for the two clauses examined here, states can also consider whether they should still be maintained and for what purpose. To this end, a prior assessment of the risks related to their insertion and the type of phrasing must be made. States could in this case assess the extent of the obligations to which they are committed through these two provisions in order to understand the possible consequences. In this respect, they can, for example, develop a number of scenarios that are plausible for their national context.
Should a state subsequently wish to maintain at least one of these provisions in an investment treaty, the following key considerations may be useful when restructuring and rewording efforts are made.
First of all, as regards the FPS clause, the negotiators should also ensure that its scope is clarified by indicating that the clause is restricted to actions of private persons that would create harm for investments, to the exclusion of actions carried out by the state. In addition, the drafters should specify that the infringements in question are those of a tangible nature, to the exclusion of legal infringements, in order to prevent the creation of overlaps between this clause and those pertaining to issues such as FET or expropriation. With respect to implementation, the drafters can specify that the required level of protection is that which the state could provide in light of its situation at the time of the events. Finally, they can choose to no longer consider the clause to be equivalent to the minimum standard of customary law but instead specify exactly what they would like it to cover. It is appropriate to do so because the two standards do not relate to the same purposes.
Next, concerning the loss compensation clause, it should be made clear that it is the state that will assess the appropriateness and the terms of compensation. The drafters should also consider limiting the material scope of the clause by clearly and exhaustively listing the exceptional circumstances in which investors may invoke it. The extent of compensation could also be limited by the non-discrimination mechanism (NT, MFNT). Finally, to ensure that the state benefits from some room for manoeuvre, the drafters can attach conditions to the exception clause to allow the state to grant compensation in accordance with its budgetary capabilities.
Last of all, in the interests of streamlining, the negotiators can consider putting a stop to the combination of these two provisions within the same treaty. Hence, it is advisable not to include a loss compensation clause if an FPS clause already exists. Since these two clauses can be invoked under similar circumstances (conflicts, internal disturbances, or similar events), the loss compensation clause creates an additional burden with regard to the obligations that states already fulfill under other provisions such as NT, MFNT, expropriation, or FET clauses.
Abas Kinda is an International Law Advisor at the International Institute for Sustainable Development.
 Unconditional compensation clauses present a stark contrast to the exceptions relating to the safeguarding of the essential interests of the state, which are sometimes referred to as security clauses.
 This is the case, for instance, in the following BITs: Mali–Algeria (Art. 4); Mali–Cameroon (Art. 4); Mali–Germany (Art. 4); Mali–Switzerland (Art. 3); Benin–Burkina Faso (Art. 2); Benin–Canada (Art. 7); Benin–China (Art. 2); Benin–Morocco (Art. 2); and Benin–United Kingdom (Art. 2).
 It is, for example, associated with the FET standard in Article 5 of the Argentina–France BIT of July 3, 1991; with expropriation in Article 6 of France’s model BIT of 2006 published by UNCTAD; and with the no-discrimination rule in the Treaty of friendship, commerce, and navigation between the United States of America and the Italian Republic of 1948 and the 1951 Supplement thereto.
 Article 10(1) of the Treaty establishing the EEC; also see the following BIT: Côte d’Ivoire–Belgium–Luxembourg Economic Union (Art. 3).
 This is the expression used in almost all of the BITs negotiated by Germany (French version of the texts).
 As is the case with the investment agreements signed by Algeria.
 As in Article 6 of the Canada–Mali BIT.
 Most agreements give little or no information about the possible content of this protection requirement.
 This is the case in Chapter 8, Section D, Article 8.10 of the trade agreement between the EU and Canada (Comprehensive Economic and Trade Partnership Agreement [CETA]).
 See Article 6 of the Belarus–Croatia BIT; Article 7 of the Canada–Croatia BIT; Article 5 of the Morocco–Bulgaria BIT; Article 4 of the United States–Cameroon BIT; Article 5 of the Côte d’Ivoire–Belgium–Luxembourg Economic Union BIT; Article 6 of the France–Iran BIT; and Article 5(4) of the Burkina Faso–Benin BIT.
 Article 7 of the Albania–Netherlands BIT.
 Article 5(2) of the Gabon–Morocco BIT.
 Article 12 of the ECOWAS Energy Protocol.
 De Nanteuil, A. (2010). L’expropriation indirecte en droit international des investissements. Université Panthéon-Assas.
 “Investments made by investors of either Contracting Party shall enjoy, in the Territory and in the maritime zones of the other contracting party; protection and security, pursuant to the principle of fair and equitable treatment referred to in article 3 of this Agreement.”
 L.E.S.I S.p.A. and Astaldi S.p.A. v. Algeria, ICSID Case No ARB/05/3, Award (November 12, 2008), paras 174–175, the Tribunal relying on the dissenting opinion of S. Asante in Asian Agricultural Products Ltd v. Sri Lanka, ICSID Case No ARB/87/3, Award (June 27, 1990)
 Saluka Invest. B.V. v. The Czech Republic, UNCITRAL, Partial Award, paras. 483–84 (Mar. 17, 2006; Asian Agricultural Products Ltd v. Sri Lanka, No. ARB/87/3, para. 50; Tulip Real Estate and Development Netherlands B.V. v. Republic of Turkey, ICSID Case No. ARB/11/28, Award, paras. 430–37 (Mar. 10, 2014), https://www.italaw.com/sites/default/files/case-documents/italaw3126.pdf; Allard v. Barb., PCA Case No. 2012-06, Award, paras. 240–250 (June 27, 2016), https://www.pcacases.com/web/sendAttach/1955; Técnicas Medioambientales Tecmed, S.A. v. The United Mexican States, ICSID Case No. ARB (AF)/00/2 Award, para. 177 (May 29, 2003), 19 ICSID Rev.— FILJ 158 (2004).
 Amco Asia Corporation and Others v. The Republic of Indonesia, Award, November 20, 1984, 1 ICSID Reports 413; Eastern Sugar v Czech Republic, Partial Award, March 27, 2007, para 203. Italics original.; Parkerings v. Lithuania, n 4, para 355.; Wena Hotels v. Egypt, Award, December 8, 2000, 41 ILM 896 (2002).
 Biwater Gauff v. Tanzania, Award, July 24, 2008, para. 730.
 Saluka Investments BV (The Netherlands) v. The Czech Republic, Partial Award, 17 March 2006, paras. 483, 484; PSEG v. Turkey, Award, 19 January 2007, at paras. 257–259; Eastern Sugar v. Czech Republic, Partial Award, 27 March 2007, para. 203.
 As for the situation involving Ukraine and Russia.
 For most countries in sub-Saharan Africa, such as Burkina Faso, Mali and Niger.
 United Nations Conference on Trade and Development. (2009). The protection of
national security in IIAs. UNCTAD Series on International Investment Policies for Development. https://unctad.org/system/files/official-document/diaeia20085_en.pdf
 American Manufacturing & Trading, Inc. v. Republic of Zaire, ICSID Case No. ARB/93/1, §§ 6.07-6.11; Asian Agricultural Products Ltd v. Sri Lanka, ICSID Case No. ARB/87/3, § 3.
 CIRDI, Biwater v. Tanzania, 2008, §. 729 ; Renée Rose Lévy de Levi v. Republic of Peru, ICSID case No ARB/10/17, Award (26 February 2014), para 406 ; Azurix Corp. v. The Argentine Republic (ARB/01/12), decision on the recusal of the President of the Tribunal, February 25, 2005 and award of July 14, 2006.
 Schreuer, C, (2013). The protection of investments in armed conflicts. Cambridge University Press, p. 10.