The article “Known Unknowns,” by Alex Debs and Nuno Monteiro (DM), newly released electronically by International Organization in advance of its publication this fall, is an important addition to the bargaining theory of war and adds new insights to the causes of the Iraq War of 2003. Since DM challenge several points in my International Security article (PDF) on bargaining theory and the Iraq War, let me respond briefly.
The key innovation in the model is that investments in military capabilities by the Target (T, Iraq in the case) produce changes in the probability of military victory only with some lag, thus openning a window for the Deterer (D, the U.S.) to launch a preventive war. The lower the costs of war, the more uncertain D is about T’s program, and the more effective the war is likely to be in eliminating the threat to D, the greater the chance of preventive war and the greater the likelihood that the war will be mistaken. This is a nice addition to the basic bargaining model with important implications that go beyond the Iraq case on which the theory is based.
From the outset, I want to clarify that we agree far more than we disagree, including about the central tension between Iraq and the U.S., the effects of 9/11 on the timing of the Iraq War, and the difference between Iraq and North Korea. Debs and Monteiro engage in the product differentiation usual in academic scholarship — highlighting differences rather than commonalities — but the latter are large and overwhelm the points of disagreement, in my view.
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 makes a lot of sense and captures something of decision-making within the Bush administration in the run-up to the war. But it remains a puzzle why Saddam could not demonstrate that he did not have active WMD programs underway. DM argue that it is hard to prove a negative, which is true. But the point is that Saddam did not try but, rather, intentionally obfuscated both his intentions and his actions throughout the 1990s and through the return of the United Nations weapons inspectors in December 2002. Given the clear intention of the U.S. to launch a preventive war by fall 2002 (see below), one might think Saddam would have had every incentive to come “clean,” but he did not. My argument, following much of the literature on the Iraq War, is that Saddam was speaking to multiple audiences including Iran and his potentially rebellious Shite majority and hoped to deter the latter by promoting continued uncertainty over his weapons programs. DM argue further than their model does not require obfuscation from the problem of multiple audiences – and this is true, analytically – but it misses a central feature of the actual bargaining failure.
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.
Second, the model hinges on uncertainty about T’s military program. 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. In turn, given the Bush administration’s belief and the publicity surrounding this belief, we are still left with the puzzle of why Saddam did not capitulate in the face of almost certain war. As I argued in my essay, why the Bush administration held this mistaken belief – one not shared with the earlier Clinton administration – and why Saddam did not capitulate at the final hour remain unexplained by bargaining theory.
Finally, for preventive war to be rational, DM correctly deduce that it cannot be “too expensive” and must be “effective” in removing the threat. But here the Bush administration grossly erred in deeply consequential ways. In particular, it defined effectiveness purely in terms of removing Saddam from power and did not consider the possibility of internal anarchy that followed. The failure to consider the costs of governing Iraq after regime change may have been intentional – to realistically estimate the costs would have made plain that the war would be expensive and less likely to be effective in removing future threats to the U.S., and thus would have received significantly less support from the American public. Or it may have simple stupidity, though I remain suspicious of this interpretation. Nonetheless, this failure to consider realistically the costs of the war and postwar reconstruction was clearly “irrational” but hugely consequential. DM are correct to focus on costs and effectiveness, but unfortunately ignore how and why the Bush administration utterly failed to produce accurate estimates of these key variables.
But that one can pose a rational model that predicts preventive war does not make it the right model or necessarily do justice to the facts of the case.
The nub of the matter is that I emphasize factors that fall outside the bargaining model: the multiple audiences of Saddam that complicate the standard two-party bargaining game, the skewed beliefs of the Bush administration and their origins in psychological dispositions, and the failure of the U.S. to consider the likely postwar governance costs. DM present a fully rationalist model that can explain the origins of the Iraq War. But that one can pose a rational model that predicts preventive war does not make it the right model or necessarily do justice to the facts of the case. As someone often portrayed as a rationalist and defender of rationalist theory, this is an admittedly unusual and possibly ironic position for me to take. Non-rationalists have always complained that the discipline privileges any rationalist explanation over other possible explanations, a stance I have often taken and for which I plead guilty. In this case, however, strict adherence to rationalism and the bargaining model as currently construed – even with DM’s important extension – still misses essential aspects of the politics and decision-making that led to the most disastrous and mistaken war of our time.
Editor’s Note: This post responds to an article published in International Organization by Alex Debs and Nuno Monteiro. Their post on the subject appeared yesterday.