Interesting comparison to invading germany after Hitler took the Rhineland as a justification to preemptive war.

Using that logic if Canada and Mexico were to attack America in a preemptive war because Bush invaded Iraq without cause and that they believed America was on a path to world domination - would that be justified?

A nitpick. The first word of the title is either mispelled or misused.

The content is high class and thought provoking. Thanks.

Society constructs certain rules, the widespread compliance with which allows for a reduction in various costs. The rule of honesty, for instance, might decrease contracting and contract enforcement costs; the rule against unprovoked violence allows people to invest less in personal protection and thus reduces the costs of transacting day-to-day business that involves walking on the streets.

The rule against attacking a nation unless attacked introduces a degree of certainty into international interactions that presumably allows everyone to reduce the cost of self defense. Once this implicit international consensus prohibiting preemptive war is broken, the costs of defense for all nations will likely rise. It is the discounted value of these future defense expenditures that Judge Posner appears to omit from his cost-benefit example. It seems that at least a plausible case can be made that this will not be a negligible term in the cost-benefit balance that might make the threshold at which preemptive action is justified (at least the first time) extraordinarily high.

Dear Justice Posner,
I am a 2L at DePaul and I just wanted to say that I think all of your legal decisions are brilliant. I think that you and Dr. Thomas Sowell are the most insightful economic minds in the world today.
Sincerely,
Charles

i just thought "preentive" was just a contraction of preemptive and preventitative, like he was introducing a new concept, hence the new catchy lingo.

i think fareed zakaria, the 9/11 commission report and the recent defence policy board recommendations make a lot of sense, tho too. in other words, don't ignore soft power and diplomacy and USAID-type development assistance at the expense of hard(thick)-headed realism.

e.g. i think n.ireland should be looked to as a model.

Dear Judge (and hopefully one day Justice) Posner,



I'm a 2L as well (at Georgia) and have had a deep admiration for your judicial opinions, which it seems, have graced almost all of my casebooks so far.




I will comment only that I think that both 1) your Rhineland example is a perfect case of when preemptive war is optimal and 2) the costs far outweigh the benefits of our present venture in Iraq. I would have said so (and did) even from an ex ante perspective. It seems to me that, even where we benefit from attacking Iraq in insolation (a notion I might also dispute) the kick-to-the-hornet's-nest effect this will have on the Arab world will only bring enormous costs to the nation in years to come.




Respectfully,
Tanner

Your analysis is surprisingly one-sided--it seems not to include any of the costs of preventive war.

Consider, for example, the costs of perceived illegitimacy, which tends to follow preventive strikes. The more distant a threat, the more illegitimate the war will appear in the eyes of (a) potential allies and (b) the citizens of the country one invades. In other words, the more distant the threat, the higher the cost of the war.

And, of course, perceived illegitimacy doesn't stop with the war itself. Post-Iraq, the US is seen as less credible and less moral by millions of people, some of whom may be in a position to help us kill terrorists.

There are other costs; why not address them?

I respectfully disagree that the Nazi reoccupation of the Rhineland illustrates the benefits of your model. The British did not attack Germany in response to the reoccupation of the Rhineland because they underestimated the threat posed by Germany. Simply put, the British believed that Hitler would stop at the German border. The French were immobilized by their internal politics, and were loath to fight without British aid. In the real world then, the anticipated costs of attacking Germany were likely lower than the anticipated benefits of such an action. Clearly the French and British governments should have reacted more forcefully to Germany's treaty violations up to and including the reoccupation of the Rhineland. However, it does not seem likely that an economic analysis would have supported invasion of Germany at that point in time.

This preventative war thing is pretty difficult. Alternative histories from 1936 are possible - and the assumption that Hitler would have been overthrown is too quick by far. (We were sure Hussein would be overthrown after Gulf War I.)

But how is this for an alternative history.

Hitler manages to portray himself as a victim. Invasion falls back to hard lines (there is no overwhelming Allied build up then and Hitler had not expended himself in Stalingrad.)

As a result Hitler never looks East until the West is thoroughly beaten. Japan never joins the war - and America thus also sits on the sidelines for way too long.

Hitler then controls Western Europe including Britain but his heart is set on making the German minorities of the East German majorities. The war ends in a cataclysmic struggle between fascism and stalinism.

You are assuming in your post that free market liberalism had wide enough support in countries like France in 1936 to make all that possible. It did not.

Being a collaborator had romantic appeal for some French in 1940. Because free market liberalism has been such a success after WW2 we should not assume that it had a natural following of scale before WW2. We should also not assume then that free market liberalism would win a war in which it could be perceived to be the aggressor.

JH

The probabilty of an attack is only certain, when the attack has occured. If that is the case, then the benefits of a preventative war, is only evident after the fact.

The answer I get from this is that we don't know with any certaintity that premptive strikes have benefits. The benefit is only seen, after the fact. Those who want to divert back to "well, what would happen if we didn't go after Hitler," should play craps in Vegas for a few hours, and see how fast their money evaporates.

Preemptive war like gambling at the craps table is just a bad way of making money. Stick with selling stereos.

One potential difficulty I see with using a cost-benefit analysis to decide whether to engage in a preventive war is that any such analysis seems (at least as a practical matter) to ignore or otherwise be biased against information contained in the other side's revealed "commodification" preferences. By waiting at least for the other side's strike to become "imminent" we gain valuable information about their LACK of willingness to fight that would not otherwise available. Rapidly changing circumstances and the imperfect lines of communication between two parties that hate each other would seem to make such information even more valuable. Although analytically we may be able to account for that in the CBA, as a practical (perhaps even pragmatic) matter I'm not so sure we can do so accurately.

Of course, if the other side has already revealed its preferences, for example, by attacking another state, then we can estimate its commodification preferences far more reliably. Germany's invasion of the Rheinland, and Iraq's invasion of Kuwait, may be examples. But was our most recent invasion of Iraq?

What do you mean by "probability" in this context? Are you trying to say that if the Nazis were to occupy the Rhineland again and again, then as the number of Rhineland-occupations tended toward infinity, then the proportion of them which resulted in war in Europe would tend toward fifty per cent? Presumably not, because that's pretty plainly incoherent.

So I'm presuming that this is a Bayesian concept of "probability as degree of belief". Which rather points up the weakness in this doctrine of pre-emptive attack; unless the attack is imminent, then what you are talking about is the subjective assessment of a policy-maker, with no real evidence to back it up. Since what a policymaker believes is most clearly influenced by what he wants to believe, then what you have here is a doctrine of wars of aggression, with its indecency covered up by a fig-leaf of pre-emption.

What other large and costly government projects would you suggest that this doctrine could be applied to? For example, if President Bush were tomorrow to wake up and decide that the USA needed a National Health Service, massive cash reparations to the descendants of slaves and to take the commanding heights of the economy into public ownership, would you regard it as sufficient reason for doing so that he had subjectively assessed the probability-weighted payoffs from doing so as positive?

The problem with your arguement is that there is no set standard as to determining if a preventitive war will in fact be preventitive of anything and if it the benefits of the war will outweigh the costs. How do you decide which countries to attack and at what point? Which countries pose a threat?

As for Iraq, you must remember that the White House withheld information about North Korea's nuclear program until after a congressional vote deciding our plan of action has taken place. If we were to take your Rhineland example into account we should also admit that both North Korea and Iran pose a much greater threat to the world than Iraq ever did. That is clear. Yet no preventitive measures are being taken now. The worst the current administration has to offer is sanctions. In fact, because of the "preventitive war" in Iraq the Bush administration can not wage a preventitive war in Iran or North Korea. International relations aside, the mere logistics of such an event would stretch our armed forces to a breaking point and would have major negative effects on the economy.

A preventitive war must be planned in good faith and the planning must deal with immediate changes in overall intelligence. Once news of the North Korean nuclear program broke out the preventitive war for Iraq should have been put on hold. A decision should have been made as to who posed a bigger threat. Intelligence concerning WMDs in Iraq was known to be shaky even then where as North Korea outright admitted to possessing weapons of mass destruction.

What you have to understand is that things like "preventitive war" and "preemptive strike" are all terms used to excuse whatever the current policy is. One should never take any such statement at face value. Take the trade embargo with Cuba, for example. The official explanation is that the embargo is still in place to punish a repressive Communist regime. Yet China remains our biggest trading partner and it is the biggest Communist state in the world.

One should never try to rationalize any action without first analyzing whatever data is at hand.

Your simple numerical example, which compares the expected costs of waiting to be attacked, military build up, and preventive war, leaves out one important alternative: waiting to collect more information about the costs and probabilities. Waiting to collect information may turn into waiting to be attacked if your standards for sufficient information are too high, but it is a distinct alternative, and an attractive short-term strategy, in situations where:

1. The threat of imminent attack is low,
2. There is a decent chance of gathering enough information to change which of the other alternatives you consider to be least costly, and
3. The costs of the other alternatives are not likely to increase significantly while you are taking time to gather information.

Given that Iraq was largely incapacitated by various restrictions and that weapons inspectors were in the country learning about its capabilities, Iraq in March 2003 seems to be a strong candidate for meeting these three criteria. Thus, continued inspections, rather than preventive war, would have been the cost-justified course of action.

(technical note: there seem to be problems with posting after previewing)

As dsquared noted above, talking about probabilities in the context of preventive war is a bit misleading. If the decision to go to war was a mechanical, predictable, well-understood process, we could perhaps assign it an a priori probability and work the expected costs and benefits from there. In this case cost-benefit accounting is both reasonable and efficient.

Take the drunk driver. We understand the consequences of impairment relatively well, so we can assign an expected cost to driving under the influence. In many cases the cost justifies a legal incentive not to drink and drive.

But how about what Knight once called uncertainty, as opposed to quantifiable risk? When we talk about unique, one-time events, we cannot assign a definite probability to them. Activity related to them is entrepreneurial in the Austrian sense, and quite possibly beyond cost-benefit accounting. This is especially relevant when the outcomes depend on people's decisions, as opposed to a mechanical process of one kind or another.

Take the person sitting next to the driver. Suppose she's drunk. Should this be criminal on the grounds that a drunk person's advice might lead to a crash? I'd say no, because ultimately the driver is the one responsible for the safety of the passengers, and there is no clear causal relationship between what the passengers say and what the driver does. For the most part the law agrees -- we regulate the conduct of the proximal actor, because he's the one making the final decision, and we cannot reasonably treat him as an automaton which can be made to operate according a nice statistical description. (OTOH incitement offences provide a notable exception to the reasoning. But should they?)

I think this is generally speaking why there are no thought crimes, but instead punishment only follows definite actions, or at least definite signalling (e.g. declaring war, breaking into a house or building a bomb). We modify people's incentives so that rational judgement of costs and benefits will make crime unprofitable but do not try to touch the mechanisms leading to the final judgment. The same then goes for preventive war.

Of course war is hazier in that it's a collective phenomenon. It is difficult to assign guilt and to put the proper disincentives in place when considerably more than one person participates in a crime. War is also messy in that there's a grave imbalance of cost between building something and waging war on it, and also between causing indiscriminate destruction and punishing only those responsible. Terrorism is even tougher because individual disincentives do not work -- how do you punish a suicide bomber?

But does this mean the basic reasoning is incorrect? Not likely. All it does is raise the potential cost of war. If it's true that the a priori probability of war cannot be reliably estimated, the only reliable disincentive we can give is to commit to responding in kind. Which is probably why mutually assured destruction worked -- even if we didn't know the expected cost, we indirectly knew that going to war would not be profitable.

Given that no theory of preventive war is acceptable, and given that opponents of preventive war will not accept anything short of absolute certainty (dead people and smoking gun in hand of the accused), then the next logical question that I can think of is: how high should costs be allowed to reach before reacting? Of course, both assumptions could be relaxed, but let's run with this ....

Ex: Since it is accepted that Saddam was paying families of suicide killers in Israel, would Israel have been justified in nuking Saddam?

Ex2: Since Chris Hitchens has documented a plethora of interesting connections between Saddam and terrorists including al Qaida, how long should those conditions have been allowed to continue before the probability of there being any kind of link between Saddam and murder of American civilians could be admitted to be 1?

In other words, did we actually have to wait for more people to die after the 1st WTC bombing and subsequent flight of the perp to Iraq? How many? Some people would actually respond, in effect, infinity, since they reject military responses altogether. I think that if you categorically reject preemptive war, then it is incumbent on you to define the conditions that must be met before you react. This requires an admission that you are willing to let people die (to incur costs) before you do so. If you still think that no action is tolerable, doesn't that suggest that you are willing to bankrupt yourself to avoid the cost of war?

Paul Eremenko comments above that the abrogation of the rule against attacks unless one has been attacked or the threat is "imminent" will raise costs of self-defense for all, and is not taken into account in Judge Posner's model. This is perhaps true as far as it goes, but Mr. Eremenko does not seem to notice that on the same logic, the existence of rogue state or related non-state actors who would attack without warning similarly introduces uncertainty and raises costs. Although estimates of such costs, especially in discussions such as this, are always heuristic and imprecise, I suspect the financial and psychological costs of living with a terrorist threat at least equal and perhaps exceed (at least for the state making the cost-benefit analysis whether to engage in preventive war) any increased uncertainy and cost resulting from the preventive attack. After, a nation state actor is duty bound to take into account the costs and benefits to its society first of all and to give them primacy over the potential costs and benefits to others (except of course as they affect its own costs and benefits over time).

The other aspect of the cost benefit critiques here, as well as the more unsophisticated uses of cost benefit analysis, which I find problematic is a sense that there is a neat continuum upon which decisions can be made. (The flaw with much criticism of elementary economic theory as well) Of course, in the real world, cost-benefit functions, like demand functions, are almost certainly not convex, continuous and twice-differentialbe. There are surely discontinuities, and I take the possiblity of a major terrorist attack with weapons of mass destruction as a clear example of a catastrophic event that would represent (at least) a point discontinuity. Such discontinuities make for (and perhaps account for historically) the great difficulty in assessing after the fact the intuitive cost benefit analyses performed by decision makers with not only imperfect, but disaterously wrong information.

Since nation states, however, are in the business of making decisions under uncertainty, it seems to me we simply have to get the "best" information we can reasonably obtain and make our analysis on the basis of it. Will the decision makers sometimes get it wrong? Surely. But in dealing catastrophic events, our tolerance for the risk of being wrong (i.e. being willing to act with less certainty) surely increases as the expected costs remain high even when the probability of their occurence decreases. Is this not elementary?

[It's been too long since I read Becker and Posner's major works and was current on the literature to know if anyone has modeled this stuff as convincingly as the guys in the '70s like Werner Hildenbrand who applied measure theory to demand and equilibrium theory -- can anyone point me to the relevant literature, preferably in English as I am no longer fluent in higher mathematics?]

You dance around the question I think everybody has in mind here: should we have invaded Iraq? You almost answer it: as our information becomes sparser, the justification for war, in expectation, becomes weaker. Without good information aggregation methods, your analysis says that we can't justify a war.

It only becomes more difficult with terrorism, which is a catostrophic event. By `catastrophic', I do not mean what Gary seems to mean (causing much damage), but what the statisticians mean: you can't predict it with any reliability. There is no easy bell curve. We've been studying earthquakes for a century or two now and still don't have but the inkling of a clue as to when major earthquakes will occur; we've been seriously studying terrorism for at most a decade or two, and we have absolutely no statistical methods that begin to allow us to aggregate information into a reliable prediction of whether a terrorist act would occur.

BK -

I disagree that my analysis suggests the Iraq war was unjustified because of our failures in creating effective ("good enough?") information aggregation methods. That would be true only if one assumed that we can accurately assess the risk that our information is bad. While I agree that issue represents a systemic issue, I do not believe we have the luxury of defering decisions in media res until we have a high degree of confidence (in the theoretical sense you are speaking) in the quality of our information aggregation. After all, inaction itself, even for the (generally admirable) purpose of improving information for decision makers, represents a decision. I do not share what I take to be your bias (again in the statistical sense) in favor of inaction.

Your point about the use of catastrophic should be well taken. Although only having the been the student of students of Professor Becker, I don't think it would be a stretch to believe he was using the word in both the ordinary language meaning and the statistical sense -- surely it is possible for an event to be both.

There is a tension in pieces such as these by Judge Posner and Professor Becker: they are writing indirectly about the real world, more or less in real time, using the analysis and language of economics that is inherently abstract. The interesting questions are not so much whether the abstraction is entirely correct (which no abstraction can be), but whether the abstraction is useful. At this point, I think the abstraction is useful for thinking about the problem, but I am not sure the analysis is yet sufficiently worked out to be applied convincingly to particular cases --- althougth that is an important aspect of testing their hypotheses.

I'm surprised, given the criticism generally of the analogy of the Anglo-French decision to militarily oppose Hitler's remilitarization of the Rhineland (which every historian knows scared the Bejesus of the the German military, who sent the two battalions of troops in without ammunition and orders to turn tail at the slightest resistance from the Entente powers), no one has mentioned the effect of the Iraq intervention on Lybia's Col. Khadaffi. Does anyone seriously believe that Khadaffi would have renounced weapons of mass destruction and cooperated with us were it not for a sincere desire on his part to avoid the fate of Saddam Hussein?

Judge Poser begins this entire discussion by sliding this at us:

Indeed, the essence of self-defense is striking the first blow against your assailant.

What?

How about: "the essence of self-defense is defending, deflecting, or otherwise avoiding, etc. the assailant's first blow."

Its nice for you to set up a continuium from pre-emptive to preventitive to self-defense, but you can't take the english language seriously and suggest that striking out first offensively is the essence of self defense.

Methinks you have been watching and listening to sports casters who cast about pithy aphorisms like "the best defense is a good offense."

But even these folks realize that they emphatically are not literally talking about defense.

I am 100% certain that you cannot be 100% certain that someone will attack you.

What you describe the "essence of" is preventive force. But it is most certainly offensive.

Remember that World War I was a preventive war, and that illustrates both why the Britain and France were unwilling to intervene in the Rhineland, as well as the reason why there should be overwhelming and reliable evidence before resorting to war.

In the case of WWI, the costs were vastly underestimated, while the benefits were overestimated. (Sound familiar?) So, a war that was intended to be short, cheap and definitive became one of the greatest man-made disasters of world history, with previously unimaginable costs in both money and human life. Moreover, the final outcome of the war was so inconclusive that yet another war was fought 20 years later, on roughly the same territory with roughly the same sides.

The tragedy of WWI was foremost in the minds of planners leading up to WWII, and for good reason. More than anything else, they wanted to avoid another catastrophe like WWI. This didn't blind them to the Nazi danger, but it did cause them to adjust their estimates of the costs (too high) and benefits (too low) of war with Nazi Germany. Eventually, as more information became available, and as the intentions of the Nazis became more evident, France and Britain decided to go to war. Of course, WWII caused massive destruction, but is considered to have been worth it, because of the costs, estimated retrospectively, of losing to the Nazis.

Ultimately, the greatest danger of going to war is the difficulty of correctly estimating the costs. This is an issue that, I think, is papered over too quickly in commentary here. It is also the reason why prudence dictates that war be considered as a last resort -- not because of its unique or intrinsic evil, but because it is a Pandora's box. Once it is opened, it is almost impossible to predict what will emerge.

In particular, the Bush administration's arguments for going to war with Saddam Hussein were almost entirely mistakes or lies. Even if one believes it permissible to mislead the public into supporting war against Saddam, it is vital to estimate, correctly and with caution and intelligence, the costs of that war. By now, though, it's absolutely clear that the Bush administration bungled that, too.

These issues -- of estimating costs, of intelligence and of prudence -- also hold of preventing terrorism, as well as state-to-state violence. However, the tool of war, being between states, is mostly useless against terrorist groups. Although the history of terrorism is shorter than the history of war, we already have good and bad examples of executing campaigns against terrorist groups. The Bush administration, however, seems to be unaware of these examples; they're also deliberately obscuring the real distinction between terrorist groups and nationalist insurgencies. Even if this makes good PR, it is a mistake, insofar as they seem to believe their PR.

In any case, and as a final thought, I wonder about the propriety of erasing the line between defensive and preventive war. Perhaps we should maintain it -- if only to maintain the line between preventive war and offensive war. After all, any arguments made in favor of preventive war may be applied to support offensive war as well, and the differences between them may be, at least partially, in the eyes of the beholder. I wonder how many people outside the US believe the war in Iraq was offensive, rather than preventive.

Just out of curiousity, Judge Poser tosses out .5 as a nice "round" number of probability for the likelihood of future attack.

I'm curious to know what folks think that actual "probability" was in the case of Iraq and why.

I too would like to see more writing on how to account for the costs of preventive war. The costs include, as mentioned above, the loss of perceived legitimacy. There is also a giant opportunity cost associated with going in pre-emptively. Now the US is stuck in Iraq, and has much less leverage, both diplomatic and military, against Iran and North Korea as a result of the Iraq war.

There is also the problem that the attacked region and country will view the preventive attack as illegitimate, leading to the creation of more terrorists or other groups that could attack the US in the future.

Your analysis does not seem to address the myriad, multiplying costs of preventive war as it is currently being played out.

Doesn't war have lots of external costs? Costs which the concept of imminence is, in essence, trying to roughly include into the calculation?

Or is the latest dead brown baby in a fallujah better considered a transaction cost, to be ignored?

First, as I didn't do so in the comments of the first official post, congratulations on the new blog. As an avid consumer of blogs, I have to say I'm excited about this one.

I would add to Judge Posner's comments that the benefit of attaining certainty is not to be underestimated. The probability of direct or indirect (e.g. funding or supplying terrorists) attack is something to contemplate, to be sure, but I can't shake the notion that this variable is an aggregation of:

A) What is the probability that the capability to harm us is present, and

B) What is the probability that the willingness to harm us is present?

What do we do when faced with an entity that we know has the willingness to harm us, but there is uncertainty about the capability? Eliminating the uncertainty in this case is a very large benefit of putting boots on the ground to find out.

Certainly better intelligence can help us make informed estimates of these two probabilities. I fear that many people have exaggerated notions about what intelligence is capable of, however. When dealing with an entrenched despot with a decades old internal security apparatus and a lot of geographical area in which to hide surprises, there is only so much we will know.

I am left with the image of the police officer who challenges a suspect to put his hands in the air. The suspect refuses and keeps his hands in his pockets. At some point, the impasse can be reasonably resolved by aggression.

To compare the erroneous (but almost universally so) cost-benefit analyses (intuitively) made by the Powers in July and August 1914 with Iraq is, in my view mistaken.

However, I think it's not unfair to suggest that an aversion to a repeat of World War I was on the minds of the British and French governments in 1936, especially in an era when the oxford Union debated and passed (if I recal correctly) a resolution to the effect one would under no circumstances be willing to fight for King and country. And a fear of unwilling troops was not entirely unwarranted, especially in light of the French Army mutiny of 1917 and its later performance in 1940. It is useful to consider the different performances of the French and British in the Second World War in light of their histories: While WWI (especially if you consider the whole Home Rule fandango as a piece with it) represented a huge upheaval in British society, it was Englands first truly major war since the defeat of Napoleon, also an English victory. France had suffered a major defeat in 1814, and again in 1870, each of which convulsed French society from bottom to top. Although the British were war weary after WWI, they were not played out, they had the national reserves of will and determination upon which they drew in 1940, they had not lost a major war since the 18th century. The French, by contrast, had nothing left, especially if one considers revanche to have been the great unifying French idea after 1871 Having achieved it at pyrrhic cost in 1919, there was simply no idea (nearly) all Frenchmen could agree was worth fighting for.

There is a case to be made that WWI was a preventive war (the whole set of arguments about the various powers calculations of when it was to their individual advantage to fight a European war some thought inevitable), I have never found that line of reasoning quite persuasive.

More likely, I would argue the series miscalulations which led to the outbreak of WWI bespeak more of a combination of failures to understand how various Powers would react to threats and the rigidity of the Powers' mobilization plans. Especially daunting was the inabilty to mobilize partially, especially on the part of the less modern Powers Austria and Russia. But we could spend entire careers debating the Kriegsschuldfrage....

A cost-benefit argument for preventive war is quite convenient when you can pick and choose which costs you think are worth discussing. As some of the comments above have mentioned, you do this discussion a disservice by not discussing other potential costs of preventive war, such as appearance of illegitimacy to other nations, the inability to sufficiently defend the decision to one's people, the loss of civilian or other life on either side, etc.

Now, if your post was meant to be an introduction to utility theory, then you hindered your intent by clouding a theoretical discussion with a political angle. It works nicely for you because if anyone criticizes your post, you can claim that they're criticizing the underlying theory, not the conclusion. But your post greatly simplifies any sort of analysis needed to justify preventive war (which I'm not claiming to be unjustifiable) by implying that the most important costs can be enumerated and computed with sufficient reliability.

For a more thorough and rigorous discussion of these matters, check out William T. Vollmann's 7 volume treatise on when violence may or may not justified, Rising Up and Rising Down. You'll find a careful analysis on the kinds of costs one can incur by the application or witholding of violence. For example, your post, Judge Posner, implies by omission that costs can be tallied linearly. But things are not that simple when considering civilian life. Is it justified killing 100 of our enemy's civilians to save 10 of ours? What about 1000 for 100? Or 100,000 for 1000? Or 1,000,000 for 100,000? Or 1,000,000 for 100? You can't just plug these numbers into your basic utility theory equations.

The concept of preemptive war based on a cost/benefit model is rather suicidal. Since that same model will be used against those willing to engage in preemptive war.

The learned scholar Prof. Becker of the University of Chicago wrote the following about those involved in these suicidal attacks:

The only really effective approach is to stop them before they engage in their attacks. This is accomplished by tracking them down and imprisoning or killing them based on evidence that they intend to engage in suicidal attacks. Those planning such acts can also be punished on the basis of intent.

So how do you prevent these kind of escalations where one professor professes his intend to kill the other? And they are even supposed to be collegues and friends, running a blog together!

Wow, that was underwhelming. I could've read this on Instapundit. All of this analysis is based on a one-time game. But intervention itself creates conditions for future, subsequent "blowback."

Also, the analysis of 1936 Rhineland is really really really weak. We can play hypotheticals all day and make the argument that if the French hadn't demanded exorbinant reparations in 1918, the German right wouldn't have used that

Or go back even further and say that if the U.S. had not intervened in 1917, Europe would have had to fight to a negotiated peace, which would've kept the balance of power more-or-less where it should've been. The U.S. intervention, and then (sensible) extraction in the 1920s, meant that the allies won the war, but couldn't win the peace.

One other factor that plays into this discussion is that war is almost never undertaken except after a major attack -- an event that the republic can accept as justification or pretext to do the job.



Names like 911, Pearl Harbor, Lusitania and the Gulf of Tonkin have all served that purpose. In the 1990's the US was attacked several times where the awfulness of the attack didn't reach the level that would allow any president to move into Afghanistan. Only after 3,000 people were killed on our shores could it happen.



The preventative war in Iraq could only be done with public approval after 911, even though the connection remains tentative. Democracies can only wage war when the people agree. That eliminates any possibility that Britain could have confronted Germany earlier. Even the early years after the invasion of Poland were known as the "phony war", since there was little appetite to pursue it aggressively. And we will not attack Iran or N Korea now, no matter what type of provocation until an actual attack begins.



So my feel is that this is a moot question. A democracy can't become involved in a preemptive war, except in very specific circumstances. Those circumstances include a violent shock to the nation, and a situation where there's a bad guy in the same part of the world that needs to be dealt with. Not a picture that will likely happen again.

Judge Posner, how would you factor into your equation the cost to global stability or peace generally. Isn't your formula necessarily state-centric? How do you interpret your formula in light of the goals of the UN Charter "to save succeeding generations from the scourge of war"?

As with MRL, I fell off the rails when I read "Indeed, the essence of self-defense is striking the first blow against your assailant."

I hope this is a phrase that would be remedied by editing - since what is described is in fact the essence of offense.

The essence of defense is to not be harmed - striking preemptively is one of many possible tactics, but is in no way the essence of the thing.

So, let's say I am Iran, or N. Korea, and I reasonably suspect, can put a numerical probability on, an invasion by the USA? By this rationale, I am justified in a pre-emptive invasion of N. America. Of course the theoretical economics argument is sound. It is the practical employment that makes no sense whatsoever.

Had you left out the references to Iraq, one might have been able to excuse this as a simple academic exercise.

Sometimes I wonder whether otherwise apparently bright individuals lose all sense of rationality when devotion to ex-post justification of the Bush Administration's war in Iraq takes over. Absolute failure to consider the situation from any other party's perspective than their own seems to be the hallmark of such thinkers.

Essentially this same rationale has been tried to justify our flaunting of the Geneva Conventions, with very likely adverse effect on the treatment of our own personnel.

I see that I am not the first or only person to see through the blatantly political agenda here.

This is a frankly embarrassing first post for two reputedly very smart people.

Concerning the ubiquitous 'blowback' principle of foreign policy, I would note that there is an assumption inherent in the argument that the choice to not act has no consequences, or 'blowback'. I can't see how that view is justified.

In general, I don't know that historical counterfactuals are useful tools in arguments about justification. They tend to be divorced from the perspectives of the affected parties. What level of reparations in combination with the loss of the WWI would be low enough that the German right wouldn't have seized on it for leverage? Who knows?

"The concept of preemptive war based on a cost/benefit model is rather suicidal. Since that same model will be used against those willing to engage in preemptive war."

Power escalations are rational in many cases for just this reason. International relations is fundamentally about the ability to inflict harm. It therefore behooves those with liberal values to be able to inflict massive harm on those without those same values. The justification may be made both ways, to be sure, but that point is moot. The reason Saddam didn't march on Washington was not that he felt in his heart it was wrong to engage in aggressive acts, it was because he was physically incapable of doing so.

To take preemptive action off the table is to surrender enormous strategic advantages, especially when dealing with a non localized threat. Every entity that can count itself as distinct from Al Qaeda gets a free shot at us in that scenario. Suppose bin Laden has a brother that runs a fundamentalist, anti West terrorist organization distinct from Al Qaeda. Do they have to blow up a couple of buildings before we are justified in attacking them? What if Al Qaeda breaks up and reforms into smaller groups with different names? If we kill or capture all members of AQ who were directly involved in attacking the US, are we done? Other members of AQ certainly didn't attack us.

CR-

Since you responded to me by name (way up in the thread at this point), I feel that I should reply to you.

First, the expected utility of a preemptive war assumes that we can calculate an expected utility, which implicitly but conspicuously assumes that ambiguity = risk -- that there's a sliding scale from bad information with hazy priors up to perfect knowledge, and we're never at perfect knowledge but we can work with what we have. But the two are not equal, which makes for some huge problems. It's not that we're using the best model that we have, and one day we'll incrementally get better; using a risk-based model to measure an ambiguous payoff is capital-W Wrong. In most cases we make do with calling ambiguity risk, because it's all our tools can handle, but when thousands of lives and billions of dollars are at risk, maybe we should be extra-careful about our assumptions. Given that there is no reasonable model of decisionmaking under ambiguity, maybe a bias toward inaction is the right thing to do.

I know it's gauche to make comments like this, but I've blogged about this; see the blog for more on risk v ambiguity.

Is there anybody who seriously believes that Libya would have given up its weapons program if Bush hadn't intervened? Yes.

"As with MRL, I fell off the rails when I read "Indeed, the essence of self-defense is striking the first blow against your assailant.""

A wise man once said, "What is true of the kung-fu fight is not true of the gun fight." If he didn't say it, he should have.

Real people don't dodge bullets, and the first bullet can kill you.

I think that states sometimes use some sort of rational calculus to go to war. But Judge Posner's very number-crunching look at the subject seems to gloss over a key point--estimation of political and military forces on a quantitative basis is extremely difficult if not impossible.

If such estimation is impossible, then any theoretical justification based on an exposition of such principles at work is, to say the least, intellectually suspect.

Furthermore, the question as posed simply evades the moral dimension, one of the critical elements of any war analysis. I guess I might be called "old-school" in these days of realpolitik, but I think that preventative war is immoral.

More importantly, the analysis' only historical example is one where preventative war was decided against, creating a huge bias towards finding preventative war as useful. But examples of preventative war that led to ruin or did not create the benefits claimed abound in history--Stalin's backdoor move on Poland in 1939, or his invasion of Finland a year later. The biggest of these examples, Germany's invasion of Soviet Russia in 1941, is not touched on at all. Barbarossa was based entirely on a belief that eventually, Nazi Germany and the U.S.S.R. would come to blows and that now, rather than later would be the time to fight that war. To argue that Barbarossa was worth it is simply ridiculous.

But I think the best illustration of the absurdity of the concept is its application to every country on Earth. Imagine a "Preventative War Index" by which the Pentagon would measure the likelihood of attack from other countries and then use this matrix for decision-making. I can think of many small countries who are unlikely to do much damage to the U.S., but whose invasion would cause little grief to the U.S. Despite the fact of low risk of attack, it would still be cost-beneficial to attack these countries. However, I think that few readers would support invasions of Bhutan or Madagascar.

Presumably, the Pax Americana of the future leads our attention to a McCarthyism which threatens everything we hold dear. On the other hand, the unstated purpose of this war brings about the predatory imperialist aims outlined by the crypto-fascist Project for a New American Century. So far, the influence of Leo Strauss belies justifications given by the world's leading apologists for the essential Western imperial interests. It appears that the pro-Sharon neoconservative cabal brings forth this calamity brought to us by a horrific onslaught, known as Shock and Awe.

Paul Eremenko,

The rule against attacking a nation unless attacked

Except there is no such rule. Posner made that point quite clearly right there in paragraph 1. Disgree with it all you want, fine--but I don't think you can fairly claim that Just War or some other widely accepted theory of international conflict supports your position.

John Hempton,

the assumption that Hitler would have been overthrown [in 1936] is too quick by far

Well, we know for a fact that officers were planning a coup if Britain and France had decided to defend Checkoslovakia. German military strength was also far weaker in '36 (this is the main flaw in your alternative history.) Because of these factors, the assumption that Hitler could have been defeated over the Rhineland has always seemed very uncontroversial to me--though of course the general's plot was not known to the Allies at the time (it hadn't even occurred yet.)

MRL,

Now that you mention it, "the essence of self-defense" statement is perhaps too condensed to be useful. How about "Without the ability to strike the first blow against your assailant if necessary, you certainly don't have self-defense, at least not as understood in the Anglo-American common law tradition"? That's certainly the mental context I supplied the first time I read his statement.

Me,

Although yours are good questions, there are some troublesome elements lurking right below the surface, most notably the apparent link you make between stability and peace. That seems pretty darned state-centric to me: quibble about the legality of deposing Sadaam, while never once accounting for all the Kurds, Marsh Arabs, and others harmed by the Baathist regime because that was all internal state affairs.

This topic is impacted by change in several areas.

Technology has escalated the destructiveness that even small groups of potential enemies can wield. This skews the complex preventative war calculations that many on this forum wish for - possibly to the breaking point. Another point of instability is the breakdown of the "nation-state mystique" that was so carefully bulit at the Congress of Vienna, namely the idea that the leadership of a nation is comprised of 'special people' who are not to equivalent to some nameless infantryman. Hence, the "no asassintaions' policy [Pay close atention: not Law, but Policy, which is a stroke of a pen].

Not only is preventative war seem to be growing more needful, its proper expression may well become the immediate destruction of a ruling elite as the standard opening move: Pawn to King4 = JDAM in the Beloved Dictator's bedroom.

Since this involves far fewer casualties and a potentially more odious target set, it is likely to be much easier to "sell" to the public.

This going to be an interesting century.

Judge Posner: I wonder if you feel the exact same logic should apply in criminal trials for assault. Should the individual's weighted calculation of the costs and benefits of the risk from not preventively assaulting an enemy be considered, rather than more traditional standards for self-defense? If not, why not?

Is this what the "Burning Bed" defense is about?

As long as we are considering historical counterfactuals, let's go back to WW I and the consequences of America's intervention and the removal of the Kaiser.
US intervention delayed the exit of the Soviet Union and tipped the scales toward the Allies.
Without it, Germany probably would have won the war and the map of Europe and the balance of power would looked slightly different. Germany wouldn't have been a threat, as the GDP of the Triple Entente was 5% or so larger than that of Germany and the Austro-Hungarian empire combined, and wouldn't have been on the hook for reparations.
Since the Kaiser was a check on the rise of a dictator, Hitler wouldn't have come to power and so WW II probably wouldn't have happened.
Churchill accepted the logic of this.
As for Japan, if FDR had pursued a libertarian foreign policy and kept the navy within three miles of the West coast, its military wouldn't have felt compelled to bomb Pearl Harbor.
Hawaii was not a US state then, just evidence of America's nascent imperialism.

I would love to see your calculations for how much a baby's leg or thousands of dead and displaced civilians in Fallujah is worth in your cost-benefit analysis. How about a 19 year-old dead marine or his grieving mother? Or three years in Guantanamo Bay for a poor Pakistani who was in the wrong place at the wrong time and turned over to the Americans for a couple thousand dollars in bounty?

Or what is the cost I have suffered for being separated from my wife for the last six months? Not much compared to the examples above, but significant when you multiply it a couple million times.

War is a bloody, horrible, messy business. Preventative war will never survive a cost-benefit analysis because by its very definition preventative war can be avoided, and since you can't put a cost on human suffering (maybe you can, but I always found your decisions to lack any spark of human feeling), preventative war is never justified because there is always an alternative.

A few "costs" that have yet to come up in the context of the Iraq discussion posts:

1) How do you account for the costs in human capital as a consequence of Saddam's extermination of his own countrymen? He had nearly a generation to squander the productive talents of the Iraqi people; his psychopathic sons were to be the ideal product of the next iteration of the regime. Doesn't there come a point where if you do not act preemptively, you may get to a point where the entity you are reacting against will never be able to recover?

2) The cost of securing total victory vs. long-term management of the problem. For all the comments re supposed facts that we did not acurately quantify prior to invading Iraq, I see far greater danger in not recognizing, or finishing conflicts to total victory, and believe that when in doubt, err on the side of securing victory. Would N. Korea have nucelar weapons today had Truman decided to "win" that war instead of accept a divided Korea? Would we have been better off deposing Saddam back in the early 1990s? What if Israel were allowed to keep and govern the territories that it won as a result of various wars initiated against it, rather than give them back to the very ones who call for the nation's destruction? What if Reagan had unleashed the full force of the Marines in response to the attack in Lebanon?

3) The cost of requiring a "rational" strategy of the U.S. when dealing with inherently "irrational" actors on the other side. Take preventative or offensive war off the table, or place a legalese framework on the decisionmaking process, you give far more leverage to your enemy.

4) Long-term strategic costs. Regarding the current threat from Islamofacism, one of the issues that needs to be addressed is attacking both its financial and ideological underpinnings. That is primarily coming from Saudi Arabia, with secondary effects (in terms of terrorist training methods and threat of nuclear capacity) from Iran and Syria. What are the comparative costs of overturning the Saudi regime directly vs. Iraq? Is it possible to have a successful regime change in Saudi Arabia when you have Saddam Hussein sitting on the border? Ditto with any revolution-from-within in Iran or Syria, were one ever to occur? Now that the US is formally entrenched in Iraq, with a presence in Afghanistan, doesn't that change the cost-benefit analysis (to render non-war options more likely) for surrounding nations vis. providing added inducements for self-reform that would never have existed had preemptive activities been off the table?

5) The "legitimacy in the eyes of others" cost is overplayed, and I think a false cost. All of the "others" are essentially free riders. Their opinions shouldn't count for much, if anything.

For what it is worth, I personally think that the problem with pure cost-benefit analyses, absent a guiding moral/philosophical component, is that cost-benefit analyses will ultimately work to the benefit of the person most likely to abuse them. Arguably France is the nation that has been most wedded to pure economic cost-benefit analyses in its Middle-East strategy; Saddam Hussein was able to manipulate that analysis all too well (see, UN Oil-for-food scandal, for starters). Is France's geopolitical strategy really the intellectual ideal we want our nation to strive for?

"Had France and Great Britain responded to this treaty violation by invading Germany, in all likelihood Hitler would have been overthrown and World War II averted."

Nonsense.

The only thing that would have happened "in all likelihood" is that there would have been unintended, unpredicted, and unpredictable consequences.

It is just as likely, for example, that Hitler would have repulsed the invasion and increased the development efforts to produce the German atomic bomb .

The point is not that preventitive war is never justified. Sometimes it surely is, but war is always a last resort solution, representing not the continuation of diplomacy by other means, but instead a monumental failure. And because the chaos that accompanies that monumental failure is equally monumental, the outcome of a war -any war- can never be accurately predicted despite Posner's apparent self-assurance that it can. Except in one regard:

That innocent people will die horrible deaths approaches a confidence level of one.

In any event, by any criteria, except Bush's, nothing in the run-up to the Iraq war justified a preventive war by the US (and its allies).

Arguably France is the nation that has been most wedded to pure economic cost-benefit analyses in its Middle-East strategy

Sheesh, do you really believe the crap you write? Yes, France over the last ten years did have the closest economic ties to Iraq. But if you want the country that looks at the pure economic cost-benefit analysis in its Middle-East Strategy, we are it baby? We overthrew the first almost democratic country in the Middle East and installed a dictator because we thought it would be too friendly to the USSR (Iran). When the Shah got overthrown we supported Iraq in Saddam's war against Iran out of spite eventhough Iraq was closer to the Soviet Union than us and we knew how awful Saddam was. You admit that Saudi Arabia is a problem yet we are their biggest backers and we didn't say a word when they delayed their elections and announced women would not be able to vote (look for them to quietly cancel them completely with no complaint from us).

Do you really think that a "revolution from within" will result in more friendly governments in Iran or Syria? Be serious. Chances are we will try in foment one and it will be tainted by our influence and result in a fundamentalist backlash. The hardliners in Iran have gained power, not lost it, since 9/11 because they have got the country to rally around the flag against us and increased oil revenues have allowed them to buy off the fence sitters. We support corrupt dictatorships in Pakistan, Saudi Arabia, the Gulf States, Egypt, and Jordan. This administration has shown no real interest in pressuring them to change nor has it pressured Israel and the Palestinians to come to some kind of accomodation.

Uncertainty does not justify a bias in favor of inaction. Inaction is a decision with consequences, just as it action. Since human behavior is inherently uncertain, all social choices, even those classified as “inaction”, must be made in the face of uncertainty. This necessitates the assignment of probabilities that are themselves only uncertain estimates. There is no alternative. Therefore, Richard Posner’s methodology is inescapable. Of course, this method can be used to justify any action or inaction.

The typical argument in favor of the US invasion of Iraq includes the following:
Saddam Hussein was a genocidal dictator who had already slaughtered more than 1 million people and would probably kill more if left in power. He had probably already assisted terrorists acting against US interests and would probably do so in the future. Containment was not effective and would probably deteriorate over time. The defeat of Saddam Hussein would probably discourage other nation states from supporting terrorism in the future. The US could probably establish democracy in Iraq as it did in Germany, Japan, and South Korea and democracy in Iraq would probably encourage liberalization among Arab nations and thereby reduce support for terrorism.

Not all of these uncertain expectations need to be true to justify the US invasion of Iraq, but it is clear that the decision to invade can not be justified with certainty.

What about the decision not to invade? This would typically include the following:
The US is not responsible for Saddam Hussein’s genocide and should therefore give it only minor weighting in any decision, but the US would be fully responsible for all deaths that occur due to its invasion. Saddam Hussein might kill far fewer people in the future than he did in the past, in part due to containment. Containment of Saddam Hussein will probably work adequately far into the future. Invasion of Iraq will probably motivate more Islamists to become terrorists than would inaction. Overthrowing Saddam Hussein will probably not deter other nation states from supporting terrorism. The US is unlikely to ever succeed in establishing democracy in Iraq. Saddam Hussein probably never had significant ties with terrorists acting against the US and probably never will in the future if the US does not invade Iraq. The US can probably prevent future terrorists acts against the US or at least keep them to a cost less than that of war in Iraq.

Clearly, the decision not to invade Iraq is based only on probabilities, not certainties. Hence, there can be no certainty. We are forced to argue probabilities.

I don't think Judge Posner omitted "the discounted value of these future defense expenditures" from his cost-benefit example. In the case of the United States, there is no such benefit. The only nation states that threaten the United States are already on the outside of international law (North Korea, Iran). There is no increase in the threat of attack from say Canada, as the first commenter so foolishly suggests.

In fact, the primary stabilizing force throughout history has not been any rule agaisnt pre-emptive attacks (which have been common forever) but rather the fear of destruction by a superior force. East and West Germany existed in relative peace not because of any international law, but because the United States and USSR assured the destruction of either, should it turn agressive.

In this context, I would argue that the United States, as the sole super-power, discouraged agression by rouge nations when it attacked Iraq (both times) and Afghanistan. A pre-emptive military policy serves the same goal as agressive police enforcement of minor violations, it cleans up the whole neighborhood. Our pull-out of Vietnam on the other hand, had the opposite effect and encouraged minor conflicts around the globe, as did the failure of the USSR to subdue Afghanistan.

"East and West Germany existed in relative peace not because of any international law, but because the United States and USSR assured the destruction of either, should it turn agressive."

That is a gross oversimplification of an exceedingly complicated situation, one in which MAD surely was an element, but which also included the utlility of institutions of international law and the deployment, after October '62, of state of the art communications technologies by world leaders.

If US "discouraged agresion [sic] by rogue nations," how come the world has Darfur? And what on earth is a rogue nation anyway?

Anonymous- your links don't work so I've no idea where you're referring me.

I've always been uncomfortable with claims that calculations of utility can be neat and part of a truly orderly continuum -- as I thought my post made clear. As economists or social choice theorists we use relatively simple models such as the 'utility calculation continuum' because we're convinced they're useful (i.e. generally good enough, not because we're convinced they're exhaustively accurate) and because they lend themselves to relatively straightforward explication. When we want to be significantly more precise in defining our functions, for example, the mathematical price of admission to the discussion shifts from the calculus one expects most undergraduates to understand (at least generally) to the sorts of advanced mathematics understood by few (a prominent mathematical economist once suggested to me in a conversation that there were less than 100 people in the world who were really capable of understanding and contributing to the literature, and half of them were mathematicians who couldn't care less).

Nonetheless, I am not persuaded by your distinction between risk and ambiguity, nor am I convinced that modelling decisionmaking under uncertainty (risk) is inappropriate to model decisionmaking under ambiguity. You may have blogged on it, but I guess I'd need to see a more fully fleshed out argument, perhaps even the mathematics you have it mind (though, as I said before, I'm very rusty with serious mathematics).

Tristero writes:

'If US "discouraged agresion [sic] by rogue nations," how come the world has Darfur?'

Let's not play dumb now. Clearly he meant aggression against the United States. Is Darfur a part of the US? "Duh".

Uncertainty does not justify a bias in favor of inaction.

Rick,

Your analysis fails to account for the fact that greater uncertainty exists concerning the prediction of the threat as opposed to calculating the negative consequences to a state engaging in preventative war. Negative consequences to a state engaging in preventative war would be far easier to calculate because their effects would be much closer in time to the decision to go to war. A threat calculus must take into account a potential threat years into the future, making prediction much more difficult.

Since it is assumed that the calculation is being made to determine if a preventative war is to be launched now, the calculation of the negative consequences is far easier than the threat.

Thus the risk of inaction will always be easier to calculate and to mention the costs of inaction in the same breath as action does not to full justice to the problem.

The factor of uncertainty of threat demonstrates that the calculus will more often than not fall on the side of "angels fearing to tread" rather than "fools rushing in."

Our pull-out of Vietnam on the other hand, had the opposite effect and encouraged minor conflicts around the globe, as did the failure of the USSR to subdue Afghanistan.

Steven,

Your analysis seems to suggest that you would think it better that the Soviet Union had subdued Afghanistan. I find this troubling. Furthermore, you provide no evidentiary justification for either position quoted above.

Cato makes a good point regarding the cost of uncertainties created by allowing terrorists to run amuck. Many analysts have said there is a $10+ dollar a barrel "terrorism premium" on the price of oil right now. This illustrates just one of the real costs of waiting to clean out the hornet's nest. (Watch this premium drop as it becomes apparent that OBL & co. are severely weakened as a global menace, probably a month or so after the successful elections are held in Iraq.)

I won't even go into oil-for-food, which makes a complete mockery of Blar's argument that Saddam was in any way "contained."

Off topic, but necesary given no E-mail contact to administrator, is the fact that the spam bot mechanism does not work with the preview feature correctly. One has to copy one's work, and then paste again into the first page, sometimes with disasterous results. Please fix. Thanks.

The premise of weighing pre-emptive war in terms of the probability of attack and the balance of cost and harms is an interesting one. While the idea might seem most pertinent to American involvement in Iraq and Afghanistan, I’m curious how the model would apply to third party conflicts such as those that occur in Africa. Does the probability that a foreign country will attack another foreign nation still carry the same weight as the probability that one would attack us? For example, was it necessary for Britain and France to wait until the invasion of Poland to declare war on Germany or could the likelihood of the attack been sufficient? What if they had invaded due to the occupation of Austria or the Sudetenland, where the eventual probability of a German attack against them appeared significantly lower or possibly non-existent? In the calculations of pre-emptive war, the question exists of what relationship there is between the probability of being attacked and the probability of conflict between foreign nations, as well as the consequences of that relationship or lack thereof. For example, can the high probability of third party conflict justify pre-emptive attack even when the probability of eventual conflict with either foreign nation is otherwise low?

To try and answer this question, we might turn to cost-benefit analysis as the Honorable Posner suggests, but now it is more complicated. The nation considering pre-emptive invasion to prevent the third party conflict has to look at a greater variety of costs and benefits due to the multiple relationships involved. Let’s assume that the Nation of A is considering a pre-emptive strike against Nations B or C (or possibly both) to avoid war between the two. Nation A would likely need to consider the C/B-analysis between itself and B as well as with itself and C. In addition, there is the C/B-analyses between B and C to consider, if not more.

One interesting result of this approach is to ask if nation A gains nor loses in its C/B-analyses, can it still justify pre-emptive invasion due to the C/B-analyses between B and C? Between individuals, the answer is yes. Even though both Bill and Tom, who are in an intense disagreement over ownership of a piece of land, intend to set fire to the property, society can intervene with criminal sanctions on both parties before either can carry out the act. This is still true even when there is no risk of escalation between them or any involvement of third parties and their property (Assume that they are both happy if neither person can use the property, which happens to be a small island near an urban location). At this point, the only significant principle perhaps for justifying these criminal sanctions is society’s “interest” in preserving wealth. While there have been many actual and potential disputes between nations over contested territory similar to the situation of Bill & Tom, historically the preservation of wealth and lives doesn’t seem enough to merit a pre-emptive invasion when a conflict between two foreign nations is likely. Therefore, it would seem that more than the preservation of wealth and lives is necessary before nations, when little cost or gain will result, will pre-emptively involve themselves in possible conflicts between foreign nations.

Rob W:

Your analysis fails to account for the fact that greater uncertainty exists concerning the prediction of the threat as opposed to calculating the negative consequences to a state engaging in preventative war.

This greater uncertainty as to threats would suggest a bias in favor of inaction only if two conditions exist: the nature of the threat must be uncertain and the imminence of the threat must be uncertain. However, if the magnitude of the threat is sufficiently great (e.g. an attack with nuclear weapons),and the reasonably foreseeable costs not at least equal or greater in magnitude than the threat of action, then regardless of the fact that the calculus is inherently risky and ambiguous, there is no necessary bias towards inaction.

Mutual assured destruction was an example where the threat of acting by an attack on a potentially threatening adversary (and unleashing MAD) was knowable and of a magnitude at least as great as the threat of inaction.

Consider that a nation state's first duty is to provide for the defense of its citizens, it must place a higher value on their lives and freedom than those of any other nation, by definition. I take it as given that there is a difference in kind between the sort of threats taken into account traditionally in determining whether preemptive action is justified (hostage takings, kidnappings, destruction of some commerical interests -- all traditional causa belli) and the possibility of an attack by weapons of mass destruction. If this is correct, I would describe such an attack as catastrophic not only in the ordinary language sense, but mathematically as well, meaning the ordinary assumptions about the orderliness of the functions no longer obtain. Or to put it another way, if the potential threat is so great that it cannot be tolerated under almost any circumstances, a decision-making rule that is biased in favor of inaction will not be acceptable to decision-makers charged with the welfare of society.

This is all rather silly. I doubt that the most brilliant economists in history could have sat down in July 1914, and crunched their cost-benefit analyses on the impending war would have even come close to predicting the outcome. Who would have predicted that four years and 12 million lives later Europe would be in ruins, four of the seven empires that participated in the war had collapsed, two of the remaining three were so damaged that they too would eventually fall apart. That the rest of the century would be taken up fighting the twin evils, fascism and communism, that the war created.

And yet you seriously want us to believe that there is some magical economic formula that can tell us if a preventative war will be "worth" the cost. What hubris and arrogance.

Freder: ummm, you're not an economist I take it....

As far as I can tell, no one has suggested there is some modern version of the felicific calculus advocated by the English Utilitarians at the beginning of the 19th century. I doubt most economists would suggest it's likely that we'll be able to construct one. The notion as I understand it is not that someone has actually and effectively constructed a real world model into which inputs can be dumped and outputs cranked (pace Wassily Leontieff and his epigoni), but that the description (model) is a useful way of describing what goes on in the real world, useful in terms of understanding how decisions are made and considering what can be learned from this.

Your Honor,

I am not a 2L and don't idolize you, but an unemployed freshly minted lawyer who recognizes grave social threat in your law and economics school, yet, grudgingly, admires your compositional style and the efficacy of some of your methods in some of your opinions.

1. You said "striking the first blow is the essence of self-defense" as though you have never considered the criminal law, nor situations in which the self-defense is proffered as a justification of violence. This has been addressed above.

In those cases, as you must know, striking first would tend to degrade the self-defense argument, portraying the putative victim as the initial aggressor, and thus the putative aggressor as the initial victim.

This is absurd.

2. Still more absurd, and shocking to contemplation, is your description of "the U.S. decision not to invade Afghanistan before the 9/11 attacks"!

Sir, when was this decision made? And by whom?

Clearly this decision must have been made "before the 9/11 attacks" and, as you have phrased it, it must have been made both in foreknowledge of those attacks, and considering them sufficient justification for the planned invasion.

Are you really saying that, oh great adjudicator?

Are United States policymakers, even now, (in)acting on the decision not to invade all the sovereignties with which we are not currently engaged in actual outright hostilities?

Before the 9/11 attacks, members of the Afghan ruling junta had 1) significantly curtailed poppy cultivation (and hence the flow of opium and heroin), and 2) been involved in negotiations with certain American concerns, mediated by certain American officials, regarding the transportation of oil. And "we" had a plan on the books to topple the regime and install our own power vacuum.

I would disagree, point by point, with the rest of your discussion, but I am too shocked and dismayed at the vertiginous spin and radical framing in your first few paragraphs to stomach it.

Where is the 9th Circuit blog?

The unstated purpose of this war is determined by capitalist interests which lead to an oil war masquerading as an endless crusade against "terrorism." Bush’s argument for war belies justifications given by the world's leading apologists for a humanitarian disaster of unimaginable scale.

This suggests that the American state, with its unelected president, venal Supreme Court, silent Congress, gutted Bill of Rights and compliant media can be seen in the light of the apparent fabrications which lead to the final subjugation of the Middle East, beginning with the $90bn invasion of Iraq. It appears that the appropriation of Arab resources brings about the resurgence of White Supremacist ideologies. On the other hand, the Pax Americana of the future is solid evidence of the seizure of the Iraqi Oil Ministry. Clearly, Donald Rumsfeld's worldview is determined by capitalist interests which lead to the predatory imperialist aims outlined by the crypto-fascist Project for a New American Century.

Presumably, the unstated purpose of this war represents the repudiation of international law in order to bring about the result of a pre-fascist atmosphere in America.

It is quite remarkable that the appropriation of Arab resources represents the crushing of internal dissent in order to propagate the flagrant lies promulgated by the political donor class. Nevertheless, Americanism as an ideology provides a pretext for the essential Western imperial interests.

It appears that the deal between the Department of Defense and Halliburton unit Kellogg, Brown & Root brings about this calamity brought to us by a horrific onslaught, known as Shock and Awe. Perhaps for the first time since the late 1940s, the influence of Leo Strauss represents the repudiation of international law in order to bring about the end of any possibility of social justice in a reactionary state. It is quite remarkable that the apparent demise of "anti-Americanism" as a respectable means of stifling recognition of American imperialism is determined by capitalist interests which lead to an act of international violence that exceeds even those of the "liberal" Bill Clinton.

From the beginning of time, whenever an individual or the leadership of a group has perceived the benefits of using violence as outweighing the costs, they have utilized violence. This will continue till the end of time, almost by definition. Organizations such as the UN are supposed to raise the costs of violence but that basic premise still holds. Therefore, I consider the distinctions between offensive, defensive, pre-emptive, and preventitive war a distraction to the real argument.

If all we're trying to prevent is a few thousand or even a few tens of thousands of U.S. citizens murdered and a few big buildings knocked down every few years, then the cost of invading Iraq may not have been worth the benefits.

But what about the value of civilization itself? It's not inconceivable that a few well placed nukes could so shake the faith people have in civilizations and its institutions that the whole thing collapses like a house of cards. For example, currency only has value because: a) people think it does and b) people think that the issuing government will be around essentially forever. The latter could possibly be called into question with an attack not all that much bigger than 9/11. Many of civilization's institutions are like that so the whole thing could unravel fairly quickly.

If civilization collapses, then what? The planet would probably struggle to support even one billion people living without the structure and efficiencies of civilization. That means at least five billion dead, of all races and ethnicities. If you estimate the probability at one in ten-thousand, that's still an expected value of 500,000 dead. Even at one in a million, that's still 5,000 dead. That's how I look at the utility analysis.

Cato,

I'd argue that if the threat is imminent, it would fall under Judge Posner's "preemptive" war, not his preventative war.

I'd also like to see the effects of internal constituencies are taken into account in this model, for example, the non-entry of Britain and France as intervenors in the U.S. Civil War.

I think I am with Quiggan here - this has to be a hoax. My history is not good - but its enough to know that this history is totally revisionist.

Alternatively BP is being too cute by half. I feel stupid even trying to comment above.

JH

CR-

This thing keeps thinking I'm a spambot; sorry about the mangled prior post. The ambiguity blog is here

http://fluff.info/blog/arch/00000101.htm

and the article from the Saban Center is here:

http://www.brookings.edu/views/op-ed/indyk/20040309.htm

Meanwhile, this isn't about doing Bayesian updating with PDFs instead of just assuming the probability of attack is a half. It's about whether it makes sense to apply any sort of expected utility framework when, as fifty people have pointed out to this point, we have no f.ing clue how to calculate the ex ante probability of an attack of a given scale. Is a diffuse prior correct? Nobody knows. Maybe we should come up with a prior distribution over prior distributions? Doing so assumes that we have a set of potentially correct models which we can somehow aggregate to arrive at the truth. Is that true about our current knowledge of terrorism? No.

Also, we know from a million lab experiments that humans do not process information or make decisions that way. So Posner and Becker's calculations don't correspond to any consistent real-life information, and aren't even a model of how humans make actual decisions.

I think the question of whether pre-emptive war can ever be justified is answered by resorting to a simple example. Hypothetically, if North Korea was physically placed were Cuba is, there would no argument that preventative war was justified to prevent North Korea from acquiring nuclear weapons.
However, Judge Posners argument makes an assumption that would be disputed by many of those who would never agree to pre-emptive war. Those disputants would argue that any cost-benefit analysis should not be limited to the costs and benefits accrued to the U.S. but rather, should also include costs to nation being attacked. I disagree with the position I have just stated but in order for Judge Posner's analysis to address those who disagree with him, he would have to explain why we should only consider the costs and benefits to the aggressor nation.

The notion as I understand it is not that someone has actually and effectively constructed a real world model into which inputs can be dumped and outputs cranked (pace Wassily Leontieff and his epigoni), but that the description (model) is a useful way of describing what goes on in the real world, useful in terms of understanding how decisions are made and considering what can be learned from this.

Well, if you read Judge, Justice, Prof, Lord (or whatever the fawning 2Ls are calling him nowadays) Posner's body of work--or at least the parts of it I was forced to read in law school-- and this little gem today, he obviously thinks he has the magical economic formula that tells us when a preventative war is cost-effective, how the British and French could have prevented World War II (Of course, why wait for the invasion of the Rhineland, they should have just poisoned Hitler's dinner at Landau).

He has an economic solution to every possible legal problem. I'm sure that if you asked him if he considered the electric chair to be cruel and unusual punishment, he would want to know if the chemicals for lethal injection cost more than the electricity to run the chair.

Cato R., Feder's point is that this technique is useless in the real world of issue politics and corporatist corruption. In fact the whole post reads like a naive make-anything-up-at-all exercise to support the Iraq war. Because isn't this at bottom a fairly banal mechanistic justification for vigilantism in general?. But it's worse than that, isn't it. In this case, brutal punishment is, in indisputable fact, collectively visited upon some large number of innocents, as well as some fraction less than one of the the presumptively guilty parties.

It's curious that the next logical step for an integrity based analysis of the situation isn't in fact taken: compare the actual outcome with what the model would predict given the inputs available at the time. We've got all the information necessary to evaluate the usefulness of this scheme.

But what if the danger of attack is remote rather than imminent?

Given that the Baathists running Iraq were routinely firing missiles at our pilots in the no-fly zone in the leadup to the war, I don't see the point in having a debate over whether the threat from Hussein was 'imminent' or not. As far as our pilots were concerned, attacks were already occurring by the time we went to war. In fact, the Iraq War Resolution states exactly that -- that is that Hussein had fired "thousands" of missiles at our pilots in the leadup to the war.

So shouldn't the question be -- after 9/11 -- were we going to go to war against other people who were attacking us or were we only going to go to war against Al Qaeda?

I would suggest that anyone who believes that we should have given Hussein a pass because his missile attacks on our pilots were not serious attacks to justify a war is someone who has learned nothing from 9/11. In fact anyone who makes that argument, in order to be logically consistent, would have to also argue that we didn't have good grounds for responding to the South's attacks on Fort Sumpter (which resulted in zero Union casualties) or that we didn't have good grounds for going to war against Hitler because Germany hadn't killed any Americans at the time Hitler declared war on us.

'War is the continuation of policy by other means'. War (for historians) does not occur in an historical vacuum. OK, so the fig-leaf changes through time - as it must; and so 'preventive' or 'pre-emptive' may now be the excuse offered.But to understand any actual, historical, war: 'cui bono?'What is the policy now being *continued* through the means of war?

Richard,

I'm so pleased that you've started a blog, but I'm so disappointed that you've made some semantic mistakes that make your argument superfluous.

As expressed elsewhere. Self-defense does not mean attacking first - you can have a pre-emptive act of self-defense (see many abuse cases) but more often, in the international arena, it means you've been attacked and are attacking back.

Further