Jeffrey LewisRebel With A Bomb

One of the major developments in the study of nuclear weapons has been a shift from pseudo-rationalist explanations to empirical research based on the decision-making style and outlook of actual  leaders.  A few years ago, Bill Potter and Gaukhar Mukhatzhanova wrote a nice review essay in International Security called Divining Nuclear Intentions that surveyed this shift, highlighting the work of Etel Solingen and Jacques E.C. Hymans.

Matt Fuhrmann and Mike Horowitz have a new paper that tests one of those hypotheses and reaches an interesting conclusion — that former rebel leaders are more likely to want the bomb.  The findings were published in the Journal of Politics, but Matt and Mike were nice enough to send along bloggy summary to tempt you into reading the longer paper.

Screen Shot 2015-11-25 at 1.40.57 PM

The Unexpected Factor That Predicts Nuclear Proliferation

Matthew Fuhrmann and Michael C. Horowitz

International concerns about the spread of nuclear weapons have seemingly subsided in the wake of the Joint Comprehensive Plan of Action between Iran, the P5+1 nations, and the EU. The world remains a dangerous place, however, in light of uncertainty over Iran’s intentions, Vladimir Putin’s moves to reassert Russian influence in Eastern Europe and elsewhere, and North Korea’s nuclear-backed threats in East Asia. These (and other) concerns have caused speculation that several countries – including Saudi Arabia, South Korea, and Ukraine – might have renewed interest in acquiring a nuclear arsenal.

How likely is it that additional countries will seek nuclear weapons in the coming years? What factors might determine whether another country joins the nuclear club?

Scholars and policymakers have long grappled with these questions. Many factors are thought to influence nuclear proliferation, including a state’s threat environment, alliances with superpowers, integration in the global economy, nonproliferation norms and treaty commitments, and technological development. These considerations are undoubtedly important for understanding the proliferation process.

Yet many analysts overlook or downplay a key fact: it is individual leaders – not countries – that make decisions on nuclear policy. And who a leader is affects the likelihood of nuclear proliferation, in addition to what country he or she represents. Many observers accept the notion, for example, that Iraq’s nuclear policy in the 1980s might have been different if someone other than Saddam Hussein had been in power. But why, exactly, is this the case?

We recently concluded a project designed to address this question that builds on new research on leaders in international relations. Our analysis, published in a recent issue of the Journal of Politics, shows that a leader’s background life experiences – what he or she did prior to entering office – sheds new light on the causes of nuclear proliferation. One experience is especially salient: participation in a rebellion against the government.

We compared how likely former rebels are to seek nuclear arsenals relative to their non-rebel counterparts. The results from our statistical model were striking: former rebels are five times more likely to pursue nuclear weapons than similar leaders who lack rebel experience.

A look at particular cases illuminates this result. A seemingly diverse set of leaders that sought nuclear weapons had one thing in common: they were all former rebels. This includes Mao Zedong (China), Ali Khamenei (Iran), Saddam Hussein (Iraq), Muammar Qaddafi (Libya), Muhammad Zia (Pakistan), Charles de Gaulle (France), and Josip Broz Tito (Yugoslavia), as well as others.

Former rebels are particular motivated to seek nuclear weapons for two main reasons. First, rebels are especially likely to place a premium on national independence and sovereignty, because they have personal experience engaging in a struggle for control of the government. This makes rebel leaders more likely to value military technologies that serve as a form of invasion insurance, like nuclear weapons. Moreover, rebels are less likely to trust promises of protection, or “nuclear umbrellas,” offered by other nuclear weapons states.

Second, rebels have a higher tolerance for risk taking. Individuals who choose to join a rebellion usually put their own lives on the line, and rebellions against the state are an inherently risk endeavor. Rebels that ultimately enter office have succeeded while taking risks, reinforcing the view that gambling pays off. Thus, former rebels are more likely to make risky choices in office, such as seeking nuclear weapons – a decision that subjects a state to significant political and economic costs including, potentially, preventive military strikes.

Our study sheds light on a policy decision that has attracted significant media attention following Russia’s annexation of Crimea: Ukraine’s decision to return nuclear weapons it inherited from the Soviet Union in the 1990s. In 1994 (the year of the Budapest Memorandum), according to our statistical model, Ukraine had probability of pursuing nuclear weapons of about 9 percent. At that time, Leonid Kuchma, who lacked rebel experience, led Ukraine.

Imagine, for the sake of illustration, that a former rebel had instead been president of Ukraine in 1994. Changing just this one thing increases the probability of Ukraine seeking a nuclear bomb to 38 percent. Substituting a rebel for a non-rebel, therefore, leads to a four-fold increase in the predicted probability of nuclear proliferation, based on our model. Kiev may still have returned the nuclear weapons to Russia in this hypothetical world, but the likelihood of it doing so would probably have been much lower had a former rebel been calling the shots.

The core insight from our study – that former rebels are particularly proliferation-prone – may be useful to policymakers interested in anticipating and responding to future nonproliferation challenges. Rebel experience is by no means the only variable that might help us predict whether countries will “go nuclear” in the future, but it is one key factor that policymakers should consider. When a former rebel rules a state, officials should be especially worried about the risk of nuclear proliferation.

Matthew Fuhrmann (twitter: @mcfuhrmann) is an associate professor of political science at Texas A&M University. He is the author of Atomic Assistance: How ‘Atoms for Peace’ Programs Cause Nuclear Insecurity and coauthor of the forthcoming book Nuclear Weapons and Coercive Diplomacy.

Michael C. Horowitz (twitter: @mchorowitz) is an associate professor of political science and the associate director of Perry World House at the University of Pennsylvania. He is the coauthor of a new book on leaders and international politics, Why Leaders Fight, and the author of The Diffusion of Military Power: Causes and Consequences for International Politics.


  1. Scott Monje (History)

    It seems that striving for a nuclear weapon qualifies as both security-seeking behavior and risk-taking behavior. That may be unavoidable, but it could invite conceptual confusion.

  2. Brian Dewey (History)

    Moral of the story…. watch out for the “loose cannons” out there…

  3. Greg Thielmann (History)

    Of the seven leaders seeking nuclear weapons, it is convenient for the thesis to list Khamenei. A more interesting question would be why the original Supreme Leader Khomenei, equally a rebel, was opposed to pursuing nuclear weapons.

  4. Jonah Speaks (History)

    I see two issues: 1) Is there a statistically significant impact from prior rebel experience to leaders seeking a bomb? Testing this hypothesis requires a correct specification of the regression equations. In my view, the relevant set of observations is not leader-years, but leaders. There needs to be only one set of observations per leader; otherwise we are observing the same leader multiple times and deriving incorrect standard errors and statistical tests. Leaders who initiate bomb programs do so during the first or second year in office (p. 25 & fn. 29). Also, only the initiation of a nuclear weapons program, not its continuation under a new leader, should be counted. Hence, only a combination of equation (7) “dependent variable: initiation” and equation (8) “one observation per leader”, and variations on these to check for robustness should be considered as appropriate tests of the hypothesis.

    2) How are rebel leaders different (on average) from non-rebel leaders, such that rebel leaders are more likely to seek a bomb? As Scott Monje points out above, it is not clear whether a nuclear weapons program should be regarded as a risky project or a prudent security measure. An alternative hypothesis would be that rebel leaders are more likely to feel that “power flows from the barrel of gun” and are thus less inclined to see nuclear weapons as “paper tigers.” In other words, rebel leaders are more likely to be militaristic hawks. One could test this alternate hypothesis perhaps by looking more carefully at leaders’ biographies, by checking whether military budgets increase more under rebel leaders, or whether rebel leaders are more belligerent to other countries.

    Miscellaneous: Table 1 does not list Zia of Pakistan as a rebel leader; the paper (p. 15) explicitly states that Bhutto (a non-rebel leader) initiated Pakistan’s nuclear program. The equations should include total GDP (not just GDP per capita) as an explanatory variable. The online appendix is repeatedly referenced, but no link is provided.

  5. Grant Christopher (History)

    I agree that there is a benefit to examining the hypothesis ‘rebel leaders tend to pursue nuclear weapons more than non-rebel leaders’ but are the methods in the (long) paper the right way? Even with when doing analysis in the ‘hard sciences’ multivariate techniques can obscure what is really going on. Putting numbers at the end of analyses where the uncertainties are not quantified is also confusing.
    I am an international relations and social sciences novice, but I find ‘thick description’ to be more helpful when looking state behaviour such as intention. We have a handful of countries that actually pursued and produced nuclear weapons. The experiences, priorities and intentions of (likely) decision makers should of course be an important element in examining a state nuclear programme. The kind of analysis in ‘The British Nuclear Experience’ by Baylis and Stoddart which takes into account the positions of leaders at both the cabinet level and in the military as well as the scientists themselves, and also examines the UK’s foreign relations etc. etc. and is my preferred method of understanding why states really do one thing or another.
    The difficultly in the kind of analysis in the paper is that of all social sciences (and astronomy) that we only have one experiment and cannot rerun. Generating fake data to test a model is not easy either. The authors were very careful in their analysis to be rigorous and try to isolate variables and I find the tested hypothesis helpful.