Editor’s Note: This is a guest post by Alexandre Debs and Nuno P. Monteiro, both of Yale University. In it, they discuss the causes of the Iraq War, a subject of some recent discussion at The Duck of Minerva. This (surprise) third installment responds to David Lake’s post, which itself was an engagement with Debs’ and Monteiro’s article–and its summary post at The Duck of Minerva.
We thank David Lake for writing a thoughtful response, Daniel Nexon for offering a platform to discuss this important issue, and readers of The Duck of Minerva and The Monkey Cage for engaging our argument.
As Lake mentions in his response, we share many views. Here, we’ll just focus on our differences, which also seem to underlie several reactions by readers in the comments to our initial post. We would also like to offer some points of clarification. We’ll center on three topics, from the most empirical to the most theoretical: how much of the Iraq War our theory explains; our contribution to the ‘rationalist’ framework; and the status of ‘rationality’ in IR theory more generally. Let us address these in turn.
The causes of the Iraq War
The first point on which we’d like to elaborate is to clarify what our theory does and does not do about Iraq. Is it a complete account of the run-up to the Iraq War? Of course not. In claiming that the Iraq war can be explained within the rationalist framework (i.e., without requiring that actors act in non-rational ways), we do not claim to capture all the features of the case. No theory — no useful theory — can provide a complete explanation of a phenomenon as complex as a war. Theories are useful when they highlight important aspects of a certain phenomenon, shedding light on dynamics that were previously in the dark and allowing for comparisons between different cases.
Like all social scientists, we constantly have to decide the proper balance between close description of a case and applicability to other (“out-of-sample”) cases. There is no magical solution to this, and reasonable people can reasonably disagree. We strove to find what was, in our view, the minimally sufficient description of the case that had the potential to generate generalizable claims about the causes of war.
Our theoretical view is that, first, as states become less certain of detecting other states’ militarization attempts, war becomes more likely; and, second, for any given level of uncertainty about this, as the cost of a preventive war lowers relative to the shift in the balance of power it is meant to avoid, war becomes more likely.
The U.S. preponderance of power that characterizes the post-Cold War lowered the cost of U.S.-led preventive wars
The U.S. preponderance of power that characterizes the post-Cold War lowered the cost of U.S.-led preventive wars. Washington can now act against most (non-nuclear) adversaries without facing the danger of possible escalation against another great power. At the same time, U.S. power preponderance increased the potential magnitude of the shift in the balance of power produced by WMD-acquisition by a non U.S.-aligned state. Then, the 9/11 attacks also had an impact on both key issues above. They lowered U.S. confidence in detecting potentially catastrophic developments on time. Plus they worsened the perceived consequences of WMD acquisition by states with which Washington enjoys poor relations.
Combining the effects of U.S. power preponderance with 9/11, then, our theory highlights why the likelihood of a preventive war against Iraq went up after 2001. Note that theories rarely make “point” predictions (i.e., predictions about specific events) so we’re not claiming that our theory retro-predicts the Iraq War. Rather, we claim that the logic above highlights the reasons why a war was more likely after 9/11 and against Iraq instead of, say, North Korea or Iran, who would be able to impose heavier costs on the United States and its allies.
Our focus on these dynamics leaves plenty of room for other views to complement our account of the war. Lake’s view of the role of misperception and multiple audiences in the run-up to the war, for example, sheds light on important aspects at play in the run-up to the Iraq War.
Our point is that given the dynamics highlighted by our theory, these misperceptions and multiple audiences were not necessary for the war. Therefore they are not really “causes” of the war. Put differently, if our aim is to describe what happened in this important historical case (and there’s nothing wrong with that), then misperception and multiple audiences must be an important part of the story. If, however, our aim is to understand what caused the war — why against Iraq? why not earlier? why did such a vast swath of the U.S. political system support the war? — then the key issue is, as we show, the lack of certainty about the absence of Iraqi WMDs conjoined with the low expected cost of a war when compared with the potential consequences of a WMD-armed Iraq.
Still, in our view, there is no reason to deny the promise of non-rational approaches to this or any other cases of war. We agree with Lake that a behavioral revolution might bear fruitful insights. To do so, however, we urge scholars who favor such an approach to use irrationality as a starting point, not a concluding remark. What biases stemming from irrationality (that are not already captured by extant work on misperceptions and bureaucratic decisionmaking) should IR theory incorporate? What is the universe of cases that such a theoretical move could better explain? Saying that war happened because President Bush chose to invade Iraq is a truism that applies to this case but has no insight to offer on other cases.
Does this mean the Bush Administration suffered from no biases or misperceptions? Of course not. We believe the Bush Administration’s assessment of the situation in the run-up to the war was based on an exceptionally high standard of evidence of the absence of WMD in order to rest its case against Iraq. Yet, support for the war in the months preceding it was vast. In fact, ten years later we still debate whether a hypothetical Gore Administration (the most relevant counterfactual in terms of U.S. decisionmaking) would have gone to war against Iraq. (Some say yes; others say no.) This indicates that the Bush Administration, much as its views may have been on one end of the risk-aversion spectrum, was certainly not alone in its analysis. Our theory provides an explanation for why this was so. Given the potential consequences of undetected Iraqi nuclearization and the inability to prove it would not happen, the war was rationalizable by most relevant actors.
At the same time, we believe the Bush Administration’s misperception about the existence of an Iraqi WMD program is often itself misunderstood. For example, Lake writes in his post:
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. Given this belief, we are back in a world of ‘full information’ -– even if that information was wrong. What mattered in the final decision was not the facts but the belief that Iraq was engaged in weapons development.
Not so, in our view. U.S. officials were not certain that Hussein was militarizing. The problem was that they were also not certain that Hussein was not militarizing. They could not rule out this possibility nor could they be certain that they would catch future militarization attempts in time — i.e., before Iraq succeeded in obtaining WMD. Fearing the consequences and the inability to prove a negative, they believed (wrongly, as it turned out) it was time for urgent action.
Certainly, the U.S. administration went to great lengths to “sell” the war and pressed on the urgency of action. Critics who “see through” these attempts may be frustrated, and with the benefit of hindsight discredit the intentions of U.S. policymakers. Furthermore, the extent to which WMD served as the real motivation or merely as a useful political line can remain a matter of debate. Nevertheless, it remains possible to account for the behavior of the key U.S. actors within a rationalist framework. 9/11 lowered U.S. tolerance to the risk if WMD acquisition by adversary states. The balance of power made war more likely against Iraq than against other likely but more powerful targets, such as Iran or North Korea. Under these circumstances, and given the cost of inaction, a rational decisionmaker could doubt the benefit of waiting for additional information.
More generally, the Iraq War case is one in which the overwhelming preponderance of analyses, both scholarly and in the popular discourse, focuses on psychological, irrational, even conspiratorial explanations that emphasize the Bush administration’s attempts to manipulate the available evidence to further the case for war. Faced with this situation in which all arguments are on one side, we think there’s much value to Burke’s logic of shifting the small weight of one’s reasons to the other side — in this case, back to rationalist explanations — in order to gain a better grasp of the case and, equally important, to be able to extract implications from this case to others.
Finally, and this is possibly the thorniest question anyone who tries to explain the Iraq War has to address: Does our theory justify the war? More sharply, are we apologists of the invasion of Iraq in 2003? Certainly not. Qua social scientists, our job is to figure out the dynamics at play in making (this) war possible. Doing so does not mean that, qua citizens, we think the war was a good idea. In fact, one could make the opposite argument with similar ease. By highlighting the mechanisms that, in our view, caused the war, we hope that our work will contribute towards finding better solutions to future cases of a similar kind. Specifically, our theory highlights how the conditions that made war an attractive solution in the Iraq case (U.S. power preponderance and a perception that Iraqi nuclearization would have dire consequences) also made the war more likely to be mistaken. An optimist might say that this could contribute to dissuading policymakers from launching a war in future cases of a similar kind.
Power shifts and the causes of war
The rationalist framework is a canon of scholarship attempting to understand the causes of war based on the strategic interaction of rational decision-makers, where “rational” means that decisionmakers have well defined preferences and choose the best action available given the information they possess.
The key insight of the rationalist framework acknowledges that war is costly and destructive, creating a puzzle: why aren’t decisionmakers able to bargain reach a peaceful settlement as an alternative to war? To explain war, we must understand why bargaining fails.
The traditional rationalist framework proposes two main mechanisms that may lead even rational decisionmakers to go to war. One is an information problem: if decisionmakers are imperfectly informed about each other’s resolve or capabilities, they may press too hard for concessions, causing war. The other is a commitment problem: decisionmakers cannot commit to refrain from exploiting their future power. Therefore, if a state expects to become weaker, it may want to strike preventively in order to forestall its own decline.
Lake writes that his explanation of the war focuses on commitment problems while ours focuses on information problems:
By endogenizing the decision to invest in new military capabilities, Debs and Monteiro convert the problem of credible commitment (on which I focus) into a problem of asymmetric information (on which they focus). In their model and empirical interpretation, the problem was not that Saddam Hussein could not commit not to develop WMD in the future, but that the U.S. was uncertain about whether a program existed and might provide an opportunity for a “breakout” that would adversely change the probability of victory in a future war.
This redescription of our argument prompts us to make two points. First, as we show in our piece, existing work in the rationalist framework cannot accommodate the commitment problem that Lake has in mind. Until now the rationalist framework focused exclusively on shifts in power — resulting from, say, different rates of economic growth — that happen exogenously and over which states have no control. Therefore, extant rationalist arguments cannot speak of the inability of a state to commit not to shift the balance of power. Power shifts, so to speak, just happen. So, in a sense, our contribution is to amend the traditional framework to allow for the type of commitment problem that Lake describes.
Second, we actually agree with Lake that the inability of a state to commit not to shift the balance of power may cause war. At the same time, we show that this requires imperfect information, which affects the incentives of both states. If states’ actions were transparent, this inability to commit not to invest in additional military capabilities would never produce conflict. War requires that the state fearing decline be uncertain about whether it could detect another state’s militarization attempts; and that the state considering militarization think that it may get away with it without being detected before its new capabilities are deployable. Put differently, it’s not that we choose to focus on information problems while Lake prefers to focus on commitment problems. Rather, in the context of militarization attempts (as opposed to exogenous shifts in, say, economic power) commitment problems can only produce war in the presence of information problems. There is no “choice” to be made.
The status of ‘rationality’ in IR theory
A final note on the role of “rationalist” explanations and the status of “rationality” in the social science and, specifically, in the study of what causes war. Many have objected to our attempt to provide a rationalist explanation for the Iraq War. The enterprise of using rationality in explaining war may be justified on at least two levels. The first, most obvious one, is a belief that states are by and large rational in their actions. The second is that assuming that actors are rational in responding to their strategic environment offers predictive insights and falsifiable propositions. We happen to believe in both. But even if one does not believe that states tend to be rational, the second reason above makes for a powerful case in support of rationalist theories. Put simply, if a theory allows for a state to have any ill-defined preferences or act in any suboptimal way based on the information it possesses, in a way anything goes. The theory becomes unfalsifiable. It also becomes useless in terms of its application to other cases. (Although it may well be better able to describe specific cases in all their facets.)
This means that we should make an effort to develop rational accounts even of tough cases such as the Iraq War — in which the role played by misperceptions, multiple audiences, and incentives to misrepresent information all make it more difficult to understand the basic features of the situation. In short, it is useful to know whether a rational decision-maker, trying to pursue his/her interest in the strategic environment, given the information he/she possesses, would engage in war.
We suspect that in the case of Iraq a rationalist explanation raises suspicion because most IR scholars (ourselves included) find the war regrettable and therefore cast a morally skeptical eye on any argument that may seem to justify it. Although we acknowledge that value-free science is a chimerical, perhaps even undesirable goal, we hope that by making clear that explanation does not equal justification we encourage more reflection on the Iraq War from a rationalist perspective (see, e.g., here and here).
At the same time, we encourage scholars to heed David Lake’s call and further pursue the avenues in which misperceptions and irrationality can improve our understanding of the causes of war. Additional work on whether there are systematic ways in which decisionmakers fall to misperceptions and make suboptimal decisions can add to the general predictive power of rationalist explanations.
Above all, we hope this debate can help move the field forward.