When former U.S. President Donald Trump indicated earlier this year that he “would encourage” Russia to “do whatever the hell they want” to countries that are “delinquent,” presumably in terms of defense spending, allies expressed alarm. This trans-Atlantic burden-sharing kerfuffle is only unusual for the crassness of the language. U.S. concerns about European free-riding date back at least to the Eisenhower administration.
The time has come for European allies to carry a larger share of the burden of European defense. The era when allies could obfuscate unequal burden-sharing by claiming, for example, that increasing defense spending is “useless” to address capability shortfalls or contributions to NATO missions is over. There is simply no empirical support for these claims — defense spending correlates strongly with any meaningful measure of defense outputs. We argue, however, that harsh public criticisms like the language above are ineffective at convincing allies to spend more, while hard-won agreements like the Wales and Vilnius pledges do in fact lead to increased spending. Moreover, the severity of current security challenges makes the domestic and E.U.-level politics of such agreements more attainable.
Now is the time for NATO Europe to truly “step up” operationally and strategically. A “European Pillar” of NATO for the current security environment can emerge from a simple defense planning principle: NATO Europe must be capable of defending the territorial integrity of European allies even in a situation in which the United States is decisively engaged in conflict outside of Europe.
Burden-Sharing Is Not Only a Trump Thing
In February 2015, Washington formally communicated to NATO allies that a significant portion of its national capabilities would be devoted to contingencies outside of Europe. Rather than address this reality by fielding the necessary capabilities themselves, non-U.S. allies simply chose to accept the risk that the portion of U.S. forces identified would not be available in a crisis. Nearly a decade later, allies should understand that demands outside of Europe have only increased, as has risk within Europe. Such risk is no longer acceptable, and European allies must be prepared to lead in the defense of Europe, regardless of who is president of the United States.
Scholars and practitioners have on occasion responded to Trumpian critiques of trans-Atlantic burden-sharing with arguments that NATO’s defense spending guidelines, to which Trump often refers, are, according to Garrett Martin and Balazs Martonffy, an inferior metric that Seamus Daniels and Kathleen Hicks argued allies should “move beyond” because it wrongly focuses on inputs. This is a mistake.
While defense outputs are surely more interesting and important than inputs like spending, they are very difficult to observe and agree upon: Are capabilities themselves outputs or merely what defense economists call “intermediate outputs?” What final outputs are countries interested in: Deterrence? Human security? These are interesting questions for economists, who continue to improve research on them. But policy should not be on pause until they are resolved.
As NATO seeks to move from burden-sharing to responsibility-sharing, defense spending remains foundational. It is clear now that there is a strong, positive relationship between defense inputs as measured by NATO and the European Union (i.e., share of gross domestic product spent on defense and share of defense budgets spent on equipment modernization, munitions, and operational activity) and a variety of important outputs. Here, we highlight three of those.
First, and perhaps most important at present, investing in readiness in the form of operating and maintenance expenditures, which also include munitions procurement and use, strongly predicts military aid to Ukraine. Not only was this the case at the outset of the Russo-Ukrainian War, but the relationship also persisted throughout 2023, as figure 1 demonstrates. A one percentage point increase in operating and maintenance spending, as a share of an ally’s gross domestic product, in one year corresponds to a 0.76 percentage point increase in that ally’s contributions to Ukrainian defense in the following year, as a share of the ally’s gross domestic product . This is a statistically significant (at the .01 level) and substantively large relationship that is not attributable to chance.
Figure 1: Operating and Maintenance Spending as a Share of Gross Domestic Product and Total Bilateral Commitments to Ukraine as a Share of Gross Domestic Product, 2022–-2023 (Source: Lanoszka and Becker, 2023)
Second, this finding echoes earlier findings that deployability and sustainability “output metrics” tracked by the European Defence Agency were also predicted by defense spending. Although NATO allies generally do not choose to release their “output metrics,” researchers are increasingly developing tools to measure defense capabilities, and those metrics also correlate well with defense spending.
Third, what figure 2 demonstrates is that material military power, defined by Mark Souva as “major naval, air and land weapons as well as nuclear weapons and ballistic missile capability,” is very well predicted by defense spending of all types, but particularly by topline defense spending as a share of gross domestic product (NATO’s 2 percent metric).
Figure 2: Defense Spending and Material Military Power (Source: Authors, based on Souva 2023 and Becker, Benson, Dunne, and Malesky, 2023)
NATO Secretary General Jens Stoltenberg’s conclusion that burden-sharing “involves cash, capabilities, and contributions” is inescapable: The more allies spend on defense, the more material capabilities they possess, and the more they have contributed to alliance priorities like support for Ukraine. Allies wishing to argue that their efficient defense enterprises yield outputs that belie their inputs should simply make their nonclassified output metricsavailable to the public.
Finally, conceptual critiques regarding fuzzy definitions and what Der Spiegel called “accounting tricks” are themselves fuzzy. NATO explains that “since 1963, [its defense spending data] has formed a consistent basis of comparison of the defence effort of Alliance members based on a common definition of defence expenditure.” We do not have to take NATO at its word, however, because the Stockholm International Peace Research Institute (SIPRI) measures defense spending independently of NATO, and the two organizations’ figures match up quite closely.
Figure 3 shows how tightly and consistently correlated NATO’s and SIPRI’s defense spending figures have been across the decades. For each year and each country, SIPRI’s figures are on the X axis and NATO’s on the Y axis. For each decade, the regression coefficient is essentially 1, meaning the figures are nearly the same, with most observations tightly clustered along upward sloping lines with slopes of about 1.
Figure 3: NATO and SIPRI Defense Spending Data, by Decade (Source: Authors, based on SIPRI and NATO data)
There are some outliers in figure 3, and they are often from a handful of countries — namely Greece, Turkey, and Portugal. Figure 4 compares the NATO and SIPRI defense spending figures by country over time and includes those three countries alongside a sampling of other countries, including Germany, recently singled out for “accounting tricks.” Even when visualized this way, and despite some gaps for Portugal and Greece, the NATO and SIPRI spending datasets are consistent with one another and become more so over time. Between the validation of NATO figures by the International Secretariat and other allies as part of the NATO Defence Planning Process, and external validation by SIPRI, analysts can be confident that these figures do indeed enable what NATO correctly refers to as a “consistent basis of comparison” both over time and across countries.
Figure 4: NATO and SIPRI Defense Spending Data, by Country: Albania, Germany, Greece, Italy, Lithuania, Netherlands, Portugal, Turkey, United States (Source: Authors, based on SIPRI and NATO data
Berating Allies Does Not Improve Burden-Sharing
Claims such as Daniel Kochis’ that Trump’s “laser-like focus on inadequate defense spending” improved burden-sharing are just as wrong as claims that European defense spending is already sufficient or that inputs are unrelated to outputs. When, as one of us observed while serving in Brussels, U.S. representatives suggested to European allies in the winter of 2016 that they might consider moving toward U.S. positions on burden-sharing prior to the Trump inauguration, allies were nonplussed. This lack of response to U.S. executive criticism appears to be borne out by the historical data.
Figure 5 visualizes that Trump’s language on NATO was unusually negative among U.S. presidents, but figure 6 demonstrates that — at best — Trump presided over a continuation of a trend of increased spending established after the 2014 Wales Summit. Conversely, mean equipment spending (NATO’s key 20 percent metric) declined from 2019 to 2020 during Trump’s presidency.
The difference in tone, between, say, former U.S. Secretary of Defense Robert Gates’ 2011 speech in Brussels — in which he shared critical “views in the spirit of solidarity and friendship, with the understanding that true friends occasionally must speak bluntly with one another for the sake of those greater interests and values that bind us together”— and Trump’s threats to abandon “delinquent” allies accused of “chronic underpayments” is distinct. Harsh language is not associated with positive effects on trans-Atlantic burden-sharing and may even provoke, as Kathleen Powers and Dan Altman found, “reactance,” in which countries galvanize resistance in response to the perception that outside pressure has infringed on their ability to make choices. In practical terms, this means that aggressive or insulting external demands make it harder for national leaders to manage the domestic politics of increasing defense spending — Trump’s singling out of Germany amid personal attacks on Chancellor Angela Merkel made her task before the Bundestag and the German people harder, not easier.
Figure 5: Positive, Negative and Ambivalent Language about NATO, by U.S. President (Source: Becker, Kreps, Poast, and Terman, 2023)
Figure 6: Disaggregated Military Expenditures by Year, Non-U.S. NATO, 1985–2023 (Source: Becker, Benson, Dunne, and Malesky, 2024)
So, despite assertions to the contrary, tough talking allies is at best ineffective and may even make it harder for allies to increase defense spending during times of (relative) peace. Until very recently, domestic political economies and institutional factors at the NATO and E.U. levels appeared to have much more significant relationships with European defense spending than external threats — whether those threats came verbally from U.S. presidents or physically from state or nonstate actors.
Threats Matter, but in Complicated Ways
It is easy to argue, as one of us has, that neither Trump nor President Barack Obama drove European defense spending up, but Russian President Vladimir Putin did. There is ample evidence that European allies are alarmed by Russia’s wars of aggression, and one might intuit that such alarm would cause increases in defense spending. But it is far from obvious how threatening behavior translates into threat perceptions, and how those threat perceptions translate into material action by threatened allies. Differences in actual behavior among allies with similar geostrategic positions provide evidence for the ambiguity of these relationships.
Canonical work in international relations focuses on threats as drivers of state behavior, including in alliances. But this relationship is very complicated to identify empirically, with threats’ effects difficult to disentangle from other factors, and often conditional on interactions with those factors. There is no evidence that defense spending correlates linearly or monotonically with threats, regardless of how those threats are measured.
A more realistic way to think about threats affecting alliance behavior is by conceptualizing threat perception at the level of individual allies rather than aggregate threat behavior and capabilities and thinking about structural breaks in those threat perceptions. It would be difficult to argue that Russia’s 2014 invasion of Ukraine and the Islamic State’s proclamation of its caliphate in the months leading up to the Wales Pledge did not help enable allies to agree to it. Such conditions were not decisive but were catalyzing, enabling the alliance to formulate a collective response, even when allies have threat “awakenings” at different times and on different scales.
Figure 7 is suggestive on this front. U.S. defense spending shifted dramatically upward from 2001, then many allies shifted upward after 2014, Slovakia after 2018, and nearly all allies save the United States after 2022. Such breaks in spending may be associated with breaks in threat perception (i.e., the United States after 9/11, and European allies after Russian aggression in 2014 and 2022).
Figure 7: Structural Breaks? (Source: Authors, based on NATO data)
Future research may evaluate such possible structural breaks in defense spending along with similar breaks in threat perception. Current research, however, indicates that allied publics do respond to information about threats by altering their preferences about fiscal tradeoffs. The Italian public, for example, is generally believed to have distinctly unfavorable views toward defense spending, and particularly “guns versus butter” tradeoffs. Italian officials indicated in 2016 that Italy — despite having agreed at Wales to increase defense spending and specifically to increase equipment spending — would exclude some new military shipbuilding projects from its accounting of military equipment expenditures in the NATO Defence Planning Process because the Italian public would find such expenditures objectionable in a time of fiscal retrenchment.
An original survey experiment, however, fielded in Italy just prior to Russia’s February 2022 invasion of Ukraine, found that Italian respondents were more inclined to favor military spending when given information about the Russian threat, and that despite general public skepticism about defense spending in Italy, such information mitigates this skepticism, even when respondents are considering high debt burdens. So, it is possible that structural breaks in the threat environment, to the extent that such breaks are occurring, will yield concomitant structural breaks in attitudes toward and behavior regarding defense spending.
Making Institutions Work
NATO allies should not, however, count on threats advertising themselves far enough in advance to prepare for them with appropriate investments. Instead, they should purchase insurance against threats before they are apparent, assuming that Russia’s initial bungling of the Ukraine invasion is an exception rather than a rule. While such preparation is not automatic, it is attainable through careful diplomacy and management of institutions. NATO the organization enables allies to coordinate defense investments, which mitigates tendencies toward free riding, rather than encouraging them.
In short, the Wales Pledge to “reverse the trend of declining defence budgets” and “aim to move towards” 2 percent (of gross domestic product on defense) and 20 percent (of defense spending on equipment modernization) guidelines “within a decade” was effective in eliciting greater spending from allies than would have been the case in its absence. Figure 8 visualizes important evidence demonstrating that Denmark and Norway increased defense spending much more significantly and rapidly after 2014 than did Finland and Sweden, which faced nearly identical (if not more challenging) international situations and threat vulnerability but were not (at the time) NATO allies.
This phenomenon strongly suggests that the Wales Pledge itself had a significant effect on its signatories. Of course, the pledge itself was at least in part enabled by Russia’s illegal annexation of Crimea. Further evidence of the effect of the Wales Pledge is the difficulty allies had in agreeing to the pledge in 2014, with reluctant nations having what one NATO official referred to as “temper tantrums” before finally relenting. If allies did not expect to be bound by the pledge, they would have instead shrugged their shoulders, agreed, and ignored it. The pledge was initiated, moreover, by then–Secretary General Anders Fogh Rasmussen himself well before Russia’s 2014 invasion of Crimea, and the differing responses of NATO and non-NATO allies make it clear that increased threat from Russia alone was not enough to change behavior.
Figure 8: Slope Charts, Nordic Countries, Pre-and Post-Wales (Source: Becker 2024)
The effectiveness of the Wales Pledge in increasing European allies’ defense spending, even when accounting for international threats and domestic political economies, points to the importance of the Vilnius Pledge — making the 2 percent and 20 percent guidelines into “a floor not a ceiling” — in preparing NATO for current and emerging challenges. Outcomes of the Russo-Ukrainian War remain uncertain — one could imagine a Ukrainian victory leading allies to believe that conventional threats in Europe have receded, while a Russian victory could lead them to seek accommodation underwritten by mutual nuclear deterrence. Neither of these specific outcomes is necessarily likely, and war in Europe may vacillate between them for some time to come.
Under such uncertainty, collective actions like the Wales and Vilnius pledges are particularly important. The Vilnius commitment means that allies have committed to collective action beyond common but uncoordinated reactions to evolutions in the threat situation in the coming years. This coordinated collective action is critically important in the face of persistent but dynamic external threats, as well as in the face of challenges posed by domestic politics within the alliance and its members.
A European commitment to the ability to lead in defending the territorial integrity of allies in Europe with U.S. forces decisively engaged elsewhere has been implicit within NATO for some time. Specifically, the NATO Defence Planning Process refers to the “50 percent” rule that “no ally should provide a contribution that represents more than half of a capability.” This has long been NATO’s capabilities “floor.” But unlike the Wales Pledge, it remains largely a non-public, implicit guideline. Best-case planning, as regards allies or adversaries, is no longer an option for NATO. Agreeing the 50 percent rule publicly at the level of heads of state and government, like the Wales and Vilnius pledges, would increase the likelihood of allies systematically complying with it, making NATO better prepared for war in Europe.
Jordan Becker, Ph.D., is an Army officer, a professor at West Point, and director of the Social Science Research Lab at West Point.
Douglas Lute is a retired U.S. Army lieutenant general and served as the U.S. Ambassador to NATO from 2013 to 2017.
Andrew Webster is an Army officer and instructor of economics at West Point.
The views expressed in this article are those of the authors alone and do not represent those of the U.S. Military Academy, the U.S. Army, the Department of Defense, or any part of the U.S. government.
Image: Glenn Fawcett
Comments are closed.