Mother of All Tail Risks

Tail risk, according to “Investopedia”:http://www.investopedia.com/terms/t/tailrisk.asp, is

bq. A form of portfolio risk that arises when the possibility that an investment will move more than three standard deviations from the mean is greater than what is shown by a normal distribution.

Got that?

There’s another, more colloquial meaning to “tail risk,” which is simply the possibility of having a freakishly bad day. There’s a “book about this”:http://www.amazon.com/Black-Swan-Impact-Highly-Improbable/dp/1400063515.

Lately, we’ve been reminded that bad events deemed vanishingly unlikely actually might not be. So unlikely, that is. The nation’s supposedly uncorrelated residential real estate markets “all started plunging at once”:http://www2.standardandpoors.com/portal/site/sp/en/us/page.topic/indices_csmahp/0,0,0,0,0,0,0,0,0,1,1,0,0,0,0,0.html. All those “credit default swaps”:http://www.financialweek.com/apps/pbcs.dll/article?AID=/20080218/REG/794188688 suddenly came into play at the same time. Two satellites collided. Two SSBNs collided.

A continually iterated potential hazard with potentially unknown correlated features is, over time, less improbable than it might seem. “Stuff happens”:http://www.totalwonkerr.net/1858/stuff-happens.

Now, per the happenstances mentioned above, you might imagine that the “financial crisis/recession/clusterf#@k to the poor house”:http://www.thedailyshow.com/video/?searchterm=clusterf#@k+to+the+poor+house is the biggest deal on the Obama administration’s plate, the issue whose management will be most consequential, for good or ill. And in the short-to-medium term, that’s awfully hard to dispute.

But In The Long Term?

Look, I know some people are into comet strikes, and the climate situation looks worrisome. But where I come from, the Mother of All Tail Risks is nuclear war.

My nightmare can beat up your nightmare, see?

From the “stuff happens” perspective, the long-term issue of greatest consequence is taking U.S. and Russian nuclear weapons off alert and keeping them that way.

Yeah, I know, STRATCOM is very clear that U.S. nuclear weapons are not on “hair-trigger alert”:http://www.armscontrol.org/act/2007_12/NuclearAlert. This seems to mean that U.S. weapons would not be launched unless A) the President ordered it or B) _[correction] someone acting in his stead ordered it after_ nuclear weapons had already detonated on U.S. soil.

I have just five or six small concerns about this. Little things.

A) How sure can we be that the President of these United States will be of “sound judgment”:http://www.amazon.com/Presidential-Leadership-Illness-Decision-Making/dp/0521709245/ref=ed_oe_p when the time comes? “Really?”:http://news.bbc.co.uk/2/hi/americas/3754199.stm “You’re sure?”:http://www.doctorzebra.com/prez/g_roster.htm “Really, really sure?”:http://www.usnews.com/articles/opinion/2009/02/19/iraq-wmd-bush-and-the-mind-of-saddam-hussein-hide-and-seek-by-charles-duelfer.html

B) After a single weapon goes off that incinerates our national leadership — origin unknown — what happens, exactly?

C) Is the Russian arsenal on higher alert than the U.S. arsenal?

D) Is the Russian early-warning system adequate?

E) Assuming the Russian arsenal is at the same “day-to-day” alert as the U.S. arsenal, and its early-warning system is adequate, how sure can we be that the “President”:http://www.businessweek.com/magazine/content/02_27/b3790047.htm, Prime Minister, and other key officials of the Russian Federation will be of sound judgment when the time comes?

F) After a single weapon goes off that incinerates the Kremlin — origin unknown — what happens, exactly?

Global financial crisis, global schminancial crisis.

Seriously, if you think none of the things hinted at above could happen — because it hasn’t happened yet, right? — try talking to someone in the financial sector.

If You Only Read One Part, Try This Right Here

Now that U.S.-RF nuclear force reduction talks are “getting underway”:http://www.armscontrolwonk.com/2235/obama-medvedev-statement, I’m as pleased as anyone to see the numbers coming down, but I would vote with both hands and both feet to make the numbers _go up_ if that’s what it took to get the G-d-damned things off alert once and for all.

7 thoughts on “Mother of All Tail Risks

  1. anon

    This statement is not true: “This seems to mean that U.S. weapons would not be launched unless A) the President ordered it or B) nuclear weapons had already detonated on U.S. soil.”

    Even if nuclear weapons had already detonated on U.S. soil, U.S. weapons could not launch without authorization from the President (or his successor in the national command authority.) There is not automatic launch for U.S. nuclear weapons, regardless of where we are in the attack cycle. The President can decide to launch before incoming warheads arrive, after the first detonation on U.S. soil, or after the attack on the U.S. is complete. The option to launch before warheads arrived was designed to ensure retaliation would actually occur, if the attack appeared large enough to disrupt command and control or defeat our retaliatory forces. In this day and age, its hard to imagine an attack of such size, and hard to imagine the President would respond tht way if the attack seemed small or the evidence were ambiguous. He has an option, he does not have to use it. If you don’t want him to use it, make sure he has lots of other options (and lots of other information) so he can make a different choice. This should be about extending decision time and offering alternative decisions, not about altering the posture of the weapons to eliminate an option you don’t like.

    And please, don’t sell me de-alerting by convincing me the President may not be of sound mind. This is just not a viable planning factor. He’s elected. We get to live that, even if he’s a nut job. Let’s assume there will enough grown-ups around him to manage the situation. Again, if this is a problem, find a real solution; don’t pick a pre-defined solution (de-alerting) and look for a problem for it to solve.

    Reply
  2. Josh

    Anon:

    Your message deserves a careful response. I will try to be more civil than you have been; even here on the wild and woolly Internet, it’s generally considered bad form to accuse a debating partner of bad faith (“pre-defined solution”). You’ve compounded that stumble by leveling your ad hominem anonymously, although I appreciate that you may not have had a real choice in the matter. Regardless, can we avoid this in the future?

    Furthermore, when speaking of solutions, pre-defined or otherwise, it generally helps if you support your argument in some manner. I’ve tried to do that. You, however, have not identified any good reasons to have U.S. and Russian nuclear weapons on alert.

    OK, on to the substance. You wrote:

    Even if nuclear weapons had already detonated on U.S. soil, U.S. weapons could not launch without authorization from the President (or his successor in the national command authority.) There is not automatic launch for U.S. nuclear weapons, regardless of where we are in the attack cycle.

    You’re right. I plead late-night blogging. To clarify: I’m not talking about a “dead hand” of the sort sometimes alleged to exist in Russia. What’s of concern is not knowing who will make the retaliation decision and on what basis. Whoever they are, they’ll have the means, motive, and opportunity to make the worst possible choice. I’d rather put more time into the process, to let cooler heads prevail. Or, to put it a little differently, this is about about extending decision time and offering alternative decisions.

    The same point goes equally to the issue of presidential inebriation (e.g., Nixon), drug addiction (e.g., Kennedy), health (e.g., Eisenhower or Johnson), and general soundness of judgment (many examples). It is simply wrong to say, as you have, that the fitness of the President

    is just not a viable planning factor. He’s elected. We get to live that, even if he’s a nut job.

    Or we all get to die, but hey, that’s our system, right? Well, pardon my snark, but obviously not. We make nuclear weapons themselves one-point safe. The launch decision should also be one-point safe. That’s a procedural variable, not a constitutional constant. The way we introduce that margin of safety is by putting more time into the process, creating more opportunities for the “grown-ups” to intervene.

    Reply
  3. anon

    So how’s this for a solution. Store the launch codes in New York. When the President wants to use them, put them in a car and send them down the New Jersey Turnpike. This will inevitably delay launch by at least 12 hours, as the codes will get stuck for hours between the toll plazas in Delaware.

    There is nothing automatic or inevitable about the launch of U.S. nuclear weapons. It takes a Presidential decison, and its hard to imagine any President doing this in less than dire circumstances. No one is so crazy as to launch hundreds of warheads when there’s ambiguous evidence of a possible incoming attack of one or a few missiles. The adults in the room would all advise delay. The prompt launch option was designed to ensure that weapons got off before they were destroyed or C3 was brought down by a massive incoming attack. A few incomin warheads would not do that, and the President would know that, and he could choose not to launch. You could argue that he no longer needs this option, since the scenario has slipped away with the end of the Cold War. But, maybe instead of messing with the technical capabilities of the missiles, you could find organizational or communications solutions that would ensure a more deliberate decision process. Give the President more options, so he’d be less likely to choose the one you fear.

    If you don’t think de-alerting is a pre-determined solution to an ever-changing problem, you must be new to the debate. In the early 1990s, the problem was that Russia might not have central command and control over the launch of its missiles, particularly those based on submarines or in Ukraine. In the late 1990s the problem was that Russia might not have enough information from its collapsing early warning system, and might launch when it didn’t need to. (Notice that neither of these problems applied to the U.S; we were just supposed to de-alert as an example, and hope Russia followed suit. The Russian response, they saw dealerting as a plot to disarm them.) The problem changed again in the 1990s (although some are still worried about Russia’s early warning system, it seems to be healing itself). Now the problem is that Crazy George or Crazy Vlad might launch weapons, just ‘cause they can. Oh, and we’ve recently added a new reason. Now de-alerting is seen as a signpost on the path to a world free of nuclear weapons. It reduces their salience, and, its a great step to take if we want to show others we are serious about climbing the mountain.

    Pick your problem, its the same solution. And never mind that the solution might cause problems of its own (re-alerting races, high value targets known as storage sights, etc.) Or pick a problem, and find a solution that really addresses a real problem. Accidental or unauthorized launch of U.S. missiles is not a real problem.

    Reply
  4. Josh

    Anon:

    I like your suggestion about New York. If the Russian launch codes are relocated to St. Petersburg at the same time, we’ll be sitting pretty.

    Yes, I am new to this debate, having never staked out a position on it before. So it’s actually possible that these are my considered views about the current and future situation. And I think there are answers to the objections you raise, mainly having to do with maintaining survivable second-strike capabilities on very low alert, as the UK and France do. But don’t quote me one that one — I haven’t penned the full manifesto yet.

    Glad to know that wrongheaded launch of U.S. missiles is not a problem. Is that also true of Russian missiles? Just asking.

    I guess Presidential advice and judgment are really very sound after all, basically because you’ve said so. Perhaps we were just imagining that we had a President who opted to invade a country in order to forestall a nuclear threat that turned out not to be there. Just for example.

    Following your example, I will further assert that whatever problem is addressed by having nuclear forces on day-to-day alert is, in fact, not a problem. Or I would, but you still haven’t told me what that problem is.

    Reply
  5. anon

    “Glad to know that wrongheaded launch of U.S. missiles is not a problem. Is that also true of Russian missiles? Just asking.”

    It depends on who you ask. Russia says it is not a problem, that their command and control structure is secure and roboust. They admit to some weaknesses in their early warning network, but it has improved in recent years. Bruce Blair says it is a problem, but his reason for why it is a problem has shifted over the years, and some of his problems has gone away. But, if it is a problem, particularly due to the early warning system, then address the problem. RAND published a book a few year back that proposed a number of cooperative measures the U.S. and Russia could to strengthen Russia’s early warning system. The JDEC was one of these proposals, as well has ideas about how we could help them fill the holes in their satellite system.

    Also, its hard to see how we can solve problems in Russia’s systems by taking U.S. systems off alert, unless you are suggesting that they’d follow our example. Except they’ve said they won’t and that by proposing they take their weapons off alert, we are suggesting they disarm. Again, there is history in this debate, stretching back more than 15 years.

    “I guess Presidential advice and judgment are really very sound after all, basically because you’ve said so. Perhaps we were just imagining that we had a President who opted to invade a country in order to forestall a nuclear threat that turned out not to be there. Just for example.”

    This is not a relevant example, unless you are arguing for the complete deactivation of nuclear weapons. Yes, Bush invaded Iraq for no reason, but he did not do so promptly. It was a delayed response. If he had done it on Sept 12, as Cheney and Wolfowitz wanted, it would be relevant. Prompt attack with ambiguous or missing information. But he waited until March 2003. In the interim, he built his case for the attack (or made it up, if you will), and pretended to play nice with the international community (let the inspectors go back to look, presented at the U.N., etc). If he had wanted to use nuclear weapons, and they had been off alert, he could have used that long time of playing nice as the time he needed to restore nukes to a useable status. The downside to this, of course, is that Russia (and maybe China) would have known we were re-alerting, and probably would have reacted in kind, which could have created a destabilizing re-alerting race (think World War I). This is one of the problems caused by the de-alerting of nuclear weapons. If you can’t monitor their status, then the steps you take are meaningless (like de-targeting), and if you can monitor it, you can detect re-alerting, which may cause a problem in itself.

    There’s another problem. If you are worried about terrorist attack or theft of nuclear weapons, then the safest place for them to be is on a missile in a silo. These things were built to withstand nuclear attack, so theft is not going to happen, and they are widely dispersed, so it would take a massive attack to destroy them. Take the warheads off, and put them in central storage areas, and you’ve created very attractive targets for small attacks, and attractive targets for theft (lets hope the security is better than it was at Minot). Again, by solving a problem that does not exist (accidental or unauthorized launch of U.S. weapons), you may create greater vulnerabilities and problems.

    So, what problem are you trying to solve? Does this problem really exist? If its a real problem, are there more direct ways to address it (better decision processes for the crazy George scenario, which is not the same as the Iraq war scenario, and helping Russia improve its early warning capability). And, if one wants to be truly analytic, even if your solution does address a real problem, might it create costs and risks that are greater than the risks from the problem you just solved?

    Now, you could also argue that, in the absence of the Cold War threat of a massive, bolt-from-the blue attack, neither side needs to maintain this posture anymore. Probably true. The British have adjusted their forces to this reality. We could to. But that’s a different argument. Its says this posture is a symbol of something that doesn’t exist anymore, and we don’t need it anymore. But, before you change the posture, you need to assess the risks of doing so, and compare them with the benefits of doing so. If the benefits are mostly symbolic, and the risks (re-alerting race, creating attractive and vulnerable targets) are real and measurable, then maybe your solution should also be symbolic, so you don’t create the risks. Or, if the risks are manageable, then proceed and change the posture of U.S. nuclear weapons, as long as you also implement the measures you need to mitigate the risks.

    But don’t tell me that the President will launch promptly just because he can, and don’t use the Iraq war as an example. It was a delayed response, not a prompt response, and the President chose to delay, even though he could have lauched promptly (he even delayed the attack on Afghanistan by about 3 weeks…)

    Reply
  6. Josh

    Anon:

    Thanks for a useful and stimulating exchange. Sorry if my impulse to sarcasm has gotten the better of me along the way.

    Just a quick response to your lengthy last one, perhaps shorter than it deserves.

    First of all, I cannot speak for any version of Bruce Blair. Let’s not bring in baggage from old debates with other people.

    Also, its hard to see how we can solve problems in Russia’s systems by taking U.S. systems off alert, unless you are suggesting that they’d follow our example. Except they’ve said they won’t and that by proposing they take their weapons off alert, we are suggesting they disarm. Again, there is history in this debate, stretching back more than 15 years.

    That’s a good question. On one hand, there is the example of the Presidential Nuclear Initiatives, which did involve unilateral and reciprocal dealerting of tactical nuclear forces. But depending on exactly what one means by dealerting, there may be reasons this model wouldn’t apply well to strategic forces in the absence of transparency measures / verification measures to build confidence that it was, in fact, bilateral and in effect. Or it might simply be hard to put into effect in the same manner. I still need to ponder it a bit, but am inclined to believe that maintaining a survivable deterrent at sea overcomes these issues. However, there are substantial benefits from creating a symmetrical situation, establishing the means to verify it and have confidence in it, and having a legal framework to govern it.

    Thus, I am strongly inclined to suggest that bilateral dealerting should be at the center of the negotiating agenda. Sorry if this wasn’t clear, but that was the intended point of the entire post.

    As an aside, you equate dealerting with disarmament (at least for the Russian side). Clearly, they aren’t the same, as emphasized in the last paragraph of the post. I’m quite pessimistic about the prospects for disarmament. In fact, I think we’ll be living with nuclear weapons for the rest of our lives, and our children’s lives, and so on and so forth. And that’s the point, right there. Whatever small chance exists of something going wrong is compounded over time. So that chance had better be vanishingly small, if it can’t be made zero. And we should not assume that it’s as small today as we would like to think it is; cockeyed optimism just has no place in these discussions. The Minot-Barksdale incident and its sequelae should be vivid indicators of the nature of the problem.

    And yes, I do think that grave Presidential errors, failures of advice, and failures of intelligence are the norm, not the exception, and this applies across all domains. We can’t make arbitrary exceptions; this is simply reality. For example, if JFK had taken the advice of his advisers concerning intervention in Cuba, it’s now pretty clear that we would have unexpectedly found ourselves in a nuclear conflict. It came close enough as it was.

    Nor does the nature of strategic ISR alter the fundamental picture; there have been too many snafus and near misses on both sides to believe that we have effectively perfect situational awareness, and always will.

    Ask the Japanese if the North Koreans launched a big rocket on April 4…

    With the passage of even just a little time after a false positive signal is received, it will become clear that an attack is not occurring — especially if we have confidence that the other side’s forces are also not on alert, which in turn will dramatically reduce the rate of false positives. And for third parties, we have a survivable deterrent.

    The nature of the problem can be shown with a fairly simple spreadsheet exercise. Where x is the chance of a Very Bad Thing per time period, and y is the number of time periods, fill out a bunch of cells with the following formula:

    =1-((1-x)^y)

    Make the time periods years and select whatever time horizon you expect for the Republic. The rest follows.

    Assuming I haven’t screwed anything up too badly in the formula, here’s what I get for a 300-year time horizon:

    When x = 0.01 bad events / year, the chance of a bad event (let’s call it z) is 0.95. Not good.

    When x = 0.001, z = 0.26. Still pretty bad.

    When x = 0.0001, z = 0.03. Better, but non-negligible.

    When x = 0.00001 (“five nines reliability”), z = 0.003. We should aim for this level of safety or better, and no, I don’t think we’re anywhere near this point today, per the above discussion and illustrations.

    But, before you change the posture, you need to assess the risks of doing so, and compare them with the benefits of doing so.

    Actually, the analysis should be symmetrical. In other words, we need to assess the risks of both action and inaction.

    I point this out because you still haven’t indicated what risks are reduced (or, if you like, benefits gained) by having the strategic nuclear arsenals of the U.S. and Russia poised to launch on a moment’s notice.

    Reply
  7. anon

    I don’t know where to start, and I really don’t want to spend more of my weekend on this, but I have to get picky on a few points. You may complain that I’m playing with semantics and words, but so be it.

    First, I didn’d say that de-alerting was the same as disarmament. I said the Russians have said that, in the past, when we have suggested bilateral de-alerting. They argued that, if you removed their ability to launch promptly, you left them vulnerable to a disarming first strike. Ergo de-alert = disarm. Russian view, not mine.

    Second, the PNIs did not take tactical nuclear weapons off alert (they weren’t on alert, at least in the U.S.) They did take the Minuteman IIs off alert, in anticipation of their elimination under START, but for the tacs, it was not de-alerting, it was deactivation and removal from deployment. We removed all our land-based nonstrategics from Europe and Asia and from surface ships and attack subs. Removed them, brought them back to the U.S., and dismantled many. You could consider this de-alerting, in the extreme, but it really isn’t the same thing. Also Gorbachev and Yeltsin promised to reciprocate, but absence any monitoring or verification, we rely on ever-changing Russian public statements to keep up with what they are doing (or what they may reverse, every time they threaten to put nukes back in Kaliningrad.)

    I like the unilteral-reciprocal model of arms control. It has strengths, and weaknesses. Bush even wanted to use it before he agreed to sign the Moscow Treaty. But using it as a quick way to withdraw and eliminate old and obsolete weapons that no longer have a role in your security strategy is different from using it to alter the posture of the weapons you still plan to rely on. It works when you agree to do something you wanted to do anyway, it may not work when it forces you to do something you would never do on your own. And this is where de-alerting is for the Russians. They would never do it on their own because they don’t think they have a problem with alert weapons and they think that proposals to de-alerting their weapons are a trick to disarm them.

    Enough

    Reply

Leave a Reply

Your email address will not be published.