Risk and Reward

Jonah Goldberg had a post yesterday wondering if we shouldn't be doing more to prevent the possibility of an earth-destroying asteroid collision. I think we should. The odds of such an impact event in any given year are low, but not really all that low, and the downside consequences would be terrible. Then, as is all-too-often the case, he follows up by quoting an email from a dumb reader complaining about Ron Suskind's book The One Percent Doctrine:

The book knocked VP Cheney for saying that even a 1% chance of terrorists getting nuclear weapons merits serious US action to stop them. But multiplying the 1% by the severity of a nuclear terrorist attack in an American city makes Cheney's statement quite reasonable. Not how it played in much of the press, though.

First, a basic note on probabilities. It's quite right to say that faced with the possibility of a Very Bad Outcome we should take the VBO seriously even if the VBO is unlikely. Cheney's actual doctrine, however, was that faced with a one percent risk of a nuclear terrorist attack against an American city, we should respond as we would were such an attack a certainty. This is obviously daft. If there's a one percent probability of 1 million people dying, the expected value is that 10,000 people will die. If, conversely, there's a 100 percent probability of 1 million people dying, the expected value is that 1 million people will die. The idea that we should treat those values as if they were the same is crazy.

In the specific case of asteroids versus invading Iraq, though, the more salient difference is the downside risks of action. Cheney's doctrine, as he operationalized it, involved simply assuming that inaction courted risk whereas action did not. That, again, is crazy. The risk of spending more money tracking asteroids and starting a pilot research program to study how you might blow them up is that some money might get wasted if your research proves useless or if no asteroids come. The risk of invading Iraq is that hundreds of thousands of people will die, America will fail to achieve critical mission objectives in Iraq, America's international alliance system will end up in tatters, and, generally speaking, the United States will find itself retreating on all major foreign policy fronts. That's the trouble with starting speculative wars -- they're quite likely to go badly awry.

Comments

And let's not forget that installing a theocratic quasi-state in Iraq will actually be worse for American national interests...

Posted by Scott Lemieux | December 5, 2006 10:15 AM

This may be over-generalizing, but I wonder what effect the modern Limited Liability Corporation has had on the risk-analysis faculties of most of the GOP elite. People who see the biggest risk the world can provide in simply losing the money you have up front, are not people who should be running your foreign policy shop.

Posted by Tony v | December 5, 2006 10:23 AM

The one percent gag is an example of obiter dicta becoming dogma. (St. Paul's Epistles are chock full of such stuff.) Cheney's baritone gruffness masks a castrati percipience. He's an ijut, but he impresses people who take secondary sexual characteristics as badges of achievement. Goldberg swans around to find a way to turn Cheney's casual remark into a functioning mode of analysis while we suffer under the tyranny of the slogan. (Does "The Laffer Curve" ring a bell? Ring the bell, hell. It put us 9 trillion dollars into debt!)

Gray's Elegy mourned the village Cromwells, but unfortunately the village Boswells got to Washington and have made us pay, pay, and pay.

Posted by Jeffrey Davis | December 5, 2006 10:32 AM

you're basically right here, MY.

We should start now to put a plan in place, so that if we ever do spot an asteroid coming, we can install Chalabi on it straightaway.

Posted by kid bitzer | December 5, 2006 10:36 AM

"The risk of spending more money tracking asteroids and starting a pilot research program to study how you might blow them up is that some money might get wasted if your research proves useless or if no asteroids come."

And, really, given the past history of space and high-technology research in general, even if there's no asteroid collision in our or our grandchildrens' lives, it won't be money wasted: the necessary advances in astrophysics, telemetry, propulsion and guidance systems would be good of their own acrod, valuable for further space exploration and almost certainly have unforseen benefits for civilian technology and industry. Whereas for Iraq, well...you covered that pretty well. The externalities will reverberate in pretty much uniformly bad ways for quite a while.

Posted by jkd | December 5, 2006 10:40 AM

The chances of an asteroid colliding with the earth are somewhat greater than the chances of finding anything worth serious consideration in a Jonah Goldberg piece, yet Matt continues his mysterious quest for such a nugget despite all evidence to the contrary. Bizarre indeed.

Posted by Farinata X | December 5, 2006 10:59 AM

If an asteroid collides with the earth it will most likely prevent terrorists from detonating a nuclear weapon in New York City, so I don't see why we should necessarily think of it as a bad thing.

Posted by neil | December 5, 2006 11:01 AM

Watch it, Matt. Posts like this give aid and comfort to the killer asteroids.

Posted by old guy | December 5, 2006 11:14 AM

"We should start now to put a plan in place, so that if we ever do spot an asteroid coming, we can install Chalabi on it straightaway."

Maybe we'll see "drop an asteroid on it" as one of the Baker-Hamilton recommendations on Iraq.

"Posts like this give aid and comfort to the killer asteroids."

If we don't give more tax cuts to the top 1%, the killer asteroids will have already won. With the tax cuts, America's investors will be free to invest those proceeds in asteroid-destroying private companies.

Posted by Urinated State of America | December 5, 2006 11:29 AM

Excellent post but you are too kind to Cheney. With ironic understatement, you list only things that have happened as possible conceivable downside risks of invading Iraq. I appreciate the irony.

To be fair and balanced however, you should also list bad outcomes which had a 1% chance of occuring but which did not occur. For example, the invasion might have triggered an wave of anti US revolutions across all Arab countries (how about all Moslem countries ?). Didn't happen but is relevant. For another, if Iraq had had nerve gas, Zarqawi would probably have gotten his hands on some if and only if we invaded. My subjective probability on that was way over 1% and that would have been rather unfortunate.

Also another note on probabilities. One of the reasons that it is crazy to treat 1% like 100% (other than, you know the fact that they are different) is that 100%*100% = 100% but 1%*1% does not = 1%. The Cheney doctrine really requires Cheneyu's craziness and is totally different from the commenters distortion for this reason. Cheney calculated 1% chance that Iraq has a bomb project that the IAEA didn's see any trace of times 1% they could keep it hidden until completing a bomb * 10% Iraq is actively supporting terrorists * 0.001% that Saddam is crazy enough to give them the bomb = 100% sure that if we don't invade Iraq we will alll die. Or in general the chance that 3 1% risks all turn out bad is one in a million which isn't close to 100% at all.

Posted by Robert Waldmann | December 5, 2006 11:30 AM

OK, then, the chance of losing a big fraction of industrial civilisation due to global climate change is nonzero. And the stakes are gigantic. Do I hear a batsqueak from Goldberg?

*crickets*

*melting ice*

Anyway, applying the Cheney doctrine to asteroids would involve immediately shutting down the rest of the space programm and firing the entire astronaut corps into deep space with a brief to "search for asteroids of mass destruction", but not enough oxygen for their mission..

Posted by Alex | December 5, 2006 11:36 AM

You guys are nerds!

You think Cheney did any actual math when he came up with that doctrine? He just said it was too risky not to treat a nuclear terrorist threat extremely seriously, and someone came up with a catchy name for it. And in fact, many credible people consider a future nuclear attack a virtual certainty. Credible people on the LEFT think this!

The current situation in Iraq (my favorite euphemism is to call it a "situation") and the threat of nuclear terrorism are two not necessarily related things. Iraq could've gone swimmingly without altering the risk of nuclear terrorism at all. Or, it could've gone as poorly as it has while somehow stopping someone (like say Zarqawi) from going on to organize a nuclear attack someday. However, the likelihood of Iraq becoming a nuclear state somewhere down the road has been (barring any bizarre, unforeseeable complete withdrawal of American military influence from the country) dramatically reduced. Yes, Saddam was nowhere near getting one, but then the whole point of this entirely unaccountable doctrine is supposed to be about long-term thinking (i.e. longer-term than the Iraq sanctions were appearing to be). Arguably, it's not about ho badly Iraq is going, rather how we've taken our eyes off of Iran and the DPRK.

So an eventual nuclear attack somewhere on the civilized world is either a statistical certainty or it isn't. I think eventual war with Saddam was a statistical certainty. I think Bush/Cheney largely failing at whatever they set out to do was a statistical certainty. Cheney's One Percent Doctrine is fine except for the word "Cheney". And yes, 1% is too high a risk of getting hit by nuclear terrorism. But this quibbling over percentages doesn't advance the debate over what to do about it at all.

Posted by Bill | December 5, 2006 12:04 PM

There's also the issue that the moral cost of all 6 billion of us dying is much greater than the moral cost of 1 person dying times six billion--the marginal moral cost of each person isn't fixed. If you kill one person you've destroyed all the potential good that existed in the rest of that one person's life, but if you kill all people you destroy all potential that could exist in the entire future history of humanity--a history that could be near infinite for all we know.

Given that we've managed to prove the existence of neither extraterrestrials nor the afterlife, it's possible that a rogue asteroid could destroy everything of meaning in the entire universe. That would be way worse than, say, Stalin conquering America and throwing half of us into the Gulag--something that sounds infinitely evil, but on cosmic reflection doesn't have the same cardinality of horror as the potential extinction of all life in the universe.

Posted by Consumatopia | December 5, 2006 12:09 PM

Kudos to Consumatopia for crystallizing how the earth getting hit by a giant asteroid would suck way worse than anything else.

But where's the dualist in you? It would potentially also destroy all the evil that exists, too. Net/net it's a wash.

Posted by Bill | December 5, 2006 12:17 PM

Arguably, it's not about ho badly Iraq is going, rather how we've taken our eyes off of Iran and the DPRK.

Then perhaps the problem is not that the one percent doctrine is applied to the problem of nuclear attacks, but applied to the problem of nuclear attacks by specific actors, which completely glosses over the danger that dealing with one specific actor makes us more vulnerable to others. So if dealing with a 1% chance of attack by Saddam ends up boosting the chance of attack by either North Korea or Iran by much more than 1%, then the 1% doctrine ends up looking pretty damn idiotic.

Posted by Consumatopia | December 5, 2006 12:18 PM

And not only that, but the notion that a "rogue asteroid" could take out everything that's good and meaninful kinda answers once and for all the question of whther there is any meaning to anything. QED.

Posted by Bill | December 5, 2006 12:19 PM

But where's the dualist in you? It would potentially also destroy all the evil that exists, too. Net/net it's a wash.

Damn, that makes a scary kind of sense. If life is mostly pain, then might this asteroid not be an improvement?

Posted by Consumatopia | December 5, 2006 12:22 PM

"Then perhaps the problem is not that the one percent doctrine is applied to the problem of nuclear attacks, but applied to the problem of nuclear attacks by specific actors, which completely glosses over the danger that dealing with one specific actor makes us more vulnerable to others. So if dealing with a 1% chance of attack by Saddam ends up boosting the chance of attack by either North Korea or Iran by much more than 1%, then the 1% doctrine ends up looking pretty damn idiotic."

Which exposes the underlying point that it was never a doctrine at all, rather just an excuse for doing something specific when one feels like it. Anyway the idea that we couldn't statistically afford not to invade Iraq has been completely discredited as far as anyone who isn't, say, an omniscient being, can see.

Posted by Bill | December 5, 2006 12:27 PM

I gather the point of the one percent doctrine (though I havent read the book) is that evidence no longer matters in the decision making process. Cheney is just gonna do what hes gonna do (so yeah, it involves no math).

Jeffrey Davis, 80% of that is gibberish.

Posted by Anonymous | December 5, 2006 12:41 PM

Cheney's actual doctrine, however, was that faced with a one percent risk of a nuclear terrorist attack against an American city, we should respond...

...by attacking a nation-state that had a zero percent chance of giving nukes to Islamist terrorists that hated the leader's regime. That's where it all breaks down, of course: the scenarios they were throwing around were actually zero percent ones, and at least some of them knew it. "There's no time to wait for the inspectors to complete their work before Osama's close buddy Saddam nukes New York with a drone aircraft!" is much more in the realm of a gigantic purple possum collision with the earth, not an asteroid collision.

Jeffrey Davis, 80% of that is gibberish.

Ah, the 20% Doctrine. If even 20% of a comment is not gibberish...

Posted by mds | December 5, 2006 12:48 PM

I think Fafblog made a similar-enough point, years ago.

Posted by JP | December 5, 2006 12:51 PM

Jeffrey Davis, 80% of that is gibberish.

I'm reminded, obliquely, of a joke in "What's New Pussycat".

-Do you have a job?
-Helping the girls change at the Folies Bergere.
-How much a week?
-20 francs.
-That's not very much.
-Yes. But it's all I could afford.

Posted by Jeffrey Davis | December 5, 2006 1:27 PM

I see that nobody has answered Farinata's more important meta-point, which could fairly extend to all regular LA Times columnists, or at least the ones with whom I disagree.

Posted by norbizness | December 5, 2006 2:28 PM

If there's a one percent probability of 1 million people dying, the expected value is that 10,000 people will die.

Lord knows I'm no statistician, but isn't this wrong? If there's a one percent probability that I will win a million-dollar lottery, the expected value is not that I will win $10,000. It's all or nothing, and the point of the "one percent" is that the nothing is far more likely than the all.

Posted by live | December 5, 2006 3:58 PM

Does anyone bother with FACTS.

http://neo.jpl.nasa.gov/risk/

It shows the risk of all known objects hitting the earth. Obviously, it can't measure the risks of unkown objects hitting the earth but it is trying to map the heavens.

There are currently 9 objects that have a better than .01% chance of hitting the earth. In other words there are only 9 objects that are less likely than 99.99% chance of missing the earth.

Most of the high risk objects are fairly slow and fairly small. One object called 1994 WR12 is fairly big at at 110 meters in diameter but the first chance it will hit us is in 2054.

Should we worry about it hitting the earth? Yes. Is there a chance it will seriously wipe out civilization? No. I am not an expert and I suppose it could kill millions if it hit a major city. But if the earth is mostly water then the odds of it hitting land are significantly lower than hitting any land. The odds of it hitting a major city are absurdly low.

So, go ahead and panic if you want. But PLEASE, get your facts close to correct before you do.

Posted by neil wilson | December 5, 2006 8:58 PM