Would the United States fight a nuclear war to save Estonia? The question would probably strike most Americans as absurd. Certainly, almost no one was thinking about such a prospect when NATO expanded to include the Baltic states in 2004.

Yet a series of reports by the nonpartisan Rand Corporation shows that the possibility of nuclear escalation in a conflict between the North Atlantic Treaty Organization and Russia over the Baltic region is higher than one might imagine. The best way of averting it? Invest more in the alliance’s conventional defense.

There was a time when it seemed quite normal to risk nuclear war over the sanctity of European frontiers. During the Cold War, NATO was outnumbered by Warsaw Pact forces, and it would have had great difficulty stopping a Soviet attack with conventional weapons. From the moment it was formed, NATO relied on the threat of nuclear escalation — whether rapid and spasmodic, or gradual and controlled — to maintain deterrence. American thinkers developed elaborate models and theories of deterrence. U.S. and NATO forces regularly carried out exercises simulating the resort to nuclear weapons to make this strategy credible.

After the Cold War ended, the U.S. and its allies had the luxury of thinking less about nuclear deterrence and war-fighting. Tensions with Russia receded and nuclear strategy came to seem like a relic of a bygone era. Yet today, with Russia rising again as a military threat, the grim logic of nuclear statecraft is returning.

The spike in tensions between Russia and the West over the past half-decade has revealed a basic problem: NATO doesn’t have the capability to prevent Russian forces from quickly overrunning Estonia, Latvia and Lithuania. Russian invaders would be at the gates of the Baltic capitals in two to three days; existing NATO forces in the region would be destroyed or swept aside. NATO could respond by mobilizing for a longer war to liberate the Baltic countries, but this would require a bloody, dangerous military campaign. Critically, that campaign would require striking targets — such as air defense systems — located within Russia, as well as suppressing Russian artillery, short-range missiles and other capabilities within the Kaliningrad enclave, which is situated behind NATO’s front lines.

Moreover, this sort of NATO counteroffensive is precisely the situation Russian nuclear doctrine seems meant to avert. Russian officials understand that their country would lose a long war against NATO. They are particularly alarmed at the possibility of NATO using its unmatched military capabilities to conduct conventional strikes within Russian borders. So the Kremlin has signaled that it might carry out limited nuclear strikes — perhaps a “demonstration strike” somewhere in the Atlantic, or against NATO forces in the theater — to force the alliance to make peace on Moscow’s terms. This concept is known as “escalate to de-escalate,” and there is a growing body of evidence that the Russians are serious about it.

A NATO-Russia war could thus go nuclear if Russia “escalates” to preserve the gains it has won early in the conflict. It could also go nuclear in a second, if somewhat less likely, way: If the U.S. and NATO initiate their own limited nuclear strikes against Russian forces to prevent Moscow from overrunning the Baltic allies in the first place. And even the limited use of nuclear weapons raises the question of further escalation: Would crossing the nuclear threshold lead, through deliberate choice or miscalculation, to a general nuclear war involving intercontinental ballistic missiles, strategic bombers and apocalyptic destruction?

So what to do? One option would be for the West to pull back — to conclude that any game that involves risking nuclear war over the Baltic states is not worth the candle. The logic here is superficially compelling. After all, the U.S. could survive and thrive in a world where Russia dominated Estonia, Latvia and Lithuania, just as it survived and thrived during the Cold War, when those countries were part of the Soviet Union. The problem is that failing to defend the Baltic states would devalue the Article 5 guarantee on which NATO rests: the principle that an attack on one is an attack on all. And given that one could raise similar questions about so many U.S. commitments — would declining to meet a Chinese attack on the Philippines really endanger America’s existence? — this failure could undermine the broader alliance system that has delivered peace and stability for so many decades.

A second option, emphasized by the Pentagon’s 2018 Nuclear Posture Review, would be to devise new limited nuclear options as a way of strengthening deterrence and dissuading Russia from pursuing a strategy of escalate to de-escalate. For example, the U.S. might develop low-yield nuclear weapons that could be used, in a relatively limited fashion, against a Russian invasion force or the units supporting it.

This approach is probably worthwhile, because it would help fill in missing steps on the escalatory ladder between conventional conflict and general nuclear war. The knowledge that the U.S. has its own “tactical” nuclear options might inject greater caution into the calculations of Russian planners. It is possible, Rand analysts note, that limited nuclear strikes early in a Baltic conflict could convince the Kremlin that the risks of proceeding are unacceptable.

The dangers here are, well, obvious and drastic. There is always some possibility — although informed analysts debate how much of a possibility — that Russia might mistake a limited strike against military targets in the Baltics for part of a larger or more dangerous nuclear strike against Russia itself. And if the plan is to use limited nuclear strikes against Russian military assets involved in an invasion of the Baltic states, the implication is that NATO would be using nuclear weapons on the territory of its own members.

A third, and best, option is to strengthen the weak conventional posture that threatens to bring nuclear options into play. The root of NATO’s nuclear dilemma in the Baltics is that the forces it has stationed there cannot put up a credible defense. Yet as earlier studies have noted, the U.S. and its allies could make a Russian campaign far harder and costlier — with a much-diminished chance of rapid success — by deploying an enhanced NATO force of seven to eight brigade combat teams, some 30,000 troops. That force would include three or four armored brigade combat teams (as opposed to the one NATO periodically deploys to Eastern Europe now), along with enhanced mobile air defenses and other critical capabilities.

Russia couldn’t claim credibly that such troops posed any real offensive threat to its territory. But the force would be large and robust enough that Russian troops couldn’t destroy it in a flash or bypass it at the outset of a conflict. It would therefore obviate many of the nuclear escalation dynamics by making far less likely a situation in which NATO must escalate to avoid a crippling defeat in the Baltics, or one in which Russia can escalate to protect its early victories there.

Developing this stronger conventional deterrent in the Baltics would not be cheap: Estimates run from $8 billion to $14 billion in initial costs, plus $3 billion to $5 billion in annual operating expenses. Yet neither would it be prohibitive for the richest alliance in the world. The best way of reducing the danger of a nuclear war in the Baltics is to ensure that NATO won’t immediately lose a conventional one.

Hal Brands is a Bloomberg Opinion columnist, a professor at Johns Hopkins University and a scholar at the American Enterprise Institute. Most recently, he is the co-author of “The Lessons of Tragedy: Statecraft and World Order.”

In a time of both misinformation and too much information, quality journalism is more crucial than ever.
By subscribing, you can help us get the story right.