July 29, 2007
Decision Theory and the War in Iraq
Decision Theory and the War in Iraq--Posner
What can economics contribute to decisions on the further conduct of the war in Iraq? I set to one side all issues concerning the initiation of the war, the adequacy of intelligence and planning, the mistakes made in the conduct of the war since the invasion of Iraq in March 2003, and the costs that have already been incurred. (See Becker's and my postings of March 19, 2006, concerning the costs and benefits of the war up to that date.) All those are bygones and should not be allowed to influence current decision making. The correct perspective is an ex ante one.
Rational decision making has the general form of cost-benefit analysis. That is, one compares alternatives and picks the one that offers the greatest surplus of anticipated benefits over anticipated costs. This requires monetizing benefits and costs and discounting (multiplying) them by the probability that they would actually be realized if the particular alternative were chosen. The challenge to the application of cost-benefit analysis to the question of what the United States should do in Iraq lies in the difficulty of monetizing many of the relevant costs and benefits and of estimating the probabilities that they will be realized by particular courses of action that should be considered and thus compared.
An initial distinction should be drawn between monetized and monetizable costs. Our expenditures on military and civilian operations in Iraq are of course monetized, but not the deaths and injuries that our troops sustain. They are, however, monetizable. The greater the risk of death or injury, the higher the wage that is demanded to enlist or re-enlist. That wage premium (as discussed in my post of June 4 of this year), to the extent that it has risen as a result of the Iraq war, provides a basis for estimating the cost of anticipated deaths and injuries to our troops from continued involvement in Iraq. In effect, those costs are impounded in the wage premium. The very slow pace at which the army is being expanded is widely considered a sign of inefficiency, but an alternative possibility is that the expansion is being deliberately slowed out of concern that the wage premium necessary for a rapid expansion would be staggering.
Another important monetizable cost of the war is, if experience with the war in Vietnam is a reliable guide, the tendency to conceal the full costs of an unpopular war by deferring maintenance and replacement of equipment, drawing down reserve stocks of equipment and supplies, and cannibalizing spare parts from equipment not in use. There is also the present cost of long-run medical and disability benefits for the thousands of permanently injured veterans of the war. When the readily monetizable costs of the war are added to the monetized costs now running at some $140 billion a year, the total monetized and monetizable costs could be twice that amount.
There are also nonmonetizable costs, of course, such as the contribution that continuing the war makes to recruitment and training of Muslim extremists who may want to attack the United States either directly, or indirectly by destroying regimes friendly to the United States or by disrupting the production or transportation of Middle Eastern oil. The presence of U.S. troops anywhere in the Middle East apparently acts as a provocation to many Muslims. Against this it is argued by the Bush Administration that if we withdraw from Iraq, the terrorists who are attacking our troops there will as it were follow us to the United States. That is possible, but there are two contrary arguments. The first is that al Qaeda is Sunni, and if we leave Iraq the Sunnis there will find themselves hard pressed by the Shiites, who control the government; so al Qaeda may continue to be preoccupied with Iraq for years. Second, if our presence in Iraq endangers us by fostering recruitment and training of Islamic terrorists, it seems contradictory to claim that our absence would act as a similar provocation.
Another possible nonmonetizable cost is the boost to the terrorists that would be given by our acknowledging defeat in Iraq. Terrorist recruiters would argue that Islamic extremism was winning its global struggle with the West and that this was proof that God is on the side of the extremists. There is also a natural attraction to being on the winning team--the winning side in history. Again, though, there is an element of paradox in arguing that our invading Iraq was a provocation and that our withdrawing from Iraq would be an equal or (the position of the Administration) a greater provocation.
The cost most emphasized by the Administration is the possibility of chaos in Iraq if we leave--an intensified civil war with interventions by Iraq, Saudi Arabia, Turkey, and possibly other countries as well. Interventions by foreign countries in civil wars are common. Another possibility would be a partition of Iraq into Shiite, Sunni, and Kurdish states, on the model of the Yugoslav brakeup, which was accompanied by great violence. The United States would be blamed, and this might well increase Muslim hostility to the United States. Against this it can be argued, first, that withdrawal of U.S. troops might induce the contending factions in Iraq to settle their differences rather than inviting intervention by the neighboring countries, and, second, that whenever we leave, there will be anarchy in our wake because we are unprepared to commit the forces that would be required to pacify such a populous, violent, and fissiparous nation as Iraq.
If the nonmonetizable costs of continuing the war are ignored, either on the ground that the best guess is that they are likely to be a wash or that they are unquantifiable because no one can predict the consequences of our withdrawal, then the case for withdrawal becomes compelling: on one side would be costs probably in excess of $200 billion a year and on the other side no calculable or even probable benefits. Moreover, there are nonmonetizable costs to our continued involvement in Iraq, in particular the distraction of our government from other foreign policy problems and perhaps domestic problems as well.
The benefits of our staying in Iraq seem in current thinking to be limited to averting the costs I have mentioned. There is little expectation of a victory that would transform Iraq and the Middle East and weaken the terrorist threat to the United States.
An intermediate approach to valuing our continued involvement in Iraq would exploit the notion of option value, an important concept of decision theory. An option is a device for deferring a transaction until more is known about its value. We can think of the many billions of dollars that the United States is currently spending on the war in Iraq as the purchase of an option to delay a decision on whether to leave until we have more information about the likely consequences of leaving. That is a prudent course when potentially very large consequences cannot be evaluated at present but may be evaluatable in six months or a year. That seems to be the thinking of the Administration.
The objection is that there is no indication that waiting is producing any information. The optimal strategy for the strongest Iraqi factions, which is to say the Shiites and the Kurds, is simply to lie low until the United States withdraws. The Sunnis have less grounds for optimism concerning their position when the U.S. withdraws, and so they are showing signs of willingness to cooperate with us. But it is unclear how that willingness translates into a forecast of what the future holds for Iraq whether we withdraw in the near term or persist indefinitely. The Yugoslav precedent suggests that when the lid on a cauldron of smoldering ethnic hatred is lifted, civil war ensues. That process is already well under way in Iraq.
A critical variable that receives insufficient attention by the media is the condition of the Iraq armed forces and police. Is it improving? At what rate? What is the desertion rate from the armed forces (a very good measure of effectiveness)? The great failure in Vietnam was the failure to create a South Vietnamese security structure that could stand up against the North Vietnamese without U.S. aid. No matter how successful the United States is in suppressing violence in Iraq, our departure will be followed by collapse if we leave a security vacuum. Since the current emphasis appears to be on quelling violence rather than on creating a viable Iraqi security structure (which may be impossible), the option value of our continued involvement seems slight.
Comment on Benefit Cost Analysis Applied to the Iraq War-Becker
Costs are usually easier to measure in modern wars than benefits. Two estimates of the past and expected future cost of the Iraq war to the United States by Davis, Murphy, and Topel, and by Bilmes and Stiglitz are discussed in my blog entry for March 19, 2006. They quantity the cost of materials and equipment used and destroyed during the war, the higher cost of attracting volunteers to the American armed forces, the cost of the many injuries to military personnel, and the cost of reconstruction aid to Iraq. They also use modern economic research on the amounts necessary to compensate individuals for taking life-threatening risks to value the cost of the number of American lives lost in the war.
The cost estimated by these two studies differs to some extent, as analyzed in my post referred to above. However, a total cost of the war to Americans in the range of $700 billion- $1.2 trillion overlaps both studies. For the purposes of my discussion, I assume the total American cost of the war will amount to $1 trillion. As Posner well discusses, some potential costs are extremely difficult to quantify, and hence they are ignored by both these studies. Such hard to quantify costs include any increased alienation of Muslims toward the United States, the experience gained by the insurgents in fighting American military power, and a potential widespread conflict between Sunnis and Shiites.
Benefits of the war are even hardier to quantify than the costs. Possible benefits would include increased skill of the American armed forces at fighting insurgencies in cities that use suicide bombers, car bombs, and other modern tactics, preventing Saddam Hussein from using the international weapons black market to acquire weapons of mass destruction, perhaps a weakening of terrorist organizations like Al-Qadda, the stabilization of the Middle East, and possibly other benefits. Since difficulties in quantifying such benefits apply to other modern wars and also to many other large-scale activities, it might seem that benefit-cost analysis is useless when applied to these kinds of issues.
Obviously it would be much easier to assess wars and other big events if benefits also could be readily quantified; maybe that will become possible some day as economists continue to make progress in finding ways to quantify various intangible benefits and costs. I say, "continue" because not that long ago economist believed that the value of life to individuals was unquantifiable. Yet advances in the theory of risk-bearing showed how the statistical value of a life could be estimated from choices individuals make in situations that increase their probability of dying, such as driving fast, or working as civilians in war zones such as Baghdad.
Still, even without direct estimates of benefits, the costs estimated for the Iraq war can be used to make a benefit-cost assessment of the war. Given a cost of the Iraq war of say $1 trillion, how big would benefits from the war have to be to exceed this amount, and is that likely? For example, $1 trillion is equivalent to a payment of a little over $10,000 by every single American family, or about twenty per cent of average family income. Would the typical family be willing to pay that much to finance the war in Iraq because they believe that they will get at least that much in benefits from the war? Almost surely the answer would be an overwhelming "no" in any poll taken about Americans' willingness to pay this amount for the war, although some people might not appreciate some long term benefits (or costs), like a possible regional stabilization or serious harm to a terrorist organization.
The estimated cost of the war can be translated into other measures that may help in determining whether or not the war is considered a success or failure. For example, if the statistical value of life is taken to be $5 million for the average young American- this magnitude is in the center of various estimates of such a value- then a $1 trillion cost of the war is equivalent to a loss of 200,000 young men and women. Since this is far greater than the actual loss of lives even in most major wars, like the Vietnam war, it is hard to believe that Americans would be willing to pay that much for the Iraq war, even if Hussein had or was likely to acquire weapons of mass destruction. To be sure, a complete translation of the total cost into a lives-lost equivalent is extreme and possibly misleading, but it does provide a way of gauging the benefits necessary to justify the huge cost of the Iraqi war.