Michael KreponWhat Went Wrong with Arms Control?

In the mid-1980s, many smart people predicted the death of arms control. Thomas Schelling was one of them. Others included Richard Haass, Albert Carnesale, Les Gelb, and Zbigniew Brzezinski.

US-Soviet relations were in terrible shape. The White House and the Kremlin were talking past each other. Uber-hawks in the Reagan administration compiled a long list of Soviet treaty violations – some true, mostly over the top – to block new accords, which seemed unlikely in any event, given the one-sided nature of the administration’s proposals. Almost everybody was surprised when President Reagan and Mikhail Gorbachev threw nuclear orthodoxy out the window to reach a deal eliminating intermediate- and shorter-range nuclear missiles in 1987.

Schelling’s pessimistic appraisal, “What Went Wrong with Arms Control?” appeared in the Winter 1985/86 issue of Foreign Affairs. His critique deserves another read because he was a founding father of nuclear arms control and because prospects for treaty making have once again ebbed. Schelling argued that arms control had “gone off the tracks” in part because “it often looks as if it is the arms negotiations that are driving the arms race.” More fundamentally, Schelling’s complaint had to do with how misshapen the process had become. In his view, the Anti-Ballistic Missile Treaty was “not merely the high point but the end point of successful arms control.”

The enterprise went astray, in Schelling’s opinion, for the lack of “any coherent theory of what arms control is supposed to accomplish.” He wrote, “I judge the proposals and negotiations on offensive weapons to have been mostly mindless, without a guiding philosophy.” A major contributing factor to Schelling’s disillusionment was “the shift of interest from the character of weapons to their numbers.” The obsession over “matching enemy capabilities whether we like them or not” made no sense. More heresies followed:

Nobody ever offers a convincing reason for preferring smaller numbers (I may exaggerate: saving money is a legitimate reason…). And some people think that with fewer numbers there is less likelihood that one will fall into mischievous hands or be launched by mechanical error; this I think is incorrect, but may not be worth refuting because it is in no one’s main motivation. For the most part, people simply think that smaller numbers are better than bigger… If people really believe that zero is the ultimate goal it is easy to see that downward is the direction they should go. But hardly anyone who takes arms control seriously believes that zero is the goal.

What, then, should strategic arms control be about? Or, as Schelling asked, “If you were to limit something, what would you want to limit?” His answer was limits to promote force postures consisting of “economical and reliable retaliatory weapons that are neither susceptible to preemption nor capable of preemption.” Stabilizing, offsetting nuclear force postures would share three “crucial elements:” an assured retaliatory capability, “restrained targeting and some capacity for war termination.” No surprise about element #1. Element #2 — counterforce targeting as the enemy of stabilizing deterrence and arms control – was ignored once these capabilities were in reach. Element #3 was hardly discussed.

Schelling thought highly of the ABM Treaty because it reinforced cautionary behavior:

A prudent restraint from aggressive violence that is based on acknowledgment that the world is too small to support a nuclear war is a healthier basis for peace than unilateral efforts to build defenses… Most of what we call civilization depends on reciprocal vulnerability.

The counterforce compulsion trumped the ABM Treaty. The United States and the Soviet Union didn’t take Schelling’s advice because decision makers were listening instead to powerful domestic constituencies, including competing laboratories, military services and missile production complexes. The three states currently increasing their nuclear arsenals at the fastest rate – China, India and Pakistan – might just succeed in avoiding the counterforce trap, despite having familiar domestic drivers. It’s too soon to tell, but so far, Beijing and New Delhi have been focusing on economic growth while competing well below their military industrial capacity.

The Cold War fixation with nuclear numerology may have been wrongly applied, but it remains our temperature gauge for nuclear dangers, our surrogate for measuring the health of the Nonproliferation Treaty regime and a key indicator of how much effort Washington and Moscow devote to reversing nuclear excess. Numbers continue to matter. They tell us whether countries are in a holding pattern, or headed in a more dangerous or saner direction. Getting to zero may well be an impossible journey, but movement in this direction is the glue that helps keep the NPT intact.

Schelling was nonetheless right about many things, especially the overriding need to focus on avoiding or reducing the war-fighting character of nuclear arsenals. Once counterforce requirements have been embraced, nuclear excess is sure to follow. Counterforce capabilities present far more of a roadblock than missile defenses in reducing the size of large nuclear arsenals. The safest harbor for stabilizing deterrence, as Schelling advised, can be found when retaliatory weapons are neither susceptible to preemption nor capable of preemption.


  1. Tom Nichols (History)

    The head of the Air Force association says that low numbers of nukes are immoral because they prevent counterforce targeting and thus require city-busting for deterrence to operate. (It’s right there in the Deterrence Maintenance Manual.)

    I loved the 1980s, but man, I never thought I’d have to live them twice.

  2. Stephen Young (History)

    Obviously, arms control has a long history of ups and downs. There was once a Secretary of State who wrote that there will “never be another treaty, of any significance, ratified by the Senate.”

    Nine months after Sec. John Hay said that, the Senate provided its advice and consent to the Hague Convention.

    Hay’s letter was dated Aug. 5, 1899, the Senate acted on May 4, 1900.

  3. Fred Miller (History)

    Since the nuclear age began, arms control and arms reduction has happened when there has been substantial pressure on the deciding politicians and an energetic awareness in the culture at large that the arms in question pose an unacceptable risk.

    From Kennedy to Reagan, arms control treaties have come from Presidents who had previously demonstrated pro-proliferation tendencies.

    Each was influenced by a large popular movement, and by cultural and artistic expressions from movies to popular music to statements by opinion leaders far from government, and acted quite contrary to previous statements, policies and budgets.

    As a young activist I helped build the antinuclear movement in the late ’70s and saw the birth of the Freeze. In the 1984 elections, I laughed at the efforts of Democratic candidates from Ted Kennedy to Alan Cranston to Walter Mondale to claim leadership of the Freeze, a movement they had ignored when they weren’t resisting it. When Ronald Reagan said he agreed with us in desiring an end to nuclear weapons, we laughed again.

    We were wrong: like Kennedy, Reagan had a deep desire for peace, and a horror of the consequences of nuclear war. We failed to understand that our path to nuclear arms reduction and elimination must include creating the political and cultural space in which political leaders can act on their desire for peace.

    The peace movement today is anything but robust. How and when it will revive is not clear. The last big surge came after a large popular movement against nuclear power created a network of hundreds of citizen activist groups. But it may be worth noting that the Nuclear Freeze itself came from an arms control wonk, Randall Forsberg, A SIPRI alum, not from someone with an activist history.

    The wonks may need to step up again, to create a formula that, like Forsberg’s Nuclear Freeze, crystallizes popular feeling and widespread unease into a political expression that is focussed and that can be applied by expert academics, diplomats, journalists and bureaucrats to motivate politicians to support workable legislation and treaties.

  4. Jonah Speaks (History)

    Element #3 “some capacity for war termination” could use some elaboration. Paraphrasing, Schelling (p. 230) complains that counter-force targeting became identified with limited, controlled war, while counter-value targeting became identified with total or spasm war. Schelling believes the exact opposite is true. Counter-force war would be uncontrollable and lead to population targeting. Counter-value war (“a controlled retaliatory capacity”) need not be utilized to the full extent, so that some population might be spared.

    Schelling provides little rationale for this opinion, except this: “But as long as the counterforce doctrine is governing, it will be hard to impose a reciprocal denial of substantial preemptive capabilities, since the capability to destroy hard targets, publicly eschewed by McNamara, has become central to the doctrine.” Reading between the lines, Schelling believes counterforce targeting leads to greater likelihood of first strike.

    The conclusion that counterforce war would be harder to terminate does not follow — unless we add some further explanation. One explanation could be that a first strike necessitates massive numbers of weapons, and a second strike also necessitates massive numbers of weapons. This leaves little left for controlled follow-up strikes, particularly if we assume command and control is eviscerated. Countervalue war would be more controllable, if we assume cities or other soft targets can be destroyed one or a few at a time, incrementally. However, Schelling did not say any of this, so the basis of his view is a mystery.

  5. Bradley Laing (History)


    —Mr. Krepon

    —Please look through pages 51-57 to see if any of the information in this document about the UK’s nuclear force should be listed here in the comments section.

  6. Nick Ritchie (History)

    Arms control has quite a narrow meaning in US-Soviet/Russia relations, meaning(s) that evolved within a specific social and historical context that still lingers. Perhaps its more useful to think of arms control in this context as a cultural practice – a set of ‘appropriate’ behaviours constitutive of a common culture of Cold War interaction the steadily evolved between Washington and Moscow. The context has changed but institutionalised cultural practices have been slow to catch up.

    I think it is more appropriate now to position nuclear weapons management and disarmament within a broader international normative project of controlling the means of political violence. Nevertheless, it is still a little too easy to revert to Cold War nuclear numerology through the argument that the US and Russia have the most weapons so they must take the lead so we need another START-type treaty that reproduces Cold War cultural practices and so on.

    One of the key issues at stake here is whether the current configuration of the global nuclear ecosystem is stable, sustainable, and legitimate. For many it is not and global nuclear risk can only be managed sustainably through disarmament under the rubric of a broader normative project. For others it is all these things, but only by virtue of American power.

    • krepon (History)


      It’s asking a lot of arms control to contribute to “a broader international normative project of controlling the means of political violence.” Controlling the most dangerous weapons is daunting enough.

      But I do think you are making an important point about norms. My thinking has gravitated in this direction, as well. Still thinking this through, and could use some help.

      Strategic arms control has focused on numbers. Sometimes, as Schelling said, the numbers crowded out larger considerations.

      Other kinds of arms control dealing with WMD focused on norms, establishing or reaffirming customary practices among responsible states. These norms required the number zero. Zero was the number embedded in the CWC and BWC. States that do not honor this number become, ipso facto, outliers. Because numbers of these weapons that are greater than zero can be hidden, suspicions are conclusively affirmed by use. The use of these weapons becomes an indelible stain on the user. Especially for the battlefield use of nuclear weapons.

      Another huge zero: zero testing in the atmosphere. And underground. Ergo, a norm now honored by all states except North Korea. Even as the CTBT languishes.

      Norms are more about use than about number. They’re also about treaties, but treaties for the foreseeable future won’t enter into force or, if they do, they will enter into force without key parties.

      We seem to be in a period of transition: arms control is becoming more about norms than about numbers and treaties. But the numbers still matter and can’t be discarded. How to meld?

      I’ll return to this subject in a future post to provoke more commentary. Greater clarity on how norms, numbers and treaties come together would be helpful.


    • jeannick (History)

      Certainly , one of the function of disarmament talks is the creation of norms and procedures for dialogue
      nothing more but nothing less either.
      anything else is political

      The recent release of official documents relating to
      “valley forge 83” give a bit of relief to Arm control
      as see by NATO at the time .
      it led to the realization that the Soviets were afraid of
      an American first strike ,
      something the politicians had dismissed

  7. magoo (History)

    Michael Arms Control as they have and are being negotiated are limited by the fact that the US and its allies are determined to maintain the supremacy that the Atom Bomb bestowed on them in the summer of 1945. All arms control negotiations are designed to, borrow a phrase from the Chinese, to be “with American Characters” They lack universality and are discriminatory. No amount of semantics or phraseology can replace the objections of other states that are expected to toe the line.

  8. Mark Gubrud (History)

    I agree we need to think about the character of the weapons (and about how they might be used, and what that implies) but such thinking gets pretty involved. I don’t think it is completely subjective, in fact I think a lot of things are very clear, but I often find that others don’t agree with me about these things, and in general it seems hard to establish a broad politically effective consensus about these things. Numbers are at least perceived to be objective, even if they may hide a lot of complexity and subjective judgments. That’s why Russia’s apparent fixation on numerical parity is not to be dismissed; it is important that arms control treaties are supported by strong political consensus, which you don’t have if everyone is arguing about which weapons are first-strike weapons, etc.

    Also, I think the context of Prof. Schelling’s remarks has changed since 1985 but in any case count me as someone who takes arms control very seriously and also strongly believes that the numerical goal (not to be weakened by the modifier “ultimate” since it need not and must not be a distant goal) is zero.

  9. Jonah Speaks (History)

    I agree with the goal of nuclear zero, even though I have heard that Schelling objects. I give it a 2/3 chance that if the world seriously tried to, we could get to zero and stay there. Even if the 1/3 chance says it can’t be done, we can still get to way smaller numbers, measured in the dozens, not in the hundreds or thousands. This would make norms of no first use (or any use) much easier to enforce: If you do this, the world will not end, and the world will definitely overthrow your regime and punish you severely.

    The problem is political will. Obama has tried, but he is spitting against the wind right now. How to organize an effective political movement in the U.S. and worldwide is a daunting challenge.

    I also do not agree with Schelling’s complacency. Just because nuclear deterrence has “worked” for 40 (now 68) years, does not mean it will work forever. We could just as easily have been very lucky during the Cold War. The coin of nuclear disaster may simply have flipped heads instead of tails. Disaster could still face us in the coming century, even though we have been lucky so far.

  10. Andrew Tubbiolo (History)

    Arms control requires one to decide what a armament system does to the world and how the world (or whatever it is you’re working for.) would be better off with that system controlled. This requires a in depth understanding of how things are now, and how to move the variable set to reflect the desired reality in order to test your ideas in the lab of the real world. This sort of foresight is gone from American politics. It’s pretty clear to me that without a a peer competitor Americans lapse into mindless pursuit of policy to a level that borders on religious. Given how much blood and treasure we’ve spent on this process, I think I’ll go further and call it religious.

    Then there’s the rest of the world. I think from their point of view, they are content to let the Americans have at it. Europe has its social democracies and nobody is threatening them anymore. Asia is content to accept the American business sector’s transferal of the American industrial sector to their nations with the added benefit of then exporting to American consumers. Who would say no to that? It’s not their place to help the Americans, they’re content to further their interests and let the Americans go broke. I would expect no vision from them until our world order falls apart. Most likely their vision of their desired world order is a flawed as ours. So I would not expect anything out of them either.

    The Americans and Soviets raced each other in engineering, science, arms, entertainment, propaganda, and sometimes even genuine peace efforts. The various power blocs of the world today don’t race each other in this way anymore. Furthermore, the worlds power blocs are not poising large military forces against each other. Without the threat of war, making peace is just not as important as it used to be.

  11. jens (History)


    one aspect that is mentioned by Michael Krepon is the tension between numbers and use. One term that might be useful is “option”. the aim, then, is to constrain certain (destabilizing) options. Certain options are only feasible, if you have a certain quantity and quality, but the term “options” is more than that quantity, because it covers the aspect of using weapons.
    Putting options to the center of the debate could also allow arms control measures that do not require parity in terms of numbers (e.g. unilateral steps or some reciprocal steps that Thomas C. Schelling described in his essay on reciprocal measures).But here is the problem that arms control has not only a technical dimension but also a political. This dimensions makes steps that do not require parity very problematic because parity in numbers is for some states equal to parity in status (India-Pakistan, for example.)

  12. Bradley Laing (History)


    Revealed: how Israel offered to sell South Africa nuclear weapons

    Exclusive: Secret apartheid-era papers give first official evidence of Israeli nuclear weapons

  13. Bradley Laing (History)

    Safety issues with UK’s nuclear subs and facilities used to repair missiles
    Cracks in reactors and nuclear discharges found in Navy’s oldest boats
    Nuclear-qualified engineers are quitting over poor pay and conditions
    Experts described latest report as the most worrying they had seen

    Read more: http://www.dailymail.co.uk/news/article-2384224/Revealed-Shock-Code-Red-safety-report-British-nuclear-subs-fleet-hit-leaking-cracked-reactors-lack-trained-staff.html#ixzz2ay3U0SF4
    Follow us: @MailOnline on Twitter | DailyMail on Facebook