Letting sleeping bears lie: Ukraine’s cautious approach to uncertainty before the war

ABSTRACT

International relations theory has long suggested that uncertainty during militarized crises pushes states to adopt escalatory behavior. However, the leadup to the Russo-Ukrainian War challenges this view, with Kyiv downplaying the risk of conflict and adopting a cautious foreign policy. I argue that Ukrainian behavior in the leadup to the war demonstrates the need to disaggregate between types of uncertainty. While uncertainty over an adversary’s intentions and capabilities can push a state to adopt escalatory behavior, uncertainty regarding a potential adversary’s military strategy and regarding the level of support it can expect from third parties incentivizes a state to assume a more cautious posture. The piece concludes with a reflection on disciplinary blind spots regarding the impact of uncertainty on state decision-making and offers suggestions for overcoming them.

KEYWORDS:

• Russo-Ukrainian war
• uncertainty
• conflict initiation
• bargaining

How does uncertainty influence a state’s crisis behavior? International relations (IR) scholarship has generally linked uncertainty over states’ intentions and capabilities to the outbreak of war (Debs & Monteiro, 2014; Fearon, 1995; Ramsay, 2017). Specifically, the literature suggests that uncertainty pushes states to adopt aggressive and/or risky foreign policies that can produce miscalculations and security spirals, as parties to a crisis begin responding to each other’s escalatory steps. This work often seeks to describe dynamics between states, but it also makes predictions regarding the behavior of individual states, contending that uncertainty pushes them to adopt aggressive, potentially escalatory actions. Strikingly, however, Ukraine did not behave this way in the leadup to Russia’s invasion of its territory in 2022. Instead, Ukrainian leadership publicly downplayed the risk and resisted the types of escalatory steps, such as large-scale mobilization or preemptive strikes, predicted in the literature.

Ukraine’s behavior raises important questions about the embedded assumptions in many of the models employed in IR to understand the effects of uncertainty. It also invites us to consider what additional factors mediate states’ behavior in the face of an uncertain security environment. This is not to suggest that, had Ukraine adopted a more aggressive military posture, war would have been averted. Rather, the point of this article is to utilize the case of the Russo-Ukrainian War to better understand the factors influencing state behavior in periods of uncertainty generally. This is important, as it highlights certain potentially overlooked factors that mediate, for example, the relative explanatory power of deterrence versus spiral models. Perhaps more significantly, by identifying areas where Ukrainian behavior deviated from the expectations of core IR models, this piece helps identify potential theoretical and empirical blind spots within the discipline that extend beyond this particular conflict.

The rest of the article is divided into three parts. The first provides an overview of the central literature on uncertainty and war. The second identifies the major decisions taken by Ukrainian leadership in the months preceding the war and explains why it adopted the positions that it did. Importantly, section two reveals that uncertainty pushed Ukraine to moderate, rather than escalate, its foreign policy. The third section considers why the existing literature has developed blind spots that led it to miss the potentially moderating influences of uncertainty and offers suggestions for how to rectify these oversights.

Ultimately, while the existing literature has much to commend it, it suffers from a tendency to inductively develop models from a small selection of cases, most notably World War One, that center conventional military conflict between major powers. This leads it to generally limit uncertainty to issues of adversary capabilities and intentions. But in Ukraine’s case, there were many other areas of uncertainty, including over Russia’s likely military strategy and over the level of support Kyiv could expect to receive from Western partners. This uncertainty induced caution among Ukrainian leadership. The nature of the uncertainty facing a state matters, and Ukraine’s experience emphasizes the importance of disaggregating uncertainty into different types to better understand its effects. To fully correct these blind spots, IR must do a better job of developing theories and models that derive assumptions not simply from crises between great powers but also from crises featuring materially weaker states. Moreover, it should more explicitly engage with literature on areas such as military planning, strategy, and decision-making to better understand the military factors that mediate states’ responses to uncertainty.

Uncertainty and war: A review of the literature

The literature on uncertainty and war can largely be divided into three categories. The first encompasses systemic theories that evaluate how variations in the distribution of power exacerbate or attenuate uncertainty and, therefore, the likelihood of conflict. Much of this work tends to fall within neorealism, of which a core claim is that the anarchical nature of the international system creates self-help dynamics that compel states to seek security. Importantly, states face uncertainty regarding the intentions of others, which can produce arms racing and war, as each state seeks to protect itself against potential predation (Waltz, 1979). Realist work also maintains that certain distributions of power are more prone to war because they exacerbate uncertainty (Schweller, 1998). Waltz (1979, p. 165), for example, contends that multipolar systems, due to the greater number of significant actors, experience higher levels of uncertainty. The greater number of actors in this configuration makes the balance of power more challenging to calculate, so each defection—real or possible—can have an outsized effect (Waltz, 1988, p. 623).

Further, some realist scholars contend that nuclear weapons temper competition because their ability to deliver extreme and immediate destruction eliminates uncertainty about costs (Jervis, 1989; Waltz, 1981). Relatedly, Jervis (1978) contends that security dilemmas are influenced by technological systems, positing that security spirals are more acute when offensive and defensive weapons systems are indistinguishable. Without the ability to discern whether buildups are being undertaken to attack or defend, states are forced to assume the worst. Related scholarship suggests that uncertainty about whether another state is expansionist can also produce security spirals, with arming being interpreted as a signal of malign intentions. As Kydd (1997) shows via formal models, “security seekers will be more easily provoked than greedy states by the escalations of others” (p. 390).

Unfortunately, this scholarship has relatively little to say about Ukrainian behavior because it is more interested in how variations in the systemic balance of power influence great powers’ behavior. Regional conflicts, and the behavior of weaker powers, are therefore largely ignored. Moreover, to the degree that this literature does apply to Ukraine’s decision-making, it incorrectly predicts Kyiv’s behavior. Ukraine was a non-nuclear security seeker facing a Russian buildup whose purpose was not entirely clear. Yet, it was surprisingly relaxed in the face of Russia’s military buildup along its border.

Bargaining models of war represent a second area of scholarship linking uncertainty to conflict. Fundamentally, these models seek to understand why states fight rather than negotiate peaceful settlements (Fearon, 1995). Work in this tradition predicts that weaker states may choose to escalate to war when the expected long-term costs of foreign interference by a stronger power are greater than the expected costs of war (Monteiro & Debs, 2020). These dynamics become especially acute when the potential adversary is led by someone who has a history of unreliability or who cultivates an image of irrationality, as they struggle to reliably commit to negotiated settlements (McManus, 2021; Seitz & Talmadge, 2020). Other work linking political psychology and bargaining suggests that highly resolved states often deliberately choose provocative actions such as mobilization to signal their commitment to fight, therefore deterring war (Cho et al., 2023). And while certain models suggest that states might deliberately tie their hands to indicate their ability to win without employing their full power, thus signaling strength, this should only apply to the stronger state in the dispute, which Ukraine was not (Reich, 2022).

Collectively, this work suggests that commitment problems and the need to show resolve can serve as rational reasons for states to escalate against stronger adversaries, just as Japan did in 1941 and Israel did in 1967. That Ukraine chose to moderate its foreign policy despite these concerns suggests that the credible commitments problem and signaling dynamics highlighted in the bargaining literature are not sufficient to engender escalation on the part of the weaker power.

The final set of literature linking uncertainty to war emphasizes psychological biases. Johnson (2004) contends that humans are biologically prone to overconfidence, which results in irrationally optimistic beliefs regarding the likelihood of success in war. As Porter (2016) points out, even when leaders acknowledge the inherent complexity and uncertainty of the world, they have false confidence in their ability to impose order. These tendencies toward overconfidence are particularly acute during periods of intense stress and uncertainty (Johnson & Tierney, 2011). Leaders can also fall victim to the strategist’s curse, which results from military planners picking between multiple operational plans. Much like the winner’s curse in auctions, the result is that the selected plan tends to be the one with the most exaggerated efficacy, leading to systematically inflated expectations of success (Altman, 2015). This literature is full of valuable insights, but it has at least one important potential shortcoming in the context of the Russo-Ukrainian War. Namely, states might be overconfident about their safety from aggression rather than about their ability to prevail in war, leading to complacency rather than escalation.

Uncertainty’s effect on Ukrainian decision-making in the leadup to the war

As the previous section reveals, scholars have identified multiple pathways through which uncertainty can produce escalatory behavior in states. Yet, these theories and models largely fail to explain Ukrainian actions in the face of growing uncertainty over Russian intentions in the leadup to war. This section illustrates this disjuncture in more detail before showing how two types of uncertainty—uncertainty over adversary military strategy and uncertainty over levels of external support—induced caution on the part of Ukraine.

Throughout 2021, Russian forces engaged in military exercises near the Ukrainian border. The first occurred in April of 2021, with the maneuvers representing the largest since 2014. Importantly, Russia left elements of the 41st Combined Arms Army (CAA) close to Ukraine after the exercises concluded (Muzyka, 2021). These maneuvers were followed in September by another major deployment of troops near Ukraine as part of the Zapad-21 exercises. After Zapad, many participating units again remained deployed near Ukraine (Deni, 2021). By July 2021, the Center for Strategic and International Studies concluded that “Russian forces were prepared to undertake deep offensive operations, rather than be used simply as a deterrent” (Muzyka, 2021, p. 5).

Moscow’s exact intentions remained uncertain. What was certain, however, was that Moscow’s material ability to inflict short notice harm on Kyiv was rapidly growing. The extant IR literature would thus tend to predict that Ukraine would, at the least, begin its own mobilization and redeployment of forces and, potentially, even actively prepare a pre-emptive strike to prevent Russia from seizing the initiative. After all, Russian intentions were opaque, and Ukraine would thus need to err on the side of caution. Yet, the Ukrainian government largely avoided taking actions that might have been construed as provocative or escalatory.

While Kyiv did carry out limited military exercises in November 2021, it restricted them to 8,500 troops and did not engage in larger military buildups to match Russia (“Ukraine holds military drills”, 2021). Moreover, Ukrainian leadership publicly deprecated American warnings regarding the imminence of Russian aggression, downplaying the risks. Kyiv’s preparations for conflict were, therefore, gradual and constrained. Military mobilization did not occur, and Ukraine’s paramilitary Territorial Defense Force [TDF] units were only beginning to fill out in the weeks before the invasion despite having been legally authorized in July of 2021. This meant TDF units possessed only a small command nucleus and lacked key equipment and vehicles. Consequently, “interoperability with other formations was negatively affected” (Bielieskov, 2023). Kyiv’s unwillingness to mobilize regular formations also presented challenges for Ukrainian planners, as they had a relatively small force with which to defend their border (Zabrodskyi et al., 2022, p. 15). The upshot was that “many Ukrainian units were not at their assigned defensive positions when the invasion began and … were not in prepared positions” (Zabrodskyi et al., 2022, p. 23).

Why did Kyiv adopt such a cautious approach? Ukrainian behavior was partly the result of skepticism about the severity of the threat. As Ukrainian President Zelensky reflected after the invasion, “No one believed it would be like this … And now everyone says we warned you, but you warned through general phrases” (Khurshudyan, 2022). Indeed, at least one senior Ukrainian intelligence official refused to believe Russia would attack until hours before the invasion began. This was partly because Ukrainian intelligence anchored its assessments around a particular set of tactical actions that Russia never undertook (Abdalla et al., 2022). Thus, despite American warnings that Hostomel airport was a likely Russian target (Altman, 2022), the facility was only guarded by around 200 national guard soldiers led by officers “more akin to finance officers than infantry officers” when Russian forces arrived. The 72nd mechanized brigade responsible for the defense of Kyiv was still moving out of garrison on the morning of the invasion (Collins et al., 2023). While it would be incorrect to suggest that Ukraine did nothing to prepare for the invasion, Kyiv did less than it could and far less than many IR theories would predict. One might even conclude that Ukraine’s survival derived more from Russian incompetence than Ukrainian preparedness. Yet, the more significant reason for Ukraine’s tepid response to Russia’s buildup was not that Ukraine dismissed the threat but that it faced at least two types of uncertainty that, in this case, created incentives against escalation: uncertainty over the type of warfare likely to be employed by Russia, and uncertainty over the level of support it could expect from Western partners.

Little green men or conventional war?

Russian forces were massing in a way that enabled Moscow to execute a large-scale conventional invasion. Yet, Russia had, in the years preceding the war, largely oriented its military operations around gray zone operations. Indeed, the Russian Chief of Staff wrote in 2013 that “the role of non-military means of achieving political and strategic goals has grown, and, in many cases, they have exceeded the power of weapons in their effectiveness” (McKew, 2017). Furthermore, Russia and Ukraine had been engaged in a limited war in the Donbas since 2014. It was, therefore, not inconceivable that Russia might simply attempt a limited escalation rather than a full-scale invasion. This produced uncertainty over Russian war plans that shaped the Ukrainian response.

In late 2021, the Ukrainian leadership’s primary fear was that Russian operatives might assassinate Zelensky or orchestrate a coup (Krebs & Chernova, 2021). They were thus wary of adopting military measures that would spread fear, even at the expense of reducing Ukrainian defenses against a conventional threat. As Zelensky put it, “If we had communicated [the risk] … I would have been losing $7 billion a month since last October” (Khurshudyan, 2022). Creating this kind of economic uncertainty would have undermined faith in the government and might have created the very crisis of confidence Russia sought in its efforts to delegitimize and overthrow Ukraine’s government. Indeed, the perceived domestic instability confronting the Zelensky government partly contributed to some Ukrainian analysts’ underestimation of the invasion threat. As Abdalla et al. (2022) put it, “Zelensky’s approval rating was low, and the political situation was unstable. Why should Russia strike now? Why not wait?”

By February, this fear had evolved, with Ukrainian intelligence estimates suggesting that Russia would concentrate overwhelming force in the Donbas to smash the Ukrainian military and destroy Ukrainians’ confidence in their military and government, producing a propaganda victory without necessitating a full-scale invasion. Kyiv’s continued fixation on asymmetric tactics pushed it to avoid mobilization, as Ukrainian leadership understood that mobilization might engender the very disruption Russia was seeking to create. Russia could simply delay its attack, forcing Ukraine to maintain mobilization for an indefinite period and creating economic disruption and political resentment. Ultimately, the unrest this would have generated would likely have abetted Russian efforts at regime change (Zabrodskyi et al., 2022, p. 22).

Further, given Russia’s stated preferences for and history of asymmetric tactics and disinformation, Ukrainian leadership was likely wary of taking any steps that the Russians could have exploited as justification for an attack. This is particularly true given that, in 2008, Russia successfully set a trap for the Georgian government, baiting it into shelling South Ossetia (Fawn & Nalbandov, 2012). The lesson of the Russo-Georgian War was that Russia sought to manipulate the political environment to abet sharp, limited wars, not engage in the type of massive conventional invasions typically envisioned in the IR literature. Thus, rather than simply face the uncertainty of whether war would occur, Kyiv also had to guess at the kind of war that might happen while ensuring that the steps it took to prepare for one did not inadvertently harm its ability to guard against the other.

Ukraine calls for aid, will the West answer?

Ukraine faced another type of uncertainty that pushed it to moderate its military policy: uncertainty over the likely reactions of the major democratic powers. As Zelensky recounted, “[security guarantees] would have also been a signal of certainty. And, of course, we did not receive anything from the point of view of security guarantees.” He went on to note that “[foreign countries] can say a million times, 'Listen, there may be an invasion.' Okay, there may be an invasion—will you give us planes? Will you give us air defenses? 'Well, you’re not a member of NATO.' Oh, okay, then what are we talking about” (Khurshudyan, 2022)? Kyiv’s inability to confidently know that it would receive substantial military support from Western democracies likely further encouraged caution. After all, Ukraine was clearly the conventionally inferior power, and without promises of substantial external support, it was simply imprudent to risk escalation if there was any chance that it might be avoided. Of course, some states, notably the US and UK, did provide increasing amounts of aid in the months before Russia’s invasion. But the lack of unity within NATO meant that the long-term sustainability and scope of this aid was not guaranteed. Moreover, as Lanoszka and Becker (2023) point out, even “those countries that gave weapons to Ukraine were very circumspect in what they provided” (p. 182).

In fact, Kyiv was not merely unsure of the support it would receive in the leadup to a potential conflict; it was also unsure of the degree of support it would receive if attacked. Part of this derived from disunity on the part of Ukraine’s European partners. When asked about energy sanctions against Russia, for example, the then-German Defense Minister told journalists not to “drag Nord Stream 2 into the conflict” over Russia’s buildup (Siebold, 2022). The US was also unwilling to offer unconditional aid to Kyiv because Washington was unconvinced of Ukraine’s ability to offer meaningful resistance. Indeed, the Chairman of the Joint Chiefs of Staff predicted that Russia would take Kyiv in around 72 hours (Heinrich & Sabes, 2022). Full, sustained aid was thus predicated on a demonstrated ability to resist, a fact illustrated by Senator Lindsey Graham’s comment to Ukrainian officials that, “if you can defy the narrative that you’re going to be walked all over … it will be much easier to rally Congress to your side” (Banco et al., 2023).

Collectively, these two areas of uncertainty constrained Kyiv’s ability and willingness to engage in military escalation. The type of escalatory behavior predicted in the IR literature did not occur both because Ukrainian leaders worried it might backfire in a scenario where Russia planned to employ asymmetric tactics and because, due to uncertainty over the degree of Western support Ukraine could expect, they feared they would lack the resources to sustain the fight. In short, it was not simply uncertainty over Russian resolve and capabilities that drove Kyiv’s behavior, it was also uncertainty over how Russia might employ its capabilities and how resolved outside actors would be.

From whence these blind spots come

None of Ukraine’s decision-making is counterintuitive, and the types of uncertainty Kyiv confronted in the leadup to the war were hardly unprecedented. So, why have these considerations not been more explicitly integrated into core IR models on state decision-making under uncertainty? Three factors seem particularly salient. The first is that much of the extant literature draws evidence exclusively from great powers. Importantly, even when smaller states are studied, their agency often gets ignored by scholars more interested in the broader systemic competition of the era or in the manner in which great powers influenced or responded to the conflicts. Many have already recognized this great power bias (Acharya & Buzan, 2019). Unfortunately, the same scholars who criticize the field’s disproportionate fixation on great powers also tend to oppose traditional security studies’ fixation on interstate war. Thus, traditional security scholars who continue to research interstate war focus predominantly on the great powers, while scholars studying materially weaker powers are largely uninterested in interstate conflict.

But smaller powers face unique pressures and incentives. For one, they may be so materially outmatched that any form of escalation is likely to be counterproductive. Weaker powers are likely also more dependent on third parties to bolster their material capabilities, as evidenced by Ukraine’s continued requests for Western support. The literature on third parties and crisis bargaining is voluminous, including Bas and Schub’s (2016) work on secret alliances’ contribution to escalatory behavior, Powell’s (2017) piece examining the determinants and effects of third-party intervention in wars, and Krainin and Schub’s (2021) article evaluating how prospective power shifts influence alliance arrangements. Nevertheless, little of this literature directly applies to the question of state behavior under uncertainty in cases like Ukraine because it either is about alliance systems rather than constituent states or focuses primarily on institutionalized alliances rather than ad hoc coalitions. As Wolford (2015) points out, “theories designed to explain [alliances] do not apply well once … states form coalitions with nonallied partner, for whom even more elements of the cooperative military relationship remain to be negotiated” (p. 19).

These blind spots are exacerbated by scholarship around military operations and strategy being largely divorced from more theoretical work around conflict dynamics. Within IR, there exists substantial work on warfare in its own right. Scholars have employed campaign analyses to assess how potential conflicts are likely to unfold (Cunningham, 2020; Talmadge, 2008), identified the determinants of military effectiveness (Biddle, 2004; Lyall, 2020), and considered how states prepare for war (Imlay & Toft, 2007). This work has generally remained siloed, however, from more abstract theorizing around broader conflict dynamics such as war initiation, reducing our ability to apply real world military considerations when conceptualizing uncertainty. As noted earlier, uncertainty over whether a war will be fought conventionally or asymmetrically may affect crisis decision-making. Thus, by failing to seriously engage with detailed literature on military force employment and strategy in favor of more abstract concepts, scholars risk omitting important variables and missing crucial mechanisms by which military art influences decision-making around war.

Finally, these issues are exacerbated by the discipline’s fixation on World War One. As Lebow (2014) notes, this conflict serves as the basis of much of the work on bargaining models, crisis dynamics, and work linking information asymmetries and uncertainty to war. There is no doubt that World War One, with its rich historiography and historical significance, is an important case. Yet, IR's tendency to use it for theory building, often to the exclusion of other wars, risks treating sui generis aspects of the leadup to that war as constant features of all militarized crises.

For example, World War One was a conventional conflict, and all actors in the leadup to the war knew this. But this is not always the case, and by so frequently returning to the Great War for case studies and theory building, scholars risk overlooking how uncertainty over the nature of warfare shapes state behavior. Similarly, the war featured tight alliance networks, which reduced uncertainty over third party support. France knew it had Russian support, and Austria-Hungary received a “blank check” from Germany (Clark, 2012, pp. 413–415, 503). The primary uncertainty thus concerned Britain’s role, and here the scholarly focus has been on how this shaped Britain’s potential enemies’ decision to escalate rather than its potential allies’ choice to stand firm. Of course, the war is enormously complex, and its causes and consequences are still disputed. Nevertheless, there are recurring motifs and biases. Just as Rathbun (2007) rightly points out that scholars’ paradigmatic assumptions influence their conception of uncertainty, so too can scholars’ use of cases shape their understanding of uncertainty’s impact on state behavior.

In the case of World War One, the IR literature has almost exclusively looked at the great powers. Whereas studies of British and German behavior in the leadup to war are legion, Serbian decision-making, perhaps most relevant in the context of Ukraine, gets scant attention. More generally, World War One is often viewed as a war that “was in some significant degree unintended or undesired” (Lieber, 2007, p. 155). This has inspired work on spiral models and generally created a perceived link between uncertainty and unintended escalation. Of course, there are competing scholarly views that challenge the “accidental” nature of the conflict’s outbreak. Nevertheless, the conventional wisdom on the war’s origins retains a powerful sway and likely shapes, at least implicitly, theorizing that links uncertainty to escalation.

None of these shortcomings render the existing literature useless, but they do produce analytical gaps that have been revealed in Kyiv’s behavior in the leadup to the Russo-Ukrainian War. It is, of course, impossible to fully eliminate these blind spots but, by broadening the set of illustrative cases and by better integrating the literature on military operations and strategy with the literature on conflict initiation, the field can go a long way in redressing these gaps and improving our understanding of the manner in which uncertainty shapes state behavior.

Acknowledgements

For useful discussion and feedback, the author thanks Andrew Payne and Giuseppe Spatafora, as well as the two anonymous reviewers.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Samuel M. Seitz

Samuel Seitz is a doctoral candidate in international relations at the University of Oxford. He can be found on Twitter @samseitz3.