PDF Version
Vol 11, Issue 6
Editorial board: Federico Casolari, Patrycja Grzebyk (editor-in-chief), Ellen Hey, Guy Sinclair, and Ramses Wessel
Jan Klabbers (University of Helsinki)



Photo by Markus Distelrath (cc)

I. Introduction

“Making accurate predictions is difficult – especially when it concerns the future.” These words, attributable to an anonymous futurologist, come to mind when reflecting on the spate of announced withdrawals from the Energy Charter Treaty, that curiously named remnant of early 1990s global optimism (a Grotian Moment, remember?). Several states, including Germany, France, Spain and the Netherlands, have announced their intention to withdraw,[1] to loud applause from the international law community. The latter is remarkable: usually treaty withdrawals are met with cries of despair, with accusations of ‘digging the grave’ of international law and ascribed to a general wave of anti-multilateralism. Not so here, and the reason is, no doubt, the realization that some treaties are actually not worth keeping, and this group includes the Energy Charter Treaty (ECT). The ECT firmly locks into place investments into the traditional energy sector (read: fossil fuels), and therewith presents a formidable practical obstacle to the transition to greener energy. One can sympathize with the nay-sayers, but what is the legal situation? I will argue that withdrawal is mostly a symbolic act of little practical legal consequence, and will suggest a possible alternative avenue.

The Energy Charter Treaty was concluded in the early 1990s, and essentially represented a bargain between the West and the resource-rich states of eastern Europe. The latter needed investments to get their economies going after years of communist paradise, lethargy, and apparatchik exploitation – but who would want to invest significant sums there without any legal guarantees? The West, in return, realized its own energy-dependency, and realized that investing in eastern Europe’s energy sector could secure energy supplies for the foreseeable future, and guarantee handsome profits as well. Clearly, this seemed the textbook example of a win-win situation: the West would secure energy, while the East would receive much needed investments. Even a few more distant geopolitical goals would be achieved: eastern European states could this way become stable market economies, integrated in the western world. And best of all, this was expected to include the Russian Federation which, after all, has huge reserves of natural resources and is of some geopolitical significance. The initial Russian signature on the ECT marked thus not merely a win-win setting, but something approximating a win-win-win-win type of situation: everybody wins, no one loses.[2]

Put like this, alarm bells should have been ringing, as it is nigh-on impossible to create a regime where no one loses: the win-win situation is a fantasy, spun by politicians and ideologues. It took a few decades, but then it turned out that the losses had always been externalized, in the best tradition of economistic thinking: the losers are the environment, and therewith you and me. Hence the general applause coming from the international law community whenever a state announces to withdraw from the ECT: pacta sunt generally servanda, but not all pacta deserve to live forever.


II. The Law of Withdrawal

The ECT drafters envisaged, as they should, that parties may at some point wish to leave the regime – the ECT duly contains a withdrawal clause in article 47. Any party may, after an initial period of five years, announce its withdrawal by giving written notification to the depositary (i.e., Portugal). The withdrawal takes effect one year after receipt of the notification, or at such later date as is specified in the instrument of withdrawal. So far so good: this looks straightforward enough, but there is a sting: paragraph 3 of the same article provides that the provisions of the ECT shall continue to apply to foreign investments done under the ECT. When drafting the ECT, such a provision made eminent sense: the return on investments usually takes time to mature, especially in such sectors as energy with a large dependence on infrastructure. Allowing investment protection to evaporate after a mere twelve months (synchronous with the withdrawal itself) would have done little to put potential investors at ease – hence the inclusion of this sunset clause. The sunset clause is clearly intended to protect investors rather than the interests of the contracting parties (although these can go hand in hand), and is well worth reproducing here, if only because it is vaguely reminiscent of a famous Marx Brothers sketch:

“The provisions of this Treaty shall continue to apply to Investments made in the area of a Contracting Party by Investors of other Contracting Parties or in the area of other Contracting Parties by Investors of that Contracting Party as of the date when that Contracting Party’s withdrawal from the Treaty takes effect for a period of 20 years from that date.”

Like so many other provisions of the ECT (just have a look at article 45 ECT…) the drafting could probably have been more elegant, but the main gist is nonetheless clear: a state may withdraw from the ECT, but withdrawal does not terminate investment protection: the withdrawing state cannot escape from protecting foreign investments for another twenty years.

Put differently, a German (or Dutch, or Spanish) withdrawal from the ECT taking effect in 2023 effectively means that Germany will be released from a number of obligations, as is ordinarily the case with withdrawal from a treaty, but will have to continue its investment protection. Obligations it will be released from include the obligation to “alleviate market distortions” in the energy sector (article 6, §1) or facilitate the transit of energy materials and products (article 7, §1). The latter in particular is not to be sniffed at: it could possibly wreak havoc to plans to build pipelines. But a withdrawing Germany will not be released from the obligation to protect investments for another twenty years, and the heart of the ECT lies precisely in its investment protection regime.

So, inevitably the question arises: can a state withdraw not only from the ECT but also withdraw lawfully from the twenty-year protection regime set up by the ECT, or is the sunset clause immune? The wording and function of the sunset clause suggest it is indeed immune: the very idea behind it is to continue investment protection also after a party’s withdrawal, as the wording makes clear. There is a twist though: article 47, §3, sees to situations where investments need protection against the actions of other states, not the actions of the withdrawing state. Put simply: when Germany withdraws, German investors in, say, Bulgaria, Moldova and Yemen continue to be protected for another twenty years, and vice versa: Bulgarian investments (and those of the other ECT parties) continue to be protected in Germany. But the current spate of withdrawals is driven – nominally, at any rate – by something else: by Germany (to stick to the example) effectively aiming to terminate the protection of German investments in Bulgaria, so as to facilitate a greener economy. Germany is less troubled by the reverse process, one may expect: the level of Bulgarian investments in Germany is probably not that high. In short: the western investor needs to be protected against their own state, not against the eastern host state.[3] The latter was the idea behind article 47, §3, but has little traction in the current circumstances – article 47, §3, was not written with a view to protecting German investors from German actions and thus, one might think, has no bearing on the current situation – except of course that investors might see their investments affected if the sunset clause were no longer lived up to, and tribunals might agree with them.

Proper legal argument is thus required to justify withdrawal from the sunset clause, and can take several forms. First, Germany could argue – ex hypothesi – that when it withdraws from the ECT, it automatically also withdraws from article 47, §3. This argument, however, is bound to fail, as it is clear that article 47, §3, was drafted precisely so as to prevent this from happening – otherwise it would be utterly useless. There is, moreover, analogous backing by some of the case-law of the International Court of Justice (ICJ) which held, in the early 1970s, that Iceland’s unilateral termination of several fisheries agreements could not affect the dispute settlement clauses in those agreements, including the possibility of seizing the ICJ.[4] That said, the Court did not (and in casu did not need to) indicate a possible time-limit, and it may be presumed that compromissory clauses will not normally last for twenty years after withdrawal.

A second variant is to suggest that the sunset clause, considered separately as something of a continuation agreement between the ECT parties, can also be withdrawn from separately. On such a construction, a state could withdraw from the ECT according to article 47, §1, and separately withdraw from the sunshine clause of article 47, §3. But how? The ECT naturally does not contain a procedure for terminating the sunset clause, and the Vienna Convention on the Law of Treaties (VCLT) also offers little guidance. The VCLT has no explicit provisions on the topic of withdrawal from sunset clauses. Instead, it by and large views treaties as contracts writ large, based on mutually beneficial bargains, the basic assumption being that treaties form an interdependent whole. Article 44 VCLT accordingly lays down a general rule (with some exceptions though) that states cannot withdraw from individual treaty provisions unless the treaty itself so provides or the parties otherwise so agree. This has no immediate application on sunset clauses though, as these continue to exercise legal force despite the withdrawal from the instrument containing them – they apply ‘from beyond the grave’, so to speak, and mirror the situation addressed in article 44 VCLT.

Should the sunset clause be seen as creating something of a mini-regime following withdrawals from the ECT, the closest general provision contained in the VCLT is article 56, relating to treaties not containing a withdrawal clause. In such a case, the VCLT cautions, the treaty concerned is not subject to withdrawal, unless the parties intended it to be subject to withdrawal or such follows from the nature of the treaty. So, construing the sunshine regime as a treaty regime in its own right, the conclusion must be obvious: the sunset regime is not in the nature of a treaty that can be subjected to withdrawal, nor did the parties somehow agree to withdrawal – the very point of the sunset clause is to guard against the effects of withdrawal from the treaty in which it is embedded.

This then entails that a justification must be looked for elsewhere: in articles 60-62 VCLT, allowing for unilteral withdrawal in certain limited circumstances: a material breach, force majeure, or a fundamental change of circumstances. The first two of these grounds warrant no further discussion, as they do not apply – no one mentions breach of treaty, and no one speaks of having to leave the ECT as the result of a supervening impossibility of performance or an act of God. That leaves the argument of the fundamental change of circumstances. This has traditionally been approached very cautiously by international lawyers, including courts and tribunals, as it stands in obvious tension with the pacta sunt servanda principle. Even the positioning in the VCLT reflects this: the pacta sunt servanda principle is contained in article 26, while the rule on fundamental change of circumstances can be found in article 62.[5]

The wording of article 62 is already limiting. The doctrine may not be invoked (note the negative formulation) unless the circumstances concerned “constituted an essential basis” for the parties to consent; unless the change was unforeseen; and unless the change in circumstances radically transforms the extent of obligations still to be performed. Courts and tribunals have been extremely hesitant to uphold appeals to the doctrine of fundamental change, interpreting these conditions as cumulative (all of them must be met), and interpreting them narrowly. The one exception, moreover, was a case where the court concerned (the European Court of Justice) managed to get both the facts and the law wrong – albeit possibly for good reasons. This was the Racke case,[6] in which the ECJ upheld the EU Council’s argument that its suspension of the EU-Yugoslavia free trade agreement following the outbreak of civil war in Yugoslavia was justified on grounds of a fundamental change of circumstances. It suggested that since treaties are concluded against a background of peace and stable state institutions, the outbreak of civil war was best seen as a fundamental change of circumstances of the sort meant in article 62 VCLT. This was problematic on many levels, and in the end, the Court’s position was best summed up in its suggestion that when one party no longer sees “the point” in performance of treaty obligations, the treaty can be terminated, suspended or withdrawn from.[7]

Yet, if article 62 is taken seriously, it seems obvious that it cannot be relied on. Perhaps it is arguable that the need to protect the environment formed an essential basis for the ECT: the ECT is filled with references to sustainable development and the need to take the environment seriously. Yet, this automatically also means that it cannot be argued that the change was unforeseen: such an argument failed in the Gabcikovo-Nagymaros case, and is bound to fail here as well.[8] Moreover, it is not easy to see why further performance of investment protection for the next twenty years would be radically transformed – it has not beome more difficult or more onerous to protect investments, only more unpopular. The risks of climate change and continued use of fossil fuels have been well known for quite some time: the parties knew very well what they were embarking on, back in 1994, and decided to go ahead regardless. In such a setting, it becomes problematic to rely on the fundamental change of circumstances doctrine as a justification for non-performance or withdrawal. What has changed are not so much the material circumstances, but the political landscape: there is considerable political pressure in western states to speed up the transition to a greener economy – more so than in the 1990s, and rightly so. But relying on the fundamental change of circumstances doctrine is quite a stretch. It should additionally perhaps also be highlighted that Italy, which already withdrew from ECT in 2014, does not seem to have added any claim to get rid of the sunset clause as well – it is still called upon as respondent in disputes with investors under ECT, most recently in a case brought in 2020.[9]


III. Alternatives?

It is easy to see why the sunset clause is considered problematic: under the sunset clause, withdrawing states will have to continue to protect investments from foreign investors for another twenty years, and this potentially covers investments from over fifty states (in November 2022 the ECT has 53 contracting parties, including the EU). On the plus side, investments done by their own investors will also continue to be protected, but this may not compensate. As so often, international law offers only blunt tools to adapt to a novel (or perceived as novel) situation or political configuration.

That said, even a blunt tool may be better than no tool at all, and another blunt tool available here might be the revision of the ECT, either through amendment or through modification between sets of parties. The amendment route seems effectively closed off. The ECT allows for amendments (in article 42), but these require unanimous approval by the Charter Conference. It would take a long and intensive political process to get all parties singing from the same hymn sheet, so not too much should be expected from this – an amendment proposal to delete or shorten the sunset clause is unlikely to meet with the approval of all 53 parties without buying reluctant states off, and even then it would upset investors who might threaten legal action.

Another option though, and a more promising one, is that of modifying the ECT regime between like-minded parties inter se. This is allowed, under article 41 VCLT, as long as it does not affect the rights of other parties, and as long as it does not relate to provisions “derogation from which is incompatible with the effective execution of the object and purpose of the treaty as a whole.” What this means concretely is that the states now announcing their withdrawal could inter se decide to de-activate the sunset clause in relations between them: it is precisely the bilateralist assumption underlying the ECT which enables this. The bilateralist assumption relates to this: relations under the ECT (as under many, but not all, multilateral treaties) can often be analyzed in terms of pairs (or dyads) of parties – the multilateral ECT in reality reflects bundles of bilateral relations.

For instance, should France and Germany decide between themselves not to apply the sunset clause, such does not hurt Turkmenistan or Yemen or Moldova – it leaves them unaffected, as it only affects France and Germany. Neither can it seriously be maintained that such a de-activation of article 47, §3 ECT between France and Germany would be incompatible with the effective execution of the ECT as a whole. It would thus meet both criteria set by article 41 VCLT.

Of course, the real problem with this construction lies elsewhere: those who suffer, in the hypothetical example, would be French investors having invested in Germany, and German investors having invested in France. And these investors are unlikely to take this lying down, quite probably relying on such doctrines as the legitimate expectations doctrine to secure their investments, likely also in court. For if civil society organizations can push governments through ‘strategic litigation’ to abide by their treaty commitments, so too can the private sector – and the latter has deeper pockets to begin with and might find a sympathetic ear with tribunals. On the other hand, the expectation when the ECT was concluded was that money would flow from West to East, not from West to West – chances are that French investments in Germany and German investments in France have remained relatively small.[10] If so, then it follows that the de-activation of the sunset clause between France and Germany alone will not save the planet, but then again, it might be better than nothing, especially if other states do the same thing inter se, and unlike relying on the fundamental change of circumstances doctrine, it does not affect the pacta sunt servanda principle.

A possible counter-argument would be that a bilateral de-activiation would amount to a reservation plus acceptance thereof in bilateral relations between dyads of states, while the ECT expressly prohibits the making of reservations in article 46. This shows how complicated the law of treaties is, really, and how much it matters whether an activity is framed as an impermissible reservation or an allowable modification inter se – there are no strict boundaries between the two, at least not under the VCLT, where both regimes (on reservations and on modification inter se) are based on reducing multilateral agreements to bundles of bilateral relations. Still, in light of party autonomy and the freedom of contract among states, it is difficult to see why, as a matter of principle, states would be precluded from modifying a treaty inter se as long as their modification meets the requirements of article 41 VCLT.


IV. To Conclude

If there is one thing the episode demonstrates, it is that while international law may operate through states, it rarely turns those states themselves into winners and losers. Instead, the winners here are the investors; the losers are all of us, due to the urgent need for green transition, and if compensation must be paid, it will have to come from taxpayer’s money. A mere withdrawal by a number of states seems too facile, and it is difficult to justify in international law terms.

The simple withdrawal from the ECT without affecting the sunset clause amounts to taking a moral holiday: grandstanding, without actually doing much and hoping that symbolic action will be politically effective, or else just hoping that the clock will tick the next twenty years away and the planet does not perish in the meantime. And one suspects that the states announcing their withdrawal are acutely aware of this, allowing them to kill two birds with one stone (though, sadly, without saving the planet). They can engage in ‘greenwashing’ by appeasing climate activists and assuring them that withdrawing is the right thing to do. They can equally well engage in appeasing the investors, knowing fully well that the sunset clauses are carved in stone. If states really want to do something for the environment, they should not withdraw, but withdraw and de-activate the sunset clauses in their bilateral relations, even if doing so may prove costly in light of the compensation most likely payable to aggrieved investors.

In the end, one thing is clear though: international law is much like Gary Lineker’s definition of football tournaments: a game in which a number of states take part and which will be won by Germany. International law involves 200 states, and is invariably won by capital.



[1] See https://www.euronews.com/my-europe/2022/10/26/what-is-the-energy-charter-treaty-and-why-is-it-so-controversial  (visited 12 November 2022).

[2] Russia never ratified though, having indicated in 2009 not to do so.

[3] Or southern: a whopping third of the cases brought to date under the ECT have been cases against Spain (51 out of 150). See https://www.energychartertreaty.org/cases/statistics/ (visited 12 November 2022).

[4] See, e.g., Fisheries Jurisdiction (Federal Republic of Germany v Iceland), Jurisdiction of the Court, [1973] ICJ Reports 49, § 29.

[5] There is as far as I am aware no authoritative monograph on the topic available in English; the closest perhaps is the broader study by Arie E. David, The Strategy of Treaty Termination: Lawful Breaches and Retaliations (New Haven CT: Yale University Press, 1975).

[6] Case C-162/96, Racke v Hauptzollamt Mainz, ECLI:EU:C:1998:293.

[7] Ibid., § 57. For further comments, see my case note in (1999) 36 Common Market Law Review, 179-189.

[8] Case Concerning the Gabcikovo-Nagymaros Project (Hungary/Slovakia), [1997] ICJ Reports 7, § 104.

[9] https://www.energychartertreaty.org/cases/list-of-cases/page/14/  (visited 12 November 2022).

[10] To date, no cases have been brought against France, and three against Germany, all three emanating from the West.


Cite as: Jan Klabbers, ‘A Moral Holiday: Withdrawal from the Energy Charter Treaty’, ESIL Reflections 11:6 (2022).

© 2022. This work is licensed under a CC BY-NC-ND 4.0 licence.