Why the Ukraine War Caught Europe by Surprise

Right Numbers, Wrong Predictions?

Can international crises – such as the war in Ukraine – be anticipated and possibly even prevented through the use of quantitative data and mathematical models? Traditionally, international crisis management has been considered an “art,” not a “science.” Foreign policy was long considered the preserve of skillful diplomats, navigating the complexity of international relations with empathy and intuition. Even in modern capitalist democracies, diplomacy retains a degree of institutional autonomy from considerations of managerial efficiency, bureaucratic planning, and utilitarian cost-benefit calculations. However, a wider trend towards “governance by numbers,” [1] diagnosed by Alain Supiot in the field of domestic policies, has increasingly come to shape international politics.

If the forward march of quantification is unmistakable, its utility is less evident. More than thirty years after the collapse of the Soviet Union, which caught governments and academics alike by surprise, quantitative tools appear to have contributed little to forecasting — and still less avoiding — the current war between Russia and Ukraine.

Prediction by the numbers

Over the past decade, a number of initiatives have been launched — by international organizations as well as national governments and their intelligence agencies — to improve the understanding and management of international crises through the exploitation of quantitative data. Examples are numerous: the United Nations has implemented “UN Global Pulse”, a data-driven project aimed at enhancing crisis-response capabilities, and it is currently developing kindred models related to the COVID-19 pandemic. The German Federal Foreign Office has set up an “early warning tool,” PREVIEW, that purportedly enables officials to act “earlier, more resolutely and more substantially” to prevent conflicts from escalating, while the CIA actively recruits data scientists capable of developing algorithms and statistical methods to detect patterns and correlations across massive databases. All of these initiatives mobilize a range of sources, from academic databases such as the Armed Conflict Location & Event Data Project (ACLED) to commercial data collected by private actors, including intelligence-oriented startups like Preligens, and data collected internally, through the use of open-source and closed-source intelligence.

Expectations concerning conflict prevention through the systematic use of quantitative data in conflict management are tremendous. Sherdon Himmelfarb, one of the most prominent researchers of the field, thus imagines that “number crunching and pattern recognition may hold the key to predicting and preventing conflicts.” [2] Such hopes reflect the rapidly growing variety and mass of available data, computational means of analyzing even unstructured data, and social scientists’ efforts to formulate causal mechanisms to explain the origins and the dynamics of violent conflicts across time and space.

Quantitative reasoning and three misperceptions of the war in Ukraine

The war in Ukraine illustrates some of the limits and dangers of policy-making supported by quantitative conflict data. Initially, Russian President Vladimir Putin’s announcement of a “special military operation” on the morning of February 24 elicited cries of vindication in London and Washington. Two decades after the run-up to the war in Iraq, Anglo-American intelligence agencies could claim to have redeemed themselves. If US and UK sources had projected the date of the offensive weeks in advance, their unusual publicity campaign began previous autumn, when aerial photos first disclosed Russian forces massing on the Ukrainian border. In November 2021, the U.S. Director of National Intelligence briefed NATO on the matter, and early the next month the Washington Post published details of an operational plan.

According to the BBC, these warnings were partly a product of the systematic analysis of open-source intelligence and commercial satellite imagery. [3] Yet Western European governments and intelligence services largely discounted the possibility of a Russian offensive, and reacted in disbelief at the outbreak of a full-scale land war in Europe. Symptomatically, a few weeks after the start of the invasion, French President Emmanuel Macron confessed that he and his European counterparts were caught off-guard because Putin had broken with the “linearity” of the long simmering conflict. [4]

Moreover, boasts in Washington over accurate forecasting of the conflagration soon yielded to a more chastened appraisal of the course of battle itself. However exact in their anticipation, US authorities admitted to having misjudged the outcome of the fighting. As Marine Corps Commandant David H. Berger commented in mid-March, computer models and virtual wargames would have forecast a rout of Ukrainian forces within three to four days; a narrow focus on the technical details of weapons systems and other indicators had obscured, he said, less tangible qualities of morale and the “human” element. [5] While US intelligence correctly predicted the outbreak of war, it came as a surprise for European and US-based conflict researchers who attempt to establish general tendencies in the occurrence of war based on painstakingly compiled and coded data. Thus, the Peace Research Institute Oslo (PRIOPRIO), one of the leading centers in Europe, has for decades forecast a global reduction in inter-state violence. For the purposes of PRIO’s dataset, which defines armed conflict according to the minimal yearly threshold of twenty-five battlefield deaths, the Russian intervention in Georgia in 2008 — for example — was coded as a civil war, pitting the government against South-Ossetia. The Heidelberg Institute for International Conflict Research (HIIK)’s Conflict Barometer, for its part, classified the conflict between Russia and Ukraine as “non-violent,” its intensity registering 2 on a scale from 1 to 5. Such assessments give an idea of the constraints imposed by reliance on aggregated historical event data. Even as tensions between Russian and NATO multiplied in the 2000s and 2010s, the major conflict databases took little note. This doubtless contributed to the Panglossian [6] notion that interstate war had disappeared, at least in Europe, yielding perhaps to what Mary Kaldor and others have labeled “new wars, ”characterized by their lack of ideological motivation, political organization, and state control. No wonder, then, that commentators and political leaders were moved to identify in the war in Ukraine nothing less than the reappearance of History itself, the bloody clash of arms and ideas thought to have been consigned to the dustbin along with the Soviet Union. [7]

Six months after the initial attack, we can identify three inherent, methodological limits on quantitative conflict analysis as a means of forecasting war: an inability to grasp singular events that change existing dynamics of violence; the difficulty of quantifying psychological factors such as intentions and emotions; and the corollary assumption of universal rationality that often informs policies of deterrence and sanctions.

Singular catastrophes

While expressions of astonishment at the “return” of war to Europe clearly reveal a degree of ethnocentrism, they also reflect the influence of quantitative studies on futurology. Attempts to predict conflict on the basis of aggregated historical indicators too often involve a linear projection of past trends into an unknown future. Rare events tend to be omitted as insignificant because they occur so infrequently, however grave their potential effects (for instance, the use of nuclear arms). Quantitative, correlational thinking struggles to accommodate the phenomenon of the so-called “black swan”; that is, a rare yet significant event that fundamentally transforms existing patterns. As the war in Ukraine illustrates, exceptional events happen; when they do, we can expect them to have more dramatic consequences than more punctual, forecastable contingencies, such as border disputes or “terrorist” attacks.

Capabilities, intentions, and morale

A second misperception advanced by quantifiers is the idea that the prospects for military confrontation can be analyzed narrowly, in terms of material forces that can be easily counted (such as troop numbers, budgets, physical movements…). The common assumption, visible in recent debates over the reappearance of “high-intensity warfare,” is that the balance of forces between the belligerent parties can be measured in terms of manpower and equipment. Victory is therefore assumed to be a simple function of one side’s superior capabilities and resources. Analysts watch for every shift in potential adversaries’ military capability, from which they read off likely intentions. The pace and scale by which states arm themselves and acquire or develop new weapons systems can thus be taken as evidence for their aggressive designs, or lack thereof. This line of interpretation, founded on the tabulations of global military expenditure produced by the Stockholm International Peace Research Institute (SIPRI) among others, does not take into account alternative ways of understanding the articulation of capabilities and intentions, including the significance of personalities, political culture, and state-society relations. [8] As Clausewitz, the great Prussian philosopher of war, argued, these human factors mean that the outcome of combat does not only depend on what can be counted, and that all wars are governed by elemental contingency and the drive to escalation.

In the current context, many of those who correctly anticipated the Russian invasion did not expect Ukraine to withstand the attack, given that the country spends about 6 billion euros annually on its armed forces, whereas Russia spends roughly ten times as much. If massive US subventions have redressed this discrepancy, the initial course of hostilities also illustrated the importance of psychological factors, such as determination, tactical creativity, and cohesion.

In spite of the experiences of Vietnam, Afghanistan, and Iraq, Western political and military leaders continue to underrate morale as a factor in determining battlefield outcomes. The lesson that the value of territory and commitment to the fight are not the same for the defender and the attacker is often forgotten. Also too often omitted from view are constraints on the aggressor resulting from logistical considerations and lack of support from the population in occupied territories.

Above all, however, the quantitative view cannot make sense of the causes of war. Taking military capabilities as a proxy for motives and interests renders cannot substitute for the more painstaking, ambiguous search for qualitative explanations. Enumeration of guns and soldiers promises a parsimonious explanation and allows for responsibility to be assigned to an identified, aggressive enemy, without overmuch consideration of the roots of any given conflict.

Illusions of deterrence

Finally, a third misperception we might associate with quantitative analysis is the assumption that all sides of the conflict make their decisions based on the same, universal rationality of cost-benefit calculation. Deterrence is a canonical example. According to this concept, the larger and better equipped your armed forces, and the more powerful your allies, the more plausible your ability to inflict damage on a potential aggressor and thus dis-incentivize belligerence. On this view, the only coherent response available to the “West” in the current crisis is to sanction Russia, furnish arms to Ukraine, and push forward NATO enlargement. Russian leaders are presumably meant to understand that continuing the war is more expensive than ending it. It is anything but clear that the Kremlin shares this analysis, which suggests indifference to the specific interests that motivate the Russian leadership and the degree to which economic sanctions and NATO enlargement are perceived as proof of Western hostility. Paradoxically, the government may even benefit from these reactions, a patriotic salve and boost to its legitimacy.

How (not) to use data in the absence of political will

Even if quantitative analysts addressed these shortcomings, would doing so automatically translate into better political decision-making? The example of the EU is a case in point.

According to the Maastricht Treaty, the EU seeks to “preserve peace, prevent conflicts and strengthen international security, in accordance with the objectives and principles of the UN Charter.” Born out of warfare in the former Yugoslavia, where the Union was tasked with responding too slowly, the EU conflict prevention mandate has underwritten a shift from reactive “crisis management” to a more proactive strategy of prevention. Although hesitancy in the Balkans resulted at least in part from the difficulty of securing a political consensus among member states, the EU’s own self-assessment has focused on the promotion of technocratic “early warning” instruments, such as the EU Conflict Early Warning System (EWS): According to its policy guidance document,

EWS is an evidence-based risk management tool that identifies, prioritizes and assesses situations at risk of violent conflict in non-EU countries, focusing on structural risk factors with a time horizon of up to four years. The EU conflict EWS seeks to identify conflict prevention and peace building opportunities through joint, shared analysis and to develop timely, relevant, coherent and conflict-sensitive responses to prevent the emergence, re-emergence or escalation of violence.” [9]

The EWS makes use of the so-called Global Conflict Risk Index, which aggregates information from open-source databases in order to identify countries and regions at risk for violent conflict. This quantitative risk assessment —based on political, military, socio-economic, and structural indicators — is designed to alert EU decision-makers to events as they develop, enabling discussion of appropriate diplomatic, economic, and peacekeeping strategies before an emerging crisis escalates. But it evidently made scant contribution to defusing the crisis in Ukraine. Up until the moment of the invasion, the EU was sidelined in the debate over the Russian military build-up along the Ukrainian borders. European governments and intelligence services could nonetheless have concluded, even in light of the imperfect information they had (apparently the same as their British and American allies), that war was likely.

Indeed, in December 2021 the European Parliament, under intense pressure from the US, adopted a resolution asking the EU Commission and member states to prepare

the adoption of severe economic and financial sanctions in close coordination with the United States, NATO and other partners, in order to address the immediate and credible threats posed by Russia, rather than wait for another invasion to take place before taking action.” [10]

But no meaningful diplomatic initiative accompanied this tough talk, and the European sanctions themselves were to be entirely reactive, beginning with the response to Russia‘s recognition of the separatist republics of Donetsk and Luhansk on February 23. [11]

The EU Strategic Compass: More of the same?

The EU’s failure to anticipate and react proactively to the crisis in Ukraine casts a revealing light on its recently published defense white paper, the Strategic Compass. Drafted in November 2021, the document is intended to outline the medium-term orientation for the EU’s security and defense policy, with the aim of making the Union a more proactive, visible, and effective international actor. [12] However, rather than investigating how divergent political interests of EU member states might be harmonized, in line with other EU policy fields, the Strategic Compass promises more of the same: more apolitical coordination, more data collection, more analysis. In bureaucratic langue de bois starkly at odds with the scale of the threats in question, the document calls for:

shared analysis to increase [the] situational awareness and strategic foresight [of the EU], building on [a new] Early Warning System and horizon scanning mechanism [in order to] strengthen [the European] intelligence-based situational awareness and relevant EU capacities, notably in the framework of the EU Single Intelligence Analysis Capacity […] as a single entry point for strategic intelligence contributions from Member States’ civilian and military intelligence and security services.

It would be naive to believe that political and military actors act only according to data and observed “trends.” In the present case, for instance, the decision to impose sanctions on Russia and arm Ukraine was unmistakably political. However, “hard data” too often serves to occult the stakes involved in decisions of this type, by presenting them as non-ideological, objective, evidence-based policies. The managerial language of the EU Commission seems to suggest that sanctions against Russia are only a technical tool to “repair” a broken order: “Restrictive measures or ’sanctions’ are an essential tool of the EU’s Common Foreign and Security Policy (CFSP). They are used by the EU as part of an integrated and comprehensive policy approach, involving political dialogue, complementary efforts and the use of other instruments at its disposal.”

Democratic debate and political will

Even the most sophisticated social science tools cannot fully grasp the inherently political nature of warfare. As already noted, governments and populations do not act according to a uniform, universal rationality. Likewise, to take seriously the political character of armed conflict means acknowledging that political will is integral to the successful prevention and de-escalation of hostilities. In the war between Russia and Ukraine, this entails a sober consideration of objectives and means. For example, what domestic costs is the EU ready to assume to “punish” Russian aggression? Is there really popular enthusiasm for a new Cold War, and skyrocketing energy prices for years to come? Might talk of energy security revivify visions of strategic autonomy?

Trying to quantify the costs and consequences of possible responses can only be a first step. Beyond the open question of how effective EU sanctions have been in accomplishing their stated objective, the decision to reduce fossil fuel imports from Russia has been taken without substantial debate concerning the political, economic, and social impact on EU member states themselves. As summer turns to autumn, it is more and more clear that essential political questions will be posed, including the distribution of costs, the relative priority of citizens’ consumption needs versus the demands of industry, and the contradiction between summons to combat “climate change” and the rush to find readily available resources, however polluting.

The production of reliable short- and medium-term scenarios using available quantitative data may be useful when it comes to informing citizens and decision-makers about the implications of these choices. However, it should not be assumed that a consensus exists with respect to the most “rational” balance between domestic costs for European societies and the desire to penalize Russia. What is urgently needed, rather, is a genuinely democratic debate, at the EU as well as the national level, and public discussion of the available options and trade-offs. Such a course would do more to favor cohesion and security than investment in new bureaucratic tools and expensive military technologies, whose sophistication can only highlight the current lack of unity and political will in the Union.