INDIRECT EXPROPRIATION AND THE PROTECTION OF PUBLIC INTERESTS

Abstract

This article analyses what is widely known as the police powers exemption found in modern international investment agreements, with a focus on mega-regional trade and investment agreements. It explores its legal nature and requirements, the burden of proof and issues of compensation. In an attempt to curb (indirect) expropriation claims, the exemption carves out non-discriminatory regulatory measures from the scope of indirect expropriation and in such situations no compensation needs to be paid. It follows that, as a rule, foreign investors are not protected against the adverse economic effects of regulatory measures. The key question addressed is whether host States can increase the level of protection given to public welfare objectives through the use of this exemption without having to compensate investors for the measures taken. The article argues that, under the proportionality test embodied in the exemption, States can provide the level of protection that they desire without incurring a risk of liability as regards expropriation claims.

I. Introduction

Protecting foreign investments from uncompensated seizure is one of the primary objectives of international investment agreements (IIAs).¹ At the same time, States have the right, and in many situations even the duty (eg under international human rights law), to regulate in the public interest.² Doing so may encroach upon proprietary rights of foreign investors. As pointed out in an UNCTAD study, ‘almost any governmental measure could be construed as an act of interference in the business of a foreign investor’.³ Differentiating expropriatory measures from regulatory ones is one of the most vexing issues of international investment law. There is an underlying tension between private and public interests, between individual entitlements and national sovereignty, which is also germane to an understanding of the fair and equitable treatment (FET) standard.

It is important to stress at the outset that no IIA prohibits expropriation per se.⁴ It is lawful to expropriate subject to certain conditions. Those conditions are: (i) pursuance of a ‘public purpose’; (ii) ‘in a non-discriminatory manner’; (iii) ‘on payment of … compensation’; and (iv) under ‘due process of law’.⁵ The conditions are cumulative (‘and’).⁶ That is, a lawful expropriation presupposes compensation in addition to public purpose, non-discrimination, and due process.⁷

At first blush, this only relates to the legality of expropriation; it does not bear upon its existence. States have sought to clarify the definition of expropriation in modern IIAs. To that effect, annexes on expropriation set out criteria to determine what amounts to expropriation and criteria the presence of which militates against measures amounting to expropriation.⁸ The ratio legis of the annexes is to provide tribunals with more guidance, or more to the point, to restrict their discretion when finding that expropriation has taken place, especially indirect expropriation.⁹ Accordingly, the provisions on expropriation ‘shall be interpreted in accordance with’ the relevant annexes.¹⁰ But the interplay of those annexes with the provisions on expropriation remains unclear.

Factors to be taken into account by tribunals are: the economic impact of the measure at issue; the extent of the government interference; the character of the measure (ie its ‘objective’ and ‘object, context and intent’);¹¹ and under the EU-Canada Comprehensive Economic and Trade Agreement (CETA), the duration of the measure. These factors are cumulative.¹² Ordinarily, international dispute settlement bodies are reluctant to probe States’ intent.¹³ That is why the Appellate Body in Japan—Alcoholic Beverages II, for instance, rejected an aim-and-effect test in the context of the national treatment (NT) obligation in Article III of the General Agreement on Tariffs and Trade (GATT).¹⁴ In the words of the Appellate Body, ‘It is irrelevant that protectionism was not an intended objective if the … measure in question is nevertheless … “applied to imported or domestic products so as to afford protection to domestic production”.’¹⁵ This differs from the position concerning indirect expropriation, as ‘intent’ is explicitly listed as a relevant factor in CETA and the United States-Mexico-Canada Agreement (USMCA). Another clarification—the fact that the measure ‘has an adverse effect on the economic value of an investment … does not establish that an indirect expropriation has occurred’¹⁶—rules out the application of the sole effect doctrine, according to which the effect of the measure at issue is the key criterion to determine expropriation, and is helpful in light of awards such as Metalclad v Mexico.¹⁷ Still, tribunals have considerable discretion, first, because the lists of factors are non-exhaustive (‘among other factors’), and secondly, in deciding whether the factors are present.¹⁸

All this forms part of a wider attempt by States to properly calibrate investment protection.¹⁹ Whether States have achieved their goal as far as indirect expropriation is concerned will be examined in this article. In doing so, it will concentrate on the provisions concerning expropriation found in the following mega-regional trade and investment agreements: CETA, the Comprehensive and Progressive Agreement for Trans-Pacific Partnership (CPTPP),²⁰ the USMCA, and the Regional Comprehensive Economic Partnership Agreement (RCEP).²¹ Because of their size and the number of participating countries, mega-regionals set trends. It should be noted that an investor–State dispute settlement (ISDS) mechanism does not yet exist under RCEP (but is part of its work programme).²² Barring covered government contracts,²³ indirect expropriation is not actionable under the USMCA,²⁴ thus only leaving State–State dispute settlement.²⁵ Also, neither the CETA provision on expropriation nor its investor–State dispute resolution mechanism are operational yet.²⁶ Although this article focuses upon mega-regionals, similar provisions are also employed in bilateral investment treaties (BITs), notably treaties based upon the India Model BIT 2015.²⁷

The provisions that remove non-discriminatory regulatory measures from the definition of indirect expropriation have major implications for the domestic legal order.²⁸ For example, paragraph 3(b) of Annex 14-B to the USMCA provides that ‘[n]on-discriminatory regulatory actions by a Party that are designed and applied to protect legitimate public welfare objectives … do not constitute indirect expropriations, except in rare circumstances’. The Annex distinguishes regulatory actions from indirect expropriation and thus restricts the scope of the latter. States do not have to pay compensation to foreign investors in respect of such measures because they are not expropriatory,²⁹ whereas for expropriatory measures compensation is a requirement of their legality.³⁰

While the agreements under consideration all use similar wording, there are differences. CETA, the CPTPP and the USMCA make allowances for ‘rare circumstances’ whereas RCEP carves out non-discriminatory regulatory measures ipso facto. This raises the question of whether States were merely codifying existing case law along the lines of Burlington v Ecuador—where the tribunal held that measures justified under the police powers doctrine are not expropriatory³¹—or whether they actually deviated from it: does the carve-out provide domestic regulators with more or less policy space than the indirect expropriation standards established by case law?

The article will proceed as follows: Part II explores the legal nature of the last paragraph of the annexes on expropriation and draws inferences from them concerning the steps of the legal analysis. Part III considers the legal requirements of the carve-out and the resulting level of scrutiny of domestic law, drawing upon a comparison with the law of the World Trade Organization (WTO). The key question addressed is whether host States can increase the level of protection given to legitimate public welfare objectives without incurring liability for expropriation. Part IV examines the ‘rare circumstances’ caveat, while Part V considers the burden of proof under the carve-out. Part VI looks at the question of compensation for regulatory measures in depth, with a particular focus on the ramifications of the already infamous ruling in Eco Oro v Colombia.³² Part VII concludes.

II. The Legal nature of the police powers doctrine and its implications

The last paragraph of the annexes on expropriation, such as paragraph 3(b) of Annex 14-B to the USMCA, embodies the police powers doctrine.³³ What is the legal nature of that last paragraph, which according to some has consequences for the distribution of the burden of proof?³⁴ Property is a fictio iuris; it is normative and needs to be fleshed out by law.³⁵ It is fair to say that international investment law impacts upon the concept of property not only through the definition of investment but also through its protection standards. Domestic property rights are protected internationally provided they constitute an investment, a threshold requirement under IIAs.

Furthermore, property is not an absolute right.³⁶ For instance, it is undisputed that proprietors may have to pay taxes on their property, such as real estate taxes.³⁷ The crux of the matter is determining when delineating the boundaries of property is tantamount to an encroachment. A domestic law which the investor considers as amounting to government interference, from the perspective of the host State, might amount to the mere delineation (regulation) of property rights, for example, the right to use the investment.

The way the last paragraphs of the annexes are formulated implies that they form part of the definition of indirect expropriation (‘do not constitute indirect expropriations’). As a result, they need to be analysed when determining whether an expropriation has taken place.³⁸ They are consequently not exception clauses,³⁹ but limit the scope of what amounts to an expropriation and, as such, constitute exemptions or carve-outs.⁴⁰ The differentiation between exemptions and exceptions is also known from WTO law.⁴¹ According to the Panel in EC—Approval and Marketing of Biotech Products, ‘the term “exemption” connotes freedom from, and hence the inapplicability of, an obligation’.⁴²

Any legal analysis must start with the various agreements’ respective annex on expropriation, including the exemption.⁴³ It is only after a finding has been made that a measure is expropriatory in accordance with the elements set out in the annex that the issue of legality, covered in the main body of their investment chapters, arises. If there is no expropriation, the question of its legality (and compensation) is irrelevant. As emphasized by UNCTAD, ‘When the expropriatory nature of the measure is being opposed, it cannot be expected that the host State makes a pre-emptive payment.’⁴⁴ In line with this, the arbitral tribunal in Allard v Barbados noted that ‘the first requirement for a successful claim of indirect expropriation is that … the investor [should] have suffered a substantial deprivation of the use or expected economic benefit of the investment’.⁴⁵ The implications of some of the legality requirements also being definitional elements will be discussed later.⁴⁶

It can be inferred from the exemptions that regulatory action may amount to indirect expropriation since the definitional carve-out only relates to regulatory action that is ‘non-discriminatory’ and ‘designed and applied to protect legitimate public welfare objectives’. These cumulative⁴⁷ elements will be considered in the next section.

III. Elements of the exemptions

A. Regulatory Actions

Before considering what is meant by the ‘non-discriminatory’ element and the ‘designed and applied to protect’ test, a word on ‘regulatory actions’ is in order. Whereas CETA merely speaks of ‘non-discriminatory measures’, the CPTPP, the USMCA and RCEP refer to ‘non-discriminatory regulatory actions’.⁴⁸ The omission in CETA does not suggest a different scope, just as there is no difference in meaning between ‘measures’ and ‘actions’, both referencing acts of State within the meaning of the International Law Commission's Articles on State Responsibility.⁴⁹ One commentator raises the question of ‘whether the term “regulatory measures” refer[s] only to generally applicable laws and regulations, or whether it encompasses both general measures and specific measures targeted at the investor’.⁵⁰ This question will be discussed in the next section on non-discrimination.

B. Non-Discrimination

For a measure to be exempt under the carve-outs, the measure must be non-discriminatory. The exemptions thus combine the concept of indirect expropriation with the concept of non-discrimination. Discriminatory State conduct is also captured by other provisions in IIAs. The most-favoured-nation (MFN) and NT obligations are concerned with nationality-based discrimination.⁵¹ A Drafters’ Note to Articles 9.4 and 9.5 of the TPP confirms that these provisions ‘seek to ensure that foreign investors or their investments are not treated less favourably on the basis of their nationality’.⁵² Non-discrimination as part of the FET standard, by contrast, is wider and includes discrimination on grounds ‘such as gender, race or religious belief’.⁵³

There appears to be an inherent tension between the elements of the exemptions, on the one hand, and the requirements of a lawful expropriation, on the other hand.⁵⁴ This is because non-discrimination and the pursuance of legitimate public welfare objectives are both prerequisites of a lawful expropriation.⁵⁵ Article 14.8.5 of the USMCA stipulates that ‘whether an action or series of actions by a Party constitutes an expropriation shall be determined in accordance with paragraph 1 of this Article and Annex 14-B (Expropriation)’,⁵⁶ which seems to suggest that the legality requirements are also germane to a finding of expropriation. The requirement of ‘public purpose’ would be met if the measure at issue protected a public welfare objective in terms of the Annex. Both requirements are equally open-ended and, therefore, equivalent. That said, their interchangeability is not undisputed.⁵⁷ It is suggested here that whilst ‘public welfare objectives’ is the modern concept, drafters continue to use the term ‘public purpose’ in the provisions on expropriation for customary reasons.⁵⁸

This duplication of requirements creates friction between the annexes and the provisions on expropriation: on the one hand, it is said that non-discriminatory regulatory measures are non-compensable (except in rare circumstances), whilst the provisions on expropriation make compensation a requirement if they are to qualify as non-discriminatory expropriatory measures for a public purpose. Put differently, as with the provisions on expropriation, compensation is required in the case of expropriatory measures which are for a public purpose and non-discriminatory. Under the carve-outs, regulatory measures are exempt if ‘designed and applied to protect legitimate public welfare objectives’—that is, for a public purpose—and non-discriminatory. Assuming there to have been due process, there appears to be no difference between both categories of measures—compensable expropriatory and non-compensable regulatory measures—except for the premise that one category is expropriatory and the other is regulatory in nature.⁵⁹ This calls the added value of the exemptions into question.

Conflicting case law illustrates the dilemma. Although it concerns direct expropriation, a good starting point is the Santa Elena v Costa Rica case. There, the tribunal highlighted the fact that because ‘public purpose’ is a legality requirement, on its own it is insufficient to demarcate compensable expropriation from non-compensable regulatory action: ‘where property is expropriated, even for environmental purposes …, the state's obligation to pay compensation remains’.⁶⁰ This pronouncement assumes expropriation. We are concerned, however, with whether there has actually been an expropriation, which logically precedes the question of legality.⁶¹ This is when the police powers doctrine comes into play.

The OECD defined that doctrine as ‘an accepted principle of customary international law that where economic injury results from a bona fide non-discriminatory regulation … compensation is not required’.⁶² The doctrine's reach is contested. On the one hand, the arbitral tribunal in Methanex v USA held that ‘a non-discriminatory regulation for a public purpose, which is enacted in accordance with due process and, which affects, inter alios, a foreign investor or investment is not deemed expropriatory and compensable …’.⁶³ On the other hand, given how rare direct expropriations are in practice,⁶⁴ the tribunal in Pope & Talbot noted that ‘a blanket exception for regulatory measures would create a gaping loophole in international protections against expropriation.’⁶⁵ By the same token, the tribunal in Vivendi v Argentina (II) observed that ‘If public purpose automatically immunises the measure from being found to be expropriatory, then there would never be a compensable taking for a public purpose.’⁶⁶

The tribunal in Saluka v Czech Republic steered a middle course and ruled that ‘States are not liable to pay compensation to a foreign investor when, in the normal exercise of their regulatory powers, they adopt in a non-discriminatory manner bona fide regulations that are aimed at the general welfare.’⁶⁷ This, of course, begs the question of what constitutes the ‘normal’ exercise of regulatory powers. In a similar vein, the arbitral tribunal in Quiborax v Bolivia opined that ‘the taking must not qualify as the legitimate exercise of the State's police powers’.⁶⁸ Case law has tried to reify ‘normal’/‘legitimate’ in this context,⁶⁹ and the exemptions have codified that case law:⁷⁰ the exercise of police powers is legitimate when it is, first, non-discriminatory, and secondly, ‘designed and applied to protect legitimate public welfare objectives’.

How is sense to be made of this? If there was no difference between the legality requirements of expropriation and the requirements of the exemptions, this would mean that the exemptions are little more than guidelines for treaty interpreters to the effect that, under normal circumstances, regulatory actions are not compensable.⁷¹ Hence, only the ‘rare circumstances’ caveat would have practical relevance, which would also entail that the RCEP exemption—by omitting that caveat—is inconsequential, as it could not resolve the internal conflict noted above.

The solution is to distinguish between non-discrimination as a requirement of the legality of expropriation and as a requirement of the exemptions. As to the legality requirement, the arbitral tribunal in Teinver v Argentina found that ‘Discrimination requires differential treatment of Claimants’ investment from other similar investments in like circumstances.’⁷² To find discrimination in this connection, something more is required than the simple designation of the investment by the host State for expropriation. An expropriatory act, such as a seizure, will always be addressed to, and in that sense target, an investor. This alone would not make the act unlawful. The salient point is whether the host State would have acted differently if the investor was someone else.

It should be noted that, in the context of the exemptions, ‘non-discriminatory’ does not assume a rational relationship between the measure at issue and the public welfare objective pursued. As will be seen later, this is covered by the ‘designed to protect’ element.⁷³ Returning to the OECD definition, it defined non-discrimination as part of the police powers doctrine as follows: ‘A state measure will be discriminatory if it results “in an actual injury to the alien … with the intention to harm the aggrieved alien” to favour national companies.’⁷⁴ However, this nationality-based definition does not add anything to the MFN/NT obligations, nor to the legality requirement for expropriation: an MFN/NT breach would automatically make the taking of property unlawful, because ‘No Party shall expropriate …, expect … in a non-discriminatory manner’.⁷⁵

The exemptions hark back to Article 11(a)(ii) of the Convention Establishing the Multilateral Investment Guarantee Agency, the relevant part of which reads as follows: ‘with the exception of non-discriminatory measures of general application which governments normally take for the purpose of regulating economic activity in their territories’.⁷⁶ The accompanying commentary defines ‘non-discriminatory measures of general application’ as including ‘taxation, environmental and labor legislation as well as normal measures for the maintenance of public safety’.⁷⁷ The ‘general application’ part was omitted in the exemptions, which suggests that it was deemed superfluous by the drafters in light of the non-discrimination requirement. The CPTPP adds to that list of measures of general application ‘competition, … health or other regulatory laws’.⁷⁸

Against this backdrop, the reference to ‘non-discriminatory’ measures in the exemptions, and by extension the definition of indirect expropriation, appears to make a distinction between measures of general application and measures that single out the claimant investor.⁷⁹ In Eco Oro v Colombia, the tribunal deduced from the fact that ‘The measures affected all mining concessionaires’ that they were non-discriminatory.⁸⁰ To be able to bring a claim under CETA, the CPTPP, or the USMCA, the investor must have suffered ‘loss or damage’ by reason of the alleged breach, making the incurrence of loss or damage an admissibility requirement.⁸¹ This has an impact on this. If others, be they nationals or foreigners, suffer the same loss or damage as a result of the measure at issue, this is an indication that the measure is of general application, in other words, that it constitutes a regulatory measure.⁸² By understanding the term ‘discriminatory’ here as meaning that the measure is not of general application—which is different from the investment protection standards where it refers to differential treatment on grounds of nationality, gender, race, etc—the principle of effectiveness is respected.⁸³

In summary, non-discrimination in the context of exemptions does not refer to nationality-based differential treatment, as in the MFN and NT obligations as well as the legality requirement for expropriation itself.⁸⁴ Rather, it implies that the measure at issue is of general application, understood as meaning the ‘non-selective application of laws and regulations’.⁸⁵ This means that such measures were not designed to apply specifically to the claimant investor. The fact that, for example, the revocation of an operating licence may be non-compensable does not militate against this finding, as the underlying law upon which the revocation was based might itself be of general application. That some IIAs have no MFN and NT obligations does not change the meaning of non-discrimination in the exemptions either, because the omission of those obligations is to be respected as reflecting the common intention of the contracting parties.⁸⁶

The next section will consider the ‘designed and applied to protect’ standard, including the proportionality test which it embodies and which is key to determining whether a measure of general application is non-compensable under the exemptions. It is key because it determines the level of scrutiny of, or, in other words, the level of intrusion into, domestic law.

As for the question of ‘objectives’, the arbitral tribunal in Oxus Gold v Uzbekistan observed that ‘in case of indirect expropriation, there is a State law or regulation, or sometimes some behaviour, – the purpose of which was not to take the property’.⁸⁷ Just like the legality requirement of ‘public purpose’, a public welfare objective's legitimacy (and hence its protectability in principle) will rarely be challenged by claimant investors, bearing in mind that the list of public welfare objectives is non-exhaustive (‘such as’).⁸⁸ The Preambles, which form part of the relevant context within the meaning of Article 31(2) of the Vienna Convention on the Law of Treaties (VCLT),⁸⁹ expand that list further.⁹⁰ This allows host States to pursue multiple public welfare objectives with the same measure.⁹¹

C. Designed and Applied to Protect

All substantive investment protection standards subject domestic law to international scrutiny. The level of scrutiny by investment tribunals depends upon the type of proportionality test embodied in the ‘designed and applied’ standard. Suitability, necessity and strict proportionality need to be considered.⁹² It should be emphasized that the issue here relates to whether an indirect expropriation has taken place, not whether the amount of compensation is proportionate.

The arbitral tribunal in Electrabel v Hungary defined proportionality as requiring ‘the measure to be suitable to achieve a legitimate policy objective, necessary for that objective, and not excessive considering the relative weight of each interest involved’.⁹³ According to De Brabandere and Baldini Miranda da Cruz, ‘Proportionality is a general principle of international law that presupposes that an examination of a possible conflict between different rights will follow some general rational rules that have the goal of ensuring that even restricted rights will not be completely suppressed.’⁹⁴ Both in domestic and international law, the principle, or technique,⁹⁵ of proportionality helps settle conflicts between competing rights and interests in a systematic manner.⁹⁶ Often, such conflicts are at the root of an investment dispute: the expansion of a mining project within First Nation land affects Indigenous rights; the construction of a hotel complex next to a national park affects the environment; conversely, climate change measures imposing emissions reductions affect investments in the energy and primary industries sectors.⁹⁷ Claimants regularly challenge the policy choices and prioritization decisions made by host States under the heads of FET and indirect expropriation.

It has long been recognized that proportionality is instrumental in determining whether there is indirect expropriation.⁹⁸ The crux of the matter is what is or is not proportionate.⁹⁹ This hinges, in particular, upon the meaning of ‘designed … to protect’. A suitability test would be a low threshold.¹⁰⁰ Bradshaw summarizes that test as follows: ‘In order to be suitable, … the measure must be capable of furthering the policy objective. This requires a rational connection between the desired end and the means (i.e. the regulatory measure) chosen to advance it.’¹⁰¹ Rajput notes that a suitability test is ‘akin to the standard of reasonableness’.¹⁰²

WTO jurisprudence has concretized the concept of necessity. It is comprised, inter alia, of a ‘determination of whether a WTO-consistent alternative measure which the Member concerned could “reasonably be expected to employ” is available, or whether a less WTO-inconsistent measure is “reasonably available”’.¹⁰³ This has been endorsed by investment tribunals.¹⁰⁴ Necessity in this context, as an element of proportionality, is not to be confused with a circumstance precluding wrongfulness in terms of Article 25 of the ILC's Articles on State Responsibility under which, in order to be deemed necessary, the measures adopted need to be ‘the only way’ for the State to protect the public welfare objective in question.¹⁰⁵

A strict proportionality test would be the most intrusive approach since it allows claimant investors to challenge the host State's decision to prioritize a particular public welfare objective. If successful, the investor's interest trumps the protected objective. An investor could, therefore, bring about a reduction of the level of protection offered to a public welfare objective in order to protect its own, competing, interest because, in the words of the arbitral tribunal in PL Holdings v Poland, the measure is ‘excessive in that its advantages are outweighed by its disadvantages’.¹⁰⁶ Under a strict proportionality test, the adjudicatory body weighs up ‘the charge or weight imposed to the foreign investor and the aim sought to be realized by any expropriatory measure’.¹⁰⁷

The ‘designed and applied’ test has two elements: the design element and the application element. The former concerns the measure itself, the latter relates to the way the measure has been implemented. These will be addressed separately. It should be noted that the ‘designed and applied’ test is not the only way in which non-discriminatory regulatory measures are carved out from the scope of indirect expropriation. Canada's 2021 Model Foreign Investment Promotion and Protection Agreement (FIPA) employs the following clause: ‘A non-discriminatory measure … that is adopted and maintained in good faith to protect legitimate public welfare objectives … does not constitute indirect expropriation … .’¹⁰⁸ So far, however, this language has not been employed in any agreement.¹⁰⁹

1. Designed to protect

The design element is indicative of a less strict proportionality test. A strict proportionality test would be irreconcilable with the much simpler ‘designed and applied’ test used in the exemptions and must be considered to have been superseded by that test. In ‘rare circumstances’, however, a stricter test may still apply.¹¹⁰ The rejection of proportionality sensu stricto is also rooted in the fact that an investment tribunal is not a constitutional court and does not have the same powers.¹¹¹ This more deferential reading of investment protection standards is borne out by pronouncements of recent arbitral tribunals. For instance, the tribunal in Muszynianka v Slovakia found that ‘it is not the role of an arbitral tribunal to “weigh the wisdom of legislation” … arbitral tribunals must pay deference to the choices States make when deciding how to implement policy objectives’.¹¹²

a) Suitability

‘Designed to protect’ implies, at the very least, a suitability test, this being the first step of proportionality. The measure at issue must pursue a legitimate public welfare objective. But this alone would not be sufficient to distinguish non-discriminatory regulatory action from indirect expropriation, because for an expropriation to be legal, it must be ‘for a public purpose’, in other words, pursue a legitimate public welfare objective.¹¹³ Returning to the Santa Elena v Costa Rica case, the tribunal said that ‘While an expropriation or taking for environmental reasons may be classified as a taking for a public purpose, and thus may be legitimate, the fact that the Property was taken for this reason does not affect either the nature or the measure of the compensation to be paid for the taking.’¹¹⁴ Henckels concludes that ‘whether a direct expropriation has been effected for a public purpose is only relevant insofar as it affects the issue of whether an expropriation is lawful or unlawful, and not the issue of whether an expropriation has taken place’.¹¹⁵ For regulatory measures, things are different: whether or not the host State pursued a public purpose (ie a legitimate public welfare objective) also forms part of the definition of indirect expropriation, and hence is germane to the question of whether the measure at issue is expropriatory in the first place.

That said, the design element requires more than just the identification of the legitimate public welfare objective. On the spectrum of scrutiny, the more stringent step would be to consider whether the measure at issue is not incapable of promoting the protection of the stated objective. At this juncture, it is useful to canvass WTO law. Inferences may be drawn from international trade law because of the systemic function fulfilled by international investment law in the global economic system, which is complementary to international trade regulation.¹¹⁶

(1) Means–end nexus

WTO law also has a design element. Under WTO law this is a preliminary to the necessity test.¹¹⁷ The Appellate Body deduces this from the phrase ‘to protect’. One could equally argue that necessity encompasses suitability.¹¹⁸ In Colombia—Textiles, the Appellate Body observed with respect to Article XX(a) of the GATT that

the phrase ‘to protect public morals’ calls for an … examination in order to determine whether there is a relationship between an otherwise GATT-inconsistent measure and the protection of public morals. If this … examination reveals that the measure is incapable of protecting public morals, there is not a relationship between the measure and the protection of public morals that meets the requirements of the ‘design’ step.¹¹⁹

The Appellate Body thus established that the design element involves an examination of the relationship between the measure at issue and the public welfare objective pursued.¹²⁰ This is reminiscent of the ‘relating to’ test in Article XX(g) of the GATT, which requires ‘a close and genuine relationship of ends and means’.¹²¹ The Panel in Brazil—Taxation sums up the case law as follows: the ‘designed to protect’ element amounts to an analysis of ‘whether the measure is not incapable of contributing to that objective’.¹²² According to the Appellate Body, this requires an assessment of the measure's ‘content, structure, and expected operation’.¹²³ This final step, the ‘expected operation’, is expressly covered by the ‘applied to protect’ requirement in the exemptions.

(2) Is the measure apt to make a contribution?

According to the Appellate Body, the design element is not ‘a particularly demanding step’.¹²⁴ In WTO law, whether the measure makes, or will make, a contribution to the achievement of the stated public welfare objective is subjected to a test of necessity.¹²⁵ Therefore, under the exemptions, a respondent could argue that it does not matter if the measure at issue actually contributes to the protection of the policy objective as long as the measure was designed for that purpose. In other words, an adjudicatory body would not gauge the effectiveness of the measure, for even ineffective measures could be ‘designed … to protect … public welfare objectives’.¹²⁶ Indeed, with respect to ‘relating to’, the Appellate Body confirmed in China—Rare Earths that the term ‘does not require an examination of the actual effects of the challenged measure’.¹²⁷

The problem with applying this approach to the exemptions is that the only requirement that a measure would have to fulfil would be the pursuance of a legitimate public welfare objective. Given that the list of such objectives is non-exhaustive, this is a high—and in the author's view inadequate—standard of review, which accords the host State too much deference in the context of indirect expropriation.

Whether the measure is capable of making a contribution to the achievement of the public welfare objective, correctly understood, is a question of suitability and should be scrutinized as such.¹²⁸ This is buttressed by the way the Appellate Body defines that element: ‘A contribution exists when there is a genuine relationship of ends and means between the objective pursued and the measure at issue.’¹²⁹ It is true that the Appellate Body also stated that for the purposes of WTO law ‘using a “contribution” test for the “relating to” analysis, is not, by itself, an appropriate substitute for a holistic assessment of whether a measure has a close, genuine, and substantial relationship to’ a legitimate public welfare objective.¹³⁰ At the same time, however, the Appellate Body found that ‘consideration of the predictable effects of a measure may be relevant’.¹³¹

Indeed, both ‘relating to’ and ‘necessity’ connote varying degrees of a means–end nexus, with the former being looser.¹³² One of the main purposes of the annexes on expropriation is to clarify the means–end nexus required for regulatory measures, and hence the degree of proportionality required. Under German law, where the proportionality test originates,¹³³ the salient point under the suitability prong is whether the measure ‘could further the desired outcome’.¹³⁴ From this it follows, that a measure is not suitable if it cannot further the desired outcome, or impedes its achievement.¹³⁵

That said, whether or not a measure makes a ‘contribution’ to achieving the objective is only one aspect of the necessity analysis.¹³⁶ Other factors are ‘the importance of the objective’, the ‘restrictiveness of the measure’ and, as mentioned above, ‘a comparison between the challenged measure and possible alternatives should … be undertaken’.¹³⁷ It is, therefore, unquestionable that the ‘designed to protect’ test gives regulators more leeway than necessity,¹³⁸ which in addition requires those scrutinizing a measure ‘to assess, in a qualitative or quantitative manner, the extent of the measure's contribution to the end pursued, rather than merely ascertaining whether or not the measure makes any contribution’.¹³⁹ This has the advantage of more easily accommodating measures that have more than one regulatory objective.¹⁴⁰

As far as scrutiny goes, the ‘designed to protect’ standard thus sits between a suitability and a necessity test. If the drafters had wanted a fully-fledged necessity test, they could have simply said so: ‘actions by a Party that are necessary to protect …’. Instead, they adopted the ‘designed and applied’ language. What the Appellate Body stated with respect to Article XX of the GATT also holds true here: ‘In the light of the different connecting words used, we consider that a mixing of the different tests under Article XX(b) and Article XX(g), absent of context, would result in an approach that ignores the important distinctions between the various subparagraphs of Article XX’.¹⁴¹

(3) Conclusion

In conclusion, the measure at issue is designed to protect public welfare objectives if it is ‘not incapable’ of protecting them. This presupposes a genuine means–end nexus, akin to the ‘relating to’ standard known from the WTO exceptions. In the author's view, it involves an assessment of whether the measure is ‘capable of contributing’ to the stated legitimate public welfare objective, judged at the time when the measure was adopted.

Because the relevant moment in time to determine suitability is when the measure at issue was designed, that is, when the measure was adopted, it is immaterial for the purposes of the design element whether it later turns out to be insufficient to protect the public welfare objective. This is again different to the necessity test, where it is not the contribution as such but the potential of the measure to make a contribution that needs to be present when it is subject to review.¹⁴² Conversely, if the measure does make a contribution to the protection of the public welfare objective, this evidences its suitability,¹⁴³ bearing in mind that for some measures, such as measures designed to combat climate change, it may be some time before any such contribution becomes apparent.¹⁴⁴ It follows that host States have a margin of appreciation when it comes to making predictions about the future. As will be seen, States can, therefore, increase the level of protection given to public welfare objectives without the worry that this might be considered to amount to an indirect expropriation.

b) The level of protection

The exemptions are underpinned by the investors’ expectations: investors know that they will be subject to, and must comply with, regulatory measures that may interfere with their investments. ‘[T]he nature and extent of governmental regulation or the potential for government regulation in the relevant sector’ delineate the investors’ horizon of expectations when making the investment.¹⁴⁵ This, of course, is equally relevant to an FET claim as it is to an expropriation claim. As Khachvani stresses,

each society strikes a certain balance between specific sorts of economic interests, on the one hand, and common interests such as environment and health on the other. When an investor acquires its investment and related property rights in the host State, it accepts that balance as found.¹⁴⁶

This is common ground among investment lawyers. It seems generally accepted that States will not have to compensate in respect of measures which seek to maintain the level of protection for public welfare objectives which existed at the time of the investment.¹⁴⁷ But what if a host State wished to increase the level of protection offered, by, for example, imposing more stringent environmental standards or by re-zoning areas where mining may occur?

An increase of the level of protection raises difficult questions. Was the foreign investor supposed to ‘expect’ this change in attitude and approach to risk on the part of the host State? Put differently, should the costs of an increase of protection be borne by society as a whole or the individual (here the foreign investor)? Whether measures intended to provide such enhanced protection amount an expropriation depends on whether that increase in protection is covered by the exemptions.

Khachvani denies that they are: ‘If the State … alters the balance [between private interest and public welfare] in favor of public welfare and thereby renders the investor's assets economically unviable, it will have to pay compensation.’¹⁴⁸ Framing the question as a risk allocation issue, Khachvani contends that

The risk that changed circumstances may render a particular activity a nuisance is a risk that an entrepreneur should assume as a business risk. In contrast, the risk that the State may enhance standards of public welfare at the expense of limiting or banning an investor's activity is a sovereign risk.¹⁴⁹

The counterarguments are the phenomenon of regulatory chill and democratic considerations. A new government that is more environmentally sensitive than a previous government would be considerably hampered by this reading. For example, the environmental concerns of a society may and do change over time. Would this not be covered by the investors’ horizon of expectations as well?

In the author's view, the answer turns on whether or not host States are entitled under IIAs to set the levels of protection as under international trade law. As a preliminary point, it should be noted that regulatory chill is largely the result of broad interpretation of the FET standard. A finding of indirect expropriation, by contrast, presupposes a substantial deprivation of property interests.¹⁵⁰ Otherwise, the measures in question would not have ‘an effect equivalent to direct expropriation’,¹⁵¹ or, in the words of the tribunal in EMV v Czech Republic, ‘the measures … must, however indirect, amount to a “taking”’.¹⁵² Every encroachment upon property which is below that threshold, a loss in profitability or value for example, would not qualify.¹⁵³ This raises the question whether only the total loss of value or control would be deemed ‘substantial’. The tribunal in Busta v Czech Republic found that ‘loss amounting to two-thirds of the total value … would represent a substantial loss of economic value’,¹⁵⁴ whereas the tribunals in Total v Argentina and Mobil v Venezuela required ‘a total loss’.¹⁵⁵ Whilst losses below that high threshold may still fall foul of the FET standard, that is a different issue.¹⁵⁶ The bottom line is that if a regulatory measure does not substantially interfere with property interests, it is simply not expropriatory for that reason alone, and so there is no need to rely upon an exemption. It is important to bear this in mind when considering regulatory chill caused by indirect expropriation.

Under international trade rules, it is settled that not only can States determine the public welfare goals that they wish to pursue (eg environmental protection) but they can also set the level of protection (eg high level of environmental protection exceeding international standards).¹⁵⁷ A complainant could not claim that the level of protection afforded is excessive. Provided the public welfare objective is legitimate, its protection cannot be diminished through trade litigation. Government measures can only be challenged on grounds that there is a less trade-restrictive measure that achieves the same public welfare objective to the extent desired by the respondent. Stopping there (ie with necessity), as was the consistent case law of the WTO Appellate Body, removes the balancing component from the proportionality test, because no competing rights and interests are weighed against each other. In sum, there is no strict proportionality test in international trade law.¹⁵⁸ This is true of WTO and free trade agreements (FTAs) alike.¹⁵⁹

It has already been seen that strict proportionality is ruled out under the exemptions, except for ‘rare circumstances’. So, under normal circumstances, no compensation would need to be paid if the host State introduced measures to increase the level of protection given to a public welfare objective. Moreover, Khachvani's view leads to a conflict with European Union law. The European Court of Justice held in the CETA Opinion that the deep scrutiny of domestic law that ensues from a strict proportionality test (namely, ‘call[ing] into question the level of protection of a public interest’) is not compatible with EU primary law.¹⁶⁰

2. Applied to protect

While the design element is concerned with the substance of the measure at issue, the second element relates to the application of the measure, which must be undertaken in the same spirit of protecting public welfare objectives as the underlying measure, and includes implementing administrative regulations and procedures. Considering the substance of a measure and its implementation by domestic authorities separately is likely to have been influenced by WTO law, where the same distinction is made in the chapeau to the general exceptions¹⁶¹ and Article X:3(a) of the GATT. With respect to the latter, the Appellate Body found that ‘a distinction must be made between the legal instrument being administered and the legal instrument that regulates the application or implementation of that instrument’.¹⁶² From this it can be inferred that the second part of the test concerns ‘the substantive content of legal instruments that regulate the application or implementation of laws, regulations, decisions, and administrative rulings’.¹⁶³

As with the definition of the design element, the measure must be applied in a manner that does not contradict the public welfare objective.¹⁶⁴ This implies the continuous furtherance of the stated public welfare objective following the adoption of the measure, though bearing in mind that the contribution that the measure makes to that protection need not be evident at the time of review.¹⁶⁵ The requirement of continuousness is made clearer in Canada's Model FIPA (‘maintained’). The good faith test in the Model FIPA does not directly allude to proportionality and might, therefore, entail a lesser degree of scrutiny. However, this alone is unlikely to lead to different results, given the overlap between the principles of good faith and proportionality.¹⁶⁶ The more significant difference with the ‘designed and applied’ test arises from the omission in the Model FIPA of the ‘rare circumstances’ caveat, which will now be considered.

IV. Rare circumstances

The exemptions limit the concept of indirect expropriation. The arbitral tribunal in Pope & Talbot warned of the risk of making that concept inutile through an overly wide application of the police powers doctrine.¹⁶⁷ For that reason, some IIAs limit the limitations themselves: in ‘rare circumstances’, the exemptions do not apply. This technique reflects the difficulty of finding the proper calibration of investment protection.¹⁶⁸

The four mega-regional agreements under consideration come to different solutions: In CETA, the Schranken-Schranke—the limitation on the limitation—is confined to one ‘rare circumstance’, namely ‘when the impact of a measure or series of measures is so severe in light of its purpose that it appears manifestly excessive’.¹⁶⁹ This means that a non-discriminatory regulatory measure would be expropriatory (and hence trigger the host State's obligation to compensate) should this exceptional circumstance be met, along with the other requirements of expropriation (eg interference ‘with a tangible or intangible property right or property interest’).¹⁷⁰ The CPTPP and the USMCA refer to ‘rare circumstances’, in the plural, without specifying what such circumstances might be.¹⁷¹ One of those circumstances has been identified by the Methanex tribunal: ‘unless specific commitments had been given by the regulating government to the then putative foreign investor contemplating investment that the government would refrain from such regulation’.¹⁷² RCEP, like the ASEAN Comprehensive Investment Agreement¹⁷³ and the COMESA Investment Agreement,¹⁷⁴ omits the ‘rare circumstances’ proviso altogether and so has the most restrictive concept of indirect expropriation.¹⁷⁵ Under the CPTPP, USMCA, and CETA, the exemptions establish a rebuttable presumption.¹⁷⁶

Looking at the CETA exemption in more detail, that exemption does not apply if, due to its severe impact, the measure is ‘manifestly excessive’ in light of its purpose. This imports a strict proportionality test.¹⁷⁷ As has been seen, under such a test States cannot set the level of protection they desire; an increase of protection without the payment of compensation would not be sanctioned. However, this presupposes that there has been a substantial deprivation of ‘the fundamental attributes of property …, including the right to use, enjoy and dispose of its investment’, having ‘an effect equivalent to direct expropriation’.¹⁷⁸ That is, for the rare circumstance exception to be relevant, either the substantial deprivation must be excessive ‘in light of its purpose’ or the impact of the measure must exceed even a ‘substantial’ deprivation (ie amount to a total loss of value/control, or close to it). Otherwise, Khachvani would be right in pointing out that ‘If the regulation has no such substantial effect, it would not in any event qualify as expropriation and the State would not be under the primary obligation to compensate.’¹⁷⁹

Other BITs give similar examples of rare circumstances. For instance, the US–Korea FTA as well as the Trilateral Investment Treaty Between Japan, Korea and China mention the circumstance ‘when an action … is extremely severe or disproportionate in light of its purpose or effect’.¹⁸⁰ The Canada–Colombia FTA limits the exemption ‘when a measure … is so severe in the light of its purpose that it cannot be reasonably viewed as having been adopted in good faith’.¹⁸¹ Finally, the India–Korea Comprehensive Economic Partnership Agreement lists as a rare circumstance ‘when a measure … is extremely severe or disproportionate in light of its purpose or effect’.¹⁸² What those caveats have in common is the requirement of ‘a very significant aggravating element or factor in the conduct of the State’ that prompts the application of a strict proportionality test.¹⁸³

In the next section, the burden of proof under the exemptions will be discussed.

V. The burden of proof

As to burden of proof, the UNCITRAL Arbitration Rules 2013 provide that ‘Each party shall have the burden of proving the facts relied on to support its claim or defence’.¹⁸⁴ The same is true for the recently amended ICSID Arbitration Rules.¹⁸⁵ The CPTPP and the USMCA stipulate that ‘the investor has the burden of proving all elements of its claims, consistent with general principles of international law applicable to international arbitration’.¹⁸⁶ Because the allocation of the burden of proof in the CPTPP and the USMCA reflects a general principle (actori incumbit onus probatio),¹⁸⁷ the same applies to the other mega-regionals under investigation which do not have such an explicit rule.

It follows that the claimant investor must make a prima facie case that the measure at issue is expropriatory. This involves making a case that there has been a substantial deprivation of property interests. What is not clear is whether the investor's burden of proof also extends to the exemptions. The affirmative view is predicated upon the belief that exemptions constitute negative elements of the claim.¹⁸⁸ As a consequence, the claimant investor would have to show that the measure at issue was not designed and applied to protect the legitimate public welfare objective pursued by the respondent.¹⁸⁹ Case law in relation to the police powers doctrine seems to confirm this. In Servier v Poland, the tribunal found that ‘the burden … falls onto the Claimants to show that Poland's regulatory actions were inconsistent with a legitimate exercise of Poland's police powers. If the Claimants produce sufficient evidence for such a showing, the burden shifts to Poland to rebut it.’¹⁹⁰ In a similar vein, the WTO Panel in EC—Approval and Marketing of Biotech Products allocated the burden of showing that the requirements of Article 5.7 of the Agreement on the Application of Sanitary and Phytosanitary Measures, which it qualified as an exemption, were not met to the complainant. The Panel held that

Characterizing Article 5.7 as a qualified right and not an exception also has implications for the allocation of the burden of proof concerning the issue of the consistency of an SPS measure with Article 5.7 … [I]n cases where a complaining party alleges that an SPS measure is inconsistent with the obligation in Article 2.2 not to maintain SPS measures without sufficient scientific evidence, it is incumbent on the complaining party, and not the responding party, to demonstrate that the challenged SPS measure is inconsistent with at least one of the four requirements set forth in Article 5.7. If such non-compliance is demonstrated, then, and only then, does the relevant obligation in Article 2.2 apply to the challenged SPS measure.¹⁹¹

The opposing view takes the ‘rare circumstances’ caveat as a starting point. As has been seen, when the requirements of the exemptions are satisfied, they establish a presumption against the regulatory measure being expropriatory.¹⁹² The claimant investor can rebut that presumption by invoking a rare circumstance (in which case the measure would be deemed expropriatory). This means that the investor bears the burden of proving that the situation amounts to a ‘rare circumstance’.¹⁹³ From this, it can be deduced that the burden of proof with respect to the exemptions rests with the respondent State. It is for the respondent to prove their assertion that the measure at issue is a non-discriminatory regulatory action rather than an act of expropriation. Otherwise, the investor would bear the burden of proof for both the rule (ie the presumption) and the exception (ie its rebuttal).

Reflecting general principles of international law, the burden of proof regarding exemptions should be allocated as follows: the State should show that the measure at issue is of general application and pursues a legitimate public welfare objective. Should this be achieved, the burden then shifts to the investor to show that there is a ‘rare circumstance’ for their claim to be able to proceed.

VI. The obligation to compensate

It is undisputed that a foreign investor is entitled to compensation even for lawful expropriation, be it direct or indirect.¹⁹⁴ As a corollary, compensation is also due when an (eg discriminatory) expropriation is lawful as a result of an exception. At first blush, this conclusion seems counter-intuitive. It is relevant in relation to RCEP, where the general exceptions capture expropriation.¹⁹⁵ In the author's view, what the Santa Elena tribunal stated with respect to the environment can be applied to other legitimate public welfare objectives, that is, potential grounds of justification: ‘Expropriatory environmental measures—no matter how laudable and beneficial to society as a whole—are, in this respect, similar to any other expropriatory measures that a state may take in order to implement its policies.’¹⁹⁶ In such a situation, ‘society as a whole [and not the expropriated owner] should bear the economic burden of achieving’ a public welfare objective; and further, ‘society cannot advance its welfare at the expense of a few individuals when the cost for such advancement should be borne by society as a whole’.¹⁹⁷

The difference between this and unlawful expropriation is the kind (and amount) of compensation owed: fair market value if the expropriation was lawful,¹⁹⁸ if not, the investor's expectation interest (ie ‘any financially assessable damage including loss of profits insofar as it is established’).¹⁹⁹ The tribunal in Bear Creek Mining v Peru confirmed, in connection with an ‘unlawful indirect expropriation’, that a general ‘exception … does not offer any waiver from the obligation … to compensate for the expropriation’, and that ‘damages for an unlawful expropriation should at least be as much as the compensation for a lawful expropriation’.²⁰⁰ Conversely, if there is no expropriation—because the requirements of an exemption are met—no compensation needs to be paid under that head of the claim. The tribunal in Eco Oro v Colombia considered this outcome an ‘inconsistency’.²⁰¹

In a nutshell, the issue is whether there is an obligation to provide reparation even if wrongfulness is precluded. Article 27 of the ILC's Articles on State Responsibility are silent on ‘the question of compensation for any material loss caused by the act in question’,²⁰² and are in any case confined to circumstances precluding wrongfulness as set out in those Articles.²⁰³ In Eco Oro, decided under the Canada–Colombia FTA, the arbitral tribunal denied that there was an (indirect) expropriation²⁰⁴ but still found an obligation to compensate for a breach of the minimum standard of treatment even though that breach was justified under an exception.²⁰⁵ On that basis, measures necessary to protect legitimate public welfare objectives are subject to compensation.²⁰⁶

In order to determine the legal effects of the exception, the tribunal drew upon the wording of the exemption in the annex on expropriation (‘non-discriminatory measures … do not constitute indirect expropriation’).²⁰⁷ Contrasting the exception with the exemption, the tribunal concluded that, because the former does not stipulate ‘the circumstances in which a measure is not to constitute a treaty breach’, its successful invocation does not rule out the claimant's entitlement to compensation.²⁰⁸ This author disagrees. In truth, the differences in wording are intended to distinguish exemptions from exceptions and by no means imply that only the former exclude liability.

A. Difference between Exemption and Exception

The tribunal in Eco Oro should have paid heed to the difference between exemption and exception: exemptions mean that there has not been an (indirect) expropriation, exceptions offer a justification for what would otherwise have been an illegal expropriation, thus making it lawful.²⁰⁹ An exception only comes into play if there is an expropriation,²¹⁰ whereas if the conditions for an exemption are met, the measure is not an expropriation and so does not need to be justified.²¹¹

The Eco Oro tribunal misunderstood the purpose of general exceptions when it said that ‘neither environmental protection nor investment protection is subservient to the other, they must co-exist in a mutually beneficial manner’.²¹² This may be correct in relation to obligations in IIAs which have to be interpreted in harmony with other international law obligations of the parties (as per Article 31(3)(c) of the VCLT); it is wrong in relation to exceptions. It amounts to a reversal of the logic of exceptions and defies their ratio legis, which is to relieve a respondent of liability.²¹³ When the conditions of an exception clause are satisfied, the public welfare objective protected under the exception prevails: ‘nothing in this Agreement shall be construed to prevent a Party from adopting or enforcing measures …’.²¹⁴ There is no place for co-existence. That ‘the FTA is equally supportive of investment protection’, as postulated by the tribunal,²¹⁵ is irrelevant under an exception. The salient point is that a justified breach is lawful.²¹⁶ In this context, expropriation has to be distinguished from other substantive investment protection standards. As regards a justified and hence lawful expropriation, it is worth recalling that, in the author's view, compensation would still be due, not to the tune of the investor's expectation interest but limited to the fair market value.

B. Permission to Act

The Eco Oro tribunal makes a distinction between the right of the respondent to take a measure (under an exception), on the one hand, and the liability that flows from doing so, on the other.²¹⁷ The fundamental fallacy is that the tribunal assumed that permission to take action ‘without the payment of compensation’ needs to be expressly articulated in IIAs,²¹⁸ whereas it is the opposite that is the ‘default position under general international law’.²¹⁹ Under the Articles on State Responsibility, the precondition for compensation is an ‘internationally wrongful act’,²²⁰ which the successful plea of an exception negates. What is more, the right of States to expropriate has long been recognized.²²¹ Contrary to Eco Oro, other tribunals have held that ‘where the bilateral investment treaty at hand contains an exculpatory provision …, compensation is not payable to the extent that such provision exonerates that party from liability’,²²² and with regard to a security exception, that ‘it is the factor excluding the State from its liability vis-à-vis the damage caused as a result of the measures adopted by Argentina in response to the severe crisis suffered by the country’.²²³

This line of case law accords with WTO law where it is uncontested that a WTO Member may infringe its WTO obligations without legal consequences if the requirements of an exception clause are met. It is surprising that the tribunal disregarded standing WTO jurisprudence, given that the pertinent FTA, in a different part, explicitly requires its consideration for the purposes of interpreting similarly worded provisions.²²⁴ In the final analysis, the conclusion must be that there is no obligation on the host State to compensate the foreign investor for measures that fall within the scope of an exemption.

VII. Conclusions

It is the prerogative of the State to regulate in the public interest through the adoption of non-discriminatory public welfare measures. Case law which promotes a broad application of the concept of indirect expropriation constitutes a threat to that prerogative.²²⁵ As so often, the solution lies in careful treaty-drafting. The parties to modern mega-regional trade and investment agreements have, consequently, sought to reduce the discretionary powers of investment tribunals by better delineating the concept of expropriation.²²⁶ To that end, they have appended annexes on expropriation to their IIAs.

The last paragraphs of those annexes are exemptions embodying the police powers doctrine. They restrict the scope of indirect expropriation. If their requirements are met, there is no expropriatory measure, and hence no obligation to compensate. Exemptions constitute negative elements of an expropriation claim.

The burden of proof to show that its actions fall within the exemptions lies with the respondent State. If the claimant investor then wishes to claim that there is, exceptionally, a ‘rare circumstance’, then this is for the investor to prove. Exemptions incorporate a suitability test, which also considers whether the measure in question has the potential to contribute towards a legitimate public welfare objective. This test is less stringent than the fully-fledged necessity test as found in WTO law and is, therefore, more deferential to State sovereignty.

The suitability of a measure should be assessed as at the time of its adoption. This gives host States ample discretion to take measures, subject to the proviso that the measures are applied to further a legitimate public welfare objective. Where the measures taken fall within the ‘rare circumstances’ exception, a strict proportionality test applies, meaning that the level of protection of a public welfare objective set by the respondent State can be challenged by the claimant investor as being excessive. This does not mean that the State would be legally obliged to reduce the level of protection; instead, if successful, the claimant would be entitled to compensation for the loss suffered.²²⁷ However, not all IIAs contain such a caveat, it being dependent upon the contracting parties’ preference for how strong a level of investment protection they wish to see. In the absence of the ‘rare circumstances’ caveat, and leaving aside a potential FET claim, foreign investors are not protected against the adverse economic effects of regulatory measures, understood as public welfare measures of general application, on their investments.²²⁸