Nuclear Deterrence

In brief

Bernard Brodie, writing in 1946 as the first atomic bombs were still fresh, drew the conclusion that has organized the subject ever since: "Thus far the chief purpose of our military establishment has been to win wars. From now on its chief purpose must be to avert them." A weapon powerful enough to end a war before it is fought changes what armies are for. Nuclear deterrence is the strategy that follows. You do not defend your cities by intercepting the enemy's missiles; you defend them by guaranteeing that any attack brings retaliation so ruinous that no rational leader would order it. The peace this produces does not rest on goodwill. It rests on a shared expectation of catastrophe. Its sharpest form, mutually assured destruction, holds that when each side can absorb a first strike and still obliterate the other, neither has any reason to start, and the standoff becomes stable precisely because both sides remain permanently vulnerable.

The full treatment

The problem it answers

Before 1945, security came from defense: fortifications, armies, the capacity to blunt an invasion and prevail in the fighting. Nuclear weapons broke that logic. A defense that stops ninety percent of incoming warheads still lets through enough to destroy a nation. When even a small fraction of a modern arsenal means the end of a country, winning the war in any traditional sense becomes impossible, and prevention becomes the only strategy left. Deterrence answers the question: how do you stop an enemy from attacking when you cannot physically stop the attack itself? The answer is to change the enemy's calculation rather than block their weapons. You make the cost of aggression so certain and so unbearable that the expected value of striking is always negative.

How it works: credibility and the second strike

Deterrence is a claim about the future made to shape a decision in the present, and its whole force depends on being believed. A threat that the adversary does not credit deters nothing. This is why the doctrine came to center on the survivable second strike. If your retaliatory forces can be wiped out in a surprise first strike, your threat to retaliate is hollow, because after the attack you would have nothing left to strike with, and the enemy knows it. So states invest enormous effort in making their forces able to ride out an attack and answer it: hardened silos, mobile launchers, and above all ballistic missile submarines hidden in the deep ocean, which no first strike can reliably find and destroy. Assured second-strike capability is the technical heart of the whole arrangement. Once both sides possess it, the situation is mutually assured destruction: neither can escape retaliation, so neither gains by going first. The stability is perverse but real, and it inverts ordinary intuition. A missile defense that protected your cities, or a first-strike force that could disarm the enemy, would be destabilizing, because it would tempt one side to strike and frighten the other into striking first. Vulnerability, shared equally, is what keeps the peace.

Schelling's logic of commitment

Thomas Schelling (1921 to 2016) gave deterrence its intellectual spine in The Strategy of Conflict (1960) and Arms and Influence (1966), work that won him the Nobel Memorial Prize in Economic Sciences in 2005. Deterrence, he saw, is a bargaining problem, and bargaining power often comes from giving up your own freedom to choose. A driver who conspicuously rips out his own steering wheel in a game of chicken wins, because he can no longer swerve and the other driver must. Commitment, credibly binding yourself to a course of action, can be stronger than flexibility. Schelling's subtlest contribution was "the threat that leaves something to chance." A threat to launch an all-out nuclear response over a minor provocation is not believable, because carrying it out would be suicidal. But a threat to take a step that raises the risk of things spiraling out of everyone's control, where no one fully decides and events might take over, can be entirely credible. You do not threaten certain destruction; you threaten to let go of the wheel a little. Deterrence, in his hands, became the deliberate and shared manipulation of risk, a competition in resolve conducted at the edge of a shared abyss.

The stability-instability paradox

The theorist Glenn Snyder named the doctrine's most troubling feature in 1965: the stability-instability paradox. The very stability of the nuclear balance at the top, the mutual certainty that no one will start a total war, can make conflict at lower levels more likely, not less. If both sides know the other will not escalate to nuclear use over a limited fight, then limited fights, proxy wars, border skirmishes, and coercion below the nuclear threshold become safer to wage. The nuclear umbrella that prevents the largest war may license many smaller ones. The India and Pakistan rivalry after both tested weapons, and the many Cold War proxy conflicts fought while the superpowers avoided direct combat, are the standard illustrations. Peace at the summit, turbulence on the slopes.

Lineage

Nuclear deterrence is the atomic-age descendant of the older idea of the balance of power, the notion that peace is kept by an equilibrium of strength that makes aggression unprofitable. The bomb transformed that logic by making the penalty for miscalculation absolute. Its analytical tools come from game theory: the field grew up partly at the RAND Corporation, where Cold War strategists funded it precisely to reason about nuclear conflict, and deterrence is best understood as a repeated game of threats, signals, and credibility. Brodie set the founding premise in 1946. Albert Wohlstetter, in his 1959 Foreign Affairs essay "The Delicate Balance of Terror," warned that deterrence was not automatic and had to be engineered through survivable forces. Schelling then supplied the general theory of commitment and coercion. By the 1960s, under United States Secretary of Defense Robert McNamara, "assured destruction" hardened into declared policy, and the acronym MAD, coined by critics, stuck to it.

The strongest case for it

The empirical case is blunt: no two states possessing nuclear weapons have ever fought a full-scale war against each other, and the Cold War, a decades-long confrontation between armed and hostile superpowers, never became a hot war between them. Deterrence theorists argue this is not luck. When the cost of total war is national annihilation, leaders on all sides are pushed toward caution they would not otherwise show, and crises that in earlier eras would have tipped into general war (Berlin, Cuba in 1962) were instead resolved short of it. Deterrence, on this view, does something conventional force never could: it makes great-power war irrational rather than merely costly. Kenneth Waltz pressed the argument furthest, contending that because the logic is so clear and the stakes so total, even new nuclear states behave with sobering care, so that the slow spread of weapons might stabilize rather than endanger the world. Deterrence, its defenders hold, has kept the peace among the strongest for the longest stretch in modern history.

The strongest case against it

The objections are serious and come from people who studied the system closely. Scott Sagan, in his debate with Waltz (The Spread of Nuclear Weapons: A Debate, 1995), argued that deterrence assumes a rational, unitary state, while real arsenals are run by large organizations prone to accidents, false alarms, and narrow institutional interests. He documented near-misses in which the Cold War standoff came close to catastrophe through error rather than intent, and warned that each new nuclear state adds fresh chances for exactly such failures. The doctrine, on this reading, is one accident away from disproving itself, and it can only be proven wrong once.

A second line of attack targets credibility itself. If retaliation would be suicidal, then the threat to retaliate is not believable, and a rational adversary should discount it. Deterrence therefore rests on either irrationality or Schelling's uneasy "something to chance," a peace that depends on the residual fear of losing control. Critics ask whether it is wise to stake civilization on that.

A third argument holds that MAD is a passing technical condition, not a law. Keir Lieber and Daryl Press argued in the mid-2000s that advances in accuracy and surveillance could restore a meaningful first-strike advantage, eroding the assured second strike on which stability depends and reviving the temptation to go first. If mutual vulnerability can be engineered away, the whole edifice wobbles.

Finally, there is the moral objection, pressed by disarmament thinkers and by scholars of the "nuclear taboo" such as Nina Tannenwald: a strategy whose stability is secured by a standing, credible commitment to kill millions of civilians is monstrous even if it works, and the fact that it holds hostage entire populations who have no say makes deterrence a permanent moral hazard rather than a genuine peace.

Where it stands now

Deterrence survived the Cold War and remains the organizing doctrine of every nuclear-armed state, though the tidy two-player logic of MAD has frayed. The world is now multipolar in nuclear terms, with more actors, regional rivalries, and asymmetries that a bipolar theory never modeled. New pressures crowd in: missile defenses, hypersonic weapons, anti-satellite capabilities, and cyber threats to command-and-control systems, each of which can undermine the confidence in survivable retaliation that deterrence needs. Analysts debate whether artificial intelligence in early-warning and launch systems will steady or destabilize the balance. The core insight endures, and it connects to the bargaining model of war, which treats war as a failure to reach a deal that both sides would prefer to fighting: deterrence is the attempt to make the bargaining range so plainly favor peace that no one tests it. Whether that logic holds as the technology and the number of players change is the open question of contemporary strategy.

Test yourself

Deterrence works only if the threat is believed, yet carrying out the threat, all-out retaliation after your country has already been destroyed, would accomplish nothing and cost everything. Would you, in that final moment, actually give the order? If your honest answer is no, has your reasoning just quietly dismantled the peace you were counting on, and what does it take to make a threat credible that you would never rationally want to fulfill?