Bryan Caplan  

Chickens for Peace

Richter Back in Print... Is the Ruling Class an Elite?...
David Friedman's dinnertime conversations with his son Bill inspired the following hypothetical:
Imagine that someone in our future is equipped with a device capable of delivering packages to the past. He makes a list of thirty or forty of the most influential people in the U.S. as of (say) the 1850's, prepares for each a package of history books, and delivers the package to the recipient's desk a week or two before some prominent natural event, such as an earthquake or eruption, is due to occur.

Each package includes a dozen identical color photographs and a cover letter. The letter predicts in detail the event about to occur and explains that the package is being sent in the hope of preventing a very bloody war. The photographs could not have been produced with mid-19th century technology; the hope is that they plus the prediction will be enough to persuade at least some of the recipients that the package really is from the future. What happens?
Father and son are both pessimistic:
Bill's guess was that the deep South states would respond by immediately seceding. My reaction--not inconsistent with his--was that what the intervention has created is a high stakes game of Chicken. Leaders in the North can tell those in the South that they might as well surrender now, since the alternative is a long and bloody war that they will lose. Leaders in the South can argue in response that the North, knowing what the cost of the war will be, will have to back down and let them go.
David adds:
It could make the plot of an interesting novel. If I were writing it--not likely to happen--I would be inclined to show the intervenors from the future as naive do-gooders who take it for granted that if only both sides had known, the war would of course be averted.
I tremble at the thought of correcting the microeconomics of a pair of Friedmans.  But here goes.  In the real world of bluff and counterbluff, it is the mixed strategy equilibrium of the Chicken game that's relevant.  In this mixed strategy equilibrium, the probability that each side plays War falls as the payoffs of (War, War) fall.  Since you only get (War, War) if both sides play War, the equilibrium probability of the bad outcome rapidly falls when the worst case scenario gets uglier.  The punchline: Basic game theory is on the do-gooders' side.

But wait - there's more.  Not only does basic game theory support the naive conclusion that pessimism about war is good for peace; so does basic history.  The standard story of World War I is that both sides thought they'd be "out of the trenches by Christmas" - and therefore blundered into carnage.  The standard story of the Cold War is that both sides realized that nuclear war would be devastating.  So they kept their fingers off their respective Buttons, and mankind weathered the storm.  You've probably seen the same mechanism in your own life: Deadlocks with bad consequences are a lot more likely to be resolved via compromise than deadlocks with trivial consequences.

I'm a pacifist and an economist.   And as far as I can tell, pacifists' idealistic efforts to teach the world - and especially the young - about the horrors of war are eminently realistic.  Reading All Quiet on the Western Front and Barefoot Gen won't turn every child against warfare.  But the more pessimistic the average person is about the effects of war, the more peaceful our world is likely to be.

Comments and Sharing

COMMENTS (13 to date)
AMW writes:

Your analysis suggests an alternative plot: the do-gooders send back photos from the South well after reconstruction, showing how prosperous the region is able to be without slavery. Their hope is to woo the Confederates into foregoing the war and embracing the progressive values of the late 20th century. This, of course, makes the Confederates feel that the war can't be all *that* bad, if they [or their progeny] will come out of it so well on the other side, even if they lose.

Hijinks ensue.

Gian writes:

An unjust court may be more evil than war.

Solzenitsyn (Red Wheel-vol 1).

JLA writes:

I'm not sure if game theory is on the do-gooders side - war is a repeated game, so doesn't the folk theorem imply that we can have an equilibrium where both sides fight some of the time?

agnostic writes:

Game theory favors no side a priori -- that assumes that you've picked the right game as your model for how humans behave in some context, and that can only be judged by looking at what the evidence suggests. So it's always an empirical matter.

In this case, you've picked the wrong game. How do we know? Because aside from the very recent Cold War, most large-scale wars were fought -- often for decades or over 100 years.

Is that because they under-estimated how ugly war could be? No way. Those involved know from personal experience, second-hand from relatives and acquaintances, and whatever history they know, that any given war can quickly devolve into hell -- raping and pillaging, occupation by barbarian hordes, starvation through blockades, impaling, crucifixion, etc.

They're not like some naive boy who thinks that the worst result of a head-on collision in chicken is a fender-bender and a few stitches. So a picture from the future of how nasty it'll be won't change their expectations much. They already knew how brutal war can be.

In fact, what if the soldiers are playing a different game? They're the ones who march off to war; they don't have to obey orders. What if sending a picture that shows an even uglier war than they'd expected would make them *more* like to run off into battle? Then the pacifists' good intentions would backfire big-time.

I'll explain an example of such a game in the next comment, lest this one get too long. But even the abstract possibility that such an anti-chicken game could exist should give the pacifists' pause.

agnostic writes:

This is like a coordination game with a mixed strategy. I'll follow the graphical approach from Chicken in Wikipedia. I call it Prove Your Worth Through Combat.

The basic idea is that potential warriors, when deciding whether or not to march off to war, will do so if it will honestly signal their skill, bravery, devotion, etc. (Or perhaps earn them a place in Warrior Heaven, if they believe in the supernatural.)

Honest signals must be costly in order to not be faked. Therefore a potential warrior is not going to go hog-wild if he expects the other side to have a low chance of also going hog-wild on the battlefield. That would be like shooting fish in a barrel, which any wimp can do, and thus earns him no recognition for skill, bravery, devotion, etc.

However, the more certain he becomes that the enemy will go hog-wild, the more convinced he is that going hog-wild himself would score him points among allies, potential mates (warrior / "men in uniform" groupies), the Gods and spirits, etc.

Referring to the graphical set-up of Chicken, let X and Y be potential warriors (we ignore those who are of low quality in arrow-shooting skill, strength, bravery, etc., since they won't march off to war in any event). Let x and y be the probability of going hog-wild on the battlefield, where 0 means they'll just sit there and take whatever you deal out, and 1 meaning they'll certainly act like a juggernaut.

agnostic writes:

Exactly contrary to Chicken, y is an increasing function of x, and x of y as well. From Y's point of view, if X has a 0 chance of going hog-wild, Y will see no point in trying to prove his worth by battling X and will stay home. If X is certain to go hog-wild, Y is also certain to do so. So, y rises from 0 to 1 at some point x = p, rather than fall from 1 to 0 as in Chicken. Similarly for X's point of view.

Further, the value of p will be pretty close to 1 -- if it were close to 0 or even 0.5, that means your opponent is not very likely to go hog-wild, and you'll stay home and wait for a battle with someone that would prove your worth.

So there are 3 equilibria as before, but now they are (0,0), (1,1), and (p,p), where (p,p) is pretty close to (1,1) anyway.

Sending them a picture from the future suggesting the outcome will be bloodier than they'd expected will push the (p,p) point even closer to (1,1), or mutually assured destruction. You've given all those young males yearning for something big and serious to prove that they're a skilled and brave man, rather than an inept and cowardly little boy, an even greater conviction that going hog-wild in battle will give them the glory they're after.

agnostic writes:

Why is this like a coordination game rather than Chicken? For the same reason that boxing matches are between roughly equals, rather than a maniac chasing after a coward. Who would agree to such a match?

If the war you chose to send a picture to was one where the soldiers were in it for completely mercenary reasons, your bloody portrayal might work. But in most cases soldiers see it as a rite of passage, a chance to prove their worth, and are more enthusiastic the greater the challenge that the enemy poses.

So the pacifists are the naive ones here, and their good-intentioned strategy would cause more bloodshed, not less. A better strategy would be to lie to the soldiers of the past -- tell them how few enemies will show up, how pathetic and cowardly they'll be, and how everyone back home will not give you guys glory and groupies, but rather look at you apathetically (not necessarily disgustedly) -- "big deal, my kid sister could've beaten those guys up."

Joey Donuts writes:


Does the probability of the glory occurring post mortem change the decision? Granted, some cultures, e.g. jihadis, post or pre mortem doesn't matter. In fact post mortem may have more value.

As an empirical matter the German invasion of Poland was an event where one side stood no chance, i.e. tanks against horses, but the Germans went ahead anyhow.

The Germans had a warrior culture. Do you think that if the citizens of Germany had seen pictures of Dresden, they would have pushed the fight? If the generals had seen the results of Stalingrad, do you think they would have pushed the fight? I think not.

Many if not most wars occur because one or both sides underestimate the costs of war.

AMW writes:

Many if not most wars occur because one or both sides underestimate the costs of war.

I think this works as an explanation for the recent Iraq War. If Americans knew how costly it would be, and how little they would get out of it, that war wouldn't have happened.

But I doubt the same can be said for the war in Afghanistan. Immediately following 9/11, I think the median voter would have accepted all the consequences of that war and then some, just for a chance to get back at Al Qaeda.

Nick writes:

Add "Johnny got his gun" to your book list

Mark Bahner writes:
I'm a pacifist and an economist.

I'm neither a pacifist nor an economist. But I've argued that Abraham Lincoln was not even a good president, let alone not one of the best:

Why Lincoln wasn't a good president

On 2/21/07 I made the following observations and comments:

1) The overwhelming majority of Southerners did not own slaves. Of the 1.6 million Southern families, only 384,000 (24 percent) owned slaves.

2) Of those 384,000 (24 percent), 88 percent owned fewer than 20 slaves. So only 46,000 families (out of 1.6 million, or about 3 percent) owned more than 20 slaves.

3) And only 3000 families (0.2 percent!) owned more than 100 slaves.

4) Therefore, if the U.S. government had paid EVERYONE who owned less than 20 slaves full market value--or even greater--that would have covered 97 percent of all Southerners. And if the U.S. government had paid everyone who owned less than 100 slaves full market value, that would have covered 99.8 percent of all Southerners. (The remaining 0.2 percent might think they got a bad deal, but who really gives a damn about a few rich slave owners?)

4) The total market value of the approximately 4 million slaves in the U.S. in 1860 has been estimated at $3.3 billion.

5) This can be compared with an estimate of the Civil War's *immediate* costs of $9.3 billion ($6.2 billion for the North, and $3.0 billion for the South) Not to mention the deaths of 600,000 mostly-young male soldiers (the value of their lifetime earnings alone could easily have exceeded $3 billion). Not to mention the pensions for veterans that easily exceeded the immediate costs.

Civil War cost estimates

Lincoln could have allowed the South to secede peacefully (except for the battle at Fort Sumter, where NOT ONE Union soldier was killed by a Confederate). He could have then urged Congress to make a reasonable offer to buy the freedom of ALL the slaves in the U.S. (including those in the Union!). If he had done so, it is certainly possible that the massive bloodshed, suffering, destruction, and expense of the Civil War could have been avoided.

You can say all you like about how the "Southern culture" precluded the acceptance of a reasonable offer (e.g., full market value, or even greater, for all owners of fewer than 100 slaves). But I'm virtually certain that men like General Lee knew early on--or even before the war--that the war was going to end badly. Any halfway rational analysis of the comparative wealth, population, and industrial might of the North versus the South would end with the conclusion that, at best, the South could have fought to a "victory" like North Korea's "victory" in the Korean War. There was simply no way the South could ever invade and conquer the North! The best the South could hope for would be to fight to a bloody draw that allowed secession. If Lincoln had embarked on a campaign to PERSUADE the South and the North (ala the Federalist Papers) that continued union with an ending to slavery was preferable to secession and war, he might very well have been successful. After all, he WAS a brilliant orator.

And y'all can exclaim all you want about what a great president Abraham Lincoln was. But the FACTS strongly suggest otherwise. He was a brilliant orator. But his role in the Civil War was NOT evidence that he was a great president. No truly great president would embark--and follow through!--on such massive bloodshed, suffering, destruction, and expense without first trying--or continuing to try--other reasonable options.

Like I wrote before, it isn't even close. Washington was by far the better president.

Rebecca Burlingame writes:

One of the most important legal concerns of some state governments in the U.S. entirely misses this game of chicken and its consequences: shared land holdings in divorce with no mortgage. Neither party has incentive by law to settle and so waits for the other to pass away. As William Easterly noted, it costs more to settle [some properties] than the properties are actually worth.

As you said, "Deadlocks with bad consequences are a lot more likely to be resolved via compromises than deadlocks with trivial consequences." By leaving the legal default set on trivial consequences, some governments and their legal systems default on one of their most basic obligations to the citizenry, and the game of chicken creates unneccessary wars in communities and families that can go on for decades.

BZ writes:

@agnostic, @joey:

I'm wondering why you both talk about war as if the batallions are some sort of emergent order. Did Hitler play no part in the WWII invasion of Poland? Was Lincoln shocked by all those drafted New Yorkers marching off to burn Atlanta? Did the Bushes sit idly by while Iraqis were being shot at by Americans in tan fatigues?

I guess I just don't buy the "they didn't have to follow their orders" and the "warrior culture" stuff. Every society has bullies -- but it takes a bully-in-Chief to march them off to mass-murder sprees.

Comments for this entry have been closed
Return to top