It’s Time to Fold America’s Nuclear Umbrella

By Stephen M. Walt

A recent report by the Chicago Council on Global Affairs, “Preventing Nuclear Proliferation and Reassuring America’s Allies,” took a small step in a rather surprising direction. The title captures its main theme perfectly: To discourage its allies from acquiring their own nuclear weapons, the United States needs to counter doubts raised during the Trump administration and reassure its allies about the strength of the United States’ commitment to their security.

Given that the report was written by a multinational group of well-known foreign-policy insiders, most of their findings and prescriptions are unproblematic. But the following recommendation caught my eye:

“Europe needs to build up the nuclear dimension of its defense efforts, including by retaining and modernizing capabilities for existing NATO nuclear missions and by France and Britain working together to extend their nuclear deterrents to their European allies.”

Why is this statement so intriguing? Because it shows the authors of this report recognize that Europe as a whole might be more secure if it could rely on a locally based deterrent instead of continuing to shelter under the U.S. nuclear umbrella. And if that is true for the nations of Europe, then it might well be true for others. Although the report’s authors are opposed to new states joining the nuclear club (Britain and France are already members), their statement clearly implies that deterrence would be strengthened if states facing serious external threats had a nuclear guarantee that didn’t depend on Uncle Sam.

This is hardly a new issue. Since fairly early in the nuclear age, the United States has used nuclear weapons to “extend deterrence” and shield some of its allies. It sought to convince potential adversaries that the United States might use its formidable nuclear arsenal if these allies were attacked, even if the United States was not. Of course, there was always some chance that a war involving one of the United States’ allies might escalate to the nuclear level, either by accident, through inadvertence, or via deliberate decision, no matter what U.S. leaders said in advance. Even so, Washington went to considerable lengths to make its nuclear umbrella credible, partly to discourage enemies from attacking but also to convince its allies not to get nuclear weapons themselves.

Accordingly, U.S. leaders made lots of public statements linking the U.S. arsenal to its core alliance commitments, and NATO drew up various plans and doctrinal pronouncements designed to reinforce perceptions of a reliable U.S. guarantee. The United States also deployed thousands of warheads on some of its allies’ territory, along with dual-key arrangements that gave those allies some say in how, when, or if these fearsome weapons got used. Lastly, and very importantly, the United States kept trying to achieve a meaningful degree of nuclear superiority to make a possible first use of nuclear weapons to defend allies more credible. Instead of acquiring a “minimum deterrent” (i.e., retaliatory forces that could survive any possible attack and then inflict unacceptable damage on an aggressor), U.S. war plans and weapons decisions always focused on trying to come out on top in the awful event of an actual nuclear war.

Why did the United States do this? In good part because convincing people you might use nuclear weapons to defend an ally isn’t easy. One might imagine a U.S. president using nuclear weapons to retaliate against a direct attack on U.S. territory or to deter the extremely unlikely prospect of a conventional invasion that threatened U.S. independence. This is the one thing nuclear weapons are good for: deterring existential threats to their possessors’ independence or autonomy. This form of deterrence (sometimes termed “basic” or “Type I”) works because the deterring side will almost certainly care more about preserving its own independence than a potential attacker is likely to care about trying to take it away. Because the balance of resolve favors the defender, even much weaker nuclear powers can deter enemies from attacking them directly. If you don’t find this argument persuasive, remember the U.S. attacked non-nuclear Iraq in 2003 and non-nuclear Libya in 2011, but it leaves nuclear-armed North Korea alone.

By contrast, deterring a conventional or a nuclear attack on an ally by threatening to go nuclear—and convincing your allies that you really mean it—is more challenging. It is one thing to threaten to use nuclear weapons to keep one’s own country from being subjugated but quite another to do so to save an ally from defeat or domination. Or, as people used to wonder back in the Cold War, would a U.S. president really risk Washington or Chicago to save Paris or Berlin? Long after they had left office, a few former U.S. officials suggested the answer was almost certainly “no.” Extended deterrence could still work because potential attackers can’t be sure about any of this, but it still isn’t as credible as deterring attacks on one’s own territory.

The solution to this conundrum—if one can call it that—is to achieve overwhelming “nuclear superiority.” If you could wipe out an adversary’s entire nuclear force in a first strike, you wouldn’t have to fear its retaliation, and using nuclear weapons to defend an ally would be much more credible. Even if a splendid first strike were not possible, perhaps you could convince a potential attacker that it will end up even worse off than you are at the end of a nuclear war to convince it not to put so much as a toe on the first rung of the escalation ladder.

Thus, the perceived need to extend deterrence is one of the reasons why the United States has long sought nuclear superiority. It’s not the only reason: A genuine first strike capability could limit damage in the event of an actual war. A few commentators have also tried to argue—not very convincingly—that superiority would enable the stronger side to coerce weaker states in crises. Chasing the holy grail of a first-strike advantage was also popular with defense contractors and parts of the armed services because it requires spending billions of dollars annually on more and more accurate weapons, more efficient and destructive warheads, improved surveillance and anti-submarine warfare capabilities, and lots of other shiny objects.

Interestingly, a number of sophisticated scholars have recently claimed that technological advances have put the United States on the brink of a true first-strike capability. Perhaps in theory, but certainly not as a usable option. To see why, ask yourself what you would do if you were president and facing a serious crisis with a nuclear-armed adversary. You’ve put the armed services on alert, and there is some danger that force might be used and fighting could escalate. Suppose your military advisors and intelligence experts tell you if you order a first strike now, you can almost certainly destroy the enemy’s entire nuclear arsenal, leaving the United States unscathed and in an ideal position to resolve the dispute on favorable terms.

Being a sensible person, you’d undoubtedly ask them: “Can you guarantee that? Are you absolutely, 100 percent sure the enemy will have zero usable weapons left, and therefore, we won’t even get our hair mussed?”

“We are highly confident of success,” you are told. “But there is a slim chance that a few enemy weapons would survive and reach U.S. soil. No more than one to three.”

Even if you weren’t troubled by the moral issues involved in ordering an attack that would kill untold numbers of people (and you ought to be), would you do it? Of course you wouldn’t, because you wouldn’t want to risk losing New York, Los Angeles, Chicago, Boston, or any other major U.S. city, which is what might happen if that first strike you authorized turned out to be just a tiny bit less effective than your advisors predicted. To issue a launch order, you’d have to believe the proposed attack would work perfectly the very first time it was executed (simulations and exercises aren’t the same), almost all of the missiles and bombs that have been sitting in silos or storage facilities for years would work as designed, and the other side wouldn’t have dispersed its own forces or hidden some extra weapons in places you had failed to detect. Based on everything the United States’ knows about complex military operations and the limits of intelligence, you’d be a fool to roll the dice in this way.

One more thing: As first-strike capabilities improve, adversaries may respond by keeping forces on higher alert or adopting “launch-on-warning” procedures that increase the risk of accidental or inadvertent war. No matter what U.S. forces are capable of in theory, in short, it’s hard to see how any president would be willing to use nukes first even if the probability of “success” was extremely high. This reality casts further doubt on the whole idea of extended deterrence, insofar as it is based on the threat to deliberately escalate to the nuclear level if a key ally is in danger of being conquered.

Extending a protective umbrella over allies in Europe and Asia may have made good sense during the Cold War, both to protect them and to discourage proliferation. But the nuclear weapons environment has changed: The number of nuclear-armed states has crept upward, and several countries (India, Pakistan, and the United Kingdom) are increasing the size of their own arsenals (though they remain far lower than U.S. or Russian levels). Moreover, the United States is not as tightly coupled to some of its traditional allies as it was during the Cold War, and serious rifts may continue to grow despite the Biden administration’s efforts to restore alliance solidarity and reassert U.S. leadership.

Which raises the obvious question: Does it still make sense to shield allies under the U.S. nuclear umbrella? Using the threat of nuclear use to protect other countries is not cost- or risk-free, and it may even be more dangerous than letting some other states acquire arsenals of their own and encouraging them to rely on “Type I” deterrence provided by their own national capabilities.

This view has been advanced before—most notably by Kenneth Waltz in a controversial Adelphi Paper 40 years ago. Waltz was not advocating giving other states the bomb or arguing that the rapid spread of nuclear weapons would be desirable; his central point was that trying to prevent the slow spread of these weapons was not without costs of its own and that in some cases, as he put it, “more may be better.” The question is: Is that becoming the case today?

To be sure, folding the nuclear umbrella might well have some negative effects. It might make states long accustomed to U.S. protection question its commitment (though there’s no logical reason for them to do so if it is still in the United States’ interest to aid their defense in other ways). It could also reduce U.S. influence or leverage if certain allies were no longer as dependent on U.S. protection, though folding the umbrella would not eliminate their reliance on other elements of U.S. power. Removing the U.S. nuclear guarantee might encourage a few states to pursue nuclear arms of their own, but it is not obvious that acquisition by Japan or Germany would be a terrible outcome from a purely U.S. perspective.

Moreover, even the possibility that these states might take over responsibility for deterring attacks on their own territory could have a sobering effect on a rising China and a recalcitrant Russia. In particular, it would remind Beijing and Moscow that their own behavior will affect the strategic calculations that their neighbors make in the near future, including decisions about nuclear arms. If China doesn’t want to face more nuclear weapons states in its immediate region, for instance, then its leaders should start asking themselves what they can do to make those neighbors feel less need for additional protection. The obvious answer: Stop harassing them in various ways, drop the sharp-elbowed approach to diplomacy, stick to agreements previously reached, and do more to resolve existing disputes on a fair-minded basis.

Whatever Washington ultimately chooses to do with its nuclear umbrella, the more important task is to move beyond the tendency to see nuclear weapons as potent signs of status, indispensable tools of statecraft, or powerful sources of leverage. Nuclear weapons are extremely useful for deterring direct and all-out attacks on one’s own homeland but not much else. For that purpose, a great power doesn’t need an enormous arsenal or some hypothetical capability to “fight and win” a nuclear exchange. All it needs is a stockpile that can survive an enemy attack and be able to respond in kind. Properly concealed or protected, they don’t need to be poised and ready to strike at a moment’s notice. Fetishizing the bomb and using it to try to protect others isn’t just expensive; it may also be dangerous.