The "Thucydides Trap" Does Not Explain Geopolitics

On social science theories as propaganda, constantly shifting to justify militarism and large Pentagon budgets

When I discuss the US-China relationship with people, I’ve found that they often turn to the concept of the “Thucydides Trap.” It seems as if few international relations books in the last decade have had as much influence as Destined for War: Can America and China Escape Thucydides's Trap? by Graham Allison, a former Assistant Secretary of Defense under Clinton and a professor at the Harvard Kennedy School. Here is the number of Google Scholar mentions I find for “Thucydides(‘s) Trap” and “China” in the same work from 1990 to 2019.

It’s a deeply unimpressive book though. Take a look at the blurbs on Amazon, and then actually read it if you want to understand the hollowness of the kinds of arguments that are used as justification for the American global empire. Allison receives praise from Kissinger, Biden, Petraeus, Michael Hayden, Ban-Ki Moon, Samantha Power, and even Klaus Schwab, who I once thought was a Twitter meme but is apparently also a real person (in their defense, I’m sure almost none of them read it).

Just some of Graham Allison’s accolades.

I explain why the Thucydides Trap fails as a lens through which to understand the US-China relationship in an article for Strategic Studies Quarterly, a journal for the US Air Force (non-paywalled). One thing that is particularly frustrating about Allison’s book is how the author manages to both make sweeping recommendations based on the reader taking his arguments seriously, while granting that his arguments really don’t justify any policy recommendations, or even provide much evidence for anything at all. As I write,

Allison gathered 16 cases over the last 500 years where a rising power challenged an established power, from the rivalry between Portugal and Spain in the late fifteenth century to the end of the twentieth century. Of these, 12 cases ended up in war. A naïve analysis therefore suggests that, if history is any guide, there is around a 75 percent chance that the United States and China will go to war in the coming decades…

As mentioned above, Allison’s work and his attempts to draw conclusions about the future course of US-China relations contain three interrelated problems, namely, unclear definitions, omitted variable bias, and selection bias. It is of note that Allison himself acknowledges his work might not withstand statistical scrutiny. His appendix 2 is titled “Seven Straw Men.” The fifth of these states, “the Thucydides’s Trap Case File offers too small a data set to support claims about laws or regularities, or for use by social scientists seeking to do so.” Allison responds, “Agreed. The purpose of this inquiry is to explore a phenomenon—not to propose iron laws or create a data set for statisticians.”

This admission is remarkable. In responding to this “straw man,” Allison creates his own, saying that he is not arguing for “iron laws.” Yet the sophisticated critique is not that 16 cases spread out over 5 centuries in a bivariate analysis does not lead to “iron laws.” Rather, it is that such an analysis provides no guidance to understanding US-China relations, a point Allison seems to agree with. His statement that he is not seeking to “create a dataset for statisticians” implies there is one standard that those who engage in quantitative analysis should apply to judging a work and another for everyone else. Presenting numbers on a phenomenon and then saying it cannot meet the standards of statisticians is like presenting an argument about genetics and saying that it cannot be judged by the standards of biology.

Moreover, this seeming humility contradicts not only how others have used Allison’s work but how he himself has promoted it. For example, in the Atlantic, after summarizing his findings, Allison writes:

Based on the current trajectory, war between the United States and China in the decades ahead is not just possible, but much more likely than recognized at the moment. Indeed, judging by the historical record, war is more likely than not. Moreover, current underestimations and misapprehensions of the hazards inherent in the U.S.-China relation- ship contribute greatly to those hazards. A risk associated with Thucydides’s Trap is that business as usual—not just an unexpected, extraordinary event—can trigger large-scale conflict.

Allison made similar points in a 2017 Foreign Policy essay. His book recommends the White House establish a Council of Historical Advisers, a group that would be analogous to the Council of Economic Advisers, and look at the past to draw lessons about the present. In April 2017, he went to the White House and briefed a group of National Security Council staffers on the Thucydides Trap.

I conclude that there are better ways to understand the US-China relationship.

A narrower historical focus on international politics since the second half of the twentieth century provides a more optimistic lens for under- standing the future of great power relations. As measured by GDP, a handful of power shifts have occurred over the last several decades. Among these are China relative to Japan and Russia, and Germany relative to other European nations. None of these cases has led to war. The disappearance of interstate conflict more generally offers hope that even if power transitions may have created a substantial risk of war in the past, they do not do so today.

In the end, however, history may be of limited utility in understanding the US-China relationship. Instead of employing historical analogies that may or may not apply or using datasets that cannot meet basic standards for establishing causal inference or reasonably predict behavior, American foreign policy should proceed by considering the interests, politics, and material capabilities of both sides. Questions such as what does China want, can the United States live with its claims, and what is worth going to war over should be at the forefront of the minds of American leaders.

The Thucydides Trap begins by assuming the two superpowers are engaged in a rivalry, all but foreclosing a more restrained American foreign policy by presenting such a view as hopelessly naïve. When it comes to power transitions, it is not enough simply to say that studying previous centuries reveals no iron laws. Rather, scholars have yet to show that conclusions about the likely course of future events derived from the distant past can withstand basic scrutiny.

It’s looking more and more like “science” as we understand it works in Western democracies the same way that a rubber stamp legislature works in an authoritarian regime. The scholarship is produced as a formality to ratify and give credibility to decisions that have already been made. This is particularly bad in the field of international relations, a strange discipline where highly cited scholars are often former government officials, like Allison himself. Recent examples of politics driving the popularity of social scientific theories are:

1. The popularity of the “end of history” narrative in the 1990s to justify democratic interventions and economic globalization (the latter was actually good).

2. The idea that terrorism is caused by a lack of democracy (which was called “neo-conservatism” to give it an intellectual pedigree but bore little resemblance to what that term once meant), used to justify the Iraq War after the failure to find WMDs in 2003.

3. Counterinsurgency theory (COIN) in the early Obama administration, promoted by generals and national security types to sell a skeptical president on the merits of continuing the war in Afghanistan.

See my forthcoming book for more elaboration on each of these examples. We see a constant pattern in which the justifications for large military budgets and an interventionist posture abroad change as circumstances do. The US formed NATO and went to Europe to deter the Soviet Union from invading Germany. When the enemy collapsed, we started hearing about democratization as a way to achieve permanent peace, along with concepts like “rogue states” and the “responsibility to protect.” Luckily for the national security state, 9/11 came along, and the actions of a few dozen men led to the US bombing, invading, or trying to overthrow more than half a dozen Muslim countries over the next two decades on the theory that anyone who shared the same faith anywhere on earth could be a potential mass murderer.

This couldn’t continue forever, so we’re back to “great power competition.” In that context, the Thucydides Trap has proved a useful concept for national security elites, as it makes what they wanted to do in the first place seem like a responsible way to prepare for what might be a historical inevitability.

Unsurprisingly, official Washington never latches on to a theory that says the US is safer from foreign threats than any other country in the history of the world, and our meddling abroad causes more harm than good. While such a theory isn’t useful to powerful interest groups, it happens to be true.

Read the whole thing here.