Reports of a possible Russia-Ukraine peace agreement continue to bubble and churn. The latest is an apparent Russian rejection of a peace plan floated by the Trump transition team. Uncertainty over the specifics of an agreement is likely to continue at least until the new administration takes office. But potential provisions are almost certainly being vetted in foreign ministries around the world. In a recent EJIL Talk post, Philipp Kehl ably addresses issues surrounding territorial transfers that would likely be at the heart of any agreement.
In this post, I would like to expand on points Kehl made regarding territorial change and address three additional issues: the reasons for the substantial legal uncertainty surrounding an agreement, potential provisions other than those concerned with the transfer of territory, and the potential role of the UN Security Council in “curing” any legal defects in an agreement. I will assume the reader’s familiarity with Russia’s occupation of Crimea and the Donetsk and Luhansk regions of Ukraine.
I will also assume that any agreement will have as parties not only Russia and Ukraine but other states that have been assisting both sides. Some rules of treaty law applicable to a potential peace agreement differ depending on whether a treaty is bilateral or multilateral.
Finding the “law” of Peace Agreements
Despite innovative scholarship seeking to create a unified legal theory of contemporary peace agreements (such as Christine Bell’s lex pacificatoria and Carsten Stahn’s jus post bellum), contemporary international law does not contain a stand-alone set of doctrines addressing peace agreements. This was not always the case. In the pre-Charter era, peace agreements were a central concern of the “law of war,” since they demarcated the formal end of that specific body of rules applicable only in times of armed conflict and the return to normal rules of peacetime. I have described peace treaties of that era as performing a “signaling function,” notifying the international community that, for example, the law of neutrality, the suspension of regular treaty relations, and the possibility of recovering captured property had come to an end. As Marko Milanovic summarizes, “[s]ince ‘war’ was a formal business, it also needed to be formally terminated.”
There are several reasons why this body of law has effectively fallen into desuetude and why contemporary international law has very little to say about peace treaties as such (the U.S. Department of Defense Law of War Manual, for example, newly updated in July 2023, devotes only 2 pages to the termination of hostilities). First, most conflicts today are non-international (NIACs) and the number of agreements ending international armed conflicts (IACs) has radically declined. The PA-X dataset records only 7 final IAC agreements since 1990 as compared to 63 final agreements for NIACs. Second, most scholars agree that international no longer recognizes “war” as a distinct legal category, rendering the signaling function largely unnecessary. The end of armed conflicts (the term largely replacing “war”) is now a factual question rather than a matter of formal termination. Third, the UN Security Council has in several cases partially or wholly taken over the settlement of conflicts, passing resolutions whose terms might, in a prior era, have appeared in a peace treaty. The most prominent is Resolution 687 (1991) for Iraq-Kuwait but there are many others.
Law bearing on IAC peace agreements now comes instead from cognate fields – the law of treaties, human rights, the jus ad bellum, the jus in bello, and the law of state responsibility (among others). Agreements ending NIACs (which have become much more common) are not generally recognized as binding treaties (because they necessarily include at least one non-state party) and as a result have a more indirect relationship to international law.
The Role of Coercion
Kehl properly focuses on a Russia-Ukraine agreement potentially being void under Article 52 of the VCLT because it would be “procured by the threat or use of force in violation of the principles of international law embodied in the Charter of the United Nations.” The critical underpinning for this means of voiding an agreement is, of course, the prohibition on the aggressive force itself in Article 2(4). Peace agreements procured by unlawful force are the “fruit of the poisonous tree” and cannot grant legal entitlements to aggressor states: ex injuria jus non oritor. The pivotal factor is that the force involved is unlawful: an aggressor state may not coerce agreement to a treaty while a state using force lawfully – in self-defense or as authorized by the Security Council – may do so.
The rule is uncompromising. Even if Ukraine were to find benefits in other parts of a peace treaty – security guarantees, no Russian opposition to seeking EU membership, compensation for wartime destruction, etc. — these cannot be weighed against the fact of coercion. VCLT article 44(5) provides that in cases of coercion, “no separation of the provisions of the treaty is permitted.” The entire treaty fails. (Since Ukraine and Russia are parties to the VCLT, there is no argument to be made that this and other procedural aspects of treaty law may not amount to custom).
On the other hand, those same procedures state that even a treaty coerced by force does not become void of its own accord. The Vienna Convention provides in Article 65(1) that only the coerced party can raise an objection of coercion (an objecting party must raise “a defect in its consent to be bound”). It must do so by sending a notice to other parties so that a process of evaluating the claim may take place. Such a burden on Ukraine to assert coercion as an affirmative defense (to use an American term) raises the prospect that it may simply choose not to do so (Dörr & Schmalenbach, p. 955). Indeed, one could make a good case that is what is likely to happen: why would Ukraine even enter into peace negotiations and do the hard work of reaching an agreement if it intended to declare the treaty void after its conclusion?
There is an alternative scenario. Ukraine might initially sign and ratify a peace treaty in order to buy time to revive and regroup its forces and rebuild infrastructure. If, later on, Russian control over the annexed territories appears tenuous and its resources for resupply limited, Ukraine might see a strategic opportunity to raise the coercion claim and void the treaty. Or there might be new leadership in Russia that reevaluates the strategic value of the territories, in which case it might not react with force if Ukraine denounced the treaty as void. Ukraine would not forfeit its coercion claim if it were to take such a wait and see approach (Dörr & Schmalenbach, p. 955).
If Ukraine made a coercion claim at any point and removed itself from the agreement, other states could continue as parties despite Ukraine’s departure. In accordance with VCLT article 69(4), voiding the consent of one party to a multilateral treaty means “that the treaty will nonetheless remain valid for the other parties.” (Aust, p. 280). As perverse as it might sound, one could imagine Russia and a Trump-led United States insisting that an agreement continue (i.e. Russian sovereign control over annexed territory continue) even if Ukraine denounced it as coerced. Article 69(4) could allow them to do so.
Notwithstanding the VCLT, viewing the treaty as void only upon Ukraine’s initiative seems at odds with fundamental principles. That the agreement should be void ab initio derives from Russia’s territorial gains themselves being void. As the Security Council declared of Iraq’s purported annexation of Kuwait in 1990, the act “has no legal validity, and is considered null and void.” (SC Res. 662 (1990)). How can a peace agreement ratifying such an illegal act and coerced through that illegal act nonetheless have legal validity? The entire reason for voiding a coerced treaty is the recognition that any consent is never truly voluntary, and making the viability of a coercion claim contingent on Ukraine’s failure to raise it would seem to replicate the false consent that the doctrine was designed to avoid.
Jus Cogens Invalidity
Russia’s use of force to annex Ukrainian territory may provide a second basis for invalidating the agreement: that it violates a peremptory norm under Article 53 of the VCLT. The prohibition on the aggressive use of force is the quintessential jus cogens norm, and the ILC has twice identified it as having a peremptory character – in commentary to its version of Article 53 (then Article 50) and in its 2022 Draft Conclusions on Peremptory Norms (Annex A). The claim would be that treaty provisions ratifying Russia’s annexation would also ratify the unlawful force giving rise to the annexation. A treaty in conflict with a jus cogens norm is void ab initio.
One objection to this claim is that, by its terms, a peace agreement would not require or even contemplate the illegal use of force. At most, the argument would go, the agreement would memorialize as a fait accompli the results of prior unlawful force. In a recent article, Ingrid Brunck and Monica Hakimi argue that the prohibition on annexation is historically and analytically distinct from Article 2(4)’s prohibitions and that the two ought not to be conflated. Indeed, the two ILC listings of peremptory norms mentioned above, while not exhaustive, both include aggressive force but do not include annexation. Anthony Aust states in his widely-cited book on treaty law that “perhaps the only generally accepted example [of a peremptory norm] is the prohibition on the use of force as laid down in the UN Charter” (p. 257).
It is unclear whether this is a distinction with a difference for purpose of treaty validity. In the Israeli OPT Advisory Opinion, the ICJ described “the prohibition of territorial acquisition resulting from the threat or use of force” as a “corollary of the prohibition of the threat or use of force” (para. 175, emphasis added). Later in the opinion, the Court stated that jus ad bellum rules “determine the legality of the continued presence of the occupying Power in the occupied territory” (para. 251), suggesting those rules extend beyond the use of force itself to sovereignty over the territory.
Amnesty for Russian Officials
The International Criminal Court has to date issued four arrest warrants for high Russian Federation officials for various crimes committed in Ukraine. Because the indictees (including President Putin himself) cannot safely travel abroad, it seems likely Russia will seek amnesty for their officials and to close to the entire ICC Ukraine investigation, which the prosecutor described in June 2024 as open and active.
While some post-Cold War peace agreements (mostly in NIACs) have included amnesties of various kinds, impunity for international crimes is increasingly disfavored in international law. Many treaties defining those crimes oblige states with jurisdiction to investigate and prosecute perpetrators. Human rights bodies, led by the Inter-American Court of Human Rights, have found such amnesties to themselves violate human rights treaty obligations. Based on this trend against impunity, the UN Secretary-General has declared that “United Nations-endorsed peace agreements can never promise amnesties for genocide, war crimes, crimes against humanity or gross violations of human rights.” (para. 10).
But these legal developments concern amnesties in national legal systems, the reform of which is often a central concern of NIAC peace agreements. Amnesty for Putin and his colleagues would be for prosecution in the ICC. The ICC Rome Statute contains no provision for amnesty. Indeed, when the Ugandan government demanded that the ICC prosecutor withdraw charges against leaders of the Lord’s Resistance Army because they were purportedly impeding a peace agreement, the prosecutor refused to withdraw the arrest warrants.
The potentially detrimental effects of prosecutions on peace negotiations or agreements could conceivably be relevant under Article 53 to whether initiating a new investigation would “serve the interests of justice.” But Article 53 is not applicable where arrest warrants have already been issued, as has happened with Putin et al. But even for future defendants, the Office of the Prosecutor strongly suggested in a 2007 Policy Paper (p. 9) that investigations’ effects on peace processes was not a factor it would consider under Article 53, stating “the broader matter of international peace and security is not the responsibility of the Prosecutor; it falls within the mandate of other institutions.”
The only mechanism for discontinuing active ICC investigations or cases is Article 16 of the Rome Statute, by which the Security Council may defer cases for 12 months through a resolution under Chapter VII. One might imagine a peace agreement provision obliging the parties to vote for Article 16 deferrals ad infinitum. The Rome Statute places no limits on the indefinite renewal of Article 16 deferrals, a possibility supported by most commentators. The incoming Trump administration’s apparent support for Ukrainian territorial concessions makes it unlikely the United States would veto an Article 16 resolution. Whether other Council members would discard the trend against amnesties and find impunity for Russia’s blatant international crimes in Ukraine worth securing a peace agreement will be a matter for their own political conscience.
Amnesty via Article 16 will not be easy to achieve. Since any one of the five Council permanent members could veto an Article 16 resolution, all of them would need to join the peace treaty. The treaty would need to require them to vote for the resolutions year after year. In addition, since support from the P-5 alone will not guarantee a resolution’s passage (a majority requires support from 3 more elected members), and the identity of non-permanent members of the Council in any given year cannot be known in advance, Russia will need to find support from new elected members each time an Article 16 resolution comes up for a vote.
Mass Expulsions
Implicit in Russia’s justifications for its invasion of Ukraine has been the erasure of Ukrainian national identity. President Putin famously declared that Ukraine “does not exist” as an historically distinct political community. In areas under Russian occupation, the Atlantic Council reports, “all traces of Ukrainian statehood and national identity are being ruthlessly purged. Ukrainian children are forced to study a Kremlin curriculum that demonizes Ukraine while glorifying the invasion of their country. Adults must accept Russian citizenship if they wish to access basic services such as pensions and healthcare.” One can well imagine many Ukrainians resisting this forced assimilation and being expelled from the territory, lest they become partisans or a fifth column of resistance across the border from the still-intact Ukraine.
Mass expulsions, sometimes referred to euphemistically as “population exchanges,” have been features of past IAC peace agreements (pp. 136-40). The 1919 Greek/Bulgarian agreement and the allies’ 1945 Potsdam Agreement are two prominent examples. Their description in strategic terms masked immense human suffering in their wake. Such agreements among political elites to denationalize their citizens and dispossess them of their property are obviously inconsistent with a wide range of human rights principles. Indeed, the legal trend for post-conflict states is to facilitate the return of IDPs and refugees and to provide restitution or compensation for their seized property. Efforts to facilitate return and restitution in the aftermath of the Bosnia and Kosovo (pp. 241-44) conflicts are two examples of collective efforts to achieve exactly the opposite results of mass expulsion.
Neither the right to return to one’s home nor the right of citizens to their property has achieved jus cogens states and could not function to void a Ukraine peace agreement. But because both are individual rights against direct infringement by the state, they are not Ukraine’s to waive and infringe indirectly in a peace agreement. Indeed, by doing so, Ukraine could incur state responsibility to its own nationals under human rights law for attempting to terminate their claims.
Can the Security Council “Cure” Defects in a Peace Agreement?
Let us assume for the sake of argument that all of the legal defects identified above are valid. Could any or all of them be “cured” and a peace agreement salvaged if it were approved by the Security Council under Chapter VII?
As a general matter, the Council has no obligation to respect states’ existing treaty obligations; indeed, Article 103 of the Charter expressly subordinates those obligations to obligations under the Charter, which include those imposed by the Council under Chapter VII. The Council has regularly altered states’ preexisting treaty obligations: in Resolution 748 (1992) it precluded Libya from exercising its right to try two of its citizens under the Montreal Convention; in Resolution 687 (1991) it substantially enhanced Iraq’s obligations under the Non-Proliferation Treaty and effectively made it party to the Biological Weapons Convention, which it had not ratified (Roberts, 1993); in Resolutions 1483 and 1546 it departed from the Hague Regulations’ fact-based test of when an occupation begins and ends by declaring those dates by legislative fiat; and in Resolution 1816 (2008) and others it altered UNCLOS’ definition of piracy as occurring only on the high seas.
It is one thing to alter or impose otherwise valid treaty obligations; it is another for the Council to require adherence to a treaty that would be otherwise invalid. Those arguing that the Council possesses such legislative power might point to agreements ending the Bosnia and Kosovo conflicts, both of which were procured by NATO bombing campaigns (in the case of Kosovo, a US government lawyer described the campaign as an attempt “to bomb Milosevic back to the bargaining table”). (Fox, Humanitarian Occupation, p. 179). But these cases actually don’t help. The conflict between the Serbian government and the Kosovo Liberation Army was not an IAC and Serbia did not contest the validity of the agreement (Forlati, p. 329). In Bosnia, the Security Council had previously endorsed the NATO bombing of Serb positions (Resolution 836 (June 4, 1993)), meaning the force used to coerce the agreement was lawful.
This question ultimately involves whether or not limits exist on Security Council legislative power to enact measures it deems appropriate to “restore international peace and security” (Charter, art. 42). Some argue human rights law as such binds the Council (echoes of the Kadi and Al-Jedda cases), but this is far from a majority view. Serena Forlati asserts the Council can impose agreements on states as long as it does not transgress jus cogens norms. The prohibitions on mass expulsions and amnesties for international crimes are not jus cogens, so at best this claim would apply only to the anti-annexation norm.
Invoking peremptory norms to limit Council action finds support among many scholars. But it runs into the problem that the Council speaks on behalf of the entire international community (Charter, article 24) and all member states agree to carry out its dictates under Chapter VII (article 48). A supermajority resolution recognizing Russian sovereignty over parts of Ukraine would thus call into question whether a norm prohibiting such recognition was “accepted and recognized by the international community of States as a whole as a norm from which no derogation is permitted.” (VCLT, art. 53). Recall Bosnia’s furtive claim in the 1990s that the Council’s arms embargo on the former Yugoslavia deprived it of the right to self-defense, an “inherent” and likely peremptory norm.
If Council power is limited by jus cogens norms, then a Chapter VII resolution obviously cannot cure a peace agreement that violates the jus cogens prohibition on annexation. If the Council is not so limited, then the resolution may succeed.
Separately from the jus cogens argument, Stefan Talmon argues that voluntary consent is a fundamental attribute of treaties and the Council, “even when acting under Chapter VII, cannot change the concept of a treaty in international law.” (p. 22). Talmon proposes a third way the Council might subject a state to an otherwise invalid treaty, which is to incorporate the substance of treaty obligations into the terms of a Chapter VII resolution (p. 23). This would be similar to the Council’s effective imposition of the terms of the 1999 Terrorist Financing Convention on the entire international community, including non-parties, in Resolution 1373 (2001).
Any of these limitations on Council power are more likely to inform the thinking of foreign ministry lawyers in crafting and justifying a peace agreement than they are to play a role in litigation testing an agreement’s validity. International lawyers (including this one) take no joy in predicting that Ukraine is likely to bow to political pressure and not challenge the validity of a peace agreement. Even if it did, the ICJ’s clearest refusal to engage in judicial review of Council actions came in the Lockerbie case, where the Council had subordinated Libya’s treaty rights to the political imperative of trying two terrorist suspects. Would any government lawyer advise that the ICJ would take a less deferential stand in a Russia-Ukraine case?
