Achmea – A Perspective from International (Investment) Law

By Pekka Niemelä

A week has passed since the European Court of Justice (ECJ) rendered the landmark Achmea judgment. A number of posts analyzing the judgment have already appeared in the legal blogosphere (see e.g. here, here, here and here). Much of this commentary has focused on describing the Court’s reasoning and on analyzing the judgment’s broader implications. Most commentators agree that there was nothing unexpected in the Court’s conclusion that the arbitration clauses in the less than two hundred intra-EU BITs have, as the Court put it, an ‘adverse effect on the autonomy of EU law’ (para. 59).

The judgment’s reception has also varied in accordance with the view one has of the underlying purposes of investment treaties – do they promote the international rule of law or narrow corporate interests at the expense of the public interest? Accordingly, those critical of investment treaties and arbitration have welcomed the judgment, whereas the proponents of investment treaties have argued that the judgment leads to less ‘rule of law within the EU’.[1] On a higher level of abstraction, the plausibility of the Court’s reasoning also depends on the view one has of the EU in general: is it an autonomous constitutional order based on the protection of fundamental rights and certain foundational values? Or should the EU demonstrate more openness towards other international law regimes, as it is just one such regime among others? Depending on the view one has over these two intertwined general questions, Achmea can either appear as a logical corollary of EU constitutionalism or as a breach of the EU’s commitment to the international rule of law.

What this blogpost strives to do is to take issue with the Court’s understanding that arbitral tribunals interpret and apply EU law in ways that pose a threat to its autonomy. The point is not to argue that the Court’s reasoning and conclusions are incorrect, but to shed light on the ways in which arbitral tribunals have actually ‘used’ EU law, and to show that the Court’s understanding (with which most commentators sympathize) that investment arbitration poses a threat to the autonomy of EU law is somewhat inflated.

Adverse Effect – Or Very Little Effect?

In a nutshell, the Court argued that if a court or tribunal potentially rules on a matter ‘covered by EU law’ (para. 55), and if that court or tribunal is situated outside the EU judicial system, the autonomy of EU law is under threat, even if EU law is only occasionally relevant to the disputes over which arbitral tribunals exercise jurisdiction. In other words, the fact that the EU has adopted legislation in fields related to intra-EU investments ‘pre-empts’ the creation of courts and tribunals that deal with disputes that may relate to those fields.[2] The Court’s construction of the autonomy of EU law has a strong normative undertone, and its reasoning relies on the idea that the EU forms a constitutional order founded on the values enshrined in Article 2 TEU. This abstract idea is complemented by the more practical concern about the ‘full effectiveness of EU law’ (para. 56), a central part of which is the availability of ‘remedies sufficient to ensure effective judicial protection for individual parties in the fields covered by EU law’.[3] This latter concern, in turn, relies on the idea that EU law provides (through the medium of member state courts) such effective remedies and that by transferring disputes to investment arbitration (or by creating the possibility of such transfer), something short of full effectiveness follows. The judgment’s references to the principle of mutual trust, and to the idea that the member states respect the EU’s foundational values by implementing EU law effectively, creates the perception that arbitral tribunals not only decrease the ‘full effectiveness’ of EU law, but also apply and implement values other than those listed in Article 2 TEU.

As noted, it is useful to test whether these concerns are justified in light of arbitral practice, so the question is how have arbitral tribunals interpreted and applied EU law? The first observation is that EU law instruments or arguments are not necessarily raised in intra-EU arbitrations, and a quick glance of the publicly available awards indicate that this holds true with respect to a large number of other intra-EU arbitrations.[4] Quite often, the challenged measures relate to policy areas where member states enjoy wide discretion (if not exclusive competence), which begs the question of whether the measures relate to a ‘field covered by EU law’ if the measures only have to comply with the fundamental freedoms and the principle of non-discrimination and have no direct connection to the requirements of specific EU acts (such as directives and regulations). For example, in the IP Busta arbitration the claimants were UK citizens who had formed a joint venture with a Czech company, and the latter had moved (or stolen) certain goods owned by the claimants, which the local police later seized and returned to Messrs. Busta. The claimants alleged that the Czech police had not returned certain goods and had also failed to make an itemized list of the goods as Czech law required. The question was whether the actions of the Czech Police breached the fair and equitable treatment standard of the Czech-UK BIT. While the EU has issued a number of directives in the field of criminal law, the above scenario is by and large ‘domestic’ as the claimants could not resort to an EU law remedy, and one could argue that the dispute does not come within a field covered by EU law, as it only relates to the mundane actions of the local police. Similarly, in some of the arbitrations where specific EU law instruments have been referred to, the tribunal’s use of EU law has clearly not had any impact on the ‘full effectiveness’ of EU law or its autonomy. For example, in Maffezini the tribunal simply noted that an EU directive required the claimant to carry out an environmental impact assessment, period. This suggests that intra-EU BITs provide an effective remedy to investors in particular in circumstances to which EU law is not directly relevant as it does not dictate the contents of domestic policy. A related observation is that intra-EU BITs have also provided more effective remedies to EU nationals in situations pertaining to breaches of the fundamental freedoms, without any adverse effects on the autonomy of EU law necessarily following (see below). The Court’s contention that EU law provides for a complete system of remedies, or at least remedies ‘sufficient to ensure effective judicial protection for individual parties in the fields covered by EU law’ (Case C-64/16, para. 34) has to be understood as a formalistic conception in the sense that BITs clearly provide more complete and effective remedies to investors than EU law or domestic law – and this understanding has been at the heart of the reasoning of arbitral tribunals in cases where they have rejected the argument that intra-EU BITs are incompatible with EU law.[5]

The Achmea arbitration itself provides a perfect example of this ‘remedial’ ambivalence: the Commission started infringement proceedings against the Slovak Republic on the basis of a complaint concerning the legislative change that prohibited Achmea (and other companies offering health care services in the Slovak Republic) from distributing its profits freely, but since the Slovak constitutional court declared the prohibition unconstitutional and thereby void, the infringement proceedings were discontinued. Now, the Court’s case law allows companies to seek compensation for alleged breaches of the fundamental freedoms (in this case, free movement of capital), but on less generous criteria than the criteria of liability under BITs, and, as is well-known, the amounts of compensation awarded under BITs typically exceed those available under EU law and domestic law. In particular, compensation for economic loss occurring as a result of general legislative acts (such as the one in Achmea) is not easily forthcoming under EU law or national law, whereas in Achmea the tribunal awarded the claimant investor some 22 million euros in damages. Depending on one’s underlying view on the EU and investment arbitration, the judgment’s implications for the autonomy of EU law could be read in two ways. First, if you’re pro-investor and pro-market, the tribunal’s decision could be seen as directly supporting the principle of free movement of capital within the internal market – as increasing its ‘effectiveness’ , to use the Court’s own word. Second, if you’re critical of investment arbitration and think that EU law provides adequate remedies to economic operators, it’s logical to argue that it’s up to the ECJ to determine the criteria of non-contractual liability in situations concerning the basic freedoms, including the question of damages. This view could be backed up with the observation that the Court’s liability criteria is based on principles derived from the legal systems of the member states, which in turn are premised on a more nuanced balancing between the attendant public and private interests (in comparison to the balancing, or lack thereof, under investment law). There are other similar cases where EU investors have challenged domestic measures breaching one or more of the fundamental freedoms, with arbitral tribunals awarding compensation which was not available to the affected investors under EU law (not to mention domestic law).[6] One relevant question is whether EU law allows the member states to provide compensation in these type of cases; if it does, then a plausible argument is that Achmea and analogous cases do not adversely affect the autonomy of EU law.

The general Role of EU Law in investment arbitration

´The disputing parties can invoke EU law related arguments in intra-EU arbitrations if this is considered a good litigation strategy and if the facts and legal materials facilitate their use, and this holds true irrespective of the wording of the arbitration clause or the content of the arbitration rules that govern the dispute. The choice of law clauses in BITs, when they exist, place no restrictions in this regard either. In other words, whether or not an intra-EU BIT contains a provision guiding the tribunal to interpret or apply EU law when deciding the case, EU law arguments can be raised. In principle, however, EU law can have two different roles in investment disputes; it can either be a factual element or part of the applicable law. But what does it mean to say that EU law is part of the applicable law? In principle, it means that EU law applies to the merits of the dispute together with the provisions of the BIT and any other applicable rules of international law. While this possibility is clearly problematic in light of the autonomy of EU law, the question is whether arbitral tribunals have done so in practice.

Generally speaking, treaty obligations stand on equal footing under international law. Hence, two distinct arguments could be made on the relationship of EU law and investment law if a member state wants to argue that EU law is part of the applicable law. First, the member state could argue that its obligations under EU law take priority over its BIT obligations, in which case the priority of one obligation over the other can only be established by applying conflict rules, but such arguments have not been raised in intra-EU arbitrations (apart from conflict arguments challenging the jurisdiction of the tribunals which is a different matter). Second, the member state could argue that its EU law and investment law obligations constitute a harmonious set of obligations, implying that its compliance with EU law cannot constitute a breach of the BIT. This argument has been raised in arbitral practice a few times, but in only one arbitration (to my knowledge) has it formed the basis of the Court’s reasoning, meaning that the tribunal accepted the primacy of EU law over the relevant investment treaty. I will discuss this arbitration briefly below. Generally speaking, it is difficult to think of other ways in which EU law could be ‘applied’ in investment arbitration.

In most of the arbitrations where the second argument (about harmonious interpretation) was raised, the tribunals held that EU law was an element of fact. This meant that the arbitral tribunals took into account the relevant EU law instruments when analyzing, to quote one tribunal, the ‘rationality, reasonableness, arbitrariness and transparency’ of the challenged member state measure.[7] However, even in these cases EU law did not influence the tribunals’ analyses because the challenged measures did not have any tangible connection to the EU act which the member state invoked as part of its defense. Moreover, in the few cases where the tribunals have discussed EU law extensively, they have typically acknowledged that if the challenged measure would have flowed from the requirements of EU law, this would have been an indication that the measure is a ‘rational public policy measure’,[8] which cannot breach the relevant BIT. In other words, the tribunals have acknowledged that the implementation of EU law, generally speaking, cannot lead to a successful claim under the relevant BIT.

This is not to say that arbitral tribunals have not engaged with EU law in ways that would not appear as problematic from the perspective of autonomy. There is no point in discussing the fascinating Micula saga, as that arbitration concerned measures adopted prior to Romania’s EU accession, and at that stage the candidate states are not bound by the acquis, even if the association agreements require them to approximate their domestic laws to EU law. However, the Electrabel arbitration, which was commenced under the Energy Charter Treaty (ECT), provides an interesting case from the perspective of autonomy. A Commission state aid decision was at the heart of the dispute. The tribunal held, inter alia, that the ECT did not protect investors from the implementation of the state aid decision as the member states’ obligations under the ECT and EU law formed a harmonious set of obligations. In other words, the tribunal acknowledged that if a member state adopts a measure on the basis of EU law, this does not breach the ECT. The tribunal also interpreted the state aid decision to see what requirements it imposed on Hungary and then concluded that the decision is attributable to the EU as a matter of international law. It is clear that the Electrabel tribunal’s use of EU law is poles apart from e.g. the Maffezini tribunal’s engagement, and its discussion on the division of competences between the EU and its member states is problematic in light of Opinions 1/91 and 2/13, where the Court explicitly stated that questions of competence are its sole prerogative. However, one could also argue that the tribunal’s analysis honored the autonomy of EU law by coming to the only plausible conclusion on the question of competence, and the tribunal’s dictum that the ECT does not protect investors from the implementation of EU law is equally respectful of the autonomy of EU law, as it is premised on the invariable primacy of EU law over the ECT.

Between Electrabel and Maffezini, there are many other ways in which EU law can relate to intra-EU arbitrations. To give an example, EU investors have lodged more than thirty claims against Spain under the ECT, which stem from the scaling back of certain solar energy subsidies between 2008 and 2014. Most of the cases are still pending, but the publicly available awards show that the 2009 renewable energy directive was discussed in some of the arbitrations. The 2009 directive contains mandatory national targets as regards the production of renewable energy, which the member states have to achieve by 2020. Interestingly, the Commission has expressly stated that the legislative changes that scaled back the subsidies did not breach the directive. The Commission reasoned that member states ‘retain full discretion over whether they use support schemes or not and, should they use them, over their design, including both the structure and the level of support’. This discretion includes the right to ‘enact changes to their support schemes, for example to avoid overcompensation or to address unforeseen developments such as a particularly rapid expansion of a precise renewables technology in a given sector’. The Commission concluded that the directive provides no grounds ‘to take legal action against Spain with regard to the legislative changes which affected the level of support given to investors in renewable energy projects’, and the affected investors were advised to seek judicial review before Spanish courts. As noted, given that the criteria of state liability for non-contractual breaches are more strict under (in this case) domestic law, Spanish courts have rejected all compensation claims brought by the affected investors, whereas some of the tribunals dealing with the ECT claims have awarded sizeable compensation to the claimants.[9]

Do the more than thirty claims raised against Spain threaten the ‘autonomy of EU law’ by preventing the ‘full effectiveness of EU law’? Clearly, a more plausible answer is that they do not, although the answer will again depend on whether or not one supports the formalistic idea that when a measure falls within a field covered by EU law (however indirectly and however much discretion member states have in that field), the EU judicial system alone has competence to deal with the dispute. Still, if the relevant directive does not dictate the policies with which Spain should reach its renewable energy goals, then changes in those policies, as the Commission itself notes, do not provide any ground for claims of compensation as a matter of EU law. When such wide discretion exists, a plausible argument is that claims of compensation under investment treaties do not adversely affect EU law. The Court’s concern in Achmea about autonomy and about the effectiveness and availability of remedies in the fields covered by EU law gives little recognition to the varied substance of investment disputes and to the relatively marginal role that EU law has had in individual investment disputes. Neither does the Court’s approach recognize that arbitral tribunals appear to respect the autonomy of EU law, as also testified by the above quotes from intra-EU arbitral awards. In one sense, the Court’s autonomy concern is based on a juristic fiction of a perfect legal system that contains no gaps or lacunae in the protective safety-net it offers to its subjects, and the glorified references to the principle of mutual trust and the EU’s foundational values sound somewhat empty in light of arbitral practice and in light of the fact that arbitrary exercises of public power continue to take place within the internal market, with EU law having little to offer to the affected investors.

However, problems remain – state aid and investment arbitration

These remarks do not entirely eliminate the Court’s autonomy concerns as arbitral tribunals may well engage with EU law in ways that have an adverse effect on its autonomy. Again, the solar energy claims against Spain provide a good example. In November 2017, the EU Commission decided that some of the legislative changes that scaled back the solar energy subsidies did not constitute illegal state aid under EU law. A central EU law principle is that economic operators cannot entertain legitimate expectations over the lawfulness of state aid they have received unless the aid scheme was notified to and authorized by the Commission before its application.[10] Hence, as a matter of EU law, the investors who benefited from the solar energy subsidies could not entertain legitimate expectations about the legality of the aid, and neither could they (as noted) challenge the scaling back under EU law as Spain enjoys wide discretion on how to reach its renewable energy goals. Given that the Commission’s state aid decision did not include an analysis of the original aid scheme, its decision noted that any compensation awarded to investors ‘on the basis that Spain has modified the premium economic [i.e. the original] scheme by the notified [i.e. amended] scheme would constitute’ state aid. The Commission also observed that arbitral tribunals are not ‘competent to authorize’ the granting of state aid, and if they do so, ‘this compensation would be notifiable State aid pursuant to Article 108(3) TFEU and be subject to the standstill obligation’.

One would assume that arbitral tribunals presiding over the claims against Spain would pay heed to the basic principles of EU state aid rules, and even more so after the Commission’s quoted decision. Commonsense also dictates that all economic operators within the internal market must, as diligent businessmen, be aware of the basic requirements of EU state aid rules and that investors should not be able to challenge changes to aid schemes when the host state has failed to notify the scheme accordingly. However, just last month, the Novenergia tribunal held that the Commission’s state aid decision is not relevant in the proceedings because the tribunal ‘was not applying EU law’ (para. 465), with the Commission’s decision being ‘entirely irrelevant to the determinations pertaining to this Tribunal’ (idem.).[11] The tribunal awarded the claimant some 53 million euros in damages, but in light of Achmea the award is clearly unenforceable within the EU as it is in direct breach of EU state aid rules, thereby breaching EU public policy which constitutes a ground of annulment of the award under EU law. Generally speaking, whether or not the original aid scheme constitutes illegal state aid (there are strong indications that it doesn’t), the tribunal should have acknowledged that the claimant investor is bound by EU state aid rules and that the satisfaction of the award is conditioned on the Commission’s finding on the compatibility of the original scheme with EU state aid rules.


Most commentators have read the Court’s judgment as implying that the same line of reasoning could be transposed and applied with respect to other intra-EU BITs and the ECT. I agree with this general assessment, but I would extend it to extra-EU BITs (i.e. treaties between member states and third states) as well. While extra-EU BITs are governed by the Grandfathering regulation, that regulation provides that the member states ‘are required to take the necessary measures to eliminate incompatibilities’ between EU law and extra-EU BITs. While the principle of mutual trust does not obligate third states, it is difficult to see how extra-EU BITs could not give rise to the same abstract concern that formed the basis of the Court’s conclusion in Achmea. If a third state investor brings a claim against a member state, it may well be that the tribunal will have to interpret (or even apply) EU law in similar ways as tribunals established under intra-EU BITs. Since those tribunals are situated outside the EU judicial system, it is difficult to see how the Court’s concern would not extend to them as well. It is unlikely, however, that the Commission or any of the member states will raise the matter given the high political premium placed on the continued existence of extra-EU BITs. Commentators have also argued that the judgment has no bearing on arbitrations carried out under the ICSID Convention, and while this is true in a formal sense, it seems equally clear that ICSID tribunals may interpret and apply EU law, with the Court’s conclusion extending, by analogy, to them as well.

It remains to be seen what the reactions of the member states and the Commission are, and what will happen in the pending (BIT and ECT) intra-EU arbitrations. If one takes the perspective of counsel representing an investor in an imagined intra-EU arbitration not (directly) related to EU law, one plausible argument is that since the arbitration stems from a purely domestic measure, the autonomy concerns of the Court are entirely immaterial to the dispute at hand. Similarly, the entire judgment and the Court’s main conclusion were premised on the situation where an arbitral tribunal interprets and applies EU law, indicating that the judgment has no impact on pending proceedings where EU law plays no (or only an indirect) role. Moreover, as the relevant BIT remains in force, and as the contracting states have not taken any other relevant action either, the tribunal cannot work from the premise that Achmea renders the BIT void or inapplicable as a matter of international law. There is clearly room for interpretation as to what the Court’s finding on autonomy means, and it is up to the contracting states to determine what its implications are.

[1] Nikos Lavranos, ‘Insight: The end of intra-EU BITs – what next?’, Borderlex News, 9 March 2018 (requires subscription).

[2] I owe this general point to Harm Schepel.

[3] The quote is from Case C-64/16, para. 34, to which the Court referred in para. 55 of the judgment.

[4] See e.g. Gavazzi v. Romania, ICSID Case No. ARB/12/25, Award, 18 April 2017; WNC v. Czech Republic, PCA Case No. 2014-34, Award, 22 February 2017.

[5] See e.g. Binder v. Czech Republic, Award on Jurisdiction, 6 June 2007; Eastern Sugar B.V.(Netherlands) v. The Czech Republic, SCC Case No. 088/2004, Partial Award, 27 March 2007.

[6] See e.g. Edenred S.A. v. Hungary, ICSID Case No. ARB/13/21; Le Chèque Déjeuner and C.D

Holding Internationale v. Hungary, ICSID Case No. ARB/13/35.

[7]AES Summit v. Hungary, ICSID Case No. ARB/07/22, Award, 23 September 2010, para. 7.6.9.

[8] Ibid., para. 10.3.16.

[9] These cases also have a state aid component, which I discuss below.

[10] See e.g. Case T-179/09, Dunamenti Erőmű Zrt. v Commission, ECLI:EU:T:2014:236, para. 104. If there are exceptional circumstances which caused the beneficiary to assume that the aid is lawful, such circumstances can play a role ‘only in resisting the possible recovery of that aid’.

[11] Novenergia v. Spain, SCC Arbitration (2015/063), Award, 15 February 2018 (available at the IAReporter news portal (requires subscription).