I was asked recently by a group of nuclear experts what I thought would be the most important principles to keep in mind if undertaking U.S. contingency planning for a world with a “very low” total number of nuclear weapons in the American arsenal. The question was not necessarily meant to imply that such a world is likely, or that it is even desirable. (Both of these points can be argued – though not always particularly persuasively – but that wasn’t the point.) Rather, assuming that a world of “very low” U.S. nuclear numbers was at least possible at some point, they wanted to discuss what I thought a “very small” nuclear arsenal and weapons infrastructure would need to look like.
I did have views on the subject, and in the interests of engendering a broader debate, I offer them here in order to elicit feedback from NPF readers. In a nutshell, I suggested that a U.S. nuclear force posture based upon the idea of “very low” numbers would have to pay careful attention to four basic points: (1) maintaining the infrastructural “floor” necessary for essentially any nuclear arsenal; (2) preserving the ability to restore or “regrow” a larger arsenal if circumstances demand; (3) ensuring resilience or robustness of our nuclear forces; and (4) ensuring that whatever capabilities we retain are as operationally flexible as possible.
I. An Infrastructural “Floor”
To begin with, there is likely to be some kind of infrastructural “minimum” we must retain, in terms of our weapons-production infrastructure, for so long as we intend to keep any nuclear weapons at all. To some extent, it has been possible to reduce the size of our nuclear weapons infrastructure (a.k.a. “the Complex”) as the size of our arsenal has declined precipitously since the end of the Cold War. In terms of the expense, staffing, and physical “footprint” of the weapons Complex, however, the returns from reducing weapon numbers taper off sharply, pointing towards some kind of asymptote: a fairly sophisticated Complex is required even if we are to keep even a single weapon, to maintain it over time, and to be able to replace it if it is used or breaks.
To keep any warheads in our arsenal indefinitely, we must retain the capability to manufacture the appropriate fissile materials and a series of sometimes exotic other substances, and remain able to do the sophisticated metallurgy, engineering, and electronics that go into a modern nuclear warhead, as well as to troubleshoot (and, if necessary, redesign and adequately test) the assembled whole. We must also maintain the ability to build, maintain, and test all of the various components of at least one delivery system capable of carrying this warhead, as well as the complex symphony of systems involved in performing the command-and-control and targeting functions necessary for successful delivery.
These requirements run not merely to physical capital, moreover – that is, manufacturing facilities and sophisticated equipment – but also to human capital: the scientists, engineers, technicians, and other experts who are needed to do such work, and whose eventual successors must be educated and trained to high standards as they age. Especially in a highly technical arena that even today is in no small part still an apprenticeship art, one does not simply shrink wrap a sophisticated industrial capability, as it were, and put it in the freezer. Maintaining capabilities is hard work.
The expenses of indefinitely maintaining even a small arsenal can also be powerfully affected by additional factors. Were we to forswear underground nuclear explosive testing by ratifying the Comprehensive Test Ban Treaty, for instance, we would also have to maintain the expensive suite of capabilities that has been developed – and is still being developed, for this is a process of decades more than years – to ensure the reliability of our stockpile without such testing. Non-testing has a high price tag, and the invention of an entirely new infrastructure to help ensure weapon reliability over time without testing has already cost us many billions of dollars since the late 1990s. If we are resolved not to return to testing, such capabilities must not only be built but also be maintained indefinitely, even for a very small arsenal.
Nor is simply maintaining an arsenal the only issue, for in addition to maintaining the capability to ensure against technical surprise in the reliability of whatever warheads we retain – monitoring them on an ongoing basis and fixing or finding clever ways to work around whatever problems might be found – our Complex would presumably also have to help safeguard against surprise by others. Whether or not we have “very low” numbers, and especially if we do, it would remain of enormous importance to have a cadre of canny nuclear weapons experts carefully evaluating all the latest intelligence and open-source technical information bearing upon what other countries are doing with their own nuclear weapons programs, and helping watch for proliferation and arms control treaty violations. Nuclear weapons design and its associated disciplines are arcane specialties, and we would need to keep a stable of experts on hand not merely for so long as we have any such weapons, but indeed as long as anyone else has any.
So there clearly is a “floor” of infrastructural capabilities that must continue to exist even if we move to a force posture with “very low” numbers. The effective unit cost of such devices, in aggregate, would become vastly higher as they become less numerous, for many such costs are inherent in any having. To be sure, a nuclear arsenal costs considerably less than a modern conventional one, but it would cost a great deal nonetheless, and having a small stockpile probably wouldn’t cost vastly less than a large one. Anyone seriously advocating a reduced force posture must acknowledge these costs and be willing to support paying them; if there is indeed any way to have a “very small” arsenal that is consistent with our national security interest, it will necessarily require the maintenance of a sophisticated and fairly extensive infrastructure.
The requirement of what I (admittedly awkwardly) call “restorability” proceeds from the fact that even if we think it likely that in the future we will no longer need significant numbers of nuclear weapons, the future is notoriously hard to predict. If it turns out that comfortable assumptions about our future threat environment, about deterrence, and about other nuclear weapons needs turn out not to be valid, what do we do? This is an important point, for the notion of strategic “hedging” has been important – indeed, critical – to U.S. strategic policy since the end of the Cold War.
As we have reduced our arsenal since the end of the Cold War – a process that has brought us down to a small fraction of what we had in 1991 – we have generally kept available a somewhat larger number of weapons than we think we actually need to have, given our immediate deterrence needs at any particular moment in time, on the plausible assumption both that it is possible that our threat projections will turn out to be wrong, and that it would be terribly dangerous to be caught with too few. (The stockpile “hedge” also represents a degree of insurance against technical uncertainty. If one warhead design develops problems, we will have kept a different, replacement weapon available somewhere in storage: at present, we have more than one warhead type for each delivery system, so that even the discovery of some dire technical problem with one design would not make that system useless.)
Furthermore, for more than a decade now, we have emphasized trying to shift this “hedge” function increasingly from having a stockpile of existing weapons (i.e., maintaining a reserve of weapons that could be reactivated and pressed into service in a pinch) into having merely what U.S. officials have called a “responsive” infrastructure. This ideal of responsiveness revolves around the hope that instead of keeping larger numbers of inactive weapons on hand, we could achieve our goal of strategic “hedging” against future uncertainty in part by building and maintaining a nuclear infrastructure that is capable of producing new weapons relatively quickly. Having such a Complex, it is hoped, will allow us to address threats in an unpredictable future without having to keep such a large stockpile on hand today.
This insight about “hedging” and responsiveness is a critical one, and should inform our thinking about “very low” numbers, for Complex modernization and genuine responsiveness are essential for reductions. This is the basis of the imperative of “restorability.” If movement toward extremely low numbers is to be done safely, we will have to develop and maintain the ability to reverse course and reconstitute a more sizeable force if our strategic environment so requires (e.g., by maintaining a productive capacity significantly in excess of what we need for the maintenance of our baseline arsenal). “Hard-wiring” ourselves irreversibly into “very low” numbers would violate the sensible principle of responsive hedging that has been vital to our strategic planning for many years. More importantly, and for the very reasons that gave rise to our hedging strategy in the first place, “hard-wiring” ourselves small would be terribly dangerous.
Planning for “low numbers,” therefore, must pay serious and sustained attention to the challenge of ensuring that our Complex becomes and remains genuinely responsive: we must retain design capabilities and production capacity sufficient to rebuild a larger force later if our “best guesses” today about future threats turn out to be overly optimistic. In the circumstances we face, being able to go back up in a contingency is a necessary part of safely going down in the first place. We forget this at our peril.
III. Resilience and Robustness
The third element that I think would be essential in any reasonable policy of seeking “very low” numbers is to ensure that what few weapons systems we retain can be relied upon to do their job. This is not merely about safeguarding against technical surprise in terms of our weapons’ ability to explode as advertised, though this would be very demanding because the error margins for device reliability would presumably have to be tighter, in a small-numbers future, even than they are at present. (If you have 10,000 devices, you can probably accept it if a certain small percentage do not work properly. If you have only a thousand, however, you have to demand a higher degree of reliability. If you have numbers that are “very low,” you’ll need to be more demanding still. At the asymptote, with a single weapon, there is essentially no error margin at all.) But there is more to it than technical reliability.
Resilience or robustness is also about making sure, with analogously stringent error margins, that our delivery systems are reliable, accurate, and highly resistant both to enemy attack before launch and to an adversary’s defensive countermeasures thereafter. If we continue to rely upon ballistic missile submarines (SSBNs), for instance, we must stay ahead of adversary efforts to crack the nut of strategic anti-submarine warfare (ASW) – a subject that preoccupied navies in the late stages of the Cold War, and which would acquire special new importance in a “very low” numbers world. To the extent that our aerial delivery systems rely upon evading radar by means of “stealth” technology (or advanced electronic warfare [EW]), we need to stay similarly ahead of counter-stealth (and counter-EW) innovations. To the extent that we rely upon ballistic missiles, they must be able to penetrate enemy defenses. Our command-and-control systems, and the intelligence support that feeds our targeting cells, would also need to be sufficiently redundant and “hardened” that our forces could continue not merely to survive but also to perform successfully in the face of whatever an adversary throws at us. Additionally, of course, we would need to preserve the kind of research and development capabilities and the broad defense industrial base that are capable of sustaining all this into the indefinite future. There’s nothing “easy” here; being a nuclear superpower committed to responsible reductions and deterrent stability is hard work.
Finally, since there’s no way of being particularly sure precisely what threats we will face in the mid-to-long-term future, a force posture of “very low” numbers would require us to ensure that what systems we do retain are ones capable of as much flexibility as possible, in terms of operational capability against a broad range of potential targets. This logic is both simple and irresistible: if you’re restricted to possessing a small toolkit but don’t know exactly what tasks you will need to be able to accomplish with it, you should choose tools that can do as many things as possible, even if this means sacrificing something in terms of ideal performance in any single task. A Swiss Army knife, for instance, doesn’t cut as well as a nice kitchen blade, turn screws as well as a real screwdriver, cut paper as well as full-size scissors, or open wine as well as a proper corkscrew. But it can do all of these things decently, and you certainly wouldn’t want to have only a corkscrew on hand if you thought it possible you’d need a screwdriver. At low numbers, flexibility matters.
In the nuclear weapons context, this would mean developing and maintaining devices and delivery systems optimized for usefulness across a range of potential future missions and target sets. The “very low” numbers challenge thus points toward new weapon development, rather than simply the preservation a handful of legacy systems optimized for Cold War missions that may well not be what we need them for a generation or two from now.
- For one thing, especially since existing U.S. designs do not seem likely to perform well against the kind of hardened and deeply-buried targets (HDBTs) that are presently proliferating around the world in places such as China, Russia, Iran, and North Korea, this clearly means finally getting serious about acquiring an earth-penetration capability, coupled with effective precision guidance.
- “Tailored output” weapons are also advisable, capable of producing especially powerful electromagnetic pulses for the kind of counter-electronics warfare that seems likely to be an important part of future conflict between sophisticated states, or perhaps an extremely high radiation flux intended to destroy chemical or biological weapons agents with as little collateral damage as possible.
- Being able to vary the explosive yield of our nuclear weapons would also be very important, allowing individual devices to be assigned to a greater range of future targets and missions. Variable yield gives more options.
- We would also want to assure maximal interoperability between individual nuclear weapon designs and whatever delivery systems we retain. We are already exploring concepts for having a “common” warhead design that could be used on more than one ballistic missile. Perhaps it will eventually be possible to have two or more designs – their multiplicity helping preserve technical redundancy and minimize the risk that a “single-point” failure collapses our entire deterrent system – each of which could be used with any of our delivery systems on an essentially “plug-and-play” basis.
- We should also maximize our ability “dynamically retask” weapons after launch, so that systems can be retargeted on the fly in an evolving warfighting scenario, and make sure we retain the kind of operational flexibility provided by bombers operating in adversary airspace (e.g., in search of elusive mobile targets, or in “hole-hunting” against delivery systems hidden in, and potentially moving within, a network of interlinked HDBTs).
A force posture of weapons and delivery systems built along these lines would be a “Swiss Army knife” force, much more able to meet the deterrence and operational challenges of “very low” numbers than would the decades-old systems we maintain today.
To be sure, there would be objections to this approach. Because some of these steps might perhaps be difficult without a resumption of underground explosive testing for developmental and design-validation purposes, and because there has developed over the years a blinkered neuralgia on the U.S. political Left (and internationally) about “new” weapon designs, the arms control and disarmament community will no doubt oppose what I am describing. Nevertheless, moving to such a sophisticated and flexible force would make small numbers a great deal more strategically sensible than would otherwise be the case. Those who desire “very low” numbers should thus ask themselves what is more important: preserving litmus-test ideological purity against testing and “new” designs, or actually reducing nuclear numbers. It is very likely we cannot do both of these things, so this is a debate worth having, and in the open.
Please don’t mistake my point. I am not making the case for a “very low” numbers force posture here, for there are good reasons to be wary of such a shift at this time. (Indeed, as Clark Murdock has suggested – and in Arms Control Today, no less – the Obama Administration’s case even for much more modest reductions is internally contradictory, and in places all but incoherent.) Anyone who purports to be serious about achieving “very low” numbers, however, should be able to explain how his or her plan answers the challenges I have posited here about an infrastructural “floor,” preserving the ability to “regrow” a larger arsenal, ensuring robustness, and ensuring maximal operational flexibility.
To my eye, a minimalist force posture, if it were to be strategically wise at all, could not be simply a shrunken version of what we have today; it would have to look very different, and it would require the maintenance of what would still be a formidable weapons Complex and defense industrial base. Unless we think through these issues carefully – and unless we feel confident of our ability both to provide what such a posture would truly require and to maintain this over time – reductions to “very low” numbers are likely to represent not some great victory for disarmament principle in the service of international peace and security, but instead a dangerous recipe for national weakness, strategic instability, deterrence failure, and ultimately conflict.
I invite reader feedback.
-- Christopher Ford