Graham Allison's Thucydides Trap framework has shaped how presidents, generals, and scholars think about great-power rivalry. But is it built on solid ground? A balanced examination of the strongest criticisms — and the defenses that answer them.
Few concepts in modern international relations have achieved the cultural penetration of the Thucydides Trap. Since Graham Allison introduced the term in 2012 and elaborated it in his 2017 bestseller Destined for War, it has been cited by heads of state, debated in congressional hearings, analyzed in war colleges, and invoked in thousands of op-eds about US-China competition. Chinese President Xi Jinping himself has publicly engaged with the concept by name. By any measure, the Thucydides Trap is the most influential geopolitical framework of the 21st century.
But influence is not the same as accuracy. The theory's extraordinary reach has also attracted extraordinary scrutiny. Political scientists, historians, military strategists, and international relations theorists from across the ideological spectrum have raised serious objections to the framework's methodology, its assumptions, and its implications for policy. Some argue that the Thucydides Trap oversimplifies complex historical dynamics. Others contend that it risks becoming a self-fulfilling prophecy — that by telling leaders war is the historically probable outcome, it may inadvertently make war more likely.
This page examines the most important criticisms of the Thucydides Trap with the rigor the subject demands. Understanding where the theory succeeds and where it falls short is not merely an academic exercise. If the framework that informs how world leaders think about the most dangerous geopolitical relationship on Earth has significant flaws, those flaws have real-world consequences. Getting the analysis right is, quite literally, a matter of war and peace.
"The most dangerous thing about a powerful theory is not that it might be wrong, but that it might be right enough to be believed and wrong enough to be misleading."
The single most frequently cited criticism of the Thucydides Trap concerns the selection of the 16 historical cases that form the empirical backbone of the theory. Allison and his research team at Harvard's Belfer Center identified 16 instances over the past 500 years in which a major rising power threatened to displace a major ruling power. Twelve of those cases ended in war — a 75% rate that forms the dramatic statistical core of the framework. But critics argue that this number is an artifact of how the cases were chosen, not a genuine reflection of how power transitions work.
The historian Arthur Waldron of the University of Pennsylvania has been among the most vocal critics on this point. Waldron argues that Allison's case list includes some rivalries that do not genuinely fit the rising-versus-ruling power template, while omitting others that clearly do. For example, the rise of Japan relative to China in the late 19th century, the rise of the United States relative to Spain in the same era, and numerous power shifts within the European balance-of-power system are either absent or categorized in ways that critics find questionable.
Hal Brands of Johns Hopkins University's School of Advanced International Studies has raised a related concern: that the criteria for what counts as a "major" rising power and a "major" ruling power are insufficiently defined, allowing the researcher to include or exclude cases in ways that support the desired conclusion. If you define the universe of cases broadly — including every significant shift in relative power over five centuries — the percentage that ended in war drops substantially. If you define it narrowly, including only the most dramatic confrontations, the percentage rises. The 75% figure, Brands suggests, depends critically on where you draw the boundaries.
The political scientist Steve Chan of the University of Colorado has conducted the most systematic critique of Allison's case selection. Chan's research, published in his book Thucydides's Trap?: Historical Interpretation, Logic of Inquiry, and the Future of Sino-American Relations (2020), argues that Allison's dataset suffers from both selection on the dependent variable (choosing cases where war occurred and then finding a rising-versus-ruling dynamic to explain them) and confirmation bias (interpreting ambiguous historical evidence in ways that support the framework). Chan's alternative analysis, using more rigorous case selection criteria, finds a substantially lower rate of war in power transitions.
This is not a trivial objection. The entire persuasive force of the Thucydides Trap rests on the claim that history reveals a reliable pattern. If the pattern is an artifact of methodology rather than a genuine empirical regularity, the framework's policy implications change dramatically. A 75% war rate demands urgent preventive action. A 40% rate — still alarming — suggests a very different set of policy responses.
Perhaps the most philosophically important criticism of the Thucydides Trap concerns its relationship to determinism. The framework presents war as the historically dominant outcome of power transitions — 12 out of 16 cases, or 75%. While Allison explicitly states that war is not inevitable and that the four peaceful outcomes prove that leaders can escape the trap, critics argue that the framing itself carries a powerfully fatalistic message that undermines the very diplomacy needed to avoid conflict.
The concern is not merely theoretical. If senior policymakers in Washington and Beijing internalize the belief that structural forces make war between a rising China and a ruling United States the probable outcome, their behavior may shift in ways that make conflict more likely. Defense budgets expand. Military postures become more aggressive. Red lines harden. Diplomatic flexibility shrinks. Threat assessments inflate. Each side interprets the other's defensive measures as offensive preparations, triggering the classic security dilemma spiral that the theory describes but may also accelerate.
"If you tell two great powers that history says they will fight, you may be giving them permission to stop trying not to." — attributed to a senior European diplomat at the Munich Security Conference
Joseph Nye, Allison's colleague at Harvard and one of the most respected international relations scholars alive, has raised this concern directly. Nye argues that while Allison's historical analysis is valuable as a cautionary tool, the deterministic packaging risks distorting how policymakers use it. Nye's own framework — emphasizing the importance of "soft power," institutional design, and narrative management — suggests that the structural pressures identified by Allison are real but far more manageable than the 75% war rate implies, provided leaders invest in the right tools.
The British historian Margaret MacMillan, author of The War That Ended Peace, has made a subtler version of this argument. MacMillan contends that historical analogies — including the Thucydides Trap — are most dangerous when they are presented as laws rather than as cautionary tales. History does not repeat in mechanistic patterns, she argues. Each situation is unique, shaped by contingency, personality, accident, and choice in ways that no statistical generalization can capture. By packaging history as a pattern with a probability attached, Allison risks giving policymakers false confidence that they understand the forces at work — when in fact the most dangerous threats are precisely the ones that do not fit the pattern.
This critique cuts to the heart of a fundamental tension in the social sciences: the relationship between structural explanation and human agency. If the Thucydides Trap is a genuine structural force, then leaders must understand it to counteract it. But if the act of naming and publicizing the force changes the behavior of the actors involved, the framework becomes part of the dynamics it claims to describe — a feedback loop that complicates any claim to scientific objectivity.
Allison has responded to this criticism repeatedly, arguing that awareness of dangerous structural pressures is the precondition for escaping them, not an obstacle to escape. He draws an analogy to medicine: a doctor who tells a patient that their condition is life-threatening is not causing the disease but rather creating the urgency needed for treatment. The Thucydides Trap, in Allison's framing, is a diagnosis, not a destiny.
Of all the criticisms leveled at the Thucydides Trap, the nuclear revolution argument may be the most devastating to the framework's predictive claims. The argument is straightforward: nuclear weapons have fundamentally altered the cost-benefit calculus that governed great-power competition for the previous five centuries. Before 1945, a rising power could rationally calculate that war with a ruling power — however costly — might produce gains that outweighed the losses. After 1945, that calculation changed permanently. A war between nuclear-armed great powers does not produce winners. It produces civilizational extinction.
This is not a marginal adjustment. It is a qualitative transformation of the entire strategic landscape. The political scientist Kenneth Waltz, one of the founders of structural realism, argued that nuclear weapons produced a "nuclear revolution" in international politics — making great-power war not merely risky but irrational to a degree unprecedented in human history. If Waltz is right, then the 15 pre-nuclear cases in Allison's dataset have limited relevance to the one case that actually involves nuclear weapons: the US-Soviet Cold War. And that case ended peacefully.
Robert Jervis of Columbia University, one of the leading scholars of nuclear strategy, has argued that the Thucydides Trap framework fundamentally underestimates the transformative effect of nuclear weapons on great-power behavior. Jervis contends that nuclear deterrence does not merely add a new variable to the traditional dynamics of power transition — it overwhelms all other variables. The structural pressures identified by Allison are real, but they operate within a constraint so powerful that comparing nuclear-era rivalries to pre-nuclear ones is like comparing chess to Russian roulette: the rules have changed so fundamentally that the old patterns no longer apply.
However, defenders of the framework point to a crucial counterargument: nuclear weapons have not prevented great-power crises. The Cuban Missile Crisis, the Berlin crises, the Sino-Soviet border clashes of 1969, and numerous incidents in the South China Sea and Taiwan Strait demonstrate that structural pressures can still push nuclear-armed states toward confrontation. The question is not whether nuclear weapons prevent war in peacetime — they clearly do — but whether they can prevent war during a crisis when escalatory dynamics take on a momentum of their own. Allison argues that the answer is far less certain than nuclear optimists assume.
"Nuclear weapons have made great-power war unthinkable. But the history of the 20th century is largely a history of unthinkable things happening."
A subtler but analytically significant criticism of the Thucydides Trap concerns the definitional ambiguity at its core. The framework rests on the distinction between a "rising power" and a "ruling power" — but critics argue that these categories are far less clear-cut than Allison's framework implies.
Consider some of the complications. When exactly does a state qualify as "rising"? Is it when its GDP exceeds a certain ratio relative to the ruling power? When its military spending crosses a threshold? When it begins to challenge existing norms and institutions? When other states perceive it as threatening? Allison's framework does not provide precise criteria, and the absence of a clear definition makes it possible to classify ambiguous cases in ways that support the theory.
The "ruling power" concept is equally problematic. In the modern international system, who rules? The United States is the dominant military power, but China already has the world's largest economy by purchasing power parity. The EU is the world's largest single market. Does "ruling" refer to military dominance, economic supremacy, institutional leadership, or some combination? Different definitions produce different case lists and different outcomes.
The international relations scholar David Kang of the University of Southern California has pushed this critique furthest in the context of East Asian history. Kang argues that the Thucydides Trap framework is built on a distinctly European understanding of international relations — one in which power transitions are inherently competitive because the European state system was defined by sovereignty, anarchy, and balance-of-power logic. In East Asia, by contrast, the historical pattern was one of hierarchical order centered on China, in which rising and falling powers were accommodated through tribute, deference, and institutional integration rather than through war. If Kang is right, then applying a European-derived framework to the US-China relationship may be fundamentally misleading.
This critique does not necessarily invalidate the Thucydides Trap, but it does suggest that the framework's universalist claims — that it identifies a pattern applicable across all times and cultures — need significant qualification. The structural pressures of power transition may operate differently in different normative, institutional, and cultural contexts, and a framework that treats all contexts as equivalent risks missing the variables that actually determine whether a given power transition leads to war or peace.
Critics do not merely argue that the Thucydides Trap is flawed — they contend that better frameworks already exist for understanding great-power competition. Several alternative theoretical traditions offer explanations of power transitions that are, in the view of their proponents, more rigorous, more nuanced, or more empirically grounded than Allison's approach.
A.F.K. Organski developed Power Transition Theory (PTT) in the 1950s, decades before Allison coined the Thucydides Trap. PTT offers a more precise and testable version of the same basic insight: that wars are most likely when a rising power approaches parity with the dominant power in the international system. But PTT specifies the mechanism more precisely than Allison does. It focuses on GDP as the primary measure of national power, identifies a specific "zone of parity" in which the rising power's GDP reaches roughly 80% of the dominant power's, and argues that war probability peaks when the challenger is both nearing parity and dissatisfied with the existing international order.
PTT scholars like Ronald Tammen and Jacek Kugler have argued that their framework is superior to the Thucydides Trap because it provides clear, measurable criteria for prediction rather than relying on historical analogies. They have also shown that not all power transitions are dangerous — only those in which the rising power is dissatisfied with the status quo. A satisfied rising power (like the United States overtaking Britain in the late 19th century) can achieve parity without triggering war, precisely because it does not seek to overthrow the existing international order.
John Mearsheimer of the University of Chicago has developed an alternative framework — Offensive Realism — that shares some of Allison's pessimism about great-power competition but rejects the rising/ruling power distinction as the key variable. In Mearsheimer's view, all great powers seek regional hegemony as a matter of structural necessity, regardless of whether they are rising or ruling. The anarchic nature of the international system — the absence of any authority above the state — compels every major power to maximize its relative power, because only hegemony provides genuine security.
For Mearsheimer, the danger of US-China competition arises not from China's rise per se but from the structural imperatives of the international system itself. China would seek regional hegemony even if it were not "rising," and the United States would resist any peer competitor even if it were not "ruling." This framework generates different policy prescriptions than the Thucydides Trap: where Allison emphasizes accommodation and diplomatic creativity, Mearsheimer emphasizes containment and military preparation.
At the opposite end of the theoretical spectrum, liberal institutionalists like Robert Keohane and G. John Ikenberry argue that international institutions, trade interdependence, and democratic governance can mitigate or even eliminate the structural pressures that the Thucydides Trap identifies. In this view, the post-1945 international order — built on institutions like the United Nations, the World Trade Organization, and the network of alliances and agreements that structure global governance — represents a qualitative break from the anarchic, self-help system that produced the wars in Allison's case list.
Ikenberry, in particular, has argued that the liberal international order is uniquely capable of accommodating rising powers because its rules-based character gives even non-dominant states a stake in the system. China has benefited enormously from the existing order — its economic miracle was built on access to global markets, technology transfer, and the stability provided by American-led institutions. This gives China powerful incentives to work within the system rather than to overthrow it, even as its power grows.
Constructivist scholars like Alexander Wendt offer yet another critique. Constructivism holds that the international system is not defined by fixed structural forces but by the ideas, identities, and norms that states collectively construct. In Wendt's famous formulation, "anarchy is what states make of it." The implication for the Thucydides Trap is profound: the structural pressures of power transition are not objective facts but products of how states interpret each other's intentions. If the United States and China construct a shared understanding of their relationship as competitive but manageable — rather than as a zero-sum struggle for dominance — the structural pressures that Allison identifies may never generate the fear, honor, and interest dynamics that lead to war.
Each of these frameworks illuminates aspects of great-power competition that the Thucydides Trap, with its focus on a single structural variable, tends to obscure. The strongest version of the critique is not that Allison is wrong, but that he is incomplete — that the rising-versus-ruling power dynamic is one of many factors that determine whether great powers go to war, and that treating it as the primary variable overstates its explanatory power.
Graham Allison has not been silent in the face of criticism. Over the years since Destined for War was published, he has engaged directly with many of the objections raised above — in academic forums, op-eds, congressional testimony, and public lectures. His responses reveal both the framework's genuine strengths and the areas where the debate remains unresolved.
Allison argues that his 16 cases were selected through a transparent, replicable methodology developed at the Belfer Center with input from leading historians. He acknowledges that reasonable scholars can disagree about borderline cases but maintains that the core finding — that power transitions carry a high risk of war — holds under any reasonable selection criteria. He notes that even Steve Chan's more restrictive case list produces a war rate substantially above what a random or benign model would predict, confirming the basic insight even if the exact percentage is debated.
More fundamentally, Allison contends that the obsession with the precise war rate misses the point. The Thucydides Trap is not a statistical claim about probability; it is a structural claim about the dangerous dynamics that power transitions generate. Whether the war rate is 75% or 50% or 40%, the pattern is alarming enough to demand serious attention from policymakers. Debating whether the number is 12 out of 16 or 8 out of 20 does not change the policy imperative.
"When someone tells you that only 50% of power transitions end in catastrophic war instead of 75%, that is not reassurance. That is a fire alarm." — Graham Allison
Allison vigorously rejects the charge that his framework is deterministic. He points to the four peaceful outcomes in his dataset as proof that war is avoidable and argues that the entire purpose of the Thucydides Trap concept is to prevent war by making leaders aware of the structural forces pushing toward it. He compares the framework to a medical diagnosis: a doctor who tells a patient that their condition is dangerous is not condemning the patient to death but rather creating the urgency needed for treatment. The trap is a call to action, not a counsel of despair.
Allison also points out that the critics who accuse him of determinism are often the same ones who argue that the case selection is flawed — a logically inconsistent position. If the 75% war rate is overstated, then the framework is less deterministic than its critics claim, not more. You cannot simultaneously argue that the framework is too deterministic and that the war rate is lower than advertised.
Allison acknowledges that nuclear weapons have transformed great-power competition but argues that the transformation is less complete than nuclear optimists believe. He points to the Cuban Missile Crisis as evidence that nuclear-armed states can come terrifyingly close to war even when both sides understand the consequences. He emphasizes that the escalatory dynamics of the Thucydides Trap — fear, misperception, entangling alliances, catalytic third parties — do not disappear in the nuclear age; they merely operate within a higher-stakes context where mistakes are less recoverable.
Allison also challenges the argument that the US-Soviet Cold War proves that nuclear deterrence automatically prevents great-power war. He notes that the Cold War involved at least a dozen episodes where the world came perilously close to nuclear exchange, and that survival was often a matter of luck as much as strategy. The fact that we survived the Cold War without nuclear war does not mean that nuclear deterrence is foolproof — it means we were fortunate. Applying the same logic to the US-China rivalry, with its own set of flashpoints and its own potential for miscalculation, is not a basis for complacency.
Allison views alternative frameworks not as competitors but as complements to the Thucydides Trap. He has praised Power Transition Theory as a more formal version of the same basic insight, acknowledged the importance of institutions in mitigating conflict risk, and agreed that nuclear weapons add a crucial variable. His argument is not that the Thucydides Trap is the only framework needed to understand great-power competition, but that it identifies the single most important structural variable — the rapid shift in relative power — and that other frameworks gain explanatory power when used alongside it, not instead of it.
After examining the strongest criticisms of the Thucydides Trap and Allison's responses to them, what is the honest assessment? The answer, as with most serious intellectual debates, is that both sides have important points — and that the truth lies in a more nuanced understanding than either advocates alone.
The case selection critique is legitimate. Allison's 16 cases are not the only defensible list, and the precise 75% war rate is debatable. But even alternative analyses that use different case selection criteria find elevated war rates during power transitions. The specific number matters less than the pattern: rapid shifts in relative power create dangerous dynamics that increase the risk of conflict.
The determinism critique is important but partially misplaced. Allison has been clear — consistently and repeatedly — that war is not inevitable. The four peaceful outcomes in his dataset, and his extensive analysis of the conditions that produced them, demonstrate his commitment to the idea that leaders can escape the trap. The perception of determinism is a problem of reception, not intention, though Allison could do more to emphasize agency and choice in his public messaging.
The nuclear weapons critique is the most substantive challenge. Nuclear weapons have unquestionably transformed great-power competition in ways that limit the relevance of pre-nuclear historical analogies. But Allison is right that nuclear deterrence is not automatic, that crises can create escalatory dynamics even between nuclear-armed states, and that the US-China relationship involves flashpoints — particularly Taiwan — where the risk of inadvertent escalation is real.
"The Thucydides Trap is not a crystal ball. It is a smoke detector. It does not tell you that your house will burn down. It tells you that conditions are present that make fire dangerously likely — and that you had better take action."
The fairest assessment is that the Thucydides Trap is best understood as a heuristic — a useful simplification that captures an important truth about the structural pressures of power transition — rather than as a scientific law with precise predictive power. Like all heuristics, it illuminates some things while obscuring others. It is most valuable when used alongside other analytical frameworks, not as a substitute for them.
What the Thucydides Trap does brilliantly is focus attention on the structural dimension of great-power competition — the systemic pressures that operate independently of the intentions, wisdom, or goodwill of individual leaders. This is a genuinely important contribution. Policymakers tend to focus on personalities, events, and short-term crises; the Thucydides Trap forces them to look at the deeper structural forces that shape the strategic landscape. Even critics who reject Allison's specific methodology tend to acknowledge the value of this structural lens.
The framework's limitations are equally real. It oversimplifies complex historical events. It underestimates the transformative impact of nuclear weapons. It lacks precise definitions for its core concepts. And its dramatic framing — 12 out of 16 cases ended in war — risks communicating a degree of certainty that the evidence does not support.
The responsible approach is to take the Thucydides Trap seriously without taking it literally. The structural pressures it identifies are real. The historical pattern it reveals is genuinely alarming. And the policy implication — that leaders must take extraordinary measures to manage the US-China power transition — is sound. But the framework should be used as one tool among many, not as the master key to understanding the most complex geopolitical relationship in modern history.
The Thucydides Trap has not been definitively debunked, but it has faced serious scholarly criticism. Critics argue that Allison's 16 historical cases involve cherry-picking, that the framework is too deterministic, and that nuclear weapons have fundamentally changed great-power dynamics. However, even critics generally acknowledge that Allison's core insight — that rapid power shifts create dangerous structural pressures — has genuine analytical value. The framework is best understood as a useful heuristic rather than a predictive law.
The most common criticism is the case selection problem. Scholars like Arthur Waldron and Hal Brands argue that Allison cherry-picked his 16 historical cases to fit a predetermined narrative, omitting power transitions that did not lead to war and including cases where the rising-versus-ruling dynamic is debatable. Steve Chan of the University of Colorado has published the most systematic version of this critique, finding a substantially lower war rate when more rigorous case selection criteria are applied.
Many critics argue yes. The framing that 12 of 16 power transitions ended in war can imply that conflict is nearly inevitable, which risks becoming a self-fulfilling prophecy. If leaders believe war is structurally determined, they may invest less in diplomacy and more in military preparation. Allison responds that his framework is explicitly designed to warn against fatalism and that the four peaceful outcomes in his data prove war is avoidable — the framework is a diagnosis meant to spur preventive action, not a counsel of despair.
This is one of the strongest criticisms. Nuclear weapons create mutually assured destruction, fundamentally altering the cost-benefit calculus that drove historical power transitions to war. The US-Soviet Cold War — the only nuclear-era case in Allison's dataset — ended peacefully. However, Allison counters that nuclear weapons did not prevent terrifyingly close calls like the Cuban Missile Crisis, and that the escalatory dynamics of power transition can still produce crises where miscalculation leads to catastrophe even between nuclear-armed states.
Several alternative frameworks address great-power competition. Power Transition Theory (Organski) focuses on GDP parity as a measurable war predictor. Offensive Realism (Mearsheimer) argues all great powers seek hegemony regardless of rising/ruling status. Liberal Institutionalism (Keohane, Ikenberry) contends that international institutions and trade interdependence can prevent war during power shifts. Constructivism (Wendt) holds that threat perceptions are socially constructed rather than structurally determined. Most scholars recommend using these frameworks alongside the Thucydides Trap rather than relying on any single theory.