In January 2018, Russia secretly launched a cruise missile powered by a small nuclear reactor at a military testing range in the northern region of Arkhangelsk. The test of this bizarre doomsday weapon was a failure—it landed in the sea just a few kilometres from the launch site. The test would have remained a secret, but in August 2019 Russian scientists attempted to lift the wreckage off the Arctic seafloor. There was an explosion—one powerful enough to be detected by monitoring stations in Finland, Norway and Sweden. Five scientists were killed and a brief spike of radiation was detected in the nearby city of Severodvinsk. Images on social media showed emergency service workers responding in Hazmat suits. The Comprehensive Nuclear Test Ban treaty Organisation, the body charged with detecting nuclear explosions, predicted that any plume of radionuclides from the accident would soon drift over monitoring stations in central Russia. Then those stations mysteriously stopped working. Viewers of the drama series Chernobyl might not have been surprised.
The story of how modern Russia found itself in a Soviet-style effort to suppress information about a nuclear accident is a story about the collapse of the post-Cold War peace. Vladimir Putin’s Russia is resurrecting Soviet-era nuclear weapons programmes, and covering them up with Soviet-era disinformation because Russia and the US are drifting back into the logic of the Cold War. Even worse, they are drifting towards the free-for-all of the early Cold War, before there were any restrictions on the terrifying competition of the arms race.
Neither nation has shown any interest in averting this outcome, despite the widespread expectation that Putin’s assistance to the Trump campaign would usher in an era of improved relations and the US president’s claim that the two men have “very good feeling for each other.” Although the Mueller investigation ended with a whimper, Trump has remained sensitive to allegations that Russia aided his election and possessed compromising information about him. A series of hawkish advisers, including the now-departed John Bolton, have been able to keep the president from engaging in the peculiar kind of deal-making with Putin that Trump seemed to relish with North Korea’s Kim Jong Un.
Instead, these same advisers have been able to continue the slow unravelling of the post-Cold War order as one by one, the arms control treaties that codified the end of enmity have fallen by the wayside amid allegations of cheating. That unravelling began many years ago when George W Bush walked away from the crucial Anti-Ballistic Missile treaty, but it has accelerated under Trump. Last August the US withdrew from the 1987 Intermediate-range Nuclear Forces (INF) treaty, the landmark Reagan-Gorbachev agreement that banned an entire class of nuclear weapons. The US also announced its intention to withdraw from the Open Skies Agreement, which allows countries to conduct unarmed flights over other countries’ military installations, and has accused Russia of violating its commitments to the testing of nuclear weapons.
The last remaining arms control agreement that limits the number of nuclear weapons that the US and Russia may deploy is the so-called New START treaty, signed in 2010, which restricts both nations to 1,550 weapons deployed on 700 delivery systems—missiles and bombers. In February 2021, the New START treaty will expire. US officials have already ruled out discussing an extension without adding China to the deal, while Russian officials have said there is not enough time even for a bilateral negotiation. When the New START treaty lapses, there will be no legally-binding limits on the two countries’ nuclear forces for the first time in nearly 50 years.
Instead, both the US and Russia are engaged in massive recapitalisation of their nuclear forces. In the US alone, this investment is estimated to exceed a trillion dollars over the next 30 years. While Russian statistics are harder to come by, Putin’s bizarre menagerie of science-fiction-like weapons cannot be ignored—as well as the nuclear-powered cruise missile, there are underwater drones armed with nuclear explosives, hypersonic gliders that can manoeuvre around missile defences, and a giant intercontinental missile that Russian media boasts could “wipe out parts of the earth the size of Texas.”
The personal connection between the two leaders is not the check on an arms race that it might have been were Trump not fixated on nuclear weapons as the ultimate trapping of presidential power. He discusses the bomb in the same way another old man yearning for lost youth might describe a fancy car—boasting of his “bigger and more powerful” red button. More fundamentally, the Trump ego kicks against any restrictions on the president and his nation, trusting in his own deal-making genius to ward off the apocalyptic dangers, as opposed to the agreements that have kept them at bay for nearly half a century. That is so dangerous that reasonable people might worry that Moscow and Washington are falling back into a pattern of hostility that led to moments like the Cuban Missile Crisis. To which, Donald Trump reportedly said, “Let it be an arms race.”
Armed and in your living roomThe oldest problem of the nuclear age was what the Nobel Prize-winning economist Thomas Schelling called the “reciprocal fear of surprise attack.” During the Second World War, both Moscow and Washington suffered a surprise attack—the Soviet Union by Hitler’s Operation Barbarossa, the US by Japan at Pearl Harbor. As nuclear powers in the post-war world, each remained fearful that the other might execute another surprise attack, or “bolt from the blue.” And each feared that, unlike in 1941, there would be no possibility of recovery.
Schelling explained complicated things using clear analogies. To make sense of the nuclear stand-off he pictured himself, awakened by noise in the dead of night, walking downstairs with a gun, and confronting a burglar who was also armed. Schelling realised the inherent danger in such a situation. Even if neither party wished the confrontation to turn deadly, it might very well end in disaster. Schelling might have preferred the burglar to leave, but he doubted the burglar would know that.
Schelling proceeded to think through the complications. “Worse, there is a danger that he may think that I think he wants to shoot. Or he may think that I think he thinks I want to shoot. And so on.” The morass of intentions, perceptions and incentives struck the economist as intractable. “Self-defence is ambiguous,” he dryly concluded, “when one is only trying to preclude being shot in self-defence.”
This may seem obvious when stated with Schelling’s lucidity, but it is the opposite of how most people think about burglars or nuclear-armed superpowers. Because at the root of Schelling’s conception is a very unusual idea. Anyone in 1958—when Schelling wrote his famous essay, “The Reciprocal Fear of Surprise Attack”—could understand that American vulnerability to Soviet nuclear weapons was a danger. But what Schelling realised was that Soviet vulnerability to American nuclear weapons was every bit as dangerous, not just for Moscow, but also to Washington. When neither side knows what the other is thinking, there will be a dangerous temptation for both to strike first—unless both sides know deeply that any first-strike will result in a ruinous retaliation.
Schelling’s idea was hugely influential. The Kennedy administration abandoned any effort to win a nuclear war, settling instead for a retaliatory capability that amounted to the “mutually assured destruction” of the enemy in the event of a surprise attack. The Johnson Administration opened talks with the Soviet Union on strategic arms limitations, talks that resulted under the Nixon administration in the Vladivostok accord, usually referred to as SALT I, and most importantly the 1972 Anti-Ballistic treaty (ABM), which limited each side to two missile defence sites, later reduced to one.
From Schelling’s perspective, and I think he was right, the most essential thing was to limit the deployment of anti-ballistic missiles or ABMs. ABM is a bit of jargon, but the term is worth considering. Today, we use the term missile defence—after all, who is against defence?—but analysts in the 1960s thought it was important to be clear about what the system actually did. ABMs shot at enemy missiles—and therein lay the problem. It would be far easier to shoot at missiles being launched by an enemy already damaged from a surprise attack. “What was worrisome was that ABMs might offer a strong advantage to a first strike,” Schelling wrote. “The idea was that ABMs might… work poorly against a prepared attack but well against a damaged retaliatory force.” There would be fewer missiles to shoot at since many would have been destroyed in the surprise attack and the launch of surviving missiles would be disorganised and scattered over a longer period of time. We ought not call them defences when, as Schelling had warned us years before, self-defence is ambiguous when one is trying to avoid being shot in self-defence.
Schelling considered the ABM treaty the only truly important arms control agreement. In the 1980s, he looked back with justified pride at his work and the work of others in that period, concluding that the 1972 treaty was “the end point of successful arms control.” While missile defences later attracted bipartisan support, he never wavered.
But there were always those who never accepted Schelling’s insight. The idea that the US might want Soviet leaders to feel secure in its nuclear deterrent was a heresy. From the beginning, hawks mocked “mutual assured destruction” as MAD.
The tension between these two schools of thought was on display within the Reagan administration. On the one hand, it negotiated some of the most important arms control agreements with the Soviets. Reagan even came close, sitting in a little cottage in Reykjavik in 1986, to agreeing with Gorbachev to eliminate all nuclear weapons. On the other hand, Reagan never really embraced the idea of living with vulnerability—hence his Strategic Defense Initiative (SDI) plan to place missile defences in space. His opponents derided it as “Star Wars,” just as his friends had called Kennedy’s policy MAD. But unlike Schelling, Reagan could not see any ambiguity in self-defence. “Wouldn’t it be better to save lives than to avenge them?” he asked. Gorbachev’s insistence on the United States also ending SDI research led Reagan to walk out of that Reykjavik cottage empty handed.
This fundamental tension was never resolved. The Soviet collapse allowed the US to negotiate a series of one-sided arms control agreements that greatly favoured the US with a more compliant Russia. These agreements allowed for progress in reducing its arsenal but also destroyed the foundation of self-restraint. Gorbachev and Yeltsin were amenable negotiating partners, but the effect of that was to leave American officials, pundits and experts feeling free of MAD and falsely believing US supremacy would last forever. Meanwhile, their disillusioned Russian counterparts bided their time.
A quiet rearmingGrowing US missile defence capabilities were on a collision course with the 1972 ABM treaty. The Clinton administration attempted to resolve the conflict by negotiating a new understanding with Boris Yeltsin’s Russia that would demarcate which systems would be permitted and which would not. The George W Bush administration took a different view. In 2002, the US withdrew from the ABM treaty, explicitly rejecting—for the first time since Kennedy—the notion that a mutual deterrent relationship was acceptable. The younger Bush stated “we are on the path to a fundamentally different relationship. The Cold War is long gone. Today we leave behind one of its last vestiges.” “Star Wars” did not end with the Soviet Union. The George HW Bush administration preserved the programme at a much smaller scale, and it continued through the Clinton administration. US missile defences were refashioned to defend against small threats from “rogue” states—a far easier technical challenge and less controversial political sell. Moscow kept complaining, but Russian concerns were brushed aside.
A smug consensus developed within Washington that all was well and the Russian grousing was meaningless. But we now know that the immediate consequence of Bush’s ABM withdrawal was that Russia initiated a variety of new programmes designed to defeat the developing American missile defences—the scope and scale of which are only came into public view on 1st March 2018, when Vladimir Putin announced a series of weapons that Russia had been quietly developing in response.
Each of the six new weapons systems Putin announced were, he explained, part of Russia’s comprehensive response to the US withdrawal from the ABM treaty. “During all these years,” Putin explained, “we have been working intensively on advanced equipment and arms, which allowed us to make a breakthrough in developing new models of strategic weapons.” Among the weapons shown off was the nuclear-powered cruise missile, named Burevestnik (SSC-X-9 Skyfall) that those five scientists died trying to recover. Another was the Poseidon, a submersible, nuclear-powered autonomous drone—a nuclear -torpedo that could deliver a multi-megaton nuclear warhead several thousand kilometres away in a US port, from where, as a briefing paper that emerged on Russian television made plain, it could contaminate populated areas with radiation.
Putin also announced a hypersonic glider that would allow Russian nuclear warheads to manoeuvre as they re-entered the atmosphere in an attempt to evade US missile defence and showed a video of a massive new missile, Sarmat, that could carry 10 nuclear warheads. This was designed to be fired over the South Pole, attacking the US from behind, where missile defence radars would not see it until far too late. Separately, Russia announced it would also be flight testing a new hit-to-kill interceptor missile, the PL-19 Nudol, that can destroy satellites in orbit—including the tracking and warning satellites on which some US missile defences depend.
While the outlandish weapons grabbed attention, perhaps the most dramatic change in Russia’s arms capability has been its embrace of long-range cruise missiles. While modern ballistic missiles are the progeny of Nazi Germany’s V2 rocket, cruise missiles are the V1 “buzz bomb.” Theses missiles are powered by jet engines, functioning like unpiloted kamikaze aircraft. When the Soviet Union collapsed, American cruise missiles such as the Tomahawk were a major US advantage, with Washington increasingly using them as a means to deliver conventional weapons, rather than just nuclear warheads.
The Russian defence industry spent the 1990s catching up, and developing a new generation of air, ground- and sea-launched cruise missiles. Russia’s deployment of a new ground-launched cruise missile led the Trump administration to allege Russia had violated the 1987 INF treaty and rationalised its own decision to withdraw. This missile was developed specifically for a system that had, as a Russian defence official said, the mission to “destroy” American missile defences.
Although the practical range of a cruise missile is limited by the size of its fuel tanks to a few thousand kilometres, we now know Russia has also been developing a nuclear-powered cruise missile of unlimited range. The US considered a similar system in the 1960s before concluding that it was a technical and environmental nightmare. Russia, however, has made a different calculation—that the unlimited range offers the ability to circumvent American missile defences.
A new start soursAt the same time, the Obama administration did feel some pressure driven by inherited deadlines. The original START treaty was set to expire at the end of 2009 and the Bush administration had refused to engage in negotiations on a successor. The New START treaty was negotiated in haste, with the American negotiators thinking of it as a temporary measure. During negotiations, however, they were surprised to discover the Russians taking a hard line on a number of issues. In particular, the Russians pressed hard to end both the monitoring at the Votkinsk Machine Plant, where Russia assembles its strategic missiles, and the exchange of telemetry data. In retrospect, it was clear that Russia was planning a new generation of strategic systems. A decade ago, few would have feared we were spiralling towards a new arms race. Obama, who set out his vision of a nuclear-free world in a speech in Prague in his first few months in office, came to power promising to improve relations with Moscow, with Hillary Clinton, his Secretary of State, presenting her Russian counterpart with a phoney red button marked “reset.” When Mitt Romney, Obama’s opponent in the 2012 election, called Russia the “greatest geopolitical foe” to the US, Obama responded that “the 1980s are now calling to ask for their foreign policy back.”
New START failed to anticipate the wide variety of new systems under development. For example, it counts each bomber as one nuclear weapon, no matter how many nuclear-armed cruise missiles it might carry. And a side-agreement to the original START treaty, one that required each side to declare the number of nuclear-armed cruise missiles on submarines, was allowed to lapse. Thus Russia has been able to deploy a new generation of air- and sea-launched cruise missiles almost entirely free of constraints. The New START treaty does not limit Russia’s potential intercontinental-range cruise missiles, or its intercontinental range-underwater torpedo. Nor does New START constrain Russia’s development of anti-satellite weapons.
None of this is to say that New START was a bad treaty: the text was always intended as a placeholder, which would be refined and improved later. But in constructing their temporary shelter, the negotiators failed to understand how various treaties were mutually reinforcing—each treaty had provisions that only worked in concert with the others. For example, negotiators had left a loophole in the 1987 INF treaty. Most intercontinental ballistic missiles (ICBMs) are missiles with three stages—one rocket stacked on top of another, stacked on top of yet another. If Moscow simply removed one of those stages and weighed the whole thing down with more warheads, the resulting missile would be precisely the kind of intermediate-range weapon that the INF treaty was supposed to prohibit.
In the 1980s, the negotiators understood this problem clearly and decided the best solution would be low limits on the number of ICBMs in the forthcoming START treaty, so that Moscow would never waste a precious ICBM slot on such a makeshift intermediate-range missile. But when negotiators haggled over New START decades later, the Americans sought a much higher limit on the number of missiles. Russia took advantage of the loophole to develop a completely legal nuclear-armed intermediate-range ballistic missile that was essentially identical to the one that had been barred under the INF treaty.
One by one, in this manner, agreements have been allowed to collapse or be hollowed out. And because the agreements had been created one atop another, each failure weakened the remaining treaties, leading to a systemic collapse in the arms control architecture.
How did this happen?An important part of the story is the failure of democracy in Russia. The original arms race was driven by Moscow’s strategy of building its security on the insecurity of its neighbours. Soviet leaders were fundamentally insecure at home, seeing conspiracies abroad to topple them from power. Not even the creation of Soviet-dominated satellite states in eastern and central Europe could reassure Soviet leaders. The Cold War only ended because Gorbachev abandoned this strategy, refusing to use force to preserve East Germany and other satellites. From the perspective of revanchists in today’s Russia, it was the abandonment of this strategy that led to the collapse of the Soviet Union, something Putin has called a “geopolitical catastrophe” for Moscow. As Putin consolidated power, he has revived this old strategy, with its fixation on spheres of influence, and backed up that fixation with force in Georgia and Ukraine.
In this way, Moscow has done its bit to recreate the arms race, but a race requires two parties. No one would confuse the United States after the Cold War with a generous victor. Whether it was the economic chaos visited upon post-Soviet Russia or the steady expansion of Nato, the west failed to provide the stability that Russians craved. Putin happily filled the vacuum.
On arms control, US negotiators were short-sighted, presuming American advantages could be preserved forever. Neither the Bush II nor Obama administrations were willing to entertain any ABM-treaty-style limits on missile defences. In some cases, it was down to sheer hostility to the idea of arms control itself—in others, it was hubris. Officials and experts who should have known better seemed to believe arms control agreements that locked in American advantages and ignored Russian interests could be maintained indefinitely. And there was cowardice. There was little reason to believe that Republicans in the Senate would ever agree to anything limiting missile defences. So why try?
After all, things seemed to be going so well. Because the withdrawal from the ABM treaty was not met with immediate collapse of the arms control framework, Washington concluded that the US could have its cake and eat it, too. Because Russia’s reaction has emerged years after America stepped up efforts at missile defence, its advocates can deny that the one led to the other.
A new threatThe old assumption was that the only thing that could kill a nuclear weapon in a surprise attack was another nuclear weapon. This, oddly, was a comforting thought. It meant you could do effective arms control by counting weapons and ensuring rough parity, leaving deterrence to work in all but the most extreme scenarios. Some of the deadly uncertainty that worried Schelling was dispelled.
But one major difference from Schelling’s time is our reliance on vulnerable computer networks. American experts each used to worry about the possibility that a Soviet attack might kill the president before he could order a retaliation, or vice versa: they called this kind of attack a “decapitation.” The real danger of decapitation as a strategy is that it creates an enormous incentive to be the first to attack. Washington and Moscow developed different strategies to deal with this problem. The Americans gave the president sole authority to order in minutes the execution of a number of pre-planned nuclear options with emergency action messages sent directly to launch units. The Soviet Union went further, creating a semi-automatic launch system called Perimeter. This system is widely called “the Dead Hand”—you can chop off the system’s head and kill it, but the dead hand will still retaliate. In a crisis, leaders could choose to delegate launch authority to a command centre underneath a mountain. If the leadership were killed or out of contact, the system—which still relied on human beings in the bunker—would retaliate against the US. In the event of a false alarm, the people in the bunker would probably not start a nuclear war. Probably.
The contemporary vulnerability of computer networks to attack complicates this situation. The US and Russian systems both rely on them to execute their doomsday strategies. For these systems to assure that decapitation will fail and lead to the destruction of the attacker, the underlying computer networks must be absolutely reliable. If there is a chance that a hacker could slow, let alone stop, an order, then there is again every incentive to go first.
The pressure on an American or Russian president in any nuclear crisis would be enormous. In even the best-case scenario, the plausible threat of an unexpected attack on computer systems would loom large. A worse case would be the existence of a known vulnerability, one that had been discovered but not yet fixed, creating pressure for a leader to use nuclear weapons before their opponent could exploit it. And the worst of all cases would be the leader who not only knew about a vulnerability, but also suspected that the adversary had knowledge of it as well.
If the US knew about a security vulnerability that would allow it to attack Russia’s nuclear command-and-control system, it would be enormously tempting to hold that secret for a rainy day. Likewise, if Russia knows about a vulnerability it is likely to keep mum. But that is the worst thing that either party could do. If Russia did, in fact, know about the vulnerability, Putin would face escalatory pressure. And if Putin knew that the Americans knew but weren’t saying so, the pressure to escalate would be overwhelming. Schelling’s great insight was that our intuition about frightening our adversaries was back-to-front. A frightened burglar was far more likely to shoot him in the dead of night. So too with a nuclear-armed superpower. If we take Schelling seriously, the implication is that if the US found security vulnerabilities in Russia’s nuclear command-and-control, it should not save them for a rainy day. It should tell Moscow about them.
That is not, to state the obvious, how the US (or Russia) looks at this problem. Instead, the Trump administration has proposed a simple solution to the possibility of a cyber-attack on its command-and-control system: threaten to nuke the hackers. The idea has been met with ridicule. But it is what it is. Sixty-odd years after Schelling’s landmark article, we have no better solution to his dilemma than to stumble down the stairs, threatening loudly to shoot.
The problem of ChinaThere is another complication. Throughout the Cold War, the only relationship that counted was the bilateral one between Moscow and Washington. Other states had nuclear weapons, but they were secondary to fundamental confrontation.
But the US-Russia INF Treaty collapsed, at least in part, because China has been deploying a small but growing stockpile of stockpile of intermediate-range nuclear missiles outside of the agreement, leading to widespread dissatisfaction within both the United States and Russia about mutual restraint.
This complicates everything from issues relating to missile defence—where China’s arsenal is arguably small enough to be threatened by relatively modest US deployments—to the security of satellites in outer space. A three-way relationship can be deeply confusing. If there were an attack on US computer networks in a crisis, how confident could an American president be in attributing the attack to Russia instead of China? What if there were simultaneous crises? Schelling’s dilemma is infinitely more complicated if there are two burglars who may not even be in cahoots.
To date, efforts by the Trump administration to invoke China’s nuclear forces have been transparently cynical. US officials have argued against extending the New START treaty unless China were a party. But China’s nuclear forces are vastly smaller than the limits the treaty imposes on the US and Russia—at most about 100 nuclear-armed missiles. Unsurprisingly, China has not been enthusiastic about Washington’s proposals to use it as an excuse for allowing New START to expire.
Trilateral arms control is not impossible. But it would require the US agreeing to discuss those agenda items that are important to Moscow and Beijing, such as finally accepting the need to place legally binding limits on missile defences. It is simply not in the interest of Russia or China to accept numerical or technological nuclear constraints on their nuclear forces if the United States is free to develop and deploy missile defences to defeat them. Both Moscow and Beijing will continue to pursue programmes to defeat American defence systems, as well as investing in their own defences. In turn, Washington will be obliged to find ways to defeat them. That is the essence of an arms race.
An agreement to prohibit attacks on computer networks related to nuclear command-and-control would also be better with three. Such an agreement might not be verifiable in the way traditional arms control agreements are. But it is the minimum necessary rule for the present era of dependency on computer networks—and it can be tested. The test comes when one of the parties discovers a vulnerability in another’s command-and-control network. If the US were to discover a weakness in the Russian or Chinese command-and-control system, Washington should tell them about it—and vice-versa. Of course, this is contrary to everything we currently think about security and secrecy. We can’t possibly share how our missile defences work with Moscow and Beijing, let alone alert them to vulnerabilities in their own computer networks. And yet that was precisely the implication of what Schelling was saying in 1958. The burglar had come to rob him, but that doesn’t matter nearly as much as getting out of the room alive. Suddenly, with that realisation, Schelling and the burglar shared a fate—and a common interest.
It is more than 60 years later, but we are still standing in Tom Schelling’s imagined living room, in the dead of night and armed to the teeth, each of us trying to avoid being shot by the other in “self-defence.” Schelling understood there was little difference in the danger posed by a frightened burglar or frightened Soviets. It is the central insight of the nuclear age—that nuclear weapons are so dangerous that they compel us to cooperate, even with our adversaries. Our survival depends on it.