Why Deterrence Is a Promise You Have to Be Willing to Keep

A smoke detector with no battery is worse than no smoke detector at all, because you go to bed feeling safe. That is the best short description I know of a threat you will not carry out. The sign on the lawn says "protected by an alarm system," the burglar walks up and tests the door, nothing happens, and now the sign is not just useless, it is an advertisement.

Foreign policy runs on exactly this logic, with higher stakes. The word for it is deterrence, and it is the most important idea in American security that most Americans have never been asked to define. So let's define it, take it apart, and see where it has held and where it has cracked. This is not a lecture in political science. It is a family question. Whether your son or daughter ever wears the uniform in a real war depends, in large part, on whether somebody on the other side of the world believed us yesterday.

The definition, in one sentence

Deterrence is persuading an enemy not to do something by convincing him that the price will be higher than the prize. That sentence has two parts, and both are required. The enemy has to believe you are able to impose the price, and he has to believe you are willing to. Capability without will is a museum. Will without capability is a bluff.

Put it as a short argument, the way a debater would.

1. An adversary acts when he expects the gain to exceed the cost.
2. Deterrence changes the expected cost, or the expected gain, in his own head.
3. What counts is what he believes, not what you intend. A promise that your enemy does not credit deters nothing.
4. Therefore deterrence fails whenever he doubts either your power or your resolve, and your resolve is the one he is most likely to doubt.

Point four is the one that gets us in trouble, so keep it in mind as we go. A democracy that argues loudly with itself is the easiest government on earth to doubt.

Two kinds: denial and punishment

Strategists, going back to the work of Glenn Snyder in the late 1950s, distinguish two ways of changing an enemy's calculation.

Deterrence by denial says: you will not succeed. You make the attack fail, or look likely to fail. A wall, a missile defense, a heavily armed garrison on the border, a well-trained reserve that can reinforce in days, all of these work by convincing the attacker that he will not take the objective. If he cannot win, he has no reason to try.

Deterrence by punishment says: you may succeed, but you will regret it. You threaten retaliation severe enough that the victory is not worth having. The nuclear standoff of the Cold War was mostly this kind. Neither side could stop the other's missiles from arriving, so each promised that the answer would be unbearable.

The two blend. A garrison that cannot stop a Russian armored column on its own, but guarantees that the column will be at war with the United States the moment it crosses the line, deters by a kind of trapdoor. That is the whole idea behind tripwire forces, and it is why the placement of a few thousand soldiers can matter far more than their number.

It matters which kind you are counting on, because they fail differently. Denial fails when the enemy builds enough to overwhelm it, or finds a way around it. Punishment fails when the enemy decides you are bluffing. And punishment has an ugly feature: if you ever have to carry it out, deterrence has already failed, and you are now deciding whether to commit a catastrophe on principle.

Extended deterrence: the promise to someone else

Deterring an attack on your own territory is hard enough. Extended deterrence is the promise to treat an attack on an ally as an attack on yourself. This is what the American nuclear umbrella means: Washington says that if a nuclear or overwhelming conventional attack falls on certain allies, the response will come from us, up to and including our own weapons.

Think about how much faith that takes. In effect, an American President says to a country thousands of miles away, we will risk our cities for yours. Charles de Gaulle is reported to have put the question in substance: would Washington trade New York for Paris? It was a fair question. It is the reason France built its own nuclear force. The solution the alliance found was to station American troops and weapons in Europe in such a way that any attack would hit Americans first, putting American blood on the table before the shooting starts. The troops in Germany were never meant to hold the line alone. They were meant to make the promise believable.

This is also why the commitments in a treaty text matter so much, and why the one invocation of the NATO mutual-defense clause, covered in The One Time NATO's Article 5 Was Invoked, and Who Answered the Call, is not just a historical curiosity. Alliances are deterrence by contract. The paper has to be backed by the people who signed it.

Case one: Berlin, where the promise was kept

The best evidence that a promise can hold comes from Berlin. In June 1948 the Soviets cut the land routes into the western sectors of the city, hoping to starve the Western powers out. The answer was not a tank column, which would have meant war. It was the Berlin Airlift, which kept two million people supplied from the air for months until the blockade was lifted in May 1949. The message was that the West would not leave, and it would not be bullied into leaving either.

The pressure came back in 1961. East Germany built the Wall in August of that year, and the Kennedy administration decided that it would not use force to tear it down, but would defend the access rights and the freedom of West Berlin without compromise. Two years later, on June 26, 1963, President Kennedy stood in West Berlin and said "Ich bin ein Berliner." Critics then and since have argued that letting the Wall go up showed weakness. There is something to that. But the important fact is what Moscow did not do. It never moved on West Berlin itself, which was the real prize. The promise to the free half of the city held for the whole of the Cold War, and a divided city is a lot better than a lost one.

Case two: Korea, where the promise was not clear, and then was

Korea is the cautionary tale and the success story at once. In January 1950 the secretary of state described an American defense perimeter in the Pacific that, in the way many people heard it, left South Korea outside the line. Historians still argue about how much that speech influenced the decision to invade; the Soviets and North Koreans had their own reasons and plans. But the lesson drawn by almost everyone afterward was that ambiguity about an ally invites a test. North Korea invaded on June 25, 1950, and the United States responded with force.

After the armistice of July 1953, the lesson was applied. In October 1953 the United States and South Korea signed a Mutual Defense Treaty, and American forces have remained on the peninsula ever since. Seven decades of a stalemate that has been cold and sometimes tense, but has not turned into another invasion, is what extended deterrence looks like when it works. It is also a lot cheaper than another war. The military garrison in Korea is the kind of tripwire I described above: its value is the certainty it gives the enemy that the United States would be in the fight from the first hour.

The credibility problem: a bluff that gets called

Now the hard part. What happens when you make a threat and the other side tests it?

If you follow through, you pay the cost of the action, but you have bought credibility for next time. If you do not, you have paid a different cost, and it keeps compounding. Every adversary on earth is watching, and every ally is too. Allies will start to hedge, with their own weapons or their own deals with the other side. That is why deterrence experts say reputation is a real asset, even though some academics sneer at the idea. Skeptics point out, fairly, that one broken promise does not wreck every other promise, because enemies judge each commitment by its own interests. That is a good point. But it is not a good enough point to bet a country on.

Take a more modern example of the principle, and take it without partisanship. In August 2012, President Obama said that the use of chemical weapons in Syria would cross a "red line" for the United States. In August 2013, a chemical attack killed large numbers of civilians near Damascus. The President announced he would seek congressional authorization for a strike, Congress did not vote on it, and the United States did not strike; a diplomatic arrangement to remove Syria's declared chemical stockpile followed. You can argue, as many did, that avoiding another Middle East war was wise, and you can argue, as many others did, that the declared line was imprudent to set in the first place. Those are separate debates. What the episode illustrates is the narrower point of this post: a stated line that is crossed without a response teaches every observer something about the next stated line.

The lesson is not "never draw a line." The lesson is "do not draw one you are not prepared to defend." Choose your commitments with care, make them public, make them few, and then keep them.

The steelman: isn't deterrence just an excuse for spending?

Here is the best case on the other side, stated fairly. Critics, including some on the right, say that deterrence reasoning is infinitely elastic. Any weapon system, any overseas base, any alliance can be justified as "credibility." If every commitment is vital to the credibility of every other, then we are required to fight for everything, and the term has no limit. The war in Vietnam was defended, in part, on this logic. They also point out that the theory assumes a rational adversary who weighs costs the way we do, which is not always true.

That critique is sound, and a conservative should take it seriously, because a conservative is supposed to be suspicious of arguments that can justify anything. The answer has three parts.

1. Credibility is about specific commitments, not all commitments. A country with a few, clear, vital promises is more credible than one with fifty vague ones.
2. The real test is interest. We keep promises most firmly where the stakes to us are real, where the alliance serves American security directly, and where the ally is willing to share the burden. That is the argument in Burden Sharing Is Not a Trump Idea, It Is an Eisenhower Idea.
3. Deterrence must be matched with honest assessments of the adversary. Sometimes the enemy is not deterrable by threats, and sometimes he is deterrable by means we did not think of. The study of the other side's fears, which is what the better intelligence work does, is part of the job. For a primer on who does that work, see Who Does What in American Intelligence, Explained in Twelve Questions.

The Cuban missile crisis shows both sides at work. President Kennedy drew a firm line, backed it with a naval quarantine, and left Khrushchev a way out, which is covered in Thirteen Days in October and What the Quarantine Teaches About Resolve. Firm and flexible is not a contradiction. It is the art.

What this means for the argument at your kitchen table

When you hear a politician say that America should withdraw from a commitment to save money, or make a new commitment to look tough, ask four plain questions.

• What exactly are we promising, and to whom?
• Is the promise public and clear to the adversary, or only to us?
• Do we have the means to make good on it, and does the other side know it?
• If it is tested, will this President, and the Congress, actually follow through?

If the answer to the fourth question is "probably not," then the promise is a bluff, and you should want it either fixed or withdrawn quietly rather than left on the books to be tested. A loud bluff is the most dangerous thing a nation can own. It can start a war all by itself.

Deterrence is also part of why the broader argument about military strength matters, which Peace Through Strength Is Older Than Reagan and It Still Works lays out in its history.

The bottom line

• Deterrence is a belief in the mind of your enemy. If he doubts your power or your will, he will test you, and the test is paid for by soldiers and taxpayers.
• Denial and punishment are different tools. Know which one a given policy relies on, because they fail differently.
• Keep your commitments few, clear and funded. A promise you are not prepared to keep is worse than no promise at all.
• Allies matter because they extend the promise, but they have to carry their weight, or the promise gets stretched thin.
• When a candidate talks tough, ask what happens when the bluff is called. Back the one who has an answer.