This week, leaders of NATO member states will convene in Washington for a summit commemorating the alliance’s seventy-fifth anniversary. Ukrainian President Volodymyr Zelenskyy will no doubt renew his plea for Ukraine’s immediate accession to NATO. And US leaders, purporting to speak for the club, will almost certainly rebuff his entreaties.

In an interview this May, President Joe Biden repeated his stance that that he is “not prepared to support the NATOization of Ukraine.” As if to remove any doubt, his ambassador to NATO more recently stated, “Ukraine will not be receiving an invitation to join the alliance this July” at the summit.

There are sound strategic arguments both for and against bringing Ukraine into the NATO fold while it remains at war with Russia, and we take no position here on the ultimate wisdom of doing so. But a major argument against extending the invitation to Ukraine—that the NATO charter’s Article 5 would effectively obligate member states to deploy troops to Ukraine—is misplaced.

Ukraine’s accession to NATO could be accompanied by a limitation on the application of Article 5 to Ukrainian territory. By cabining how NATO’s Article 5 would apply to Ukraine—confining it to a narrower scope than its typical interpretation—Kyiv’s accession may also provide it with the leverage required to broker a sustainable armistice without dragging the rest of NATO into war. NATO’s no-strings-attached commitment to assist attacked members makes it an anomaly among alliances. Its lack of conditions has been a boon to the alliance’s deterrence, but it may be a luxury NATO can no longer afford.

To protect its eastern flank and provide Ukraine with the negotiating leverage it needs to stabilize the region, NATO could do what alliances regularly do: admit Ukraine but carefully specify the casus foederis—that is, the conditions that would trigger its collective-defense obligations to Kyiv. A carefully cabined accession of Ukraine to NATO could achieve a durable settlement to the Russo-Ukrainian War without sacrificing the deterrent value of NATO on which its current members depend. As competition for US dollars, materiel, and attention among allies stiffens, such a settlement is within the US interest, and potentially within NATO’s grasp.

Alternatives to NATO are Unlikely to Achieve and Maintain a Russo-Ukrainian Armistice

At the NATO summit in Vilnius in July 2023, NATO decided not to decide when to invite Ukraine to join the club. The decision to kick the can down the road reflected, in part, members’ uneasiness about bringing in a country at war. If Ukraine joins NATO while the conflict persists, there’s every reason to believe Zelenskyy would immediately invoke Article 5 of NATO’s Washington Treaty, the clause that triggers NATO states to assist an aggressed member.

The United States was reportedly the lead holdout (alongside Germany) on Ukrainian accession in Vilnius. This marked an inversion of the parties’ historical positions. In past proposals for NATO enlargement, several European states opposed US efforts to extend invitations to Georgia and Ukraine. More recently, Turkey and Hungary threatened to spoil attempts to bring Finland and Sweden into the NATO fold before both Nordic countries became members by unanimous consent. In July 2023, however, Biden remained “immovable in his opposition” to inviting Ukraine to join NATO.

What the president did offer, according to National Security Advisor Jake Sullivan, was a “bridge to NATO” in the form of “the Israel model.” The Israel model refers to the arrangement between the United States and Israel dating back to the administration of President Ronald Reagan, whereby the former pledges its military support to the latter in memorandums of understanding updated every decade or so. Though designated by US law as a “major non-NATO ally,” Israel does not share a mutual defense treaty with the United States, meaning Washington has no obligation to defend Jerusalem when attacked. Because an arrangement along similar lines vis-à-vis Ukraine would ensure material support without risking the lives of US service members, it is attractive to the Biden administration. However, this approach is almost certainly inadequate to bring about an end to Ukraine’s war. This is so for three reasons.

First and most obviously, the “Israel model” is hardly an exemplar of deterrence. The events of October 7, Iran’s concerted airstrikes on Israel in April, and the nonstop barrage of Hezbollah rockets impacting northern Israel reminds us that, for all its merits, US support to Israel has not deterred Iran from waging war against Israel on multiple fronts. This may be because the principal objective of the Israel model is not deterrence. Indeed, the US State Department defines the Israel model’s goal as “helping Israel maintain its Qualitative Military Edge” over its “regional adversaries.”

Second, it is precisely the Israel model’s limited aim—helping Israel maintain its qualitative advantage over its Middle Eastern adversaries—that makes it a poor template for Ukraine. The less than $4 billion the United States provides annually to Israel is reasonably calculated to enable Israel to maintain its military-technological edge during relatively stable times. By contrast, US interests in Eastern Europe cannot be secured with so meager a commitment. Ukraine has no qualitative military edge to maintain, and gaining any such edge over Russia would prove exceedingly more difficult than maintaining Israel’s edge over, say, Hamas, Syria, or even Iran. This strategic difference makes the Israel model a poor guide for the US-Ukraine relationship.

Last, as much as the Israel model underdelivers what Ukraine needs, it also overpromises what Washington can deliver. Ironically, NATO membership may be the more modest and realistic promise. If the Israel model is “the bridge to NATO,” as Biden told Zelenskyy it was, Congress has made clear it is a treacherously creaky bridge. Sustaining the Israel model has depended on a decades-long bipartisan consensus that the United States should aid Israel. Even this previously unassailable consensus is beginning to show cracks, with Congress’s latest vote on aid to Israel dividing Democratic members of Congress.

If Israel cannot reliably depend on the Israel model to receive aid in time of need, how can Ukraine? Ukraine has never enjoyed the kind of bipartisan support that Israel once did. Though fairly robust at the outset of Vladimir Putin’s full-scale invasion in February 2022, Republicans’ support for Ukraine aid has waned steadily as the war has progressed. Biden thus cannot credibly promise that Washington will continue to provide the support Kyiv needs to maintain a negotiated settlement with Russia.

To be sure, Ukraine’s accession to Ukraine would also require congressional support. Article 11 of the Washington Treaty provides that its “provisions [shall be] carried out by the Parties in accordance with their respective constitutional processes.” In the United States, this means that a two-thirds majority of the Senate must pass a resolution ratifying Ukraine’s accession. Still, the prospects that the Senate would do so are appreciably better than the chances that a much more fractious House of Representatives will continue to appropriate aid to Ukraine as the war drags on. The Senate, after all, voted 95-1 in favor of admitting Sweden and Finland into NATO in August 2022 and put its overwhelming support behind the April 2024 aid package for Ukraine.

Notwithstanding the Israel model’s inadequacy, the Biden administration proposed it for understandable reasons. Policymakers, professors, and pundits all seem to agree that “accession is not possible as long as war is raging.”

But it is unclear how a sustainable armistice—something Washington badly needs in order to focus resources elsewhere—is possible as long as accession is withheld. Clearly, Ukrainian military capability and America’s “unwavering” commitment to Ukraine’s sovereignty proved insufficient to deter Putin’s bid on Kyiv in 2022. Why, then, should we expect Russia to now abide by the terms of a peace deal unless it’s underwritten by a more credible deterrent like NATO’s nuclear umbrella?

At first blush, the problem of Ukrainian accession appears to pose a Catch-22: there’s little chance that NATO members would welcome Ukraine to the club until there is a sustainable armistice, but little hope for a sustainable armistice until Ukraine is in NATO.

But this conclusion is unwarranted. NATO’s Article 5 need not present an intractable obstacle to admitting an ally at war. To see why, we must understand NATO’s founding Washington Treaty, how it can be amended to add a casus foederis qualification, and how the introduction of casus foederis qualifications can be balanced against the concomitant loss to NATO’s deterrent value.

Article 5 Obligations: Legal Terms vs. Strategic Expectations of Compliance

Fueling the tepid approach to Ukrainian membership exhibited at the 2023 Vilnius summit was worry that NATO’s Article 5 mechanism—the trigger that obligates members to assist one another—will make Ukraine’s war our war.

Strictly speaking, however, an invocation of Article 5 compels no such outcome. The rub is that NATO’s credible deterrence arguably does. Although Article 5 does not require that members commit combat forces to aid an attacked ally, the expectation that members will do so substantially raises the costs of attacking NATO territory.

A closer review of the Washington Treaty’s provisions and their historical applications suggests, however, that the Article 5 guarantee, as applied to Ukraine, can be cabined to accommodate Ukraine’s accession without subverting NATO’s deterrence value. Doing so, like all policy decisions, involves tradeoffs, and in this instance, concessions from Ukraine.

Article 5 of the Washington Treaty provides that

an armed attack against one or more [members] . . . shall be considered an attack against them all and . . . if such an armed attack occurs, each of them . . . will assist the Party or Parties so attacked by taking . . . such action as it deems necessary, including the use of armed force, to restore and maintain the security of the North Atlantic area.

On the one hand, the plain language of Article 5 permits each country to decide for itself what action it “deems necessary” to effectuate the treaty’s object of restoring and maintaining security. Such action may “includ[e] the use of armed force,” but need not. A member may conclude that what is necessary to restore security today is that which NATO countries are already doing for Ukraine: providing materiel and sanctioning Russia.

On the other hand, a fair reading of the treaty in light of its historical use does suggest that if Ukraine were admitted into NATO without qualification, Kyiv’s Article 5 invocation under present circumstances would create, under international law, binding obligations for members to commit combat forces to Ukraine. On the sole occasion that Article 5 has been invoked—by the United States following the terrorists attacks on 9/11—each NATO member did ultimately contribute troops to the NATO-led International Security Assistance Force (ISAF) in Afghanistan. A NATO member would be hard-pressed, especially after sending troops to assist the US war effort against nonstate extremists in Afghanistan, to argue that what it “deems necessary” to “restore and maintain” regional security is anything short of its “use of armed force.”

The Vienna Convention on the Law of Treaties governs how treaties are interpreted. The convention holds that a treaty “shall be interpreted in good faith in accordance with the ordinary meaning” of its words “and in the light of its object and purpose.” The ordinary meaning of the obligation to “assist the Party . . . so attacked,” both at the time the treaty was ratified by NATO’s founding members in 1949 and today, has usually meant (the threat of) sending combat forces to the victimized ally. This form of assistance was the intent of the parties when contemplating Article 5, and the intent they sought to communicate to adversaries. Indicative of this understanding is Biden’s oft-repeated vow that “we will defend literally every inch of NATO,” or NATO Secretary General Stoltenberg’s similar pledge that members will not merely assist one another, but will “protect and defend every inch of allied territory based on our Article 5 commitment.”

Whatever we may make of Article 5 in the abstract, if Kyiv were admitted to NATO today and invoked the collective-defense clause, both international law and NATO’s credibility would seemingly obligate members to deploy combat forces to Ukraine.

Cabining NATO’s Article 5 Guarantee as Applied to Ukraine

NATO need not accept this all-or-nothing proposition, though. The Washington Treaty contains the tools necessary to accommodate Ukraine’s situation. Article 5 obligations can be delimited geographically and substantively to better specify the casus foederis that would trigger them. Defense pacts, including US bilateral ones, often contain scope conditions. Indeed, NATO is exceptional for the lack of conditions that trigger its collective-defense obligations.

The Washington Treaty’s Article 12 provides a mechanism for amending the treaty. In practice, the only amendments thus far have stemmed from the several protocols announcing that, upon ratification, a new member will join the club. The very first such protocol—announcing the accession of Greece and Turkey—effectively amended Article 6 of the treaty, which defines NATO’s geographic area. The protocol enlarged NATO’s geographic orbit: in addition to “Europe [and] North America” (as well as, originally, “the Algerian Departments of France,” which are no longer within NATO, per an informal amendment), “the territory of Turkey” was now also under NATO’s umbrella. (Because the parties agreed that Greece was in Europe, no wording changes were necessary to effectuate Greece’s accession.)

Since the Treaty can and has been amended to extend NATO’s territorial scope, there is no reason it cannot be used to extend that scope over only part of Ukraine’s territory. In fact, NATO did just that when West Germany acceded to NATO while East Germany remained under Soviet occupation. NATO could likewise bring under its wing all of Ukraine except the areas presently occupied by Russia—the Donbas region and the Crimean Peninsula. The casus foederis that would then trigger NATO’s Article 5 obligations would be an attack against the Ukrainian territory currently under Kyiv’s control.

A geographically delimited Ukrainian accession is not entirely free of risk, of course. First, there is the danger of signaling a de facto concession of territory from Ukraine to Russia. Second, it risks introducing a scheme of tiered membership in NATO, where current members enjoy the full suite of protective benefits over all their sovereign territory while newly admitted states do not. The risks, however, may not only be worth assuming; they might even provide a template for future accessions.

Zelenskyy’s government has continued to define victory as the restoration of Ukraine’s international borders as they stood before Russia’s 2014 invasion. An accession of Ukraine that does not encompass the Russian-controlled areas of the Donbas and Crimea could be viewed by Moscow as NATO’s tacit acknowledgment of Russia’s territorial claims. For three reasons, NATO leaders should nonetheless consider a geographically delimited accession for Ukraine.

First, bringing the territory that Kyiv controls into NATO in no way limits the assistance individual NATO members can provide to help Ukraine prosecute its war in territory outside NATO. Just as those states have aided Kyiv since February 2022, they may continue to do so at their own discretion. The difference is that west of the Russian lines, Ukraine would enjoy the full suite of protections afforded to any NATO member.

Second, NATO as an entity could invest its resources as it deems necessary to protect its territory while individual members commit what they deem prudent to Ukraine’s fight in the contested regions. In hushed tones, leaders of NATO states have expressed concerns that Ukraine’s war aims are less than realistic, maybe even undesirable. Both the Donbas and Crimea have been under Russian occupation for nearly a decade now. Even if reclaimed, the Donbas is neither easily defendable nor all that profitable for Ukraine to retain. Mostly destroyed by Russian brutality and largely populated by Russian loyalists today, retaking the Donbas would require Kyiv to fund a massive reconstruction effort and probably prosecute a protracted counterinsurgency campaign. While Crimea remains key to Ukraine’s economic viability as a state, last summer’s counteroffensive came nowhere near seizing it. The best Kyiv may hope for in the foreseeable future is using long-range fires to deny Moscow’s productive use of it.

Last, trading this territory for a NATO guarantee may be the only way to sufficiently reduce the risk to NATO members to gain their unanimous support to admit Ukraine. A geographically delimited accession could be used by NATO to force an armistice that freezes the current position of Russia’s troops. Alternatively, the geographically limited accession could follow an armistice, freezing the line of Russian troops that Putin would almost certainly unfreeze but for the threat of a NATO reprisal. Either way, NATO would likely need to deploy a sufficiently sizeable tripwire force along Ukraine’s Donbas line of control to make credible its threat to honor Article 5 if Russia violates the armistice.

The second general risk of geographically cabining Article 5 is the precedent set by opening the door to admitting what might be perceived as second-class NATO members. The NATO model has served its members well for nearly eight decades. Change carries risk. But since its inception, NATO has adapted, in ways small and large, to accomplish its mission amid geopolitical change. A NATO policy against admitting militarily contested areas like the Donbas would retain NATO’s core character as a defensive alliance against charges that it intends to provoke aggression. This Ukraine model could be applied elsewhere, too. Among non-NATO European states, Ukraine and Georgia are the most obvious candidates for admission. The sticky point for current NATO members is that in Georgia, as in Ukraine, Russian troops occupy two territorial areas, Abkhazia and South Ossetia. Should NATO decide that but for this fact Georgia would be a strategic asset to NATO, the alliance could, under the Ukraine model, admit Georgia, less the Russian-occupied territories. Again, this would reinforce the nature of NATO’s defensive alliance, signaling that it does not acquire new members to join fights, but to establish red lines against future encroachment.

Instead of, or in addition to, a geographic delimitation, NATO could also agree to cabin its Article 5 commitment substantively. That is, the parties could specify the particular adversarial actions that would entitle Ukraine to invoke Article 5. Here, not all incidences of Russian aggression would qualify for Article 5; only those that cross a threshold agreed upon by the members would. Such substantive qualifications would require Ukraine and current NATO members to draw conceptual distinctions between strategic threats, such as a tank column heading toward Kyiv, and nonexistential threats.

There is no precedent for substantive qualifications within NATO, though the Treaty’s Article 4 consultation mechanism has been used to discuss what constitutes an attack on a member. But several US defense pacts with non-NATO allies are explicitly or implicitly qualified in this way. For example, South Korea occasionally faces North Korean aggression along the Korean Demilitarized Zone and in the Han River estuary. Though the United States has almost thirty thousand tripwire troops stationed on the Korean Peninsula and a treaty obligation to defend Seoul from Pyeongyang’s attacks, Washington does not react to all such nuisances by declaring war on the North Korean regime.

Limiting Ukraine’s invocation of Article 5 would mean Kyiv receives different treatment than other NATO members. But whether NATO qualifies its Article 5 obligation to Ukraine geographically or substantively, such a decision need not be permanent. If, down the line, Ukraine successfully regains some or all of its territory, NATO members can decide whether to add that territory to that already covered by Article 5 by going through the same process that would govern the initial, limited accession. This process would allow NATO to assess the risks and benefits of throwing its full Article 5 weight behind defending that regained territory.

Some, including even alliance experts, worry that specifying the casus foederis that would trigger NATO’s Article 5 obligations would undermine NATO’s deterrence value. It is the assumption, the argument goes, that the alliance’s determination to “defend literally every inch of NATO” with all its conventional (and even nuclear) might that has deterred adversaries from violating allied territory. But deterrence has always been a function of the three Cs: an actor’s capability, its credibility, and importantly, its communication. Accordingly, by communicating exactly what NATO would defend in Ukraine (and by implication, what it would not), cabining NATO’s commitment to Ukraine need not diminish the alliance’s deterrence value.

In fact, research shows that not only is it “very common for treaties to specify limits to obligations,” but that those limits are often key to effective deterrence. Treaties are often limited to particular adversaries, to conflicts in particular theaters, or even to a specific ongoing conflict. In the history of alliances, NATO remains unique for its lack of obligatory limits. For nearly seventy-five years, NATO has been an all-weather, deterrence-driven, “open” alliance.

This has been a virtue. Few institutions have done more than NATO to provide regional stability and prosperity. But research also indicates that when an alliance specifies the casus foederis that would give rise to its activation, those “alliance contracts” successfully deter aggression precisely because they “are specific enough to provide information to a potential attacker about when assistance is likely to be forthcoming.” Being specific about when and where NATO would fight in Ukraine would communicate clear red lines to Russia and perhaps make those lines more credible.

A Closing Window of Opportunity

The Ukraine aid package Congress approved in April will likely enable Ukraine to survive the year. But Putin has powerful incentives to keep fighting beyond this year, including Americans’ declining appetite to fund a war of attrition, a fractious Congress torn between competing priorities, and the prospect of electorally driven and potentially abrupt shifts in policy toward Ukraine.

Zelenskyy will no doubt say all of this at this week’s NATO summit in Washington. While there are strategically defensible reasons for denying his pleas for Ukrainian accession, fear that Kyiv’s accession would necessarily convert Ukraine’s war into our war is not one of them. On the contrary, a carefully cabined accession may be the alliance’s final opportunity to draw red lines in Ukraine that restore peace to Eastern Europe and protect NATO’s eastern frontier for the next seventy-five years.

Luke J. Schumacher is a JD candidate at Stanford Law School, a PhD student in political science at the University of Virginia, and a combat veteran. Follow his work on his website.

Spencer A. Segal is a JD candidate at Stanford Law School. He holds an MA in public policy and a BA in economics, both from Stanford University. He previously was a missionary in the West African nations of Togo and Benin.

Image credit: President of Ukraine