Now that, thanks to Kieran Healy and the Medium Lobster, we’ve all had our fun with Richard Posner’s case for pre-emptive war, complete with toy numerical example, it’s time for me to play straight man.
Posner’s starting assumption is consequentialism: that we should evaluate an action based on whether its probable consequences are, on balance, good or bad. I broadly agree with this, so I’ll try to explain why it shouldn’t lead to conclusions like those derived by Posner.
I’ll ignore a range of more complex objections and come straight to the first distinction learned by beginning students of the subject. Should we evaluate the consequences of general rules such as “don’t engage in pre-emptive wars” (rule-consequentialism) or should we evaluate each action on a case by case basis (act-consequentialism)
For perfectly rational decision makers, following the rules of Bayesian decision theory, the answer is easy and, in fact, trivial. It’s best to make the optimal decision on a case by case basis, and an optimal set of rules would be so detailed and precise as to yield the optimal decision in every case. Posner routinely assumes this kind of perfect rationality, which is why he doesn’t see any big problems with toy examples, or with claiming that this kind of reasoning can usefully be applied to improbable catastrophes with incalculable consequences.
There are two objections that can be made here
* Human beings are not perfectly rational and do not follow the rules of Bayesian decision theory
* Since war is a negative sum game, rational decision makers do not fight wars
The first point is well known, and has been demonstrated by the work of economists like Allais, and psychologists like Kahneman and Tversky (yes, we’ve got Nobel prizes on our side of the argument as well!).
But it’s the second point that’s really critical. Posner’s one-sided discussion of pre-emptive war, with no consideration of the opponent’s motives, is broadly equivalent to the Medium Lobster’s assumption that the other side of the proposed war consists of malign Moonmen (Anglais Casse has more on this). In this respect, and in his failure to consider the possibility that the pre-emptive war might end in defeat, Posner illustrates two characteristic biases of human beings on this subject.
* we tend to overestimate the malignity and underestimate the rationality of people different from ourselves[1]
* we tend to overestimate our own competence, and ignore the likelihood that our plans will fail
So, in relation to war, the case for adhering to rules that would discourage resort to war, even when a Posnerian calculation suggests that it looks like a good idea, is particularly strong.
In particular, experience suggests that the case against pre-emptive war is remarkably strong. Lots of pre-emptive wars have been planned, and many commenced. In many cases where pre-emptive war was suggested but not undertaken, the danger has resolved itself peacefully. In many cases where a pre-emptive strike has been planned, the intended quick knockout has turned into a long grinding war, often ending in defeat.
The example almost invariably cited here by supporters of pre-emptive war, and chosen by Posner, is that of Hitler. Posner suggests that a pre-emptive war to overthrow Hitler would have been an appropriate response to the reoccupation of the Rhineland. For the reasons put forward by many of the commenters on Posner’s blog, I don’t think this is clear even with the benefit of hindsight, and it certainly couldn’t have been justified on the basis of what was known about Hitler in 1936. An invasion of Germany, based on a violation of a treaty regarded by all Germans and many others as grossly unfair, would have been a highly dangerous undertaking and could easily have failed. Much the same would be true of an invasion in response to the annexation of Austria.
Certainly, the British and French governments should have learned more than they did from these events, and should have taken a stand at Munich. But the resulting war would have been one of collective self-defence like the actual one, though with a better starting point.
fn1. Of course, Al Qaeda and similar terrorist groups are indeed both irrational and malign. But such groups rarely control the kind of state against which a pre-emptive war can be undertaken. Afghanistan was an instance, but a rare one.
John Quiggin 
Good libertarian argument there Q. 🙂
If you give the government the power (in the hope that they do good) you have to consider all the likely possible ways that they will use that power (both good and bad). Simply assuming that the government will only act to make the world a lovely peaceful happy place with rainbows and Australian cricket victories seems odd to me… but this line of thinking is common.
This is true whether applied to war or welfare (or anything in between). But people seem to have selective faith in government (read: faith in politicians, bureaucrats and John Howard). Weird.
A final note is that, at least according to my calculations, the Iraq war did not pass the consequentialist test either… even on assumptions made before the war (existence of WMD, association with terrorists). My paper indicating this was published by that evil right-wing (sic) think-tank, the CIS.
JQ, I’m interested in your comment that because war is a negative sum game, rational decision makers do not fight wars. This seems to be a non sequiteur. Rational decision makers may fight wars if they think they will gain, but their opponents will lose more. Recent wars have not been particularly kind to the victors, but that doesn’t mean that your logic is correct. It depends on whether the “rational decision makers” have as their main concern the outcome for their nation, or the outcome for all people.
Again, Pr Q’s discussion collapses the vital distinction between:
preventive war against remotely adverse political states (which involves the use of main military force and occupation of enemy territory)
pre-emptive strikes against imminently threatening military forces (which are limited in aim and specified in target)
Pre-emptive strikes are simply instinctive self-defence against the sucker punch, eg Israels destruction of Iraqs reactors denied a dictator nuclear weapons. The Predator’s attack on Al Queada terrorists was also perfectly ok.
I certainly hope that Mr Howard does not think twice about cleaning out any terrorist nests that are being built in the Badlands to our North.
Preventive wars are a illegal and a fools errand for the reasons Pr Q gives. Iraq is an existence proof of that thesis.
I agree with Alex. Furthermore, it’s not at all clear that war is always a negative sum game.
The suggestion that we underestimate our opponent and overestimate our own abilities doesn’t necessarily mean that we will lose. Nor does it mean that we won’t extract benefit from military conflict greater than the costs we incur.
Incidentally, I would have thought the prime example of a pre-emptive war that was actually fought would be the 1967 Six Day War. I think it would be hard to argue that Israel didn’t gain from striking pre-emptively.
Note that Becker-Posner’s argument assumes utter selfishness on the part of the decision maker. No weight at all is given to the costs imposed on the enemy nation. (why not cut costs to zero forever by committing genocide – don’t laugh its been done before.)
This may be realistic, but it is not moralistic. And the degree of empirical realism that they employ can be ascertained by their failure to measure ex-ante with ex-poste costs of the Iraq adventure.
I am also surprised that market fideists like Becker do not at least bother to consider employing traders or futures markets to asess future costs of war. No doubt military security is an example of market failure, but the collective wisdome of the markets and cybersphere was definitely anti-war.
The whole discussion suffers from the Platonic fantasy of omniscience that afflicted know-it-all socialists, which economists like Posner and Becker have so ably criticised in the past.
Finally, what about the opportunity cost of military expenditure? This does not even get a look in!
The whole thing is a perfect example of the kind of reasoning that gives economists a Bad Name.
Fyodor at December 8, 2004 11:06 AM provides a neat illustration of the Just Pre-emptive Strike:
The IDF was even gracious enough to return the Sinai to its rightful owners, thereby protecting itself from the charge of territorial aggrandisement. But Israel is a rather special case illustrating the wisdom of William S Burroughs aphorism:
Subscription to the platonic fantasy of omniscience is a necessary pre-condition for donning the mantle of hegemon.
On the other hand, the value of a victory sometimes outstrips the cost of achieving it. Would Vietnam (the country, not the war) have been as united and stable a country had the Americans proven themselves to be easy-beats? Or is that too platonic?
Incidentally, I would have thought the prime example of a pre-emptive war that was actually fought would be the 1967 Six Day War. I think it would be hard to argue that Israel didn’t gain from striking pre-emptively.
yeah, i hear that invading and then occupying the gaza strip and the west bank has worked out very well for israel, without any unexpected negative consequences.
“it’s best to make the optimal decision on a case by case basis, and an optimal set of rules would be so detailed and precise as to yield the optimal decision in every case” . This can’t be ture of a probabilistic set of decision rules ; they are concerned with long run optimality. i.e. if you keep on using these sets of rules then you are quarantee of an optimal outcome . However the use of these rules in a particular instance may result in a disastrous outcome ( gamblers ruin ) , but in the light of uncertainty and inability to compute all possible outcomes they are the best athat you can do .
Snuh unearths the other questions that Posner fantasy ignores : for a fanatical Zionist pretty much any cost will be outweighed by the benefit of settling some more bits of the promised/holy/home land.
Likewise with Osama’s mob – the benefit’s of martyrdom, from their POV, are bloody hard to match.
Posner is free to imagine his set of costs and benefits – but so is everybody else – that’s why the decision can only ever be a purely political one, made by those in power and for whatever reaons they choose to reveal or hide.
Posner is dreaming up grossly simplified reasoning algorithms to cut the present USA regime some slack and take the heat out of the grave and gathering historical judgement that Cheney/Wolfowitz are war criminals.
I think it was Gabriel Garcia Marquez who declared that Britain and Argentina waging the Falklands War was like two bald men fighting over a comb.
This is about as definitive a simile as possible for the negative sum aspects of war. And certainly the British State and the taxpayers who funded it spent much more on that war than they will ever recoup.
But there is another aspect of war which this form of accounting overlooks. That is the socialisation of loss and the privatisation of profits aspect. Margaret Thatcher and the Thatcherites established a decade of political dominance in Britain. Victory in the Falklands anchored this dominance. Using it, the British state overturned the political economy of Britain.
A mundane example of this form of transfer of wealth by national project is any spectacle: Olympic Games, Grand Prix. The state bankrolls the event, entrepreneurs take the profits. Punters get to bask in five minutes of reflected glory and everyone is happy: a positive sum game!
snuh,
The Israelis could well argue that it is better that they occupy Gaza and the West Bank, than the Palestinians, Egyptians, Syrians and Jordanians occupy Israel. You can criticise Israel’s strategy all you like (and I have), but the Six Day War probably saved more Israeli lives than it cost. I doubt many Israelis believe that war had a “negative sum outcome” for them.
Katz, it was in fact Jorge Luis Borges who said ‘The Falklands thing was a fight between two bald men over a comb’ in Time Magazine on 14 February 1983.
Thanks Doc. Memory can be a slippery thing.
Alex says, ‘Rational decision makers may fight wars if they think they will gain, but their opponents will lose more’ and Fyodor agrees.
You both miss the point: this statement can’t be true of both sides. So at least one side must be acting irrationally. Therefore a country that opts for war is thumbing its nose at a statistic that should give it some pause: at least half of all countries that have chosen to wage war in all of history have done so irrationally. This proportion is higher when we consider only countries initiating the wars. Countries under attack tend to opt for rational gambles.
Not so, James. They could both think they will gain, but one (or both) will be wrong. However, this doesn’t make the initial decision irrational. A rational decision can often prove to be wrong in an actual case. The overall statistics for gainers and losers are irrelevant, because the outcome of a particular war depends on circumstances unique to the particular situation. For example, there may be a great disparity in technology, manpower and strategic advantage. The main reason why war these days is less successful for aggressors is simply that they don’t fight with the same ruthlessness as aggressors did in the past.
Alex
Ur I think the Interhamwe, the Lords Resistance Army , and the Burmese Military might disagree with your last post as a matter of ‘honour’ – about being ruthless I mean.
Probable consequences don’t have short-term payoffs exclusively, it all goes into the flux of what this is, like our genes go from individual expression into the stream of life. What were the consequences of World War I? They’re still playing out. The immediate beneficiaries and casualties are all gone, though some of the fortunes that were made, and many that were augmented, continue. Men died as soldiers before they reproduced as husbands, others stayed behind and had children with women who would have married and bred with the men who died before they could. The outcome of that is impossible to measure afterward, let alone predict beforehand. And yet it’s a real consequence. So the forecast gets truncated, reduced to its limits, and the artificial boundaries of “a war” and its arbitrarily curtailed consequences are talked about as though they were something real, instead of just a convenient piece of something real.
There are elements of the Crusades in the body counts of Iraq today, and of the destruction of Solomon’s temple 2500 years ago in the decimation of Fallujah. Those consequences were probable, actually inevitable, the way anything that happens can be said to have been a direct result of the circumstances which preceded it, but without extra-sensory perception they weren’t foreseeable by the human actors who initiated and prosecuted them. So they’re discounted, pragmatically, not because they don’t matter but because we can’t discuss them accurately.
Applying logic and game theory to wholesale death is dangerously close to capitulation, it’s like talking calm sense to a psychopath. It becomes a narrow tightrope between submission and chaos, very narrow, while the audience is armed to the teeth and provably murderous.
Meanwhile, sports fans everywhere debate the efficacy of the latest weapons systems and scorn each other’s choice of team and favorite players, gloat and cringe with the shifting tides of victory and defeat, while children die in the rubble of their ancestral homes.
I’m not saying waging war shouldn’t be discussed as a rational decision – consequence proceeding from cause – but it needs to be emphasized that turning real lives into statistics makes it much too easy ignore the inhumanity of the decisions that get made; selfishly in confusion and panic, or altruistically and calmly but with the freakish heartlessness of the sociopathic warrior.
We aren’t all soldiers, and we shouldn’t be lulled into thinking we are, or thinking as though we are.
James,
Alex is right on the (ir)rationality of war. Having imperfect information about your own strength and that of your opponent does not make you irrational. The decision may be wrong, but it may still be a rational decision. Remember also that many countries don’t have a choice about going to war as they’re not the aggressor. All it takes to produce a war is for the aggressor to believe (rationally or irrationally) that they can produce a positive outcome for themselves from the conflict.
If all wars had a negative sum outcome for both sides, then they would clearly always be irrational. However, it’s not at all clear that wars always produce negative sum outcomes for both sides. This is unproven by JQ, and moreover it is possible to provide examples that challenge the assertion. Therefore it is not always irrational to start wars.
One of the tests for the success and benefits of globalisation is that no two countries hosting McDonalds franchises have ever gone to war.
Given that sometimes the decision for war is sometimes rational, what does that imply about the effect of eating McDonalds on rational thought?
Readers might be interested in reading this brief chat with Karl Sigmund, in which he discusses various game strategies, including Tit-for-Tat, Generous Tit-for-Tat and Pavlov’s strategy. Also deals with indirect reciprocity and assessment hardwiring. BTW, if you’re not already an Edge subscriber, I recommend it. Notifies you each time a new edition is released.
http://www.edge.org/3rd_culture/sigmund04/sigmund04_index.html
“One of the tests for the success and benefits of globalisation is that no two countries hosting McDonalds franchises have ever gone to war.”
About three months after Friedman published this claim, NATO bombed Belgrade, a city which contains several MacDonald’s franchises
I guess we all agree that, at the aggregate level, war is irrational. If all parties acted rationally, war would be averted. In cases where the expected outcome could be foreseen, the likely losers would accede to the likely winner’s terms. Where a likely outcome couldn’t be foreseen, the sides would agree not to fight, on the same basis that a rational person doesn’t put large sums in poker machines.
I’m not sure about Alex, but Fyodor seems to be observing that, for some leaders, war is more like betting on the TAB. Everyone knows that the expected gain of the average participant is zero, but the shrewd ones expect to come out ahead because there are plenty of mug punters out there to subsidise them.
This is unobjectionable as an empirical statement. But as I understand him the Captain is making a normative argument, namely that the TAB model is an unsound basis on which to opt for war.
This relies in turn on a positive claim that war is not like recreational gambling. If the minimum TAB bet was everything you own, the recreational betters would disappear, and the remaining participants would be facing a much lower expected return on astute calculation. In fact you would only stay in the game if you were confident the others were deluded.
By the same reasoning, you would only rationally opt for war if you assumed the opposition was stupid or irrational. This assumption is rarely vindicated, so rational leaders should revise the assumption much more often than in practice they do.
John, that was Katz’s point. You must be in a rush today.
Five possibilities:
1. Globalisation isn’t rational
2. McDonalds isn’t rational
3. Milton Friedman isn’t rational
4. The Balkan War wasn’t rational
5. JQ can’t help but be rational
On second thoughts, it wasn’t exactly the point of the joke. But that’s just because were too many bases to cover.
OK James, I *think* I get what you are saying. But there’s occasionally going to be a nation that thinks it runs the TAB.
Oh, and another point is that occasionally (just occasionally) a nation will enter a war for altruistic reasons. East Timor might be a reasonable, although small scale, example.
James,
You’re suggesting that if all actors had perfect information [where have I heard this line before…?], the states that knew they would lose a war would avoid it, e.g. by surrendering what the victor wants. But that’s not very helpful given wars are highly contingent events with uncertain outcomes. Your assumption isn’t even simplifying it’s so far from reality.
The TAB analogy is also inappropriate – poker is a better comparison. You know the hand you have, but you only partially know the hand your opponent has. The situation is decision-making under uncertainty, not a near-random punt by an over-confident speculator.
It was Thomas Friedman, not Milton.
Also, while both sides can rationally decide they’ll win the war… only one side can be correct. 🙂 It is not lack of rationality to blame, but imperfect information (which is a fundamental part of life).
This goes to the use of the word rational, and I’ll note (slightly off topic) that economists generally use the term meaning something less strict that the lay use of the term, which is one reason for much talking at cross purposes between economists and non-economists.
Iraq: The USA’s great game of risk
There have been quite a few posts around the blogosphere in the last couple of days on the subject of the reasons, motivations and legalities of the Iraq invasion, and related topics. Since we discovered that there were actually no WMD still in Iraq …
Conservationists and conservatives
Don Arthur had an interesting response to my pieces on the precautionary principle and wars of choice1. Don correctly observes that this kind of argument can be used in opposition to reform, and is therefore inherently conservative. He mentions, as…