Will the new US-Russian arms treaty blunt the nuclear threat?

On 15 January 1986, the Soviet Union's leader, Mikhail Gorbachev, introduced Kremlin plans to eliminate all of the superpower's nuclear weapons by the year 2000. The proposal was visionary, but also a bit of a propaganda ploy. One that immediately caught Ronald Reagan's attention. Later that day, when Secretary of State George Shultz went to the White House, Reagan asked him: "Why wait until the end of the century for a world without nuclear weapons?"

Reagan's question still echoes today. He and Gorbachev reversed the direction of the arms race, but in the end they didn't eliminate all nuclear weapons. Last week, a new strategic arms treaty entered into force between the United States and Russia. The treaty will lead to further modest reductions in the arsenals, which are far smaller than at the peak of the Cold War. But this is no time to heave a sigh of relief. We should be asking whether the atomic bomb and its dangers will have as much sway over the world in the 21st century as it did in the last.

Nuclear deterrence is based on the idea that a rational attacker will be restrained by the credible threat of retaliation. The concept was credited with keeping the peace in the years of superpower competition, but it also led to some of the most risky moments of the era. Repeatedly, there were harrowing false alarms of missile attack, including one in 1983 when a Soviet early-warning system lit up with signs of an impending onslaught – signs that turned out to be incorrect. Thanks to a knowledgeable watch officer, there was not an overreaction.

The need to create a credible deterrent drove planning sometimes to the absurd. For example, in the 1980s, the Soviet leadership struggled with how to guarantee retaliation if the Kremlin itself were suddenly wiped out. If the leaders were gone, they reasoned, they could not be certain to retaliate, and this left them feeling vulnerable. So they designed and built a semi-automatic Doomsday machine, known as the Dead Hand, which would be able to launch a retaliatory strike with land-based missiles even if the Kremlin were destroyed in a flash. The system was semi-automatic because the final decisions about whether to press the button were left to a few duty officers in a deep underground bunker. Oddly, the Soviets kept the entire contraption secret during the final years of the Cold War, meaning that it had very little or no value as a deterrent, since no one outside knew about it.

Today's security threats are far more diffuse and elusive than the tense stand-off of the Cold War. Nuclear-tipped missiles are not likely to stop suicidal terrorists or cyber-attackers – deterrence has less value against opponents who are nearly impossible to identify. These are the adversaries of today, and tomorrow.

Despite President Barack Obama's pledge to seek a world without nuclear weapons, both the United States and Russia are trapped in a Cold War mindset, and cling to nuclear arsenals that are far larger than they need to be. Both countries tend to regard warheads as assets, as chips in a power game. The doctrine of mutually assured destruction – the cocked-pistols approach that came to symbolise the era – has lost its relevance but, strangely, remains in effect.

It is time for a serious reconsideration of our reliance on nuclear deterrence, starting with the US and Russia, which together possess 95 per cent of the nuclear weapons on the globe today, but eventually also bringing in Great Britain, France and China, which have far smaller arsenals.

One of the first things that both Washington and Moscow could do would be to take nuclear weapons off launch-ready alert. What is the point of keeping land-based, nuclear-armed missiles ready to fire in four minutes, as they are in the United States today? It is not appropriate or necessary for the kind of relationship both countries aspire to. Russia can be independent and aggressive in pursuing its own interests, but it is certainly not an adversary in the same way that the Soviet Union was. The current missile-alert status could be modified, with a bilateral agreement to build in delays, say hours or days, before a launch could be carried out. At the very minimum, both sides should revive an idea that has been kicking around for a decade: to build a joint data-exchange centre to avoid surprises.

It seems odd, but the new US-Russia treaty does not cover thousands of existing nuclear weapons. Left out of the pact were about 2,000 strategic or long-range warheads in a special US "reserve", and thousands of Russian tactical or shorter-range nuclear warheads, as well as the West's smaller stockpile of them. The next stage of negotiations ought to reduce or eliminate these overhangs from the Cold War, which are subject to no verification, and there is not even a decent warhead count on each side. It is also urgent that Russia and the United States agree on a way to co-operate in missile defence, a divisive issue that could otherwise stall hopes for further arms-control negotiations.

For the 21st century, we need to rethink the basic concepts of deterrence. For example, the US and Russia still roughly calculate how many weapons they need based on targeting the other side's weapons. This is known as counterforce: each side aiming largely at the other side's military targets. With such calculations, the two countries reached the agreement in the new treaty on 1,550 strategic nuclear warheads and 700 launchers for each. But there is no magic in these numbers. Deterrence could be based on a different measure: the minimum necessary to dissuade the other, which would be far fewer. One study by US Air Force thinkers concluded that a stable nuclear deterrence could be maintained by the United States with as few as 311 warheads.

Nuclear weapons cannot be uninvented. But even if the knowledge continues to exist, must we keep all the hardware, or would it be sufficient to preserve just the designs on paper? Consider that chemical and biological weapons have been banned, but the knowledge of how to make them has never been eradicated. Could the same be done with nuclear weapons? Some experts are discussing an idea known as "virtual deterrence", based on the credible threat of assembling a given number of warheads. The deterrence value would be the knowledge and means to assemble them, rather than the actual weapons.

In a world with far fewer nuclear weapons, we will also need to think of new methods to verify that no one is cheating or has the ability suddenly to restore an outsized arsenal. Perhaps this could be a new task for the brilliant minds at the nuclear weapons research facilities: invent a way to build-down, safely. Obama's vision of a nuclear-free world won't come soon, but it won't come at all if there isn't careful preparation for it.

In a very real sense, nuclear deterrence, as it was known in the Cold War, has lost its overwhelming potency as the backbone of security. While the bombs won't go away, there's an urgent need to create modern instruments for getting results in international relations: persuasion, coercive diplomacy, defence and resilience against foes. Just building the threatening weapons of nuclear destruction is no longer enough.

'The Dead Hand: Reagan, Gorbachev and the Untold Story of the Cold War Arms Race' by David E Hoffman (Icon £20) won the 2010 Pulitzer Prize for General Nonfiction. To order a copy for £16 (free P&P), call Independent Books Direct on 08430 600 030, or visit www.independentbooksdirect.co.uk.