In his presentation at Yale University in the spring of 2013, Henry Kissinger admitted that he preferred reading history to political science. The lack of enthusiasm for political science, by a former political scientist and former policy-maker, now determined to build a nuclear-free world, should be concerning for political scientists interested in nuclear studies.
Early on, nuclear studies played an important role in the development of international relations and the social sciences more generally. Efforts to analyze potential nuclear crises with the Soviet Union generated important insights for the study of international security, game theory, and political economy.
More recently, the field of nuclear studies has rather been the recipient of insights and methodologies from other branches of social science. The current H-Diplo roundtable discusses two quantitative papers in nuclear studies. I urge anyone interested in understanding international relations to join this fascinating conversation, as it touches on important questions of methodology, standards of evidence, and the promises and challenges of multi-method approaches.
In my view, the new wave of quantitative studies has not lived up to its promise because it has focused too much on questions of methodology. In order to move forward as a field, nuclear studies should pay greater attention to theory, history … and politics.
The State of the Art
Until 10 years ago, qualitative methods were the preferred approach in nuclear studies. While such an approach was very fruitful, it risked generating a proliferation of theories, each built upon a small number of cases, with no unifying framework. The quantitative studies of the last decade can generate useful insights in documenting general correlates of nuclear outcomes, yet it is also important to think about the next steps.
Many of the contributors of the roundtable acknowledge that correlation is not causation, but even establishing causation is insufficient. It is not enough to know whether an independent variable causes a particular outcome, it is also important to understand how it does so. The independent variables in quantitative studies are many steps removed from the outcome of interest. They affect the strategic incentives of multiple actors, sometimes in competing directions. To understand nuclear politics, we need solid theories explaining the political mechanisms at play. In my work with my colleague Nuno Monteiro, I have tackled such issues in the context of proliferation (see here and here). Let me now comment on the two studies of nuclear crises being discussed in the roundtable.
In his article, Matthew Kroenig studies the effect of nuclear superiority on crisis outcomes, concluding that superiority correlates with success. In the roundtable, Kroenig laments the empirical nature of Gavin’s critique when, in his view, his main contribution is theoretical. Now if we look at his theory, Kroenig assumes that a state fares better in a nuclear war if it possesses a superior nuclear arsenal, and concludes that states with a superior nuclear arsenal can coerce other states into making concessions.
This conclusion is sensible given the assumption. Yet this assumption is unacceptable if one believes in a nuclear revolution, where the number of nuclear weapons does not affect war outcomes under conditions of mutually assured destruction. The empirical tenor of the H-Diplo debate is symptomatic of the fact that Kroenig did not convincingly explain why nuclear superiority matters. Perhaps important pieces of conventional wisdom were derived from prominent crises between the US and the USSR, which represented an anomalous strategic setting. Additional efforts in building a theory of escalation can advance old debates about the effect of nuclear superiority.
The claim presented by Matthew Fuhrmann and Todd Sechser is less controversial. Fuhrmann and Sechser establish that the possession of nuclear weapons does not improve the success rate of compellent threats, which conforms with a general expectation that compellence is ‘difficult.’ Yet critics are nevertheless reluctant to accept the paper’s conclusions. In my view, this is because there isn’t sufficient attention to theory.
To critics who point out that strategic selection issues are overwhelming, Fuhrmann and Sechser respond that their results are robust to concerns for strategic selection. They show how crisis outcomes correlate with the ‘stakes’ of a crisis and in their Online Appendix, they estimate a selection model, assuming that a dummy variable for the post-1945 period satisfies the exclusion restriction, i.e. it affects the likelihood of making a threat but not the success of the threat.
In my view, additional work should be done to understand the strategic foundations of compellent threats in the nuclear age. Given the nature of their contribution, which is to code compellent threat and see how they affect nuclear crises, understanding when and why states choose to make threats is not a robustness check, but the heart of the political matter. How do leaders choose to make compellent threats, instead of making deterrent threats or no threat at all? When did leaders make nuclear threats and why did they think that mentioning nuclear weapons would affect the outcome of a crisis? When and why were they mistaken? To answer these questions, we need a greater attention to the theory, history and politics of nuclear coercion.
Quantitative studies may very well make a lasting contribution to nuclear studies. Fifteen years ago, the field of democratization studies was in a similar situation, favoring qualitative approaches and area studies. In her brilliant piece, Barbara Geddes concluded that the field had generated few generalizable insights, and suggested that the field should organize the evidence not by geographic areas but by non-democratic regime types. Her dataset was used in numerous studies, including in international relations. Yet the lasting legacy of her work was the fact that, building on the work of Samuel Huntington, she introduced a new way of thinking about the politics of democratization. She suggested different causal mechanisms, which could later be tested and refined, and significantly improve our understanding of democratization.
The H-Diplo roundtable suggests that quantitative studies have not convinced scholars to think differently about nuclear politics. Nevertheless, multidisciplinary and multi-method discussions are crucial in producing a healthy dialogue between theory and empirics. Greater attention to theory can help the field of nuclear studies become, once again, a pioneer in international relations and, hopefully, a better guide for policy recommendations.