Michael KreponDeterrence Stability is a Hoax. The Delicate Balance of Terror is, Too.

Quote of the week:
“It has in fact become abundantly clear… since the dawn of the nuclear age, that the balance of terror is decidedly not delicate.”
—Bernard Brodie, War & Politics

Can we hold two contrary notions in our heads at the same time? Can we acknowledge that the delicate balance of terror and deterrence stability are both hoaxes? Sure, deterrence stability works for mid-sized nuclear-armed states like Great Britain and France, which have no beef with each other and have limited financial means. They have counterforce capabilities but they aren’t pursuing damage limitation strategies toward Moscow and Beijing. Instead, they have opted for survivability and assured destruction. British and French strategic analysts don’t jump through hoops at the latest Russian or Chinese (or North Korean) missile flight-test.

The same cannot be said for nuclear-armed states that seek advantage in the event deterrence fails or even to escape deterrence altogether by means of counterforce capabilities. Under these circumstances, counterforce capabilities are destabilizing; they raise anxieties, rather than diminishing them. Washington and Moscow drank this Kool-Aid in the 1970s and have been addicted ever since. Strategic arms reduction treaties have reduced force levels, but counterforce targeting remains king. No matter how much these capabilities are refined, there’s nothing delicate about crossing the nuclear threshold, regardless of yield and weapon effects. Escalation control will remain a cosmic roll of the dice. Nonetheless, true believers in the political and military utility of nuclear weapons remain hooked. They seek one-upsmanship, or at least to offset troubling moves.

There was a blessed lull in this competition at the end of the Cold War, which lasted until the Russian Federation revived from losing its empire and recovered from a great depression. Now Moscow and Washington are back to business as usual. As long as operational warheads with counterforce capabilities can cover targets, neither competitor feels comfortable—even in the absence of national missile defenses. The founding fathers of arms control—a high-powered group of defense intellectuals immersed in cost-effectiveness equations, IR theorists, scientists conflicted by their association with the Manhattan Project, and the Kennedy Administration’s brain trust—conceptualized a state of deterrence stability ensured by vulnerability. Deterrence stability would help define and sustain a mid-point alternative between disarmament and arms racing. They didn’t reckon adequately with the counterforce compulsion and damage limitation targeting strategies which negated the concept of deterrence stability, even in the absence of national missile defenses. For true believers in counterforce capabilities, deterrence works best when you have some sort of advantage. Advantages come in handy if deterrence fails.

Arms controllers reached the apogee of their influence in the Pentagon during Robert McNamara’s tenure as Secretary of Defense who, along with his “whiz kids,” forced the Air Force to pare back absurd plans for building intercontinental ballistic missiles. McNamara also took issue with national ballistic missile defenses, but this was an alien concept to the Kremlin. President Nixon and Henry Kissinger were no fans of strict BMD limits either, but they bowed to the realities of domestic politics, high costs, and technological constraints. When the construct of deterrence stability was then imperfectly put into practice in the first Strategic Arms Limitation accords, it satisfied no one. Hawks chafed at the Anti-Ballistic Missile Treaty, and Doves railed against an Interim Agreement that let MIRVs run free.

The ABM Treaty was the most remarkable achievement of the conceptualizers of arms control, but it didn’t prevent an intensified nuclear competition, nor assure deterrence stability. The revolution of counterforce capabilities proceeded apace, with no regard whatsoever for the ABM Treaty. After McNamara’s departure, men like Melvin Laird and James Schlesinger took the reins. Those who championed deterrence stability did not make decisions on new weapon systems, counterforce requirements, and targeting plans. Arms controllers didn’t fare much better in the Pentagon during Democratic administrations. Rising counterforce capabilities made a hash of deterrence stability even though their practical effect was to clarify that there was nothing delicate about the balance of terror.

The best that could be said about the ABM Treaty in the first quarter-century of its existence was that the offensive competition would have been even worse in its absence. The ABM Treaty finally served its intended purpose of enabling and backstopping deep cuts when two disbelievers in nuclear orthodoxy, Gorbachev and Reagan, decided to pursue them. When President George W. Bush decided to cast the Treaty aside as an impediment to the sole superpower’s freedom of action, he also made deep cuts in force structure unlikely.

Deterrence stability is a sensible and wise concept. It is achievable when decision makers in competing states are not in thrall to nuclear weapons. But it is a hollow incantation for competitors wedded to counterforce targeting to limit damage in the event of a nuclear war. Deterrence stability is anathema to those who seek advantage or, better yet, to escape from deterrence. Deterrence stability is also of no use to those who exploit instability for the prospect of gain or self-defense in a deep crisis.

Arms controllers deserve credit for the conception of deterrence stability, but not blame for its failure. The IR theorists, academics, and strategists who conceived of deterrence stability—and the companion constructs of strategic stability, arms race stability, and crisis stability—had only fleeting, or at best, peripheral influence on U.S. choices relating to offensive nuclear capabilities. In Washington and Moscow, these choices remain controlled by true believers in the political and military utility of nuclear weapons. If Beijing and New Delhi, which are now on the cusp of the counterforce compulsion, follow down this beaten path, they will have learned nothing from Washington and Moscow.

The aspirational constructs of deterrence stability, strategic stability, arms-race stability, and crisis stability remain in our lexicon because they continue to make sense in the abstract, and because arms controllers rightly remain wedded to them. They offer far preferable alternatives to the twin assumptions of escalation dominance and escalation control that are embedded in the embrace of counterforce capabilities and damage limitation targeting strategies. True believers in nuclear orthodoxy dare not lend clarity to their preferred abstractions; to do so would prompt revulsion by those on whose behalf a nuclear war would be fought. Instead, every modernization program and every refinement in offensive nuclear capabilities is defended generically in terms of deterrence, which sounds reasonable enough.

As nuclear dangers and dangerous military practices grow in several regions at once, the concepts of deterrence stability, strategic stability, arms race stability, and crisis stability remain extremely relevant. But these abstractions are not organizing principles. This terminology, conceived by eggheads—God bless them—does not connect with the general public. Nor do these constructs inform the procurement decisions of those who remain in thrall to the Bomb’s powers.

Arms control has been a process whose achievements have been widely taken for granted. Diplomacy accomplished what deterrence alone could not – to keep the Cold War from becoming hot while capping and then reducing strategic forces. Those who denigrate diplomacy have systematically set out to demolish the accomplishments of arms control, and they aren’t done yet. Their proposed remedies to reduce nuclear dangers consist of strategic modernization programs, which cost a great deal, and U.S. freedom of action to deal with proliferators, including the prosecution of wars of choice, which cost even more.

Nuclear dangers are now outpacing the ranks of those committed to their reduction. We owe much to those who laid the foundation for the practice of arms control at the outset of the Kennedy administration, but we can longer rely on their intellectual capital. One of the many challenges facing arms controllers—aside from figuring out what we now call ourselves—is to craft new terms of debate and new imagery to build popular support for our work. I’m not convinced that nuclear abolition is a banner that will take us very far and very fast. A process of nuclear arms reductions can, but it is stymied by Republican opposition and by poor relations between nuclear-armed states. Besides, this scope is too narrow, leaving many kinds of nuclear dangers unaddressed. What I do know is that progress in reducing nuclear dangers will require demystifying the counterforce compulsion and clarifying its hidden assumptions.


  1. Jonah Speaks (History)

    The title and first paragraph do not fit with the rest of the blog. You strongly favor deterrence stability as a primary goal of arms control; this is inconsistent with claiming deterrence stability is a hoax. You are strongly opposed to counter-force targeting; but counter-force targeting is real and presumably contributes to deterrence instability (a delicate balance of terror).

    Maybe the points you are trying to make are: that a “delicate balance of terror” is curable only through arms control, not through additional counter-force measures; and that deterrence stability does not happen automatically, but must be nurtured through the appropriate type of arms control measures. Neither point requires positing a “hoax.”

  2. Bradley Laing (History)


    With the 2020 Olympics and Paralympics in Tokyo just three years away, the government is stepping up efforts to prevent terrorist attacks using nuclear and other radioactive materials.

    The Japan Atomic Energy Agency has developed a device capable of detecting nuclear materials during airport baggage screening and is enhancing its nuclear forensics analytical technology.

  3. Jon Davis (History)

    I think technology is the factor that will change our reliance on nuclear weapons. Lasers, Anti-ballistic missile systems, hypersonic warheads – once we develop these technologies to the point that they are near 100% effective and reliable, then we can begin to walk away from nuclear weapons. I’ve always said that weapons are only retired once they become obsolete. Perhaps the arms control wonks should be encouraging missile defense and conventional prompt global strike? Or if this is going against the grain, then the work to be done in arms control is to make future treaties more encompassing by putting restrictions on more aspects of nuclear weapons. For instance, one area that has to be addressed is the total number of warheads in each country’s stockpile. Also, the hedge and inactive stockpiles need to be eliminated. And, reducing the allowable yield of weapons to maybe 100kt as a start. Maybe the hardest thing to overcome is limitations with inspections? On a good note, the field isn’t going to disappear anytime soon, unless there is some incredible AI algorithm being developed to solve the problem.

    • Andrew Locke (History)

      Developing the capability to conduct a Prompt Global Strike is a very bad idea. How would a targeted nation be able to tell if the incoming RV’s showing up on their warning systems would be carrying nuclear warheads or not? What about the plumes from missile launches being viewed by a satellite-based launch detection system? PGS makes a Launch-On-Warning scenario much more likely.

    • Jon Davis (History)

      I am imagining a future without nuclear weapons. At some point in time we will retire our nuclear weapons. We don’t use propeller fighters anymore. Also, you’re point only supports keeping nuclear weapons. Conventional PGS is a much better option.

    • Jonah Speaks (History)

      Another response to Andrew is that we definitely need to retire the whole notion of launch on warning. A country should never respond with nuclear weapons to an attack that is only maybe nuclear, but could just as easily be conventional. Arms controllers should think of ways to assure and/or verify that a country has adopted a no first use and no first strike policy. This can start with a verifiable physical inability to launch a prompt nuclear strike, and measures to ensure that conventional and nuclear weapons are always distinguishable.

    • Jon Davis (History)

      Jonah, I do agree with you on removing launch on warning. But it would probably too difficult to implement with Russia. However, I think a much better option is to focus on removing land-based icbms. The arms control community should focus on this. The argument has to overcome “responsiveness.” The triad is based on the belief of responsiveness, survivability, and flexibility is what provides to deterrence. If we would rely on just a dyad of submarines and bombers, we can remove any need for launch on warning. Also, we really don’t have a triad. We have a dual dyad of two kinds of ballistic missiles and two kinds of bombers. The “new” triad should be based on one ballistic missile(slbms), one cruise missile, and one gravity bomb. This is what truly gives us the most options. Land-based icbms are only a back-up.

    • J_kies (History)

      Mr Davis; various gee wiz weapon concepts (I won’t dignify lasers or CPS as being actual weapons) are not replacements for actual weapons in the eyes of the military. Hypersonics is a marketing meme (going really fast in a straight line is the simplest engagement problem known) and the vehicle concepts proposed for CPS began life as nuclear delivery systems in the 1980s (the SNLA SWERV MARV).

      If you want to talk to more effective conventional deterrent weapons (tanks, ships, planes and artillery) tossing high explosives, that’s fine and provides a notable advantage to the large nation state militaries such as the Russians, Chinese and ourselves. The US with our overwhelming advantages in conventional military systems is the largest beneficiary of reduced / eliminated nuclear inventories.

      “Tech no fix the weapon issue to help eliminate nukes.”

  4. Richard Pizzoferrato (History)

    Mutual Nuclear Deterrence MND is not MAD,,, does not require nuclear parity and is better known as Civil Deterrence, Civilized Deterrence, Civilizing Deterrence,
    nukes make venues unsafe for war….

  5. Jon Davis (History)

    The biggest mistake I see from arms control advocates is that they totally do not address military utility. Like the latest push to eliminate the LRSO/W80-4. Steven Pifer would say, we have to get rid of this new nuclear cruise missile because our strategic bombers already have a new weapon with the B61-12 gravity bomb. And his argument goes further to say that a conventional cruise missile can perform the job just as well as nuclear cruise missile. No one will address the issue that a nuclear cruise missile is the best weapon against mobile icbms. And this is why we will have a new nuclear cruise missile.

  6. Andrew Locke (History)

    A future without nuclear weapons? It would be great, like a future without hate, or climate change, or cancer, but it won’t happen anytime soon. Technology is wonderful in many ways, and it does progress, as it did in the early/mid-20th century, with the realization of how to harness the Strong Nuclear Force for explosive purposes. The enormously destructive nature of nuclear weapons, what makes them so horrible and unthinkable in use, also makes them so difficult to give up for nations that yearn for them, develop them, and posses them. They are seen by the nations that have them as a means to security, to freedom from intimidation and aggression, and for that reason are not likely to be given up. No one, not France, not Russia, not China, and certainly not the US or North Korea (especially with their current administrations), are going to give up their nukes for precision-guided conventional warheads riding on very expensive missiles. There has been only one nation that developed nuclear weapons that then got rid of them (South Africa), and the reasons they did had everything to do with the internal politics of that country. I like the idea of a total warhead yield limit, but how do you enforce a yield limit when what the US calls CNWDI (Critical Nuclear Weapon Design Information) is a closely-held secret, as it no doubt is for every country that develops such devices? It would also be wonderful to be free from fear that some nation will launch a nuclear strike based upon sensor data alone, but once again, human nature has to be taken into account. Fear of what might happen is what drives nations to acquire nuclear weapons in the first place, and that fear does not subside when the competition starts with an “enemy”, whether it be in accuracy, yield, range, or whatever. Who knows what real Launch-On-Warning policies India, Pakistan, or even Russia really have right now?

    • Jon Davis (History)

      You are only focused on the here and now.

  7. Andrew Locke (History)

    No, Jon, I am cognizant of the past, primarily focused on the present, and keenly interested in what the future may hold. I feel that developing and deploying new weapons systems that potentially add instability to an already unstable deterrence regime is utter madness. In your post to ACW made on 25 Feb 2015, you say that “I believe Prompt Global Strike will be most effective against non-nuclear states where we want to destroy a target currently assigned to a nuclear warhead”. You seem to be unaware of the dangerous uncertainty that the development of an alternative military use of ICBMs, utilizing very similar flight profiles to nuclear-armed roles, introduces into how some other nuclear-armed nations may respond to what they may see with their remote early sensing systems. To them, it may very well appear as if the US has initiated a nuclear strike against them or one of their mutual defense partners. The US, remember, has repeatedly, and publicly, declared that it reserves the right to conduct a nuclear first strike against whomever it feels deserving of such action. Say you wanted to conduct a PGS against a very high-value, time critical target located in a non-nuclear state. Would you launch just one missile, or two, or maybe even three to ensure the target is prosecuted with good effect? Remember, we are talking a conventional warhead of relatively limited yield. The MGM-31B Pershing II, as far as I know the most accurate nuclear-tipped ballistic missile that the US ever fielded, had a 100ft/30m CEP, thanks in large part to it’s terminally radar-guided MARV. This was an acceptable CEP for the type of command-decapitation missions planned for the Pershing II when carrying a nuclear warhead. When you substitute a conventional warhead for a nuclear one in even a precision guided missile like the MGM-31B, you end up with, to say the least, a massive reduction in effectiveness. Therefore, the chances of destroying the intended target go down accordingly, necessitating redundant targeting to achieve the desired effect. You see where this is headed; multiple ballistic missiles being launched to achieve the desired goal. Not a good thing when your potential enemy is watching with early-warning satellites, and may not have an independent confirmation of launch means (radar, telemetry) that allows time to pause, absorb and integrate the context of what they are seeing. The warning they perceive via satellite may be all they feel they need to respond. Numerous close calls during the Cold War resulted from erroneous computer and sensor data. They should serve as a lesson to all of us as to what can, almost did, and may still yet happen.

    • J_kies (History)

      Brilliantly and insightfully stated – bravo sir.

      Just to add ammunition to your masterful discussion; the Chinese and Russian hypersonic systems are publically asserted to be nuclear armed and the trajectory shaping is expressly to evade US BMD capabilities. In a further context of similarity; the vehicle behind the design for the US CPGS (now CPS) began life as a nuclear MARV research effort, the SNLA SWERV vehicle of the early 1980s.

      So the CPS people can pound sand; absolutely nothing about the trajectory provides distinct separation from nuclear tipped vehicles and CPS is astoundingly easy to misinterpret in use as a nuclear strike that just hasn’t landed yet.