INTERNATIONAL CENTRE FOR SETTLEMENT OF INVESTMENT DISPUTES
In the arbitration proceeding between
MATHIAS KRUCK AND OTHERS
Claimants
and
KINGDOM OF SPAIN
Respondent
ICSID Case No. ARB/15/23
[Page 1]
1. The dispute in this case can be reduced to a single question: does Spain incur international responsibility under the ECT for changing the level of the state subsidy enjoyed by the Claimants' PV facilities under RD 661/2007? This in turn raises an issue of fundamental importance about the circumstances in which a State must compensate a foreign investor for a change to a regulatory regime when that change is not tainted by any of the traditional grounds for impugning a State under international law (arbitrariness, discrimination, bad faith, lack of due process, and so on). The Tribunal is unanimous that no such grounds apply to Spain's conduct; the sole question is whether the change is a violation of the Claimants' legitimate expectations as protected under the FET standard.
2. The reason that the “solar cases” against Spain, the Czech Republic, Italy and other countries have provoked controversy is precisely because liability has centred upon the doctrine of legitimate expectations and the contours of that doctrine are far from settled. A close reading of the awards and dissenting opinions reveals that the division in the jurisprudence rests upon two different conceptions of legitimate expectations. The first attributes liability based on a notion of strict liability. If it can be said that the State has made a “promise” of an immutable level of subsidy over a particular period of time in a public regulation, and the investor has “accepted” that “promise” by investing, then liability follows if that “promise” is
[Page 2]
“breached”. This conception, as the terminology suggests, rests upon an analogy with contract law, which also attributes liability for a breach of contract on the basis of strict liability. The State's public policy reasons for “breaching” its regulatory “promise” are irrelevant to establishing liability and proportionality-type arguments concerning the nature and impact of the change provide no defence.
3. The second and rival conception of legitimate expectations is based upon fault. The State is liable for breach of legitimate expectations only if the public interest pursued by the change in regulation imposes a disproportionate burden on the investor or offends some other criterion of fault (abuse of power, manifest unreasonableness, or the like).
4. The essence of my dissent is that a State can only be condemned to pay compensation on the basis of fault under Article 10 of the ECT (or indeed under any other investment protection obligation). I thus prefer the second conception of legitimate expectations as outlined above and the first part of my dissent will explain why.
5. In summary: the concept of legitimate expectations has no textual embodiment in the ECT and in investment treaties more generally. Unless it is accepted that arbitrators have the power to legislate for new grounds for liability under investment treaties (i.e. the power just to make it up), then the only path to a doctrine of legitimate expectations through the FET standard is by interpretation. The terms “fair and equitable” cannot be interpreted to mean the arbitrators' subjective notions of what is fair and equitable on any given day. But the FET standard can conceivably be interpreted as encompassing general grounds for liability in damages against public authorities as recognised by the major legal systems of the world (or by “civilised nations” if the more archaic language of Article 38 of the ICJ's Statute is to be preferred). No legal system imposes strict liability in damages for a breach of legitimate expectations; it is always based on fault. There is no reason in principle for international investment law to take a radically different approach.
6. The failure to incorporate an element of fault into the assessment of legitimate expectations in many decisions has led to the creation of a unique no-fault compensation scheme for foreign investors disappointed with the profitability of their projects. There is no evidence to suggest that state parties to investment treaties ever envisaged such a possibility and there is no legal or moral justification for tribunals to redistribute public funds on this basis.
7. The second part of my dissent addresses the point as to whether the Claimants in this case could have had a legitimate expectation that the state subsidy allocated by RD 661/2007 was immutable over the period in question.
8. It would have been more logical to address these points in reverse order because if there was no legitimate expectation on the facts of this case, then the more general question relating to whether liability is strict or based on fault would not arise. But as this general question is critical to the coherent development of investment law and its sustainability in the future, it is defensible to address it first.
[Page 3]
9. The majority has interpreted the text of RD 661/2007 as giving rise to a legitimate expectation on the part of the Claimants that the “fundamental characteristics”¹ of the regime for renumerating investment in PV facilities would not change. The critical factors that led the majority to that conclusion are set out in the following passage:
At least in circumstances where the explicitly declared purpose of legislation is to invite investors to commit capital to projects in reliance upon guarantees of stability in a regulatory regime, specific commitments can be made by provisions in general legislation. This is particularly the case in circumstances where, as here, the great majority of capital costs in an investment are incurred rights at the beginning of the project and are to be recouped over the operating lifetime of the project.²
10. The sense of the majority's use of the concept “reliance” is akin to the Claimants' “acceptance” of an “offer” of a certain subsidy regime in RD 661/2007. In the words of the majority: “RD 661/2007 was indeed intended to induce investments in the renewable energy sector by promising attractive and stable regulated tariffs and premiums, and [...] it constituted an invitation to potential investors to rely upon that promise.”³ This is obviously the language of contract law and the analogy with the contractual position is made explicit in the majority's reasoning:
The question here is whether Spain had committed itself to refrain from exercising its undoubted power in a particular manner, so that if it chose to exercise its power in a manner that breached that commitment it may incur liability for losses suffered by those who acted in reliance upon the commitment. There is nothing particularly arcane about this question: it is essentially the mirror image of the question whether a State can, consistently with the FET provision in ECT Article 10, bind itself to perform a contract, in a manner that cannot be undone by the State enacting legislation that purports to abrogate its contractual obligations.4
11. The majority's conception of an action for breach of legitimate expectations is actually indistinguishable from an action for breach of contract. Both rest upon a notion of strict liability: if a promise is made and is then breached, then liability for any losses follows from that breach, and the reasons for the conduct resulting in that breach are irrelevant. The majority thus does not admit of any defence to this action: if the investor has suffered loss as a result of the breach of the commitment or
1 Decision on Jurisdiction, Liability and Principles of Quantum, §202.
2 Decision, §189.
3 Decision, §190.
4 Decision, §199 (emphasis in the original).
[Page 4]
promise, then liability is established. This is strict liability. No public interest or proportionality defence can therefore be raised; indeed, elsewhere in its Decision, the Tribunal says this:
The incentives offered by the RD 661/2007 scheme were not compatible with the (equally reasonable) Spanish policy on prices to be charged to electricity consumers and on the balancing of the budget. Something plainly needed to change; and the new regulatory regime was in its own terms an inherently reasonable and pragmatic response to the exigencies of the urgent and unsustainable position.5
12. The measures adopted by Spain to change RD 661/2007 were thus unimpeachable based on the traditional grounds of liability under international law (arbitrariness, discrimination, bad faith, and so on). Liability in this case is instead justified by the majority on the basis of a breach of a “promise” of stability in respect of the “fundamentals” of the regulatory regime.6
13. The majority's conception of an action for breach of legitimate expectations is also indistinguishable from an action for breach of contract in terms of the assessment of damages. The remedy that follows from the breach is to put the Claimants in the position that they would have found themselves in if the commitment or promise in RD 661/2007 had been performed.7 That is the contractual model of compensation for expectation damages.
14. It is also important to be clear about the conception of “reliance” deployed by the majority. It is not detrimental reliance: there is no dispute that the Claimants' capital costs will be recouped over the lifetime of their projects together with a return despite the abrogation of the regime in RD 661/2007 (although the extent of that return is hotly contested by the quantum experts). In other words, the Claimants' reliance interest is protected despite the change in regulation. The only type of interest at stake here is the Claimants' expectation of a certain level of profits based upon a certain level of subsidy.8
15. The difference between the protection of a reliance interest and the protection of an expectation interest is fundamental. Investment treaty law, like administrative law, must search for a balance between the State's duty to change a regulation when the public interest demands it and the interests of individuals who have relied on that regulation to plan their affairs. In administrative law, courts are far more likely to award a remedy for breach of a legitimate expectation to protect a reliance interest because that can be more easily reconciled with preserving the State's mandate and discretion to act in the public interest. (As the State's taxpayers will not bear the burden of compensating parties for their losses based on expectations as opposed
5 Decision, §259.
6 Decision, §§190, 200, 202.
7 Decision, §357.
8 See generally: D Barak-Erez, 'The Doctrine of Legitimate Expectations and the Distinction between the Reliance and Expectation Interests' (2005) 11 European Public Law 583.
[Page 5]
to detrimental reliance, the State is less likely to be fettered in acting in the public interest.) A public authority is, for example, more likely to refuse to renew a licence to operate a noisy restaurant that disturbs the neighbours if the damages payable are limited to the operator's wasted expenditure on the lease of the premises rather than expected profits over the period of the licence renewal.
16. No authority is cited by the majority for the central propositions of their decision as set out above. There is no explanation as to how, by applying the standard rules of treaty interpretation, this conception of legitimate expectations is embedded (implicitly) in the FET standard in Article 10 ECT. Where does the majority's conception come from and how is it compatible with the basic architecture of an investment treaty and the fundamental principles of state responsibility towards foreign investors? Their decision is silent on these matters.
17. These questions are not raised in order to split hairs. In the absence of a power to legislate, arbitrators must defend their interpretations of the investment protection standards by reference to the normal canons of treaty interpretation as set out in Articles 31 and 32 VCLT. This is a point made expressly in the award in Philip Morris v Uruguay—the principal authority referred to by the majority in respect of the FET standard.9 How does one construct a bridge between the text of Article 10 ECT and the concept of legitimate expectations as a ground for liability in damages?
18. The only bridge that seems plausible, and it is far from uncontroversial, is to say that the doctrine of legitimate expectations is a “general principle of law recognised by civilised nations” in the sense of Article 38 ICJ Statute that can then qualify as a “relevant rule of international law applicable in the relations between the parties” for the purposes of Article 31(3)(c) VCLT. The controversy stems from the fact that general principles of law are not necessarily rules of international law, and the doctrine of legitimate expectations may not qualify as a general principle of law in any case because it is far from being universally recognised (Australia and France, for instance, do not recognise a doctrine of substantive legitimate expectations and in many other jurisdictions, such as England, there would be no remedy in damages for a breach). I will assume that these hurdles can be overcome. There are compelling pragmatic reasons to do so: if inspiration cannot be drawn from comparative law in giving content to the FET standard, then the popular alternative, which is for arbitrators just to make it up, may prove to be irresistible. Indeed, it was precisely this potential mischief that motivated the drafters of the Statute of the PCIJ to include the possibility of having recourse to “general principles of law recognised by civilized nations” in the first place:
[I]t would impose on the judges a duty which would prevent them from relying too much on their subjective opinion; it would be incumbent on them to consider whether the dictates of their own
9 Philip Morris Brands Sàrl and others v. Oriental Republic of Uruguay, ICSID Case No. ARB/10/7, Award, 8 July 2016, §317 (“Philip Morris”).
[Page 6]
conscience were in agreement with the conception of justice of civilized nations.10
19. Drawing inspiration from comparative law in developing international rules for the protection of foreign property and other interests has a distinguished history. The celebrated Harvard Draft Convention on the International Responsibility of States for Injuries to Aliens¹¹ uses the following formulation in several articles to delineate lawful from unlawful state conduct: “it is not an unreasonable departure from the principles of justice recognized by the principal legal systems of the world”.12 More specific references to comparative law are also tailored to particular types of wrongs, such as an unlawful arrest or detention: “if the cause or manner of the arrest or detention unreasonably departs from the principles recognized by the principal legal systems of the world”.13
20. A different approach was taken by the tribunal in Philip Morris, which appeared to suggest that a doctrine of legitimate expectations could form part of customary international law and customary international law is undoubtedly part of the “relevant rules of international law applicable in the relations between the parties” for Article 31(3)(c) VCLT.14 The ICJ may have subsequently poured cold water on that idea by stating that “in general international law [there is no] principle that would give rise to an obligation on the basis of what could be considered a legitimate expectation”.15
21. My working assumption for the purposes of this dissent, however, is that the FET standard in Article 10 ECT does include a concept of legitimate expectations. But the legitimacy of referring to that concept depends entirely on fidelity to how it is employed in comparative law. One cannot legitimately have recourse to a general principle of law in the interpretation of the FET standard and deploy that principle as a ground for liability and at the same time denude it of its essential characteristics in comparative law. That is no different to just making it up under the shade of a fig leaf. It is legitimate to have recourse to a general principle of law providing a ground upon which a public authority might be liable in damages in the context of interpreting and applying the FET standard because the state parties to the investment treaty can be taken to accept that their conduct be adjudged by reference to a general principle of law that they both recognise in their own legal systems. That does not exclude the possibility that adjustments may need to be made so that the general principle of law can fit the architecture of the investment treaty and general rules of responsibility in international law. But those adjustments need to be justified in a principled and transparent fashion. And they cannot transcend the consent of the state parties as manifested in their adherence to Article 10 ECT.
10 Permanent Court of International Justice, Advisory Committee of Jurists, Procès-verbaux of the Proceedings of the Committee June 16th-July 24th 1920, with Annexes (1920), p. 311.
11 LB Sohn & RB Baxter, 'Responsibility of States for Injuries to the Economic Interests of Aliens' (1961) 55 AJIL 545, 548.
12 Harvard Draft Convention, Arts 9(2)(c), 10(5)(c), 12(1)(c), 4(b).
13 Harvard Draft Convention, Art. 5(1)(b). See also: Arts 6(b), 7(f), (k), 8(b), 13(1)(b).
14 Philip Morris, §§317-324.
15 Obligation to Negotiate Access to the Pacific Ocean (Bolivia v Chile), Judgment, ICJ Reports 2018, p. 507, §162.
[Page 7]
22. One adjustment that cannot be justified is to transform the doctrine of legitimate expectations from a fault-based ground for liability in damages to a ground of strict liability. This is where I definitively part ways with the majority.
23. The majority has not provided reasons for its preference for strict liability and hence I cannot join issue on this point. What I can do, however, is demonstrate why the majority's conflation of a regulatory relationship in public law with a contractual relationship in private law is misguided.
24. Suppose a private contract between an investor and a state party were to include the following clause: “this contract is subject to unilateral modification by the government”. We would be surprised to encounter such a clause and the question would arise as to whether this were really a contract at all. The conferral of a power upon one party to alter the contractual bargain at will is incompatible with the whole purpose of the institution of contract, which is to give the force of law to promises made by the parties in their future relations.
25. Suppose now a public regulation were to include the following provision: “this regulation is subject to unilateral modification by the government”. Such a clause would also raise eyebrows but for a very different reason. The perplexity this time would be the superfluity of such a provision: of course the regulation can be modified by the government from time to time; the government would fail in its public duties if it did not modify regulations to take account of new developments relating to the thing being regulated. And as the government is the only “party” to the regulation, both the enactment and subsequent modification of the regulation is by definition a unilateral act.
26. The truth is that the same text appearing in a contract and in a public regulation may engender completely different expectations. In other words, the formal aspects of the instrument containing the text plays a critical role in our understanding of the meaning and significance of that text and, most importantly, our expectations about the stability of the undertakings, promises or commitments in that text.
27. These differences play out in domestic legal systems by the ubiquitous division between private and public law (the substantive doctrines, the procedures for claims, the remedial responses), the establishment of different courts and tribunals to deal with civil disputes as opposed to administrative disputes, and so on.
28. The majority has failed to attach significance to the fact that the text that is said to give rise to a legitimate expectation is contained in a public regulation. There would apparently be no difference if the text were contained in a contract with the Spanish Government. Indeed, the majority has appropriated the basic features of an action for breach of contract and applied to it a breach of an undertaking in a public regulation under the guise of legitimate expectations. That turns the private/public division on its head and suppresses the basic expectations that distinct institutions of law engender in every legal system. A contractual promise to pay a certain subsidy is absolutely not the same thing as an undertaking in a general public regulation to pay a certain subsidy. When a government decides to memorialise an undertaking in a private law contract it is relinquishing its powers as a public regulator and is
[Page 8]
putting itself on an equal footing with its private counterparty.16 That undertaking cannot then be unilaterally modified by the government (pacta sunt servanda) and only the exceptional doctrines of contract law that render an undertaking inoperable due to a fundamental change of circumstances (clausula rebus sic stantibus) will provide a defence to a breach of that undertaking.
29. In contrast, the same undertaking in a public regulation can in principle be modified or revoked at any time by the government. Someone impacted by that modification or revocation is not without any remedy: public or administrative law ensures that the procedure leading to that decision is fair and the substantive outcome is reasonable by reference to a test of proportionality or rationality. But no national or regional legal system contemplates that a government should be strictly liable for modifying or revoking an undertaking in a public regulation as if it were an undertaking in a private law contract.
30. For these reasons, it is erroneous to conflate legitimate expectations that may arise from a public regulation and a private law contract. They are distinct institutions of law and they engender very different expectations. It is impermissible to “privatise” public regulatory relationships by pretending that they are no different to private contractual relationships. The majority's approach to the contrary is misguided. It will be recalled that they equated the question of a breach of undertaking in a public regulation and a contract with the following observation:
There is nothing particularly arcane about this question: it is essentially the mirror image of the question whether a State can, consistently with the FET provision in ECT Article 10, bind itself to perform a contract, in a manner that cannot be undone by the State enacting legislation that purports to abrogate its contractual obligations.17
31. Where a State binds itself to perform a contract, and then uses its sovereign power to abrogate its contractual obligations, it has committed an actionable wrong that has an ancient pedigree in international law. The essence of the wrong is that the State has generated expectations based upon a private contractual relationship, only to then step outside the boundaries of contract law and invoke its sovereign power to undermine that relationship. The State's fault in this situation is manifest. Far from being a “mirror image”, the scenario in the present case is entirely different: the Spanish Government has chosen at the outset to regulate a particular matter on the basis of a general public regulation and there has been no deviation from that position. That must have significance for the expectations that can legitimately arise for those parties who have relied upon the regulation in some way. They cannot have ever legitimately assumed that they would be treated as if they had entered into a binding contract with the Spanish Government.
32. No Spanish lawyer would have advised the Claimants that RD 661/2007—a general public regulation enacted by the Spanish Government pursuant to the authority delegated by the Spanish Parliament under the Law on Electricity—gave them the
16 I leave to one side the hybrid institution of administrative contracts.
17 Decision, §199.
[Page 9]
equivalent of a contractual right to a particular subsidy. No Spanish lawyer would have advised the Claimants that RD 661/2007 was immutable or that it implicitly indemnified the Claimants against lost profits associated with a reduction of the subsidy if it were to change. There is no contemporaneous opinion from any Spanish lawyer in support of these propositions because, as a matter of Spanish law, they are plainly absurd.
33. The important question for now is whether there is some justification for international investment law to disregard the public/private law distinction and impose strict liability for a change to a public regulation and award compensation based upon a contractual model of damages. As previously stated, the only bridge between the FET standard and a concept of legitimate expectations is that the latter is a general principle of law. It is, therefore, important to investigate how that general principle works in various legal systems.
34. It is useful to start with the position of EU law because the EU courts were in the same position as investment tribunals in the sense that a set of principles had to be developed from scratch to determine when public authorities (the EU institutions) should be liable in damages to private parties for changes to regulations. They did so by drawing inspiration from the comparative law of the Member States (and from German public law in particular).18 It is also an appropriate starting point because the two States implicated in this case (Spain as the Respondent and Germany as the home State of the investors) are both EU Member States.
35. The EU courts have developed a three-stage test for determining when an EU institution should be liable in damages for a breach of legitimate expectations.
36. First, the claimant has to establish a legitimate expectation by reference to a course of conduct or assurance to the effect that the regulation would not change. If the claimant can establish a legitimate expectation in this manner, then the prima facie label is attached to it because there is then a second stage, which is the balancing exercise to determine whether an overriding public interest should trump the legitimate expectation. The test applied at this second stage is described by some commentators as the “significant imbalance test”,19 and by others as a test of proportionality.20 If the claimant prevails at this stage, then an administrative-type remedy would be available, which is typically the annulment of the offending regulatory provision or decision. In order to get to an award of damages, the
18 P Craig, UK, EU and Global Administrative Law: Foundations and Challenges (The Hamlyn Lectures) (CUP, 2016) p. 342. A detailed account of how the related concept of proportionality was developed by the German courts and then migrated to European law (both the European Court of Human Rights and the EU courts) and then to many other legal systems of the world is provided in A Barak, Proportionality: Constitutional Rights and Their Limitation (CUP, 2012) pp. 178-210. See also: J Usher, 'The Influence of National Concepts on Decisions of the European Court' (1976) 1 European Law Review 359.
19 S Schonberg, Legitimate Expectations in Administrative Law (OUP, 2000) 150.
20 P Craig, EU Administrative Law (OUP, 2018, 3rd edn) 638.
[Page 10]
claimant must overcome a third hurdle, which is that the breach of legitimate expectations is sufficiently serious to justify damages liability.
37. The principles are best illustrated by the “milk” cases before the Court of Justice, as they demonstrate the difference in outcomes depending upon illustrative changes to the factual matrix. In an effort to deal with the problems of surplus milk production, which drove down prices to a level that made it unsustainable for farmers, the EU Council introduced a scheme by Regulation no 1078/77 of 17 May 1977, which, in essence, paid farmers not to produce milk for a certain period of time. Mr Mulder was one such farmer who joined the scheme by agreeing with the relevant authority not to deliver milk or dairy products for a period of five years from 1979 to 1984. Just before the expiry of that period, and having made some investments in anticipation of returning to milk production, Mr Mulder applied for a “reference quantity” under the Council's new Regulation no 857/84 of 31 March 1984 (this required that an additional levy be paid on quantities of milk delivered in excess of a “reference quantity” as a further measure to limit excess milk production). That application was rejected on the basis that the “reference quantity” was to be calculated by reference to the farmer's milk production in 1983 and, given that Mr Mulder had agreed not to produce milk at that time under the scheme, he was not entitled to be allocated a “reference quantity”.
38. The Court accepted that Mr Mulder could not expect to resume production under the same conditions as those that applied at the time he voluntarily entered the scheme because it was to be expected that the regulatory regime might evolve over the period in question. This principle was expressed by the Court in a subsequent case in the following terms:
[W]hilst the protection of legitimate expectations is one of the fundamental principles of the community, economic operators cannot have a legitimate expectation that an existing situation which is capable of being altered by the community institutions in the exercise of their discretion will be maintained; this is particularly true in an area such as the common organisation of the markets, the object of which entails constant adjustments to make changes in the economic situation. It follows that economic operators cannot claim a vested right to the maintenance of an advantage which they derive from the establishment of the common organisation of the markets and which they enjoyed at a given time.21
39. Nonetheless, the Court found that the new Regulation no 856/84 effectively prevented a farmer who had signed up to the scheme from resuming milk production and that there was nothing in the original Regulation no 1078/77 to suggest that that would be the case. According to the Court: “Such an effect therefore frustrates those producers' legitimate expectation that the effects of the system to which they had rendered themselves subject would be limited.” Regulation no 856/84 was deemed to be invalid.22 In a case involving another farmer, Mr Spagl, the Court then declared that
21 Case C-402/98 ATB v Ministero per le Politiche Agricole [2000] ECR I-5501.
22 Case 120/86 Mulder v Minister van Landbouw en Visserij [1988] ECR 2321.
[Page 11]
the subsequent amendment to Regulation no 856/84 was also invalid: in response to the Court's decision in respect of Mr Mulder, the Council had attempted to fix the “reference quantity” to 60% of the quantity of milk delivered or sold by the farmer prior to entering the scheme. That was also deemed to be a breach of Mr Spagl's legitimate expectations because the “discount” was far higher than for producers who had not entered the scheme and the public interest justification for the differential reduction was not necessary to achieve the overall purposes of Regulation no 856/84.23
40. In these two cases, the Court awarded a public law-type remedy for the violation of a legitimate expectation as a public law doctrine: it declared the offending provisions of the Regulation to be invalid. But EU law recognises, like many other legal systems, that a public law ground for challenging an administrative decision does not ipso facto supply the basis for compensatory damages. This is almost invariably overlooked by tribunals in the investment treaty context, even if they profess fidelity to the fault-based concept of legitimate expectations in public law by applying the “balancing test” at the second stage.
41. The EU courts have constructed a bridge between a wrong in public law (legitimate expectations) and the remedy of compensatory damages by introducing a higher threshold for entitlement to the latter. The test developed by the ECJ is whether there has been a “significantly flagrant violation of a superior rule of law for the protection of the individual”.24 A superior rule of law for the protection of the individual includes legitimate expectations. But a further element is required to shift losses from the EU to individuals and that is the additional threshold of a “significantly flagrant violation”.
42. Mr Mulder and Mr Spagl subsequently pursued an action for damages based upon these principles. The first succeeded, whereas the second failed. Mr Muller succeeded because the original Regulation no 856/84 completely failed to take into account farmers in the situation of Mr Mulder without invoking any higher public interest: “the Community legislature manifestly and gravely disregard of the limits of its discretionary power, thereby committing a sufficiently serious breach of a superior rule of law.”25
43. Mr Spagl failed because, unlike the original Regulation no 856/84, the amended Regulation that introduced the 60% rule did not make it impossible for farmers who had entered the scheme to resume milk production and the Council had also explained the public interest reasons behind the compromise in the amendment. The Council, therefore, “took account of a higher public interest, without gravely and manifestly disregarding the limits of its discretionary power in this area”.26
44. This is not the place to provide a comprehensive account of the doctrine of legitimate expectations in comparative law. The position under EU law is likely to be broadly reflective of how the doctrine works in the legal systems of most of the
23 Case C-189/89 Spagl v Hauptzollamt Rosenheim [1990] ECR I-4539.
24 Case 5/71 Aktien-Zuckerfabrik Schöppenstedt v Council [1971] ECR 975.
25 Cases C-104/89 and 37/90 Mulder and Heinemann v Council and Commission [1992] ECR I-306, [19].
26 Case C-189/89 Spagl v Hauptzollamt Rosenheim [1990] ECR I-4539, [21].
[Page 12]
Member States. It is also considered to be more favourable to individuals than public authorities in comparison with French and English law at least.27 It will suffice to note that no legal system, according to my research, applies a doctrine of legitimate expectations on the basis of strict liability. In each legal system, the inquiry begins by asking whether a legitimate expectation can be established, and, if it can, the court must determine whether an appropriate balance has been struck between the public interest grounds for changing a policy and the frustration of a legitimate expectation based on that policy.
45. The English High Court in Solar Century Holdings, a case involving a challenge to the earlier-than-promised repeal of a subsidy programme for renewable energy, described this balancing exercise as follows:
And even if a sufficiently certain promise or representation has been made that a policy will continue in force and not be changed until a fixed date there is always a balance still to be struck between the retention of that policy and the strength of the (ex hypothesi) rational grounds which have arisen and which now are said by the Government to necessitate a frustration of that prior representation or promise. The test laid down by the Courts is whether the change of policy and the concomitant thwarting of the prior expectation amount to an abuse of power.
(Emphasis in the original.)28
46. The invocation of “abuse of power” in this test makes liability conditional upon fault. In other legal systems a test of “significant imbalance” or “proportionality” is preferred. But the common denominator is that there is always an element of fault in the second stage of the analysis when the doctrine of legitimate expectations is applied.
47. A total of 24 decisions on liability in the Spanish solar cases have been made available to the Tribunal in this case. There is a clear division in this jurisprudence between tribunals or majorities that have adopted a strict liability approach in interpreting the concept of legitimate expectations under the FET standard and those that have found or rejected liability based on fault. The touchstone of fault that is generally adopted is proportionality, which is assessed primarily in terms of whether the investment continued to earn a reasonable rate of return after the regime of RD 661/2007 was abolished. The division between the two strands in the jurisprudence is rather neat: there are 12 decisions favouring strict liability as against 12 decisions adopting a fault-based approach.
48. In the lists of cases that follow, an asterisk designates a dissent on the critical issue under discussion.
27 S Schonberg, Legitimate Expectations in Administrative Law (OUP, 2000) 142-146, 150-1, 222-224.
28 Solar Century Holdings Limited v Secretary of State for Energy & Climate Change [2014] EWHC 3677 (Admin), [73] per Justice Green.
[Page 13]
49. First, the following are the 12 decisions of a tribunal or majority in favour of the strict liability approach that had been adopted by the majority in the present case. I have been careful to include in this list decisions that refer to “proportionality” or a “balancing exercise” but that in reality uphold liability simply on the basis of a significant change to the regulatory regime that causes loss and thus reflect a strict liability approach.
49.1. [1] 9REN Holding S.a.r.l v. Kingdom of Spain (ICSID Case No. ARB/15/15, Award of 31 May 2019, §§212, 253, 267, 268, 311); [2] Cube Infrastructure Fund SICAV and others v. Kingdom of Spain (ICSID Case No. ARB/15/20, Decision* on Jurisdiction, Liability and Partial Decision on Quantum of 19 February 2019, §§427-428); [3] Eiser Infrastructure Limited and Energia Solar Luxembourg S.a.r.l. v. Kingdom of Spain (ICSID Case No. ARB/13/36, Decision on Jurisdiction and Liability of 17 March 2021, §§393, 398-400); [4] Foresight Luxembourg Solar 1 S.A.R.L., Foresight Luxembourg Solar 2 S.Á.R.L., Greentech Energy System A/S, GWM Renewable Energy I S.P.A and GWM Renewable Energy II S.P.A v. Kingdom of Spain (SCC Case No. 2015/150, Final Award of 14 November 2018, §§390, 397-8); [5] InfraRed Environmental Infrastructure GP Limited and others v. Kingdom of Spain (ICSID Case No. ARB/14/12, Award of 2 August 2019, §§368-369, 418, 449-451); [6] Infrastructure Services Luxembourg S.à.r.l. and Energia Termosolar B.V. (formerly Antin Infrastructure Services Luxembourg S.à.r.l. and Antin Energia Termosolar B.V.) v. Kingdom of Spain (ICSID Case No. ARB/13/31, Award of 15 June 2018, §§560, 562-3, 568-570); [7] Masdar Solar & Wind Cooperatief U.A. v. Kingdom of Spain (ICSID Case No. ARB/14/1, Award of 16 May 2018, §§520-522); [8] NextEra Energy Global Holdings B.V. and NextEra Energy Spain Holdings B.V. v. Kingdom of Spain (ICSID Case No. ARB/14/11, Decision on Jurisdiction, Liability and Quantum Principles of 12 March 2019, §§596, 598-600); [9] Novenergia II - Energy & Environment (SCA) (Grand Duchy of Luxembourg), SICAR v. Kingdom of Spain (SCC Arbitration (2015/063, Final Arbitral Award, 15 February 2018, §§657, 694-5 [although the tribunal paid lip service to a “balancing exercise” it actually adopted a strict liability approach on the facts]); [10] OperaFund Eco-Invest SICAV PLC and Schwab Holding AG v. Kingdom of Spain (ICSID Case No. ARB/15/36, Award* of 6 September 2019, §§485, 488-9); [11] SolEs Badajoz GmbH v. Kingdom of Spain (ICSID Case No. ARB/15/38, Award of 31 July 2019, §§419, 439, 443-4, 458, 461-2 [although the tribunal refers to “proportionality”, it applies a strict liability approach]); [12] Watkins Holdings S.à r.l. and others v. Kingdom of Spain (ICSID Case No. ARB/15/44, Award* of 21 January 2020, §§569, 597, 601-3 [the tribunal refers to “proportionality” but rejects Spain's defence on the basis that the disputed measures were “not an appropriate solution to the problem”— which is a merits review that has no place in a proportionality analysis—and needless to say that tribunal did not offer its own solution...]).
50. Second, in 9 of the 24 decisions, liability was found but on the basis of fault in the sense that the changes to the regulatory regime were held not to be proportionate as judged by reference to a legitimate expectation of a reasonable rate of return:
50.1. [1] BayWa r.e. Renewable Energy GmbH and BayWa r.e. Asset Holding GmbH v. Kingdom of Spain (ICSID Case No. ARB/15/16, Decision* on Jurisdiction, Liability and Directions on Quantum of 2 December 2019, §§463, 471, 477-480, 496, 500); [2] Cavalum SGPS, S.A. v. Kingdom of Spain (ICSID Case No. ARB/15/34, Decision* on Jurisdiction, Liability and Directions on Quantum,
[Page 14]
31 August 2020, §§619-621, 626, 632, 666); [3] Eurus Energy Holdings Corporation and Eurus Energy Europe B.V. v. Kingdom of Spain (ICSID Case No. ARB/16/4, Decision* on Jurisdiction and Liability, 17 March 2021, §§314-319, 358) [4] Hydro Energy 1 S.à r.l. and Hydroxana Sweden AB v. Kingdom of Spain (ICSID Case No. ARB/15/42, Decision on Jurisdiction, Liability and Directions on Quantum of 9 March 2020, §676 [it is not clear whether the tribunal applied the fault-based test that it articulated]); [5] Infracapital F1 S.à r.l. and Infracapital Solar B.V. v. Kingdom of Spain (ICSID Case No. ARB/16/18, Decision* on Jurisdiction, Liability and Directions on Quantum of 13 September 2021, §§528, 529, 531, 601-2, 687, 755); [6] RREEF Infrastructure (G.P.) Limited and RREEF Pan-European Infrastructure Two Lux S.à.r.l. v. Kingdom of Spain (ICSID Case No. ARB/13/30, Decision* on Responsibility and on the Principles of Quantum of 30 November 2018, §§464-5, 468, 471-2, 550 [although the tribunal mistakenly considers proportionality as part of the assessment of damages]); [7] RWE Innogy GmbH and RWE Innogy Aersa S.A.U. v. Kingdom of Spain (ICSID Case No. ARB/14/34, Decision on Jurisdiction, Liability and Certain Issues of Quantum of 30 December 2019, §§550, 553-4, 576-8); [8] The PV Investors v. Kingdom of Spain (PCA Case No. 2012-14, Final Award* of 28 February 2020, §§638-9, 648); [9] Sevilla Beheer B.V. and others v. Kingdom of Spain, (ICSID Case No. ARB/16/27, Decision* on Jurisdiction, Liability and the Principles of Quantum of 11 February 2022, §§ 715, 717, 872).
51. Third, in 3 of the 24 decisions, liability was not found on the ground that the State was not at fault, either because the measures were proportionate or not otherwise unreasonable in view of the public interest:
51.1. [1] Charanne B.V., Construction Investments S.A.R.L. v. Kingdom of Spain (SCC 062/2012, Final Award* of 21 January 2016, §§514, 517, 533-536); [2] FREIF Eurowind Holdings Ltd. v. Kingdom of Spain (SCC Case No. 2017/060, Final Award of 8 March 2021, §§525, 531, 571, 589-590); [3] Stadtwerke München GmbH and others v. Kingdom of Spain (ICSID Case No. ARB/15/1, Award* of 2 December 2019, §§317-320, 325 [approach based on “reasonableness” of measures]).
52. The fact that there is an even split in the Spanish solar cases on this issue of fundamental importance is unfortunate as legal certainty is fundamental to the rule of law and international investment law is failing on that score. But what is more disquieting is the almost complete lack of engagement with the arguments in favour of the fault-based conception of legitimate expectations in the decisions that support the alternative strict liability approach. Moreover, outside of the Spanish solar context, the fault-based approach appears to be predominant in the solar cases involving the Czech Republic²⁹ and Italy.30
53. This tendency is further disquieting because the early investment cases on legitimate expectations as an element of the FET standard clearly envisaged a fault-based
29 Eg: Antaris GMBH et al v. Czech Republic (PCA Case No. 2014-01, Award, 2 May 2018, §360).
30 Eg: Blusun S.A., Jean-Pierre Lecorcier and Michael Stein v. Italian Republic (ICSID Case No. ARB/14/3, Award, 27 December 2016, §319); CEF Energia BV v. Italy (Award, 16 January 2019, §§236-243).
[Page 15]
conception of liability.31 So did the academic commentary on those cases.32 The shift to strict liability is thus a break with the early trend in the investment jurisprudence and a departure from the general principles of law recognized by the major legal systems (which as a source of international law supplies the only rational link to incorporation of the concept into the FET standard by interpretation). And all this has happened in a void of justificatory reasoning.
54. I conclude this section by taking issue with the central tenet of the majority's reasoning, which is replicated in many of the decisions favouring a strict liability approach to legitimate expectations. That is the notion that a foreign investor has a vested right to the maintenance of a particular regulatory regime under the FET standard. Following this approach, liability is imposed simply on the basis of a change to the “fundamentals” of the regime in place when the investment was made. It matters not that the change was otherwise unimpeachable as being in the public interest and not discriminatory or arbitrary or imposing a disproportionate burden on the investor. This is an extraordinary proposition with far-reaching consequences.
55. First, it fails to calibrate liability by reference to the impact of any changes to the regulatory regime to the economic position of the particular investor in question. Liability is instead established in the abstract: if there is a radical change to a regulatory regime upon which an investor has relied, then there is a breach. But investors cannot have a legal interest in the design of a regulatory regime; their only concern is how it affects their bottom line. If RD 661/2007 were replaced by a radically different regime that increased the subsidy to the Claimants, they would have no complaint before this Tribunal. If RD 661/2007 were replaced by a radically different regime that decreased the subsidy to the Claimants but increased it for others, the majority's approach would result in a declaration that the introduction of the new regime is internationally wrongful in the abstract, and, logically, in respect of everyone.
56. Second, if a radical change to a regulatory regime is sufficient to trigger liability, then the ramifications for countries experiencing political upheavals are rather serious. Take the United Kingdom. Countless foreign investors relied upon regulatory regimes developed and enacted by the EU authorities and incorporated into British law when they invested. The United Kingdom had also given the most robust assurance possible under the British Constitution that it would apply those regimes in perpetuity by enacting the European Communities Act 1972. The advent of Brexit was far less foreseeable than the repeal of RD 661/2007. Are all those foreign
31 Eg: Saluka v Czech Republic (Partial Award, 17 March 2006, §306); International Thunderbird Gaming Corporation v United Mexican States (NAFTA, Separate Opinion of Prof. Wälde, 1 December 2005, §§2, 27, 30, 48).
32 Eg: E. Snodgrass, 'Protecting Investors' Legitimate Expectations: Recognizing and Delimiting a General Principle' (2006) 21 ICSID Review 1, 45 et seq. A recent ‘retrospective' on this issue is provided in F. Ortino, 'The Public Interest as Part of Legitimate Expectations in Investment Arbitration: Missing in Action?' in C. Brower, J. Donoghue, C. Murphy, C. Payne & E. Shirlow (eds). By Peaceful Means. International Adjudication and Arbitration: Essays in Honour of David D. Caron (2022).
[Page 16]
investors in the United Kingdom entitled to compensation for breach of their legitimate expectation to stability if the European regulatory regimes are replaced by (sufficiently different) British ones? Would similar considerations apply for countries transitioning from a planned economy to a market economy with the large-scale reforms that entails?
57. Third, in an era where radical reforms in the energy sector are not only desirable but necessary for the sustainability of human life on Earth, it seems incongruous now to be interpreting the FET standard as requiring States to buy back their right to implement those reforms at the highest price. In this case the investment is in renewable energy. But if the strict liability approach is generalised (and there is no principled basis for limiting it to cases involving renewable energy), then the public interest considerations behind any reform are by definition irrelevant to the determination of liability in damages. The regulatory chill that inevitably accompanies this approach will be no consolation to a warming planet.
58. RD 661/2007 allocated a subsidy to the Claimants' PV projects. The majority considers that “RD 661/2007 contained express assurances that its fixed tariffs would apply to qualifying, registered PV facilities for a fixed period, and that changes to the regulatory regime would not apply to facilities already registered.”33 The majority does not, however, identify the precise language in RD 661/2007, or in any other text, that is said to lead to that conclusion. The Claimants, for their part, rely on Article 44(3), and in particular the underlined text set out below:34
3. During the year 2010, on sight of the results of the monitoring reports on the degree of fulfilment of the Renewable Energies Plan (PER) 2005-2010, and of the Energy Efficiency and Savings Strategy in Spain (E4), together with such new targets as may be included in the subsequent Renewable Energies Plan 2011-2020, there shall be a review of the tariffs, premiums, supplements and lower and upper limits defined in this Royal Decree with regard to the costs associated with each of these technologies, the degree of participation of the special regime in covering the demand and its impact upon the technical and economic management of the system, and a reasonable rate of profitability shall always be guaranteed with reference to the cost of money in the capital markets. Subsequently a further review shall be performed every four years, maintaining the same criteria as previously.
33 Decision, §190.
34 C's Memorial, §179.
[Page 17]
The revisions to the regulated tariff and the upper and lower limits indicated in this paragraph shall not affect facilities for which the deed of commissioning shall have been granted prior to 1 January of the second year following the year in which the revision shall have been performed.35
59. The underlined paragraph undoubtedly says that the adjustment mechanism in Article 44(3) (“this paragraph”) would not be used to change the level of subsidy set out in RD 661/2007 for projects that had a deed of commissioning before the relevant date (this included the Claimants' PV projects). The question is then whether it was reasonable to interpret this language as preventing the amendment or wholesale repeal of RD 661/2007 if the fundamental assumptions underlying RD 661/2007 were to change.
60. This basic distinction between “micro” stability within the framework of applying particular regulation and “macro” stability requiring that regulation to remain in force at all costs and regardless of the circumstances is not lost on anyone familiar with legitimate expectations in public law. In the words of the English High Court in the Solar Century Holdings case:
When what is objected to is the abrogation of a policy or a change of policy the starting point is that once a policy is promulgated and said to be settled there needs to be a rational ground for terminating it... But there is no presumption that policy cannot change; on the contrary it plainly can do so and frequently does. So the issue becomes whether there can be identified a representation of sufficient certitude that the policy will not be changed regardless of surrounding circumstances. As to this a representation that a policy will continue until a specified date is not the same as a promise that it will never be changed even if circumstances change. If it were otherwise then an intention to pursue a policy for a fixed period would become set in stone and permanently unyielding to changes in relevant circumstances however compelling they might be.36
61. (The English court then, in the passage quoted previously, stated that even where a promise of “macro” stability had been made in this sense, “there is always a balance still to be struck between the retention of that policy and the strength of the (ex hypothesi) rational grounds which have arisen and which now are said by the Government to necessitate a frustration of that prior representation or promise”. This is the fault-element for liability for breach of a legitimate expectation that I have already examined in this dissent.)
62. I repeat an earlier point: the language in Article 44(3) RD 661/2007 is language found in a public law regulation and not a private law contract. When such language is found in a public regulation, the relevant context is that a government has the power in public law to repeal that regulation if the circumstances change and the
35 C-98 (improved), PDF pp. 49-50 (emphasis added).
36 Solar Century Holdings Limited v Secretary of State for Energy & Climate Change [2014) EWHC 3677 (Admin), [72] per Justice Green.
[Page 18]
public interest requires it. One cannot read the underlined text as if it appeared in a private law contract where no such power exists. This is precisely what the Claimants ask us to do. They say this: “Absent an express provision [in RD 661/2007] allowing Spain to modify the economic regime as it applied to existing plants, it had no right to do so.”37 That is manifestly the wrong starting point for interpreting a public regulation. (And it would mean the British Government would have much to lose by failing to include a clause such as “this regulation is subject to amendment or repeal should the UK leave the EU” in every regulation it enacted between 1973 and 2020.)
63. Does the language of Article 44(3) really say that the Spanish Government agrees to absolve itself of the power to repeal that regulation if the circumstances change and the public interest requires it? That it agreed to stability in the “macro” sense? There is certainly nothing in the text itself to suggest that. And even a rudimentary understanding of the Spanish legal order and the legislative history flatly contradicts such a proposition.
64. RD 661/2007 was enacted in furtherance of Law 54/1997 of 27 November 1997 on the Electricity Sector (Law 54/1997). Law 54/1997 is the umbrella legislation for RD 661/2007. A Royal Decree Regulation is subordinate to a Law and a Royal Decree Law, which, in turn, are subordinate to the Spanish Constitution. A Royal Decree Regulation (referred to simply as “Royal Decrees”) cannot contradict Laws or Royal Decree Laws.38 In Spanish legal practice, a Law enacted by the Spanish Parliament contains the general principles regulating the subject matter in question, whereas the detailed provisions giving effect to those general principles are set out in subsequent Royal Decrees promulgated by the Government. In the present case, RD 661/2007 was enacted by the Ministry of Industry and Trade.
65. Chapter II of Law 54/1997 creates the “Special regime for electricity production” that applies, inter alia, to solar electricity production. Article 30 of Chapter II is entitled “Obligations and rights applicable to energy producers operating under the special regime” and contains the following provisions relating to the remuneration of energy producers:
3. The remunerative regime applicable to electrical energy production plants operating under the special regime shall be adjusted pursuant to provisions outlined in sub-section 1 of article 16 for electrical energy purposes.
4. The payment regime applicable for electricity production facilities under the special regime shall be supplemented by the earning of a premium, under the terms set by regulations, in the following cases:
a) Facilities referred to in letter a) of section 1 of article 27 [which includes solar plants such as those operated by the Claimants]
37 C's Reply, §234.
38 R-35, Organic Law 6/1985, of 1 July, on the Judiciary, Article 6 (“The Judges and the Courts shall not apply regulations or any other provision contrary to the Constitution, the law or the principle of normative hierarchy.”).
[Page 19]
[...] To determine the premium, the voltage level of electricity delivered to the network must be considered, along with the actual contribution to improvement of the environment, primary energy savings and energy efficiency, the economically justifiable production of usable heat, and the investment costs that have been incurred, for the purpose of achieving reasonable rates of return with respect to the cost of money in the capital market.39
66. The underlined passages in Article 30(4) of Law 54/1997 are important. First, it is envisaged that the premium to which energy producers are entitled will be established under the terms set by regulations (i.e. Royal Decrees). Hence Law 54/1997 does not itself establish the terms of the premium: consistently with Spanish legislative practice the precise terms of the premium are to be fixed in subordinate legislation in the form of Royal Decrees. Second, the overriding principle for the determination of the premium is to achieve “reasonable rates of return” by taking into account, inter alia, the investment costs that have been incurred and the cost of money in the capital market. Any Royal Decree giving effect to this provision would have to calculate the premium consistently with this overriding principle.
67. RD 661/2007 expressly recognises these core principles of Law 54/1997 in its preamble:
The economic framework in the present Royal Decree develops the principles provided in Law 54/1997, of 27 November, on the Electricity Sector, guaranteeing the owners of facilities under the special regime a reasonable return on their investments, and the consumers of electricity an assignment of the costs attributable to the electricity system which is also reasonable.40
68. Given that Article 44(3) of RD 661/2007 must be read consistently with Article 30(4) of Law 54/1997—in case of inconsistency Article 44(3) would be invalid— and given that the preamble to RD 661/2007 expressly refers to the same fundamental regulatory objectives as in Article 30(4) of Law 54/1997, is it legitimate to assume that the fixed tariff or premium rates in RD 661/2007 would never be modified by a subsequent regulation even if owners of facilities were earning a manifestly unreasonable rate of return and consumers were bearing unreasonable costs? The stability provided for in Article 44(3) of RD 661/2007 would no doubt exist so long as the fundamental assumptions underlying RD 661/2007 pertained. But if those assumptions were no longer tenable, can Article 44(3) really be read as a guarantee that RD 661/2007 would never be revised or repealed?
69. The prior legislative history categorically refutes this proposition of immutability. RD 661/2007 was preceded by RD 2818/1998, RD 436/2004 and RD Law 7/2006. This demonstrates that the terms of the premium envisaged in Article 30(4) of Law 54/1997 would be fixed by regulations as promulgated from time to time.
39 C-31, Article 30.
40 C-98 (improved), PDF p. 4.
[Page 20]
Moreover, the preamble of RD 661/2007 specifically justified the repeal of the then in force RD 436/2004 on the basis that the fundamental economic and technical assumptions had changed:
The modification of the economic and legal framework which regulates the special regime existing to date has become necessary for various reasons. First of all, the growth seen in the special regime over recent years tied to the experience accumulated during the application of Royal Decree 2818/1998, of 23 December and Royal Decree 436/2004, of 12 March, has shown the need to regulate certain technical aspects in order to contribute to the growth of those technologies, while maintaining the security of the electrical system and ensuring the quality of supply, and minimising the restrictions on the production of electricity generated in this manner. In view of the behaviour of the prices in the market, where certain variables which were not considered in the cited compensation system for the special regime have, over recent times, acquired greater importance, the economic circumstances established by Royal Decree 436/2004, of 12 March, make it necessary to modify the compensation system and de-link it from the Mean Electricity Tariff, or Reference Tariff, which has been used to date. Finally, it is necessary to include the changes in the legislation deriving from European law, and from Royal Decree-Law 7/2006, of 23 June, Adopting urgent measures in the energy sector, Introducing major amendments in respect of the law applicable to the combined heat and power business.41
70. This is deeply significant because the repealed RD 436/2004 contained almost precisely the same language that the Claimants say introduced a stabilised regime in Article 44(3) of RD 661/2007. Article 40(3) of RD 436/2004 reads:
3. The tariffs, premiums, incentives and supplements resulting from any of the revisions provided for in this section shall apply solely to the plants that commence operating subsequent to the date of the entry into force referred to in the paragraph above and shall not have a backdated effect on any previous tariffs and premiums.42
71. And yet, Article 40(3) of RD 436/2004 could not have produced the stabilisation upheld by the majority because it was repealed and replaced by RD 661/2007. The Claimants were aware of this when they invested. The Claimants simply observe that “Crucially, RD 661/2007 retained and clarified RD 436/2004's guarantee against non-retroactivity.”43 But if the Claimants and the majority are correct, RD 661/2007 did
41 C-98 (improved), PDF pp. 3-4.
42 C-75, PDF p. 40.
43 C's Memorial, §179.
[Page 21]
far more than “retain and clarify”, it introduced a radically different regime that, for the first time, was stabilised. According to the majority:
The Tribunal accordingly decides that RD 661/2007 set out assurances on which it was intended that potential investors could and would rely, concerning the stability of the compensation framework established by RD 661/2007.44
72. But the problem for this theory of immutability is that the previous RD 436/2004 contained almost exactly the same “assurances” and yet it was repealed and replaced by RD 661/2007 before the Claimants invested. Not only that, the Supreme Court had expressly ruled that the language in Article 40(3) of RD 436/2004 could not give rise to a legitimate expectation that the regulated tariff and premium would be immutable (more about that later). And the Claimants themselves accept that there could be no legitimate expectation based on the continuation of RD 436/2004:
Claimants do not argue that investors under RD 436/2004 had legitimate expectations that they would benefit from a defect in the law that would cause the tariffs to increase after investments were completed in unexpected and artificial ways.45
73. Neither the Claimants, nor the majority, explain their position that RD 661/2007 established a radically different form of stabilisation as compared with RD 436/2004, in circumstances where the text is virtually identical. This is certainly not how the industry organisations and leading commentors on the solar energy market viewed the changes. The Association of Renewable Energy Generators (the “APPA”) attacked the Spanish Government for repealing RD 436/2004 precisely because of the stabilisation language in Article 40(3), calling it a violation of the legitimate expectations of existing investors in circumstances where there was no “urgent or unavoidable matter of public interest”.46 The APPA said that the draft of RD 661/2007 if implemented “is disastrous and devastating for future investments”.47
74. The Claimants quote the following extract from the authoritative Powering the Green Economy: The Feed-In Tariff Handbook in their Reply,48 which also flatly contradicts the idea of a radical change in stabilisation between RD 436/2004 and RD 661/2007:
Up until 2004, the tariff level could change every year, thus making the financing of renewable electricity projects largely unpredictable (Jacobs, 2008).
The legislative breakthrough came in 2004, with the Royal Decree 436/2004 and the amendment in 2007 (BOE, 2007). By 2003, the Spanish regulator CNE had already established a transparent
44 Decision, §195.
45 C's Reply, §246.
46 R-0287, PDF p. 5.
47 R-0287, PDF p. 7.
48 C's Reply, §295.
[Page 22]
methodology for calculating tariffs based on their average generation costs (see Section 2.3). From 2004 on, tariff payment was guaranteed for at least 15 years, in the case of some technologies even for the whole lifetime of a power plant. In 2007 [through RD 661/2007], the tariff payment was eventually disconnected from the average electricity price and based on a fixed tariff payment per kilowatthour.49
75. If RD 661/2007 had really ushered in a radically new stabilised regime, then one would expect to read about it in the preamble that explains at length the objectives of the new regulation. But the opposite is true: the preamble emphasizes continuity with the repealed RD 436/2004:
The present Royal Decree replaces and repeals Royal Decree 436/2004, of 12 March, establishing the methodology for updating and systematisation of the legal and economic framework of the activity of the production of electrical energy under the special regime, while maintaining the basic structure of the regulations therein.
The economic framework established in the present Royal Decree develops the principles provided in Law 54/1997, of 27 November, on the Electricity Sector, guaranteeing the owners of facilities under the special regime a reasonable return on their investments, and the consumers of electricity an assignment of the costs attributable to the electricity system which is also reasonable [...]
To this effect, a system which is analogous to that provided in Royal Decree 436/2004, of 12 March, is maintained, in which the owner of the facility may opt to sell their energy at a regulated tariff, which will be the same for all scheduling periods, or alternatively to sell this energy directly on the daily market, the term market, or through a bilateral contract, in this case receiving the price negotiated in the market plus a premium.50
76. Indeed, there is no reference whatsoever to the “stabilisation” language relied upon by the Claimants in Article 44(3) in the dense five-page preamble of RD 661/2007 or to any heightened or superior stabilisation as compared with the previous RD 436/2004.
77. Also critical is the fact that the Spanish Supreme Court specifically rejected the argument that RD 436/2004 created a stabilised regime immune to revision in its Judgment of 25 October 2006. In that case, the appellants contended that RD 2351/2004, which changed the system for calculating the premiums under RD 436/2004 violated the principles of legal certainty, legitimate expectations and good faith contrary to Article 9.3 of the Spanish Constitution. After noting that the amendments did not violate Article 30 of Law 54/1997 (the appellants had not
49 RL-088, PDF p. 118 (emphasis added).
50 C-98 (improved), PDF p. 4 (emphasis added).
[Page 23]
submitted otherwise), the Supreme Court dismissed this argument based on legitimate expectations with the following reasons:
The same consideration applies to the principle of legitimate expectations, increasingly but unduly deployed as an argument against quite a few regulatory changes that some economic operators have deemed more or less harmful to their interests. The appellants argue that their investments in the production of electrical energy under the special regime were made at a particular time “when they trusted the Government not to change the legal conditions that decided them to (...) build their facility”. This premise leads them to conclude that the reduction of premiums subsequent to Royal Decree 2351/2004 with respect to those in Royal Decree 435/2004 is contrary to the principle.
We cannot agree that this reasoning can be applied to an incentive mechanism such as the premiums in question. Until it is replaced by another, the above outlined legal regulation (Article 30 of the Electricity Law) allows the respective companies to expect that the fixing of the premiums can be included as a factor relevant to their obtaining “reasonable rates of return with reference to the cost of money in the capital market” or, to put it again in the words of the preamble to Royal Decree 436/2004, “reasonable compensation for their investments.” However the payment regime under examination does not guarantee to special regime electricity producers that a certain level of profits or revenues will be unchanged relative to those obtained in previous years, or that the formulas for fixing the premiums will stay unchanged.
Just as in terms of an economic policy with many different aspects (the promotion of renewable energy but also the planning of electricity networks, and other considerations regarding energy saving and efficiency) grants and incentives for the production of electricity under the special regime may increase from one year to another, but they may also decrease when those same considerations warrant it. We stress that these changes have remained within the legal limits that regulate this type of promotion, and the mere fact that the annual adjustment or the level of the premium goes up or down does not of itself constitute grounds for revocation or affect the legitimate expectations of their recipients.
Companies that freely decide to enter a market such as electricity generation under the special regime, knowing that is largely dependent on the setting of economic incentives by public authorities, are or should be aware that they may be modified within legal guidelines, by those same authorities. One of the “regulatory risks” to which they submit and which they must take into account, is precisely the variation of parameters for
[Page 24]
premiums or incentives, something which the Electricity Sector law limited... but does not preclude.51
78. This Judgment was a matter of public record at the time that the Claimants invested in Spain. The subsequent judgments of the Supreme Court applied the same principles.52
79. An investor seeking to make an investment in the renewable energy sector in Spain on the basis of RD 661/2007 would thus have been on notice that:
79.1. The umbrella legislation, Law 54/1997, required the premium to be calculated in order to generate a reasonable rate of return;
79.2. Law 54/1997 also envisaged that the terms of the premium would be established by regulations;
79.3. Several regulations preceded RD 661/2007 and their promulgation, amendment and/or and repeal was justified by the Ministry of Trade and Energy as necessary to give effect to changing market conditions for the calculation of the premium;
79.4. RD 436/2004 contained virtually identical “stabilisation” language as RD 661/2007 and yet RD 436/2004 was repealed precisely because there were distortions in the calculation of the tariff for the special regime that needed to be addressed;
79.5. The Supreme Court had specifically rejected an argument to the effect that a provision in the earlier RD 436/2004 that purported to disapply the modifications to the calculation of the premium introduced by that Royal Decree to existing installations resulted in the immutability of RD 436/2004. The Supreme Court held that the disapplication only had effect so long as RD 436/2004 itself was not amended or repealed;
79.6. In the preamble to RD 661/2007, the Ministry of Trade and Energy justified the promulgation of the Royal Decree as necessary in order to take into account changing market conditions and to preserve the principle of a reasonable rate of return set out in Law 54/1997.
80. My conclusion is that the Claimants had no legitimate expectation that RD 661/2007, and the tariffs that it fixed for the special regime, would continue in force regardless of whether there were significant changes to the underlying economic assumptions behind that regulation. That does not mean that Article 44(3) had no stabilizing effect and was deprived of any utility (as the Claimants maintain), nor does it mean that the Spanish Government was free to adjust the tariffs whenever it wished to bring them in line with a “reasonable rate of return” (as the Respondent
51 R-0133 PDF p.4, Judgment of the Spanish Supreme Court, 25 October 2006 (emphasis added).
52 Eg: R-0137, Judgment of the Supreme Court, 9 December 2009; R-0248, Judgment of the Supreme Court, 1 June 2016.
[Page 25]
maintains). These two extremes of complete stability versus no stability present a false dichotomy. I will address this further below.
81. The conclusions relating to the interpretation of the text and legislative history of RD 661/2007 provided in the foregoing section are reinforced if the context of what was being regulated is added to the picture. What was being regulated was a state subsidy.
82. When the Claimants invested in the renewable energy sector in Spain in 2007-2008, it was impossible to make a commercial return from the production of solar energy without a state subsidy. This was because fossil fuels were still a much cheaper source of energy. Today, that may no longer be the case. In 2007, however, state intervention was critical to develop what was otherwise an unprofitable sector of the economy and the broad purpose of that intervention was to subsidize the price for renewable energy so it could compete with non-renewable energy sources. An investment in renewable energy was not an investment that made money without the involvement of the State; to the contrary, it was an investment that could only be profitable if the State was involved. Indeed, the level of profit that the investor could expect to earn depended almost exclusively on the level of subsidy that the State was prepared to pay.
83. A state subsidy to private investors in the solar energy sector is financed by taxpayers and electricity consumers. For a subsidy to be economically rational, the State must ensure that the level of the subsidy is no more and no less than is necessary to make solar energy competitive in the market for energy. More specifically, this means that the subsidy must cover the higher investment and/or operating costs of renewable energy producers as compared with non-renewable energy producers. If the subsidy is inadequate in this sense, then investment into the solar energy sector will not materialise. If the subsidy is too generous, then it will distort competition and trade and place an unjustifiable burden on taxpayers and consumers.
84. The gravity of the negative consequences that flow from a subsidy that violates these basic economic principles has inspired regulation at the international and regional level: the most prominent instruments are the WTO Agreement on Subsidies and Countervailing Measures and Articles 107-109 of the Treaty on the Functioning of the European Union (“TFEU”). Under Article 107 TFEU, State aid is generally prohibited unless it falls within one of the enumerated exemptions. One of the requirements for lawful state aid is that it is proportionate; in other words, the amount of aid per beneficiary must be limited to the minimum needed to achieve the particular objective for the production of energy.53 If the subsidy is too generous in this sense, then the EU Commission will declare it to be unlawful state aid.
85. This is important context for interpreting whether RD 661/2007 contained a promise of immutability concerning the level of the subsidy regardless of any
53 The European Commission promulgated 'guidelines' for state aid in support of renewable energy sources that makes this point expressly: R-0047, 'Guidelines on State aid for environmental protection and energy 2015-2020' (2014/C 200/01). The first such guidelines were issued in 2001: R-0347, Community guidelines on state aid for environmental protection (2001/C 82/01).
[Page 26]
changes to the market for energy. The concept of a “reasonable rate of return” encapsulated in Law 54/1997 is not there by accident: if a subsidy is “unreasonable” because it goes further than ensuring that a producer of renewable energy can be competitive in the market vis-à-vis a producer of non-renewable energy, then it constitutes unlawful state aid.
86. Given this background, it would be unreasonable to interpret RD 661/2007 as compelling the Spanish Government to maintain the level of subsidy fixed by that regulation even if it were no longer proportionate to the objectives pursued by the subsidy in the first place. It is rather unlikely that the Spanish Government would have committed itself in RD 661/2007 to violate EU and national law on state aid if the circumstances changed. There is certainly nothing in RD 661/2007 to suggest otherwise.
87. I have already stated that the parties in this case have presented a false dichotomy between complete stability and no stability at all under RD 661/2007. These extremes are encapsulated in the following passage of the Claimants' Reply:
Spain's argument would render Article 44, and especially the last sentence of Article 44.3, completely superfluous. If Spain could revise the economic regime at any time for any reason, then it was unnecessary to specify in Article 44 what kind of revisions could occur and when, and then restrict the application of the quadrennial “policy revisions” to plants commissioned in the future.54
88. The majority also opines that without the alleged promise of immutability in RD 661/2007, the Claimants' future income flow would be a matter of “pure speculation”.55
89. So long as RD 661/2007 remained in force, Article 44(3) operated to ensure that the Claimants' tariffs under the special regime would only be adjusted in accordance with the mechanism set out in that provision. I reject the Respondent's case to the contrary that the Spanish Government could make other adjustments to bring the tariffs into line with a “reasonable rate of return” while the regime in RD 661/2007 applied.
90. Does that mean Article 44(3) was devoid of utility, given the Spanish Government always retained the power to repeal RD 661/2007?
54 C's Reply, §235.
55 Decision, §352.
[Page 27]
91. Of course not. And the failure of tribunals to recognise the limitations upon the Spanish Government's power in these circumstances has contributed to the false dichotomy that I have alluded to.
92. If the Spanish Government were to repeal RD 661/2007, as it ultimately did, then its action would be subject to the constraints on public power contained in Spanish constitutional and administrative law, as well as in the ECT and international law. Under international law, the repeal could not, in respect of protected investments, be arbitrary, discriminatory, expropriatory, or violate due process and so on. If the Spanish Government, for example, were to have repealed RD 661/2007 on the basis that it no longer considered renewable energy worthy of support in rejection of the scientific evidence of global warming, then such repeal might well attract the opprobrium of international law.
93. The general constraints on public power in international law provide a certain measure of stability to foreign investments. They do not guarantee complete stability by creating a strict liability regime that compensates for any change.
94. In the present case there is no dispute that the Spanish Government's repeal of RD 661/2007 was made on the basis of legitimate public interest grounds and the Claimants' investments continue to make a return. So its actions pass the test of the traditional grounds of review under international law. This case is about whether a concept of legitimate expectations enhances the protected realm of stability further than those traditional grounds.
95. The majority says that it does but it is forced into the logic of strict liability to uphold the Claimants “all or nothing” approach. The public interest reasons for the repeal of RD 661/2007, and the extent of the impact of the repeal on the Claimants’ investments, are irrelevant to the majority's assessment of liability. It has treated the language of Article 44 RD 661/2007 as if it were contained in a private contract save that the Spanish Government is actually worse off because even in contract law there are exceptions to pacta sunt servanda (force majeure, frustration, and so on).
96. But that is not the only way to approach this case.
97. It may be that the concept of legitimate expectations is not the right vehicle for enhancing the protected realm of stability under the ECT and international law in the circumstances of the Spanish solar cases. Its usefulness is probably confined to cases involving individualised decisions by state authorities rather than general regulations.
98. The problem for legitimate expectations in these cases is that one has to accept a proposition that is patently absurd to get over the first hurdle. That proposition is that RD 661/2007 contained an undertaking of immutability when such an undertaking, if given, would have violated Spanish constitutional and administrative law as well as EU state aid law. The Spanish Government cannot, in a Royal Decree, promise not to amend or repeal that decree if the public interest requires it, nor can it bind the Spanish Parliament not to amend or repeal Law 54/1997, which provides the legislative authority for the Spanish Government to have enacted RD 661/2007 in the first place.
[Page 28]
99. So can an investor have a legitimate expectation that is inconsistent with the fundamental principles of the Spanish legal order? Some tribunals have ducked the question by relying on the circular argument that it is the ECT and not Spanish law that applies. But the ECT does not create legitimate expectations; it can only serve to protect them once established. And so the question is whether the ECT should protect an expectation that flatly contradicts the law of the host State. The majority thinks it should:
No-one questions that Spain had the legal capacity to change the regulatory regime. The question is whether it was fair and equitable to do so in a manner that caused harm to certain investors who had relied on representations that Spain would not exercise that undoubted power in relation to their investments.56
100. Of course I do not agree that the Spanish Government represented in RD 661/2007 that it would not exercise its power to change the regulatory regime in the future come what may, in violation of its own constitutional and administrative law. But leaving that point to one side, this tension between the plausibility of an expectation of immutability under the law of the host State and the protection of that expectation under international law has never been addressed in a satisfactory way.
101. It also generates a rather surreal discussion about whether the investor undertook proper due diligence on this question of local law. The consensus is that such due diligence is necessary to establish that the expectation was legitimate. But that puts the investor in an impossible position: unless it can find a local lawyer who is professionally negligent, no one is realistically going to opine that, as a matter of Spanish law, the Spanish Government gave a valid assurance of immutability in RD 661/2007. Many tribunals have commented upon the absence of such legal opinions in the Spanish solar cases as a point against the claimants. But this is at best a rhetorical observation. In the present case the Claimants have been compelled to say that, despite receiving written due diligence reports on a host of issues relating to their PV projects, the legal opinion on the immutability of RD 661/2007 was given orally by a Spanish lawyer.57 Alas there is no documentary record for it. Unlike the majority,58 I cannot give any weight to this. The fundamental point, however, is that the quest to produce such an opinion was a fool's errand to begin with.
102. The path to a more coherent analysis of liability in these cases is to abandon the “all-or-nothing” approach to stability that is premised on an artificially reconstructed promise of immutability on the part of the host State and the investor's concomitant expectation of the same. The concept of proportionality, rather than legitimate expectations, might be the better vehicle.
103. Proportionality as a ground for administrative review is likely to have a better claim to being a general principle of law recognized by the major legal systems of the world than legitimate expectations. It has also been adopted by the EU courts. The question of whether there was a promise of immutability that might generate a
56 Decision, §193.
57 2nd WS Kruck, §§3-6.
58 Decision, §191.
[Page 29]
legitimate expectation is avoided altogether. Instead, in accordance with the most developed conception of proportionality in German law, a three-stage test applies: (i) whether the disputed measure was suitable for the attainment of the desired objective; (ii) whether the disputed measure was necessary, in the sense that the state authority had no other option that was less restrictive of the individual's interests; and (iii) whether the measure was disproportionate to the restrictions thereby applied.59
104. At first blush, this appears to be a very invasive test (and indeed, the English courts have resisted adopting a proportionality test largely on the ground of the risk the courts will conduct a merits review of administrative decisions). The reality, however, is that courts apply proportionality with relatively low intensity in cases like this one where there are policy choices requiring the weighing of complex variables. The EU experience confirms this. In these cases, administrative decisions are quashed only where the change of policy is “manifestly disproportionate”.60 Courts recognise that they are neither entitled, nor have the expertise, to second-guess the careful weighing of competing political, economic or social factors undertaken by state authorities and thus apply a higher threshold for annulling such decisions. In contrast, there is higher intensity of the review in cases involving the fundamental rights of individuals.
105. If proportionality were to be adopted as stand-alone element of the FET standard, then a further adjustment would have to be made in recognition of the fact that the remedy is for damages in investment arbitration rather than the typical administrative remedy of annulment of the decision. That may, in common with the EU approach to awarding compensation for breach of legitimate expectations, require a higher threshold for imposing liability.
106. The adoption of proportionality as the criterion for liability in disputes like the Spanish solar cases would give a new lease of life to the voluminous exchanges between regulatory experts whose efforts are largely ignored in the awards (if strict liability is imposed then there is little point in considering the reasons for the adoption of the disputed measures and the other alternatives available to the state authorities). But caution must attend this exercise. An argument set out in the Claimants' Reply will suffice to illustrate the point:
Spain had other methods at its disposal to gain control over the tariff deficit, rather than cut incentives to PV facilities. The most obvious measure was to raise access tariffs for consumers, so that the prices that they paid for electricity more accurately reflected the actual costs of generating that electricity. As [the Claimants' regulatory expert] explains, Spain's failure to impose accurate costs has kept consumer prices artificially low.61
107. There is no doubt that, as a matter of economic and regulatory theory, Spain could have preserved the tariffs in RD 661/2007 and addressed the tariff deficit (which by
59 J Schwarze, European Administrative Law (Sweet & Maxwell, 2006) 685-7.
60 P Craig, EU Administrative Law (OUP, 2018, 3rd edn) 644.
61 C's Reply, §554.
[Page 30]
2010 exceeded €25 billion) by raising prices for Spanish consumers. The problem is that when the disputed measures were introduced in 2013, the Spanish economy was in a deep recession triggered by the Global Financial Crisis, and the official unemployment rate in Spain was 26.1%.62 That is higher than it was for the United States during the Great Depression of the 1930s. The unemployment rate among those under 25 in Spain was reported to be above 50%. So as a matter of reality, there is unlikely to have been a great deal of capacity among Spanish electricity consumers in 2013 to pay for a considerable increase in electricity prices so that investors in PV facilities could avoid any haircut to their tariffs under RD 661/2007.
108. If a concept of proportionality were to be applied, international tribunals would, like the EU courts and national courts, have to calibrate the intensity of the review in circumstances where a state authority has had to weigh complex and competing interests to reach a compromise that by definition cannot give full force and effect to the demands of any particular group. They would also have to bear in mind that their judicial mandate is to achieve corrective justice when a State has committed a wrong rather than to indemnify foreign investors against the impact of a fair and transparent redistribution of scarce resources by a State in crisis.
109. Other tribunals in the Spanish solar cases have sought to avoid the “all-or-nothing” approach to stability by sticking with the doctrine of legitimate expectations and interpreting RD 661/2007 as not generating an expectation among investors of the immutability of the tariffs for the special regime but as rather an expectation that they would continue to earn a reasonable rate of return.
110. In my assessment, the fact that the tariffs set in RD 661/2007 have to be based on the concept of a “reasonable rate of return” as set out in Law 54/1997 is one of the reasons that Article 44(3) RD 661/2007 cannot be interpreted as guaranteeing the immutability of the tariffs. If those tariffs, for whatever reason, fall out of line with the concept of a reasonable rate of return, then the Spanish Government would have to repeal RD 661/2007 and start again. This is precisely what had happened with the previous RD 436/2004 (and the Claimants admit that no legitimate expectations could have arisen from that regulation given the extent of the unexpected hikes in the tariffs).
111. It is another thing altogether to interpret RD 661/2007 as guaranteeing a reasonable rate of return but not the precise tariffs set out therein such that an expectation of the former would be deemed to be legitimate. I agree with the majority that the precise tariffs in RD 661/2007 reflect what the Spanish Government at the time thought was a reasonable rate of return. It is wrong to interpret RD 661/2007 as providing the Spanish Government with a discretion beyond the mechanism in Article 44 to adjust the tariffs to reflect a reasonable rate of return while RD 661/2007 was in force.
112. By trying to avoid the trap of strict liability, tribunals have thus walked into another, which is to place an artificial construction on RD 661/2007. They have started from the wrong premise—immutability of something—and tried to restrict its scope by
62 AMG First Report, p. 16.
[Page 31]
limiting it to a reasonable rate of return rather than the precise tariffs in RD 661/2007.
113. In my preferred approach, the concept of a reasonable rate of return would play an important role in the proportionality assessment. In determining whether the disputed measure placed a disproportionate burden on the investor, it would be highly relevant to assess whether the investor continued to earn a reasonable rate of return after the disputed measure came into force. This again would breathe life into the voluminous exchanges between the quantum experts in this case as to what the “target” rate of return was for PV investments under RD 661/2007, what the average rate of return is for such investments on a comparative basis, and what rate of return the Claimants' projects are now earning under the new regime.
* * *
114. I dissent from the majority's decision upholding liability based on the FET standard in Article 10 ECT as well as its decision on the principles of quantum.
Signature
Prof. Zachary Douglas KC
Arbitrator
Date: 13 September 2022