Reducing Alert Rates of Nuclear Weapons

Presentation by Professor the Hon Gareth Evans, Chair, ANU Crawford School Centre for Nuclear Non-Proliferation and Disarmament (CNND), to Side-Event sponsored by Switzerland and New Zealand, Nuclear Non-Proliferation Treaty (NPT) Review Conference Preparatory Committee, Geneva, 24 April 2013

To those not initiated into the mysteries of nuclear policymaking, maintaining close to 2,000 weapons on dangerously high launch-on-warning alert seems the ultimate absurdity of nuclear deterrence – more than twenty years after the Cold War, when political, economic and security relations at least among the five NPT nuclear weapon states, render deliberate nuclear attack virtually unthinkable.

De-alerting seems to be one of the most obvious and least difficult steps to take on the road to making the world safer and saner, with four reasons standing out as to why – on the face of it – this should be fairly straightforwardly doable.

First, the issue is confined to the U.S. and Russia – only their weapons are able to be launched within minutes of an alarm being raised that the country is under attack, not those of France, UK, China – or for that matter India, Pakistan, North Korea or (probably) Israel.  So there is nothing inherent in the theory and practice of nuclear deterrence which requires this alert status.

Second, having launch decision-times measured in minutes creates an extraordinary stressful situation for decision makers:  some people might be at their icy-calm best in crisis situations but all human experience teaches us that these are not situations in which, cool, objective, balanced judgement is guaranteed. Zbigniew Brzezinski at the Council of Foreign Relations last year gives a wonderful account of the pressures involved, describing a situation which arose when he was National Security Adviser in 1979:

I got a message from my military assistant, a general, who simply woke me up at 3:00 a.m. at night on the red phone and said, "Sorry to wake you up. We're under nuclear attack." That kind of wakes you up. And he adds 30 seconds ago, 200 Soviet missiles have been fired at the United States.

… I had three minutes in which to notify [the President]. During those three minutes, I had to confirm it in a variety of ways. And then he would have four minutes to decide how to respond. And then 28 minutes later … [if we decided] to respond with everything -- just let it roll… then 85 million would be dead and we'd be living in a different age

…Then the confirmations did not come in… within  two minutes, prior to me calling him on the third minute, it was clear that this was a false alarm. So I did nothing. I went back to sleep.

INTERVIEWER: And if the confirmation had been a little late, could we have had a problem?

BRZEZINSKI: We might have had.

Those decision times, and presidential response times, continue to this day. About the only consolation I have ever derived from all of this was the news that emerged three years that Bill Clinton, for several months of his presidency, completely mislaid the nuclear codes he was supposed to carry in his pocket at all times which means that during that period a US retaliatory nuclear strike could not in fact have been authorised even had anyone wanted to! 

Third, the basic strategic premise of the argument ­– that if you don't immediately use your weapons in a retaliatory strike you are likely to lose them – is fundamentally unfounded. Even if all the fixed site weapons and missiles of the U.S. or Russia could be destroyed in a surprise attack by the other, Russia would have more than enough mobile ICBMs and U.S. would have more than enough sea and air-launched missiles to destroy the other.

It is sometimes conceded by opponents of de-alerting that both countries’ second strike capability is robustly survivable against any imaginable first strike – who then use this to argue that the fears of de-alerting supporters about bad decision-making under pressure are completely ill-founded, because that survivability means in practice that a decision to launch within minutes of warning would never have to be made.  To which the obvious retort is  - Well if that is so, why maintain a high-alert status with all the catastrophic risks that come with it?

The fourth reason why de-alerting should be readily doable is that there is now a good answer to the last remaining argument of those who want to keep the present high-alert status alive, viz. that re-alerting, in a crisis or emerging crisis situation, creates a more dangerous and destabilizing situation than that involved in keeping weapons on high alert all the way along.

For the reasons spelt out by Hans Kristensen and Matthew McKinzie in their important recent UNIDIR paper, the re-alerting race argument really is a complete straw man.[1]

So we are left scratching our heads as to why, more than 20 years after the Cold War, this potentially catastrophic ultimate absurdity lingers on?   We explored this in the ICNND report in 2009, and were forced to conclude that it was because nuclear decision-makers in both Russia and the U.S.  are stuck in a Cold War time-warp, in which the only focus is on capability, not the much more positive story about intent; where the only scenarios that matter are the absolute worst-case ones, not those bearing any relationship to real world probability; and where the only language of analysis is arithmetical.

So we have Russia making the judgment that de-alerting its own missiles would leave them highly exposed – with the prospect of its retaliatory capability being further diminished by the future development of more effective U.S.  ballistic missile defence and long range precision guided conventional weapons.  The U.S. won’t move on de-alerting, not because it feels especially vulnerable, but because Russia won’t; and Russia sees this high operational readiness in turn as further reason to fear the possibility of a sudden U.S.  first strike.

So as General (Ret) James Cartwright and Ambassador Thomas Pickering put it in Global Zero testimony to the U.S. Senate in July 2012, “both U.S. and Russian forces are kept on quick-launch alert because the other side does the same.”  And as a recent NTI paper by Des Browne, Sam Nunn, Igor Ivanov and Wolfgang Ischinger stated, the process is now completely circular: “so long as Russia and the United States can launch hundreds of ballistic missiles on short notice against each other, both must maintain a similar capability”.[2]

The underlying issue keeps on coming back to mindset, what’s been described by Browne, Nunn, Ivanov and Ischinger as  “Cold War autopilot”. There is reflex acceptance of all the old arguments about the strategic utility of nuclear deterrence, with a complete unwillingness to accept the possibility of alternative explanations for such strategic stability as we have experienced; a reluctance to accept the extremefragilityof any peace based on nuclear deterrence, given the possibility of human error, system error, cyber sabotage and anything else that could go wrong; and either indifference, scepticism or outrighthostility to the argument that de-alerting is an important element in the larger disarmament context – an important step in reducing the role and salience of nuclear weapons in national security strategies, in enabling reduced weapon numbers, and ultimately their complete elimination.

It all reminds of my first official briefing, as a young Australian minster back in the early 1980s, on United States nuclear strategy. It was given to me, in the bowels of the Pentagon, by a man with a white dust jacket and pointer looking uncannily like Woody Allen. It was a dazzling account of the logic of nuclear deterrence and mechanics of mutually assured destruction, as both the United States and Soviet Union saw and applied it throughout the Cold War. The language was disengaged, technical, all about throw-weights, survivability, counterforce and countervalue targets. And none of it showed the slightest feeling for the millions real human beings who would be vaporized, crushed, baked, boiled or irradiated to death if ever a nuclear war ever broke out:  the scale of the humanitarian suffering was, after all, what made countervalue targeting so valuable in nuclear deterrence theory.

Breaking out of that disembodied calculus, that Cold War mindset, switching off that Cold War autopilot, is not going to be easy. While de-alerting looks at first sight to be a no-brainer, in fact as we said in the 2009 ICNND report it “may involve a process almost as complicated as numerical weapons reduction, needing to be operational and technical in character, equal to both sides and implemented in a phased way.” [3]

There are ways forward that quietly, step by step, could achieve the objective of taking at least the vast majority of weapons presently on launch-on-warning alert off that status: they have been mapped both by Browne, Nunn, Ivanov and Ischinger, and in more detail by Hans Kristensen/Matthew McKinzie in their UNIDR paper. A great deal can in fact be done short of formal treaty-level negotiations, with in effect parallel unilateral actions – made with a good understanding of what the other side is prepared to do each step of the way – being the key driver, as happened on a very large scale in the early 1990s, and in a strategic environment (in the former Soviet Union in particular) very much more fluid and uncertain than is really the case today between these two countries.

But that is going to require the right signals coming from the political leadership level, which has what has been missing until now, certainly from Russia, and even on the U.S. side where de-alerting has disappeared from at least the visible Obama agenda.  I don't know whether sustained further nagging from our sponsors today and the other states leading the charge on this issue in the UN General Assembly, and from the NGO community, is really going to make any real difference.  But the stakes here continue to be very high, and we have no choice but to continue trying.