Jeffrey LewisThe Next START

What fan of the hapless Chicago Cubs doesn’t look at a sunny day and think of Ernie Banks? His “Let’s play two!” expressed optimism in the face of fatigue, off-field worries and the inevitability that a doubleheader meant the luckless Cubs would receive two merciless drubbings by the visting team.

Come to think of it, Ernie was sort of masochist, wasn’t he?

A lifetime rooting for the bumbling Chicago Cubs is perfect preparation for a career in arms control, where the successes are few and far between, and always a bit too little, too late.

Yet hope springs eternal!  So, while bruised Administration politicos despair at the thought of going another round with the Russians and Jon Kyl, those of use in the arms control community are thinking: What should the next nuclear arms reduction treaty look like?

A kind of consensus is forming around the idea that the next treaty will include a further reduction in deployed strategic nuclear weapons (the ones covered by START) as well as a limitation on the total number of nuclear weapons (everything else including reserve and tactical nuclear weapons).

Former Secretary of Defense Bill Perry described such an approach in his testimony on New START, albeit somewhat elliptically.  (For a nice articulation of this view, see Steve Pifer, The Next Round: The United States and Nuclear Arms Reductions After New START and After New START: What Next?)

People seem to like the numbers 1,000 (deployed strategic) and 2,500 (total), though I think 1500/3000 would be a monumental achievement.  Current US levels, for reference, are 1,968 and 5,113.

The advantage of this approach — lets call it 1500/3000 — is that a single category of nondeployed warheads offers a solution to the disparity in tactical nuclear weapons, which currently “favors” Russia.  As it happens, the Russian advantage in tactical nuclear weapons is offset by the US advantage in strategic “reserve” or “hedge” warheads.  Here is how Bill Perry explained the situation:

Dr. PERRY. Yes. In my testimony I express the hope that the confidence-building that would develop from this treaty and the ongoing dialogues it would have would lead to improvements in many other areas, not just further nuclear treaties, but in the other areas of disagreement between the United States and Russia, but in particular it would lead to a follow-on treaty dealing with the tactical nukes and also dealing with the thousands of reserve warheads that we have. I might mention that the asymmetry in tactical nuclear weapons is primarily in favor of the Soviet Union, but the asymmetry in strategic weapons in reserve is primarily in the favor of the United States and is a very sore issue with the Russians that I speak to. We have the capability of rapidly uploading thousands of nuclear weapons onto our strategic forces if we choose to do so.

“Very sore.”  That’s octogenerian for “pissed off.”

The main reason for the asymmetrical force structure is technical: As former Secretary of Defense (and Energy) Jim Schlesinger pointed out twice in the same hearing, the United States keeps a large reserve when Russia does not because of differing approaches to stockpile stewardship.  We maintain existing weapons indefinitely — so we like to have lots of backups.  The Russians appear to just keep churning out replacement warheads before the old ones go bad — “turning over” the entire force every 10 years or so.  (I like our approach better.)

If US-Russian relations took a turn for the worse — “Joe Stalin comes back to life” in Arnie Kanter’s colorful phrase — we could also upload those  same reserve warheads on our missiles and bombers, giving us a pretty serious breakout capability that Moscow can’t match.  Which is what pisses off the Russians.  Er, makes them “very sore.”

So we each keep thousands of reserve weapons for different reasons: The US as a strategic hedge against technical and geopolitical uncertainty; the Russians as a tactical option in the event that things turn south with NATO or, more probably, China.

This isn’t as well understood as it ought to be.  For example, the New York Times, in an otherwise sensible editorial, noted that “Russia and the United States are each estimated to have around 2,000 stored weapons.”  I don’t know who made that estimate (passive voice!), but I am pretty sure that’s not the US IC estimate.  The Russians MIRV the hell out of their missiles and probably only have a few hundred strategic warheads in reserve to finish off an upload.  US officials don’t usually come out and say it, but you can sort of see it in various US documents.  There is, for example, the tendency to report the number Russian operational strategic warheads, rather than deployed warheads.  And the number (4,000 in 2004) turns out to be a bit lower than the number of START accountable warheads.  It makes sense when you think about it: If Moscow doesn’t have a bunch of upload and is regularly remanufacturing warheads, why bother with a hedge?

As a result, I would guess US and Russian nuclear stockpiles after the New START reductions (in 2017) will probably look something like this:

Warhead Type United States Russia
Strategic Deployed ~1,500 ~1,500
Strategic Nondeployed ~2,500 ~500
Tactical (All Nondeployed) ~500 ~3000-5000
Total ~4,500 ~5,000-7000

Now, these numbers are purely illustrative.  I probably ought to have just have written a “few thousand” and “few hundred” for non-deployed US and Russian strategic nuclear weapons and “several hundred” and “several thousand ” US and Russian tactical nuclear weapons.

Two other cautions: This accounting doesn’t use the New START accounting fiction for bombers — instead, I tried to depict what the actual stockpile will look like and placed the bomber warheads as nondeployed. And, then, of course, there are the weapons in queue awaiting dismantlement — a status differentiated in the US largely by absence of batteries and other so-called “limited-life components” that could be restored to return the weapons to active status.

But the bottom line is this: Any agreement that constrains all nuclear weapons is probably  going to allow the US and Russia to retain thousands of non-deployed nuclear weapons for their respective national reasons.  We might trim the hedge, as it were, but there is little prospect of eliminating either it or Russia’s tactical nuclear weapons stockpile in the next round of negotiations.

The implications of this are interesting.  In the near-term, one wonders if this approach will discourage efforts to transform the “hedge” from extant weapons to a “responsive” production capability (virtual Swords, anyone?).  After all, if we plan to use the hedge as a bargaining chip, why make unilateral stockpile reductions beyond implementation of the Bush Administration’s last cut in December 2007?

On the other hand, an effort to reduce the hedge means having to prepare the technical option to consolidate the stockpile of backup warheads.  So the W88/W78 FrankenLEP looks like an inevitability, even if it seems to be a solution in search of a problem.  And, for at least the foreseeable future we might end up with both: lots of interest in newish warheads and a big backlog of old ones while Antonov attempts to slowly talk Rose to death in Geneva.

The verification challenges with this approach are also intimidating.  The United States, early on in New START negotiations, sought to actually count warheads in storage at bomber bases.  (The Bush Administration had been chagrined to learn that the Russians didn’t consider any of their bomber warheads under the Moscow Treaty to be “operationally deployed.”) The Russians said “nyet,” which is how we ended up with the accounting fiction that each bomber counts as just one nuclear weapon. Given the very large number of general purpose forces that could employ tactical nuclear weapons — the US plans to procure 2,457 F-35s — any scheme for limiting tacnukes will have to make an actual count of warheads.  That means learning if the Russians have a word for “yes” with on site inspections of airbases, among other sensitive locations.

Then, of course, there is the question of what to call this new treaty.  START, SALT and SORT, all started with S because these treaties concerned strategic nuclear weapons.  I suppose we can’t call it the Comprehensive Arms Reduction Treaty, without Heritage accusing arms control proponents of putting the CART before the horse.  And, for obvious reasons, the Follow-on Arms Reduction Treaty is inappropriate.

So here is a novel idea: Let’s dispense with the cute nicknames.  Other treaties are named after places (Geneva and The Hague do well for themselves, though even Dayton got an agreement), saddled with lame acronyms (NPT, ABM, INF) or just called what they are (Outer Space Treaty, Open Skies Treaty).  I don’t care if we call this next one the Reykjavik Treaty, the NWSR or the just the Nuclear Stockpile Reduction. But the details will be hard enough without spending a lot of time on clever names.


  1. John (History)

    I am in total agreement that the number of warheads is the critical point: controlling them is the key to dismantling the largely artificial disconnect between strategic and tactical “weapons”, when in fact the only meaningful difference between “strategic” and “tactical” are the delivery systems (which may actually be the same, given long fighter-bomber ranges with refueling in the wake of significantly weaker air defenses).

    The next round should simply be called the Nuclear Weapons Reduction Treaty: NWRT. No more differentiation between strategic and tactical, no more endless discussions of what is a delivery system and what is not, simply count the number of warheads and reduce them.


  2. FSB (History)

    Ah yes, big dreams about the golden future!

    I wonder if you or Steven Pifer followed the comments on Krepon’s last post — e.g. this excerpt of one there:


    “Russia and the West must reach a suitable agreement on the issue of missile defense in the next decade otherwise Moscow will have to adopt and deploy new strategic weapons, President Dmitry Medvedev said on Tuesday.

    “We are facing the following alternative in the next ten years: either we reach an agreement on missile defense and create a fully-fledged joint mechanism of cooperation, or if we fail to do so, a new round of arms race will start, and we will have to adopt decisions on the deployment of new strategic weapons,” Medvedev said.”

    Now, back to rose-lensed dreamy-land…

  3. anon (History)

    Jeffrey, you’ve offered up some interesting arithmatic, but its just arithmatic. In the old days, we’d design the limits in an arms control treaty to enhance or ensure crisis stability and strategic stability. We banned land-based MIRVed ICBMs because they were lucrative targets in a first strike (and because the Russians had lots of them and that scared us); we capped SLBM warheads in START II because Russia argued they could pose a first strike threat (and we had lots of them and that scared Russia), etc.

    We didn’t simply pick a round number with lots of zeros and say it was the right number. What’s the strategic rationale for 1,000 warheads? How would we structure, deploy, and operate our forces at that level? How would the Russians structure and operate theirs? Would those operations enhance or undermine stability? Would we have to alter our deployment patterns and our employment plans at that level? If yes, shouldn’t we determine whether those changes would enhance stability before we pick a number that would force us to make those changes? We avoided all these questions in New START because we don’t really have to eliminate any (or many) delivery vehicles. We picked an RV number and outlined counting rules and elimination rules that protected our current deployment patterns and operations.

    One could ask the same questions about the aggregate limit. Can we structure our stockpile, at that level, to maintain confidence in the reliabilty of our deterrent?

    I’m not saying that your numbers or your strategy for the next treaty are wrong. But I have no way of determining whether they are right. Its just arithmatic. And, until I hear a strategic rationale for them, and hear about how they affect our forces and operations, and determine whether those changes will enhance or undermine stability, I think you really are putting the cart before the horse…

    By the way, when we signed the Strategic Offensive Reductions Treaty in 2002, nonone in any offical position called it SORT. It was called the Treaty of Moscow (or Moscow Treaty). SORT was a name attached to it by the arms control community, and was somewhat derogatory because that 3-page thing was just “sort of a treaty.”

    • Jeffrey (History)


      I started with the NPR conclusion we could go as low as 1300 now, and the planning assumption that you want 1 backup warhead per deployed warheads.

      That’s where I got 1500/3000. Assumes no further cuts to deployed forces.

      Pifer and others have a good argument for 1000/2500, in that we can go lower than 1300 if the Russians do, which means you got 1000 deployed, 1000 in reserve and 500 to fool around with.

      (I typed 1100, when I meant 1300. Doh! Thanks K.)

    • Kingston (History)

      Jeffery, wasn’t it 1300?

      I thought Mort Halperin’s take on this in the May 2010 issue of Arms Control Today (excerpted below) was interesting, though as Jeffrey points out, it’s not clear to me that we would need fundamentally new guidance for the 1500/3000 limit. I don’t see why we couldn’t maintain the triad (and deterrence in both oceans) at this level either.

      “The most disappointing part of the NPR is the section dealing with decisions about the size of the deployed force and the continued reliance on a version of mutual assured destruction as the basis for determining force size and posture. The business-as-usual approach resulted in part from the way that the NPR was conducted. In order to permit the negotiations with the Russians to begin early in the administration, and to do so without opening itself to the charge that it was negotiating before completing its own review, the Pentagon was instructed to do an early minireview that focused on the issues relevant to the New START negotiators. To do this quickly and without intense dispute, the minireview proceeded on the basis of existing guidance from the Bush administration. The strategy worked in the sense that the reductions consistent with that guidance were more than sufficient for the modest changes envisioned in New START.

      Further reductions, however, will require significant follow-on study and new presidential guidance that could take years to complete. With the treaty text settled and the NPR released, the administration must now undertake a variety of actions, including “[c]omplet[ing] the Presidentially-directed review of post-New START arms control objectives, to establish goals for future reductions in nuclear weapons, as well as evaluating additional options to increase warning and decision time, and to further reduce the risks of false warning or misjudgments relating to nuclear use.”

    • Mark Gubrud (History)

      “We didn’t simply pick a round number with lots of zeros and say it was the right number.”

      How can you be serious? How do you know what the correct theory of stability is, and how does anything about operations dictate the correct number of nuclear weapons? How do you decide that This Story about why we need this many weapons is The Truth while That Story about why we need that many is Just Plain Wrong?

      I say The Truth is it’s completely arbitrary, and it only seems that one has a way to calculate that some particular view is correct when one is steeped in some particular political context with its balance of political forces and their deployed arguments.

      There are always enough rationales out there for any number you choose. For some people, the right number is always a larger one. For others, the right number is a lower one.

      I say the best number of nuclear weapons, and the only one that deserves particular respect, is zero.

    • Anon (History)

      The magic number has more to do with bureaucratic infighting and $ to be distributed (witness NNSA inflation) and what the other big dude (Russia) has, than with any rational arguments — if any rational arguments about nuclear weapons even exist, which is up for debate.

      If Russia and China go to 10 nukes tomorrow, 50 nukes will be enough for the US.

      Why do we have a triad? Because the AF needs something to do.

      For deterrence, a sub based system suffices maybe with 50 nukes.

      Didn’t you write something on min. deterrence?

    • John Schilling (History)

      A sub-based system with maybe 50 nukes, might work just fine right up to the point where someone makes a real breakthrough (technical or otherwise) in antisubmarine warfare. At that point, they are looking at a brief window in which they can maybe defeat all of their enemies and/or rivals, achieve all of their geopolitical goals, establish themselves as the One True Superpower. Or maybe they can’t, but they are now also looking forward to each of their enemies/rivals facing the same decision in a few years when they come across the same breakthrough.

      This strikes me as very, very destabilizing. Obviously, you do not believe this is going to happen. Are you truly prepared to wager a hundred million lives on it? Because I’m not, and I doubt I’m alone. The logic behind the triad remains, even if we reduce the warhead count in each leg from thousands to hundreds or dozens.

    • Mark Gubrud (History)

      The question you have to ask is whether such a “breakthrough in antisubmarine warfare” scenario is technically realistic, and if you think it might be, is whoever makes this “breakthrough” going to be so confident of a clean first strike, and so motivated to carry it out, that they are willing to wager 100 million lives on it? I think that what we have here is a comic-book technology wielded by a comic-book enemy, and we can in fact be confident enough that this would never happen. Remember that the next question, after you do the first strike, is what do you do next? Which is one reason I don’t worry about the “bombs hidden in caves” objection to abolition.

      At worst, if one nation acquired the ability to keep track of another’s boomers with high confidence, the other would be likely to know about this in time to take corrective action. It would certainly involve some type of massively proliferated sensor system which would not be unobservable.

      But if you doubt this, it only once again illustrates my point that the calculation of any particular number of weapons needed for stability and security is pure pretense. The only uniquely safe number is zero.

    • Scott Monje (History)

      The Russians claim to be working on a satellite that will detect submerged submarines. Would that come within the range of the technically realistic, or is it likely to be posturing?

    • John Schilling (History)


      There has been a great deal said about possible breakthroughs in strategic antisubmarine warfare. I gather you would say technically you don’t think anybody in the world knows how to do such a thing, and you feel confident it will not be done for a very long period of
      time to come. You think we can leave that out of our thinking. You wish the American public would leave that out of their thinking.

      We are having this discussion because several comic-book technologies came together into robust, operational weapons systems in a surprisingly short time and with little clear warning to the targets.

      As for what you would do with such a technology (or non-technological solution), the first thing is you worry about an enemy being maybe only a month behind you in developing the same technology and implementing his “what do you do…” solution. Then you start making hasty decisions on the basis of incomplete information, to be implemented with nuclear weapons. I do not share your confidence that this will end well.

    • Mark Gubrud (History)


      A little more specificity would greatly improve the credibility of your argument. As it stands, it isn’t even clear whether you are claiming that some technical basis exists for anticipating one or more “breakthroughs in strategic antisubmarine warfare,” or whether that would in fact mean that our SLBM-carrying submarines would suddenly be vulnerable to surprise attack when at sea, or whether you are only referring to previous instances in which “technologies came together into robust, operational weapons systems in a surprisingly short time and with little clear warning to the targets.” If the latter, you do not say what those examples are, or why they suggest the possibility that ballistic missile submarines would suddenly become vulnerable, in the absence of a specific technological concept as to how this might be achieved.

      Again, I am not denying it is possible for submarines to become vulnerable, but given the immense effort that has been expended on this problem over the years I don’t see any reason to think that this would happen as a result of a sudden unforeseeable “breakthrough” or that we would not see it coming if, for example, it became possible to fill the oceans with enough cheap sensors to ensure detection of even the quietest subs with the kind of certainty one would want before gambling one’s civilization on an optional nuclear strike — one which would still leave you with the question “What do you do next?” to prevent the victim from simply rearming and coming back at you with a vengeance.

  4. Pavel (History)

    Russia sort of decided that the “upload potential” is not really an issue (if it ever was a real issue). As the New START showed, Russia is clearly okay with the U.S. capable of increasing the size of its strategic force to about 4,500-5,000 warheads from the New START level of 1,550. There is no reason why Russia would not be okay with the U.S. being able bring the number to 4,500 from the level of, say, 500 (or 100) operationally deployed warheads.

    I wouldn’t expect Russia trading off its tactical nukes for the upload potential. There is no reason for it to. Most likely, tactical nukes will be linked to CFE, NATO (including U.S. weapons in Europe), missile defense in Europe and things like that.

  5. Pete (History)

    First the Russian word for “yes” is “da”. Second, I agree with Anon (as does everyone else I’m guessing) that 1,000 is just a pretty number. Does anyone know how other pretty numbers were handled with previous arms (not necessarily nuclear) reductions? Was there a debate about going below 10,000 warheads? 5,000?

    • anon (History)

      The “5000” in the original Reagan START proposal was just a pretty number. It was 50% reduction from existing U.S. warhead levels. But it included only missile warheads, no limits on bombers at all. Everyone in the Reagan Admin knew the Russians would reject this, and they did, so they saw no reason to propose anything with real analysis behind it. (By the way, Perle’s original proposal for a zero option for INF was also intended to be cute, not real. Imagine our surprise when Gorbachev said yes.)

      Anyway, old START ended up at 6,000, to allow room for bomber weapons, with the 4,900 sublimit on missile warheads. But it took several years of negotiations to agree on this number, so the U.S. had time to rethink the triad (we cut Tridents from 24 to 21 to 18 boats, kept Peacekeeper at 50, rather than 200 planned missiles, cut planned warheads on Trident missiles from 12 to 8, etc). Some of these changes in force structure had nothing to do with the pending START limit, but, because the treaty took 7 years to negotiate (more than a FYDP), there was time to build the number into future plans. We also changed our targeting and employment policy right before START entered into force — there were major studies going on at SAC and DOD (Lee Butler has a great quote about the day he threw thousands of targets out of the war plan with one stroke of a pen). Some of this was helped by the collapse of the Warsaw Pact and Soviet Union (taking many countries out of the war plan), but it also resulted from changes in targeting theory and strategy. Basically, the target base had grown to accomodate a force with 10,000 warheads; there was lots of redundancy and it wasn’t too hard to shrink it if you were only going to have 5,000. The same may not true at the levels we are at now, particularly if you start cutting Tridents so much that you can’t maintain two-ocean basing. You’d basically need another theory or strategy of targeting, like a switch to minimum deterrence (which is how the above-mentioned Air Force Officers get to a force of 312.)

      And there was a debate about going from START I to lower levels (5000 or lower) in the 1992-1993 time frame. There were many “deep cuts” studies by lots of organizations. And, inside the Pentagon, they decided they could go lower if they combined the lower number with the de-Mirving of land-based missiles and an additional reduction in Trident submarines (18 to 14), if the Tridents all carried D-5 missiles (i.e. all went through the backfit program). That was the conclusion of the 1994 NPR.

      My point is not that we have to stick with the kind of employment policy we have now. My point is that it makes far more sense to figure out if you want to change your employment strategy, then decide how many warheads you need to implement that strategy, than it does to pick a cute number for your warheads, then find yourself forced to identify a consistent employment strategy. In other words, arms control should do more than just reduce warhead numbers, it should enhance stability and security at lower numbers of warheads.

  6. FSB (History)

    Three Air Force thinkers, James Wood Forsyth Jr., Col. B. Chance Saltzman (chief of the Air Force Strategic Plans and Policy Division) and Gary Schaub Jr. concluded that “America’s security can rest easily” on a comparatively small nuclear force:

    The United States, they wrote, could “draw down its nuclear arsenal to a relatively small number of survivable, reliable weapons dispersed among missile silos, submarines, and airplanes.” They said such a force might number only 311 nuclear weapons.

    Another pretty number.

    I say we need 30, or 33 if we use the unreliable RRWs

  7. P.E.T. (History)

    I would say it’s because some here seem to think Arms Control is about getting to “zero” ASAP. For those who do, I would suggest the next round of START be referred to as “Future Arms Reduction Treaty”.

    • Jeffrey (History)

      It’s “Follow-on” Arms Reduction Treaty.

  8. Fred Miller (History)

    The political momentum for nuclear arms control has always depended to a significant extent–sometimes almost entirely–on a surge of public opinion. Right now, that surge is not apparent.

    The public, and the activist public, is worried about other matters. Afghanistan, the military budget, and Israel/Palestine are of much greater concern than nuclear arms control, and none of them are at all competitive against employment, the economy, government spending and other domestic concerns.

    There is precedent for a groundswell of antinuclear opinion, especially the Nuclear Freeze movement launched in 1980 by Randy Forsberg, which quickly became a major force in national politics and forced the START Treaty. But Forsberg had an advantage: the late 70’s had seen large protests, often including thousands of arrests, at nuclear power plants nationwide.

    Building the kind of popular momentum needed to push politicians ahead on significant disarmament negotiations seems to depend on some simple, symbolic formulation like the Freeze, or on someone stealing a technological march by launching an antinuclear social networking phenomenon. It may take both, but it will also take a lot of hard work building levels of concern about nuclear weapons in a world where they have ceased to scare people.

  9. FSB (History)

    Argument for reliable replacement plumbing:

    U.S. Nuclear Command Headquarters Goes Cold After Water Main Break
    Friday, Jan. 7, 2011

    By Elaine M. Grossman

    Global Security Newswire

  10. Pavel (History)

    My comment got lost somehow, so I’ll try again (a brief version): The “upload potential” does not seem to be an issue for Russia (if it ever was a real issue). New START showed that Russia is okay with the U.S. having the capability to increase the numbers to about 4,500 from the 1,550 New START operationally deployed warheads. There is no reason it would be different at the operationally deployed level of, say, 500 or even 100.

    P.S. It’s really interesting to see that there are people who believe that these numbers are anything but arbitrary.