Commitment Problems and Preventive War

By Matthew Adam Kocher

Among the few very serious threats to interstate peace in the world today is the possibility of preventive war, in which one state attacks a second state because it fears the second state will become much more powerful in the future. The 2003 U.S. invasion of Iraq was a preventive war, aimed at stopping Saddam Hussein’s acquisition of nuclear weapons. A war launched by the U.S. to prevent the Iranian regime from gaining nuclear weapons, which many policy intellectuals in Washington currently advocate, would also be a preventive war. A less immediately pressing, but more serious, possibility is that the logic of preventive war could bring the U.S. and China into armed conflict as the size of the Chinese economy surpasses that of the U.S. in the near future.

For most of the past two decades, political scientists have analyzed preventive war as a species of “commitment problem.” In game theory, a commitment problem arises when two actors would be better off in the present by committing themselves to a cooperative relationship in the future. But, if the actors know that they will prefer to renege on their agreement in the future, the benefits of cooperation in the present cannot be realized, and even a mutually beneficial agreement cannot be struck.

In a preventive war scenario, the rising state (the one that is becoming more powerful) would like to guarantee that it would not use its powerful position to exploit the declining state in the future. The declining state would like to accept such a guarantee. Both states would prefer such a guarantee to risky and costly fighting. Yet both states know that the guarantee would be worthless once the rising state achieves a dominant position. Hence, the declining state may launch a war now in order to avoid being exploited in the future.

My Yale colleagues Alex Debs and Nuno Monteiro have a very interesting paper on preventive war that just appeared in the journal International Organization. The paper makes two novel points. The first is that the kinds of power shifts in international politics that are large and rapid enough to create a serious danger of war are almost always “endogenous”; that is, they are almost always the product of a state’s decision to rapidly build new military capabilities, rather than being a function of uneven rates of economic growth. Consider, for example, the rapid rise of the Japanese and German economies relative to that of the U.S. during the 1980s. Both countries decided not to build powerful militaries or go nuclear. The exogenous power shifts implied by their high growth rates were mild and gradual. The U.S. never considered them security threats. Iran, on the other hand, is not rising economically relative to the U.S. and is if anything declining. Yet a decision to go nuclear (an endogenous power shift) would radically transform its relationship with the U.S., rendering it secure from an American invasion for the first time.

The second novel contribution of the article stems directly from the first. If power shifts significant enough to create a danger of war are the result of state decisions, then we need to analyze those decisions and understand why states would go for big power shifts sometimes and not others. Debs and Monteiro endogenize this decision by thinking through the possibility of deterrence. They observe that military buildups involve a delayed return on investment. For example, once a state decides to go nuclear, it can take many years to build the material and intellectual infrastructure to deploy actual weapons. If more powerful states are fully aware of the investment, they are in a position to strike before the new weapons systems are deployed. Consequently, they are also in a position to deter the investments in the first place. This is a key insight: if military investments are fully observable, the commitment problem described above can never come into play, because the power shift that gives rise to it will be deterred in advance. In other words, the commitment problem that is typically thought to cause war – the inability to commit not to exploit power – actually never causes war when power comes from (fully observable) military investments with a delayed return.

Debs and Monteiro show that war requires uncertainty about power shifts. States that decide to attempt to radically alter the balance of power with other states will want to keep their investments secret or ambiguous. For example, Iran steadfastly proclaims that its nuclear program is for peaceful purposes, which is exactly what they would say if their program were intended to shift the balance of power with the U.S. Uncertainty creates the possibility of preventive wars both when the rising state is trying to shift the balance of power and when it is not. Debs and Monteiro devote a great deal of their article to showing how the 2003 Iraq War can be thought of as an instance of a preventive war created by the uncertainty over whether or not Saddam Hussein really had a nuclear program. We now know that he did not, but the Bush Administration believed he did.

For IR theory, it is fairly big news that the classic commitment problem cannot explain preventive war. Does this paper also offer something for policymakers? One area in which it seems to me to have an immediate payoff is in allaying fears about a U.S. preventive war against China. Nuclear weapons acquisition is one of the easiest ways for a state to generate a rapid and massive power shift, but a state can only cross that threshold once; China got there in 1964. China arguably has had the capacity to beat the U.S. in an Asian land war already, so there is no obvious prospect for a rapid power shift there. To my mind, that leaves one key scenario: a Chinese attempt to build a true blue-water navy that can project its power globally. If Debs and Monteiro are right, however, the U.S. will be in a good position to deter such a move for the foreseeable future. A shift of this magnitude in naval power should be pretty easily observable, and the investments needed to produce it would pay off over a decades-long time frame.

Note: Debs and Monteiro’s article has been extensively discussed recently at Duck of Minerva.


  • There are other factors to consider. For example, the type of state.

    The United States is far less disturbed by the idea of a powerful E.U. even though it would (if politically unified) also be nuclear and have a modern army and navy. The difference there is that the E.U. would be democratic. In addition during the 1980s West Germany was democratic and Japan at least had democratic trappings. In contrast Iraq and China could not (and really still cannot) claim any democratic credibility and in Iran it’s been made very clear that the winner of the vote only wins if the ayatollah decides that he should win.

    After that is the issue of bias. More than a few high ranking government officials reacting to 9/11 were hardline opponents of Saddam Hussein who had urged Clinton to attack Iraq in the late 1990s. In other words, they were predisposed to believe that Iraq was attempting to acquire nuclear weapons even though it was not.

  • “Debs and Monteiro show that war requires uncertainty about power shifts.”

    “Debs and Monteiro devote a great deal of their article to showing how the 2003 Iraq War can be thought of as an instance of a preventive war created by the uncertainty over whether or not Saddam Hussein really had a nuclear program. We now know that he did not, but the Bush Administration believed he did.”

    But there was no uncertainty about Saddam Hussein having WMD. This is well-documented:




    Perhaps they should amend their theory to state that it is the appearance of uncertainty that is required for war. The reality is that the Iraq War was a war of aggression, not a preventative war.

  • Julie, it is interesting that you use similar language to make precisely the opposite point from David Lake (see the link above to his post on Duck of Minerva): ” But where was the uncertainty about Iraq in 2002-2003? Even if the Bush administration was “fixing” the intelligence around the policy, it remains true that nearly everyone and certainly everyone in the administration was absolutely convinced that Iraq was developing WMD.” I’m no expert on the case, but it seems to me that the guy mentioned in your links could be misrepresenting what happened. Alternatively, the Bush people might have worried that the source inside the Hussein regime was a double agent. Finally, they might have weighed what he said against other evidence and assumptions that they found more compelling. It’s interesting evidence, but hardly dispositive.

Leave a Comment