Jeffrey LewisIke, Dubya and Nuclear Forces

On December 31, 2012, when the Moscow Treaty both takes effect and expires, the United States will have the lowest force levels since the Eisenhower Administration, or at least that is what the latest State-Defense-Energy white paper says:

In 2001, President Bush directed that the United States reduce the number of operationally deployed strategic nuclear weapons from about 6,000 to 1,700-2,200 by 2012 – a two-thirds reduction. Corresponding reductions in the nuclear stockpile will result in the lowest level since the Eisenhower Administration.

This claim is worth our consideration in detail, as it is one of the regular talking points delivered by Bush Administration officials and defenders of our current stockpile plan. To wit:

  • “When we have completed this task, our nuclear arsenal will be at about a quarter of its size at the end of the Cold War, and will have reached its lowest level since the Eisenhower Administration.” Chris Ford, United States Special Representative for Nuclear Nonproliferation, March 17, 2007
  • “You should not hesitate to remind critics that President Bush negotiated the Washington-Moscow Treaty to cut the number of weapons to the lowest level since the Eisenhower administration.” Senator Pete Domenici, April 18, 2007
  • “By 2012, the stockpile will be lower by nearly one-half from the 2001 level – down by roughly a factor-of-four since the end of the Cold War and the lowest level since the Eisenhower Administration.” Linton Brooks, then-Administrator, National Nuclear Security Administration, March 3, 2006
  • “It’s the lowest level since the Eisenhower administration.” National Nuclear Security Administration spokesman Bryan Wilkes, February 22, 2007.

This talking point—“lowest level since Ike”—is interesting because of what it says about secrecy and nuclear weapons policy.

When Ike left office in January 1961— as you can see from the NRDC chart that I slightly modified—the United States nuclear stockpile totaled about 20,000 nuclear warheads. Every President since George H.W. Bush has been able to claim that the stockpile was at its lowest since the Eisenhower Administration—which, frankly speaking, isn’t really an accomplishment.

But what else can they say? Although the Bush Administration does appear to be making reductions, it also continues the long-standing secrecy policy that treats stockpile size and composition as classified information.

As a result, Administration officials are reduced to blabbering about the Eisenhower Administration, when they should be talking hard numbers.

What Kind Of Reductions?

Here is what we know about the size of the US nuclear stockpile and reductions since Bush came into office.

Sometime after the 2001 Nuclear Posture Review (released in January 2002), President Bush signed NSPD 10, a nuclear weapons stockpile plan that implemented the reductions starting in Fiscal Year 2002. Then, in May 2004, Bush signed NSPD 34 “Fiscal Year 2004-2012 Nuclear Weapons Stockpile Plan.”

Although these documents are classified, Linton Brooks stated that the arsenal would be 1/4 the size at the end of the Cold War and about 1/2 the size in 2001. Assuming that NRDC is right about the number for 1989 and 2001, then the United States should have about 5,000-6,000 nuclear weapons in 2012. (Hans Kristensen estimates 5047)

Almost all of the remaining 5,000-6,000 weapons will be strategic. That is interesting, because it means the US stockpile of strategic nuclear weapons—those carried by SLBMs, ICBMs, and heavy bombers—will be at its lowest level since, well, the late Eisenhower Administration.

That’s an accomplishment, but 5,000 nuclear weapons is still an awful lot. Ike or not, the point is that US nuclear forces grew quite dramatically in number after Truman.

Moreover, numbers tell only part of the story. Eisenhower was important not just because he presided over a huge increase in the number of nuclear weapons, but also because of organizational changes to emphasize military custody of weapons and predelegation of launch authority.

That is the important legacy of the Eisenhower Administration that came to define the Cold War and remains with us today.

Secrecy and Democratic Oversight

Americans know very little about US nuclear weapons stockpiles and policies. The University of Maryland Program on International Policy Attitudes (PIPA) surveyed 1,311 Americans. PIPA asked:

  • “How many nuclear weapons do you think the
    US has in the US, or on submarines, that are ready to be used on short notice?” The median answer was 200.

  • “How many nuclear weapons do you think the US needs to have to make sure other countries are deterred from attacking it?” The median answer was 100.

Public opinion surveys should be viewed cautiously and, of course, secrecy is just one of many barriers to a proper public debate about the role of nuclear weapons in US security policy.

But, honestly, it would help to be able to use real numbers. We’ve seen some efforts at declassification. DOE, for example, released some not very insightful information on stockpile size—the total stockpile numbers stop in 1961 and there is no information about composition.

Yup, the number of US nuclear weapons during the Cuban Missile Crisis is apparently classified. (And don’t even get me started on reclassification of delivery vehicles.)

More serious efforts at declassification seem to be held up by the Department of Defense. In May 2000, former STRATCOM commander Gene Habiger, then-director of the DOE office of security and emergency operations, wrote a letter to DoD proposing “declassification of total nuclear weapon stockpile quantities (past, present, and future) and subcategorization of those quantities by purpose, delivery system, and active/inactive status, but not by location, or specific weapon type.”

Seemed like a good idea to me, but DOD disagreed. Arthur Money responded that the “proposed declassification and subcategorization would reveal sensitive information pertaining to the pace and scope of changes in nuclear weapon stockpile quantities and status [and] could provide significant information on stockpile modernization, international treaty compliance and negotiation positions.”

Uh huh.

FAS has all the correspondence and documents if you are interested.

Update: I meant to note that, in 1992, then-Chairman of the Joint Chiefs of Staff revealed stockpile numbers during Congressional Testimony:

Now let me take you back to 1990, just the year before last, when we had deployed with our forces some 13,000 individual strategic nuclear warheads, distributed as you see: approximately 4,500 on bombers, another 2,500 on intercontinental land-based ballistic missiles, and another 6,000 that you see down here on submarine-launched ballistic missile—missiles aboard our submarine fleet. The START treaty would have taken this 13,000 and will take that 13,000 down to 9,500. That puts 4,600 in the bomber fleet, 1,400 ICBMs, and 3,500 on SLBMs. Those of you who are familiar with the START accounting procedures know that some of these bombers have discounted values, so they only count up to here with respect to the bombers for START purposes.

The initiative last night—50 percent of this number, 40 percent of this number, and about a third of the submarine numbers—bring you down to here. So if we could go to closure on the President’s initiative last evening we would be talking about numbers in the range of 4,700 overall and roughly 3,600 with respect to START-accountable systems distributed as you see here, somewhere in the neighborhood of approaching two-thirds reduction.

Let me move this slide—chart over and show you another way of looking at this. And on this chart, I have added the tactical nuclear weapons that were the subject of the President’s announcement in September. If you take the strategic warheads that you saw in the previous chart, add to them the tactical warheads which are being reduced as a result of the previous initiative, we are going from 21,000 overall, just in 1990, a few months ago.

Through all the initiatives that have been announced, we would bring this number down to 6,300. This is an historic change. This is unprecedented, and it’s truly, I think, a remarkable achievement which shows the Pentagon’s willingness to adjust to the new reality that is facing us with respect to strategic forces.

As far as I am aware, the world did not end. I’d love to get a copy of this chart, by the way.


  1. Neil Couch (History)


    Let me make a few points contrary to your position. As one who once served as the OSD representative to the Moscow Treaty’s Bilateral Implementation Commission, I know that the Treaty’s reductions are real. Even without an inspection and verification regime, the Moscow Treaty really is effectively reducing deployed weapons (the only ones that really matter). I no longer have access to the aggregate number of operationally deployed strategic nuclear warheads (the total number of ICBM, SLBM, and bomber weapons), but that number is unclassified, it is reported in the annual unclassified implementation report to Congress, and is significantly lower than the number often cited based on START attribution math. As for the stockpile, from an employment point of view, the size of the stockpile is almost irrelevant since those weapons can’t be used unless they are mated to delivery vehicles. Those weapons are kept as insurance against the failure of currently deployed weapons and even this number could be greatly reduced by the Reliable Replacement Warhead. The classification of the number in the stockpile and the disaggregate number of deployed weapons is an important aspect of our nation’s deterrent.



  2. CKR (History)

    As the numbers go down, deployed weapons are no longer the “only ones that really matter.” Stored weapons are more vulnerable to theft. An inordinate number of them may be, in effect, a breach of any arms control treaty. Missiles may be one-time delivery vehicles, but bombers are not.

    As the numbers of warheads go down, verification needs to be more intensive.

    If Neil Couch can assure us that numbers are going down, why not make the process more transparent? Assuring the world that the overall numbers are going down can decrease the perceived need for increased arsenals, by China for instance.

    Couch asserts that classification of the numbers is an important aspect of the nation’s deterrent. If the approximate numbers in the goal are known, and if he can assure us that progress is being made toward those numbers, how does classifying the exact number at a given point in time add to the deterrent? Does it really matter if the number is 5603 or 5592?

  3. Jeffrey Lewis (History)


    Thanks for your comment.

    1. I do agree that the reductions outlined in NSPD 34, and reflected in the Moscow Treaty, are significant and helpful.

    2. Non-deployed weapons exist not merely as a hedge against technical problems, as you correctly observe, but also as a hedge against geopolitical changes. Linton Brooks made the point very succinctly:

    The reduced stockpile the President approved in 2004 still retains a significant number of non-deployed weapons as a hedge against technical problems or geopolitical changes.

    3. Until the 2007 edition, the unclassified implementation report did not contain aggregate information for the number of operationally deployed strategic warheads. The 2007 edition does. This is a welcome change and I should have noted it.

    The number of U.S. operationally deployed strategic nuclear warheads was 3696 as of December 31, 2006. The classified version of this Report contains the numbers of U.S. operationally deployed strategic nuclear warheads, by category of system, and estimated numbers of Russian Federation strategic nuclear warheads as of December 31, 2006.

    Obviously, this omits the number of tactical nuclear weapons, as well as the non-deployed forces maintained as a hedge against technical problems or geopolitical changes.

    4. I agree that the stockpiled nuclear weapons “can’t be used unless they are mated to delivery vehicles.” But the capability to regenerate large force levels in response to geopolitical change is a capability that the United States retains and, therefore, remains relevant to Russian and Chinese planners. Adam Hebert with Air Force Magazine wrote an excellent story about the responsive capability:

    “The responsive capability would be able to augment that [active] force,” [J.D.] Crouch explained, “and it essentially will be additional warheads that could be uploaded back onto that force if necessary and, obviously, if the president were to make a decision to do that. And that would take weeks, months, even years to do that, depending upon the system and the character of the threat.”

    Such decisions would not be made lightly, he added. “What we’re talking about is a responsive capability that would take, at the very least, weeks–but likely months and even years–to be able to regenerate.” He added that the US would not take such a step except in response to “a major change in the security environment.”

    Pentagon officials emphasize that almost all major issues remain undecided.

    Maj. Gen. Franklin J. Blaisdell, who was Air Force director of nuclear and counterproliferation operations until late May, said the responsive force would add to the flexibility of the US triad because the military will be able to draw weapons out of storage if the security environment changes.

    If Washington decided to embark on a major expansion of the nuclear arsenal, bombers would likely be the quickest vehicle for doing so. Different times would be needed to increase weapons available to bombers, submarines, and ICBMs, but “it takes little time to bring responsive weapons to the bomber force,” Blaisdell said, noting that new weapons could be available for bomber use in a matter of days.

    He went on, “It would take some more time–maybe … months”–to increase the warheads available to the submarine force, while it would probably take “a year or so” to alter the ICBM force.

    As best I can tellthe Bush Administration plans to maintain some number of ICBMs with multiple warheads and preserve, or perhaps improve, the capability to upload additional warheads to the ICBM force.

    6. The purpose of my post was not to complain about the Moscow Treaty — although I have my complaints — but to complain about stockpile secrecy that inhibits public debate. And although I am pleased to see the the aggregate number of operationally deployed strategic warheads for 2007, I believe DOD could support the release of the information outlined by General Habiger with little risk to national security. That information would include: past, present and future aggregate numbers, subcategorized by purpose, delivery system, and active/inactive status, but not by location, or specific weapon type.

    If the inactive stockpile is irrelevant, as you say, then that information should be made public.

  4. MEC

    CKR brings forth a strong argument. In addition to those comments, would transparency (accompanied by actual reductions) not also go a long way towards fulfilling the US’ international obligations? One of the biggest — and legitimate — grievances NNWS have is that NWS have not taken Article VI of the NPT seriously.

  5. tBaum

    I thought I was smart— I am a rocket scientist, and all— but Neil Crouch’s statement that the “classification of the number in the stockpile and the disaggregate number of deployed weapons is an important aspect of our nation’s deterrent” leaves me scratching my head. This is just like the argument that the budget for the intelligence community must be classified. Why? How does the classification of either of those numbers improve national security? And how would declassifying them hurt national security? If I’ve only got 1000 weapons, does it matter that anyone knows that? After all, it ain’t exactly, um, rocket science to reprogram the coordinates of the target for any of our weapons. Why, then, does it matter who knows how many there are?

    Further, senior NNSA officials have repeatedly said that we are dismantling weapons at an unprecedented pace. Prove it— especially to our allies who don’t believe us, and who think that the RRW program is an excuse to increase the size of our arsenal.

    Declassifying the size of the arsenal— in aggregate or not, would increase national security, not decrease it.

  6. Eric Hundman (History)

    I tend to agree with Jeffrey and CKR that the benefits of disclosing numbers, purpose, delivery systems, and status of our nuclear warheads would outweigh the costs.

    However, given what people like Hans Kristenson have been able to glean from what public information we have, perhaps there’s a valid argument for keeping such facts hazy. It seems possible that we’d be able to deduce at least some locations from the new information, for instance, which could conceivably pose a security risk.

  7. Jeremy (History)

    tBaum – you beat me to it. I may just be a lowly grad student trying to wrap my head around this stuff, but its beyond me how it would compromise our national security to reveal the number of warheads we have. Can anyone actually provide a good argument in its favor?

  8. Amyfw (History)

    My brain is in vacation mode, so it probably is not a good idea for me to jump into this discussion, but I’ve got a few incoherent points to make. First, I’m surprised at the secrecy surrounding the leg-by-leg warhead loadings, as this information is readily available in the START data base. I recognize that START counts accountable warheads, not actual deployed warheads, but a little bit of arithmetic can offer a lot of corrections. Second, I don’t really like the Hebert article, and had to do a lot of reading between the lines to get to the points I needed, but he’s got one thing in your excerpt that really bothers me. Bombers are not likely to be used to expand the force rapidly. Lots of cruise missiles are due to be retired, and the bombers at Barksdale are going conventional. There’s not much force to upload left in the bomber force. Also, as Jeffrey noted, one can glean from lots of public statements that at least some Minuteman missiles will remain MIRVed. Gen. Cartwright noted that the reduction to 450 missiles did not necessarilly mean a reduction in warheads, and General Deppe (who until recently was the Commander of all the ICBMs at the 20th Air Force) said in public comments that Minuteman III would carry 1, 2, or 3 warheads. And there will be upload capability. I know, but can’t tell you, the planned numbers, but arithemetic and deduction (and something in the Hebert article) can help there.

    Finally, it is quite true that the number of deployed warheads is declining. But this is not because it is mandated by the Moscow Treaty. The Treaty confirmed plans already outlined in the NPR. We are reducing deployed warheads because we decided to do so before the press release from the Crawford summit was turned into a “Treaty.” Also, the shortcoming of the Treaty is not that it does not address stockpiled warheads, but that it does not require any reduction (or even accounting for) delivery vehicles. This is how it differs from previous treaties. And this is where the upload capability comes from (stored warheads can only return if there are empty spaces on missiles, and there will be lots of empty spaces under the Moscow Treaty.)

  9. Michael

    Re: secrecy. If another state has detailed numbers about the breakdown of warheads, it becomes easier to find them, and ascertain that you know (with certainty) what and how much is where?

  10. spacemanafrica

    This is just like the argument that the budget for the intelligence community must be classified. Why?

    Essentially the sort of dollar figures needed to complete your “average” CIA operation touches in the 10-20 million dollar range. A declassification would give anyone with a day to burn enough to make a reliable count on how many serious and ongoing operations we had running as well as which programs and efforts stayed alive over the course of any given fiscal year. As well, one could track increases/decreases in spending and tie those to known or suspected programs. Depending on how detailed a release it all was, lets say they gave station budgets, one could learn a hell of a lot. I would personally defer to the agency itself when it says it does not want it’s budget made public. I also see very little benefit being derived from telling the world that we spend x amount of money paying foreign traitors, supplying arms, manipulating media and sometimes conducting wetworks, etc. My guess is it would become a statistic thrown in our faces just like our military budget and would do nothing at all to dissuade fears of an overbearing US, it would probably aggravate them. Just my opinion though.

  11. Alex W. (History)

    Historically I think it could be argued that the ambiguity about warhead stockpiles has been more hurtful than helpful to national security. It is just such ambiguity which drove the political impetus behind the ramping up of stockpiles to astronomical levels that the chart so aptly illustrates, and at the same time gave U.S. politicians carte blanche for pushing any sorts of visions of imbalance upon the public (bomber gap, missile gap, etc.).

    Who are we keeping the stockpile numbers from? It would seem to me to only be relevant to maybe two nations—Russia and China. I would be surprised if the Russians did not actually know our stockpile numbers, what with all the treaties they have been party to with us, but perhaps I am being naïve about that. As for the Chinese, from what I have read on here it does not look like they are trying to obtain first-strike strategy whatsoever, and so I can imagine the ambiguity would only seek to reinforce the feeling that the U.S. could have a first-strike against them? In any case I would be surprised if the Chinese did not already have a pretty good idea of what our stockpiles are—they seem to be doing quite well in their knowledge of the U.S. from a nuclear intelligence point of view.

    Anyway, I confess as well to not really understanding how any national security considerations outweigh the many downsides of having stockpile numbers known. Their classification seems to have very limited national security potential, while their declassification seems to me—again, an admitted political naïf—to be better from both a domestic and foreign political situation.

  12. Patrick Rogers (History)

    Is this graph correct? Right where the offiensive & defensive lines cross it looks like they switch places, with the offensive weapons numbers adopting the trend that the defensive weapons number had been following and vice-versa.

    Is this an error in the graph, or just a coincidence?

  13. Carey Sublette (History)

    I thought I’d weigh in with an observation about the declassified “number of U.S. operationally deployed strategic nuclear warheads” (3,696).

    If one takes the NRDC’s estimated active warhead count, and assume that availability for ICBMs is 95%, for SLBMs is 66% (due to one-third being off-patrol), and the availability for active bombers in 90% (but are otherwise fully loaded), then this matches the operational count exactly (subject only to round-off error).