Good morning, everyone.  Thank you to LSUS, Hudson Institute (where I spent several happy years), Peter Huessy, and all the event’s sponsors for the chance to participate, and for your patience with me on this remote video link.  I’ll only be offering my personal views today, which won’t necessarily represent those of anyone else.  Nevertheless, I’m glad of the chance to join my old State Department colleague Chris Yeaw in discussing the role, dynamics, and implications of “hedging” in U.S. nuclear strategy. 

I.               Nuclear Hedging and its Discontents 

Few of you here today will need much introduction to the basic concept of nuclear “hedging.”  This idea dates back at least to the first Nuclear Posture Review (NPR) issued under the Clinton Administration in 1994, in connection which it was said that the United States sought to “lead and hedge.”  That is, the 1994 NPR wanted the United States to lead the way toward nuclear disarmament but also to hedge against unexpected threats along that path by keeping more U.S. nuclear capability in existence than was actually needed at the time, as a kind of an insurance policy in case things turned out to be more challenging than expected.

This concept was articulated in the context of the post-Cold War nuclear drawdown then underway by both the United States and the Russian Federation.  Russia was at the time in the middle of enormous changes, of course, it being then not long after the Soviet Union’s collapse and the fragmentation of Moscow’s empire into a kaleidoscope of constituent parts, and hopes were still high that the Russian Federation would pull off a successful transition to both democracy and capitalist prosperity. 

It was thus hoped that things would keep getting better as we built an enduring post-Communist relationship with the Kremlin, but no one could rule out Russian backsliding or other problems.  Hence strategic hedging: we were optimistic about movement toward ever-lower nuclear tensions and thus force levels, but not so confident about this that we didn’t feel the need to keep more weapons around than we felt we actually needed at that moment vis-à-vis the Russians.  This “hedge force,” as one might call it, was not operationally deployed, and couldn’t actually be used with any rapidity: it just sat around “on the shelf” for a rainy day, as it were, as a kind of insurance policy against unexpectedly bad strategic weather.

And that concept has been a key part of U.S. strategic nuclear planning ever since, at least until the Biden Administration.  Now, however, the 2022 NPR has declared that we are “[e]liminat[ing] ‘hedge against an uncertain future’ as a formal role of nuclear weapons.” 

As I noted in a paper published earlier this year by the National Institute for Public Policy, it’s not entirely clear what this means.  One possibility – not at all inconsistent with how I’ve heard U.S. officials describe the new NPR at conferences – is that this declaration doesn’t mean much of anything at all.  After all, the NPR itself says that our nuclear capabilities need to provide “credible deterrence even in the face of significant uncertainties and unanticipated challenges,” and after describing how nuclear threats from our great power adversaries are increasing, it promises to maintain a “flexible stockpile” capable of “pacing” strategic challenges by “respond[ing] in a timely way to threat developments.”  So maybe the NPR’s new declaration is just empty but performative rhetoric that sounds vaguely like something good, from a pro-disarmament perspective, without representing any real change in policy.

Another possibility is that the Biden Administration understands hedging in purely numerical terms, and is quietly acknowledging that in the face of Chinese threats we’ll have to increase our nuclear numbers – which, in the short term, means drawing down the hedge of non-operationally-deployed weapons.   After all, if the Biden Administration really means what it says about maintaining U.S. nuclear forces “responsive to the threats we face” – even as the NPR details how the nuclear threats we face are today growing – it sounds like we might well have little choice but to increase our nuclear numbers. 

At least in the short term, however, such a numerical increase would have to come from reactivating weapons that we have long retained as a hedge and which are not currently in active service.  This would therefore necessarily mean drawing down the hedge of non-operationally-deployed weapons that exists under current nuclear force planning, precluding our maintaining a numerical hedge over and above operational force limits in the ways we have done for the last generation. 

A third possibility is that the nature of hedging is changing – perhaps requiring new terminology – as indeed it would probably have to change if we draw down our current numerical hedge in order to meet increased operational needs.  And this may force us to rely more for our nuclear “insurance policy” on productive capacity as a kind of “insurance policy” against strategic circumstances even worse that we currently foresee. 

Whether or not you call this “hedging” in the traditional sense, we already need more ability to rely upon nuclear weapons productive capacity to help bolster deterrence (i.e., our ability to build more weapons and to design and produce new ones rapidly and in meaningful numbers if we should need to do so) than we presently do.  And if we have to draw down the numerical hedge as a result of the Chinese and Russian nuclear build-ups, we’ll need to rely on productive capability even more. 

In this regard, however, there will be a lot of bipartisan work to do for a long time.  Since we currently don’t have the ability to produce more or new weapons on any meaningful scale, it will require a formidable recapitalization of our nuclear infrastructure: a heroic effort to make that infrastructure once again robust, responsive, and resilient in ways that it has not been since the late Cold War.  It remains to be seen whether either the Administration or Congress will be willing to take the steps that this will require – and to provide the requisite funding and consistency of strategic focus over time – but doing so is essential.

Anyway, between these three interpretations, therefore, what does the Biden Administration actually mean by eliminating hedging as a “formal role”?  My money, at this point is on explanation #1 – that is, that the hedging comment in the 2022 NPR is just an empty gesture intended merely to look vaguely like progress, made by people who promised to reduce the role of nuclear weapons in their planning but who, for various fairly obvious strategic reasons related to escalating nuclear threats, cannot really do so.

But I guess we’ll have to see.

II.             Hedging and Arms Control

But let me now offer a few thoughts on what these various developments might mean from the perspective of arms control.  These implications could be quite profound. 

The Biden Administration still wants to be seen as dedicated to arms control, and indeed to keeping alive the dream of a world without any nuclear weapons.  Nevertheless, it’s been dealt simply a terrible hand by history, in the form of a dangerous expansion of Russian nuclear capabilities and a positively huge expansion of Chinese capabilities. 

Whereas for decades a central motif of U.S. nuclear strategy was the debate over how to balance fidelity to present-day deterrence needs against the desire for continued movement along a presumed teleology toward an abolitionist nuclear “Zero,” we’re now in a situation in which that teleology – the trajectory of supposedly “irreversible” movement toward disarmament – has run completely off the rails.

For years, U.S. nuclear policy rhetoric was designed to send two central signals to reassure different constituencies within our policy community: (1) reassuring “doves” that we really are moving steadily and inexorably toward nuclear weapons elimination; and simultaneously (2) reassuring “hawks” that we aren’t moving in this direction so fast that we will fail to meet deterrence needs.  For so long as the threat environment remained relatively stable, or improved, these messages were coherent and reasonably persuasive. 

Now, however, things are quite different.  Today, as the 2022 NPR and the Biden Administration’s other main national security guidance documents make clear, nuclear (and conventional) threats are increasing, and not slowly either.  This necessarily means that if the United States is to maintain a nuclear arsenal that is, as the NPR puts it, “responsive to the threats we face,” we will need more strategic deterrence against Russian and Chinese aggression than before. 

On one level this is not unprecedented, conceptually at least.  After all, it was always implicit in the idea of “hedging” that there might come a day when we would need to reverse course in response to escalating threats, departing from the post-Cold War path of nuclear reductions for a new course of bringing some or all of our “hedge force” back into operational service.  No one wanted that day to come, but even by the standards articulated in the Biden NPR, it apparently now has.

From an arms control perspective, shifting to a policy in which we need to make claims against the insurance policy represented by years of numerical hedging will certainly present complications.  With such a transition, we would thus find ourselves back in the business of increasing the size of the U.S. nuclear arsenal, problematizing an arms control process that for many years focused primarily on numerical reductions.

One such potential complication involves the New Strategic Arms Reduction Treaty (“New START”), which is scheduled to remain in force until February 4, 2026.  Among other restrictions, that treaty restricts the United States and the Russian Federation to having no more than 800 deployed and non-deployed intercontinental ballistic missiles (ICBMs), submarine-launched ballistic missiles (SLBMs), and heavy bombers equipped for nuclear weapons delivery.  (Within that limit, the number of deployed systems cannot exceed 700.)  New START also restricts both sides to 1,550 deployed strategic warheads each.

Russian and Chinese developments, however, are making these limits look increasingly problematic.  Russia, after all, is now again in the business of building shorter-range nuclear weapons systems not covered by New START – including missiles it began building and deploying in violation of the Intermediate-range Nuclear Forces (INF) Treaty – as well as nuclear delivery systems of a sort (e.g., its Avangard air-launched ballistic missile and Poseidon underwater drone) also not covered by New START.

Worse still, China – unconstrained by any arms control regime whatsoever – is now adding literally hundreds of new ICBMs to its arsenal, at a furious pace which the Pentagon estimates will enable it to field some 1,500 warheads by the year 2035.  Nor is there any sign that Beijing intends to stop at 1,500: its “sprint to parity” may well be a “sprint to superiority.”

Both of those powers, moreover, have been ever more willing to use their non-nuclear military capabilities to cow and intimidate – and in Russia’s case, actually to invade and annex – their neighbors.  This calls increasingly into question the degree to which they seem to be “deterred” from such adventurism by current U.S. and Allied military posture.  Accordingly, one threshold question for arms control is whether, or the degree to which, this worsening strategic environment will require the United States to exceed current New START limits even before that treaty expires. 

To be sure, there is at least somewhat more “flex” in New START than the layman might assume.  Under the limit set forth in Article II(1)(b) of 1,550 warheads on deployed ICBMs, SLBMs, and heavy bombers, for instance, Article III(2) of the Treaty specifies that only “[o]ne nuclear warhead shall be counted for each deployed heavy bomber.” 

Under these counting rules, therefore – as I am sure most of you in this audience of nuclear deterrence professionals knows well – each heavy bomber is deemed to be a single “warhead,” under the overall limit of 1,550 deployed warheads, no matter how many actual warheads that bomber in fact carries.  Since it has been reported that both U.S. and Russian heavy bombers are capable of carrying more warheads than they generally do, this means that it would be possible to “upload” such aircraft with additional weapons drawn from our reserve stockpile (that is, our hedge) – increasing the total number of nuclear weapons we deploy without violating New START.   

To the degree that we could meet evolving operational needs with such bomber uploading, therefore, we could do so without breaking New START limits before the Treaty’s expiration in February 2026.  After that point, uploading would be permissible on an even greater scale if we were to bring non-operational warheads out of storage in order to load ICBMs, SLBMs, and heavy bombers to their maximum capacity, thus producing produce a very significant increase in deployed U.S. warhead numbers.  (Such maximal uploading could also be done lawfully before February 2026, if we chose to withdraw from New START – or if Russia’s announced “suspension” of New START compliance earlier this year were deemed to be a material breach of that Treaty within the meaning of Article 60 of the Vienna Convention on the Law of Treaties, thereby releasing the United States from its own obligation to comply.)

Such maximal uploading could produce a very dramatic increase in U.S. nuclear numbers.  According to the Federation of American Scientists, for instance, even though all 400 currently-deployed U.S. Minuteman III ICBMs now only carry a single warhead, about half of them use a reentry vehicle that can hold three such warheads and couple be uploaded accordingly.  There are also reported to be an additional 50 “warm” U.S. ICBM silos which could be reloaded with missiles if necessary.  All in all, one could thus increase the number of U.S. ICBM warheads from 400 to perhaps 950.

Similarly, we could upload our Trident SLBMs to their maximum capacity, which has been reported to be eight per missile from the four or five with which they are thought to be routinely deployed today.  This would let us perhaps double our number of deployed SLBM warheads to about 1,920 – and that’s even without reconstructing and reactivating the two Trident launch tubes per Ohio-class ballistic missile submarine that we deactivated in order to meet New START limits on SLBM launchers.

To be sure, maxing out the warhead load for every missile might come at the cost of reducing these systems’ ability to carry “penetration aids” (PENAIDS) or other possible non-nuclear payloads that might serve important functions.  Nevertheless, as a technical matter, as a result of our longstanding policy of nuclear “hedging,” we clearly do possess a significant – and still unused – upload capability.

Using some or all of this capacity wouldn’t by any means be a panacea, of course, especially vis-à-vis the Russians.  Moscow, after all, also operates under New START counting rules and the Russians also have a significant upload capacity in existing systems that carry fewer warheads in practice than they are capable of carrying.  In that sense, a full-scale “upload race” with Moscow would probably just result in us both having far more deployed warheads than before, but with neither having gained any particular strategic advantage from all that effort and expense.

This has long been an argument for the United States sticking to New START limits, with the likely futility of engaging in a dyadic “upload race” with the Kremlin being a compelling argument against trying it.  Nevertheless, China’s massive nuclear buildup now at least somewhat complicates that calculus.  In theory, at least, such an uploading binge by the two current parties to New START would allow them to significantly outpace China’s current nuclear weapons build-up and thus allow both the United States and Russia to counter and defeat Beijing’s apparent sprint toward numerical strategic parity or superiority, at least for some while longer. 

This wouldn’t necessarily make such a push a good idea, of course.  Nevertheless, it is logically inevitable that debates about how to respond to China’s nuclear build-up will include at least some discussion of such options. 

Indeed, this is already what was recommended recently by a bipartisan group of former experts in a study group convened by the Lawrence Livermore National Laboratory, which recommended “that the United States … upload weapons once it is no longer bound by the constraints of the New START Treaty (NST), presumably in February 2026.”  The study group didn’t have a specific target in mind for such upload capacity, noting that “[w]e cannot provide a specific number of weapons upload to maintain the ability to deter strategic attack, assure allies, and achieve objectives if deterrence fails because that must be derived from classified guidance and threat analysis.”  It made clear, however, that we should it prepare and practice such uploading, and develop such guidance and analysis, promptly.

And all this, of course, means it’s going to be increasingly difficult to talk about numerical arms control limits of any sort – much less reductions, the very idea of which sounds increasingly silly in today’s strategic context – at least unless and until China joins the arms control conversation in ways for which it has hitherto shown nothing but contempt.  This doesn’t necessarily mean that arms control is entirely dead.  Nonetheless, if arms control is to have any future in the current environment, we’ll likely have to think about it in ways very different from those to which we have become accustomed since the end of the Cold War.

III.           Learning from the Past

To put it bluntly, if negotiated nuclear weapons reduction of the sort seen in the INF Treaty, START, the Moscow Treaty of 2002, and New START represents your only conceptual reference point, you can pretty much forget about arms control for a while.  But such reductions needn’t be your sole reference point, for historically, nuclear arms control didn’t start with arms reduction treaties, or even with force limits at all.  It began much more modestly.

Specifically, nuclear arms control between the United States and the Soviet Union began in the early 1960s in response to two issues quite independent of the number of weapons and the types of delivery system possessed by each of the superpowers.  The first issue was the realization – in the wake of the “near miss” of the nuclear war that almost broke out in 1962 over Moscow’s deployment of nuclear-armed ballistic missiles to Cuba – that the leaders of the two powers needed a more systematic and reliable way to communicate with each other in a crisis.  The second issue was basically an environmental and ecological concern, related to growing global worries about the worldwide accumulation of radioactive fallout from aboveground nuclear weapons testing.

In response to these two concerns, the nuclear superpowers reached two agreements in 1963.  First, the Hotline Agreement established a secure teletype link that could be used directly between the U.S. President and the Soviet premier.  With this new connection – which has been upgraded over the years to incorporate progressively more modern technology – it was anticipated they would be better able to manage crises in the future, thus hopefully avoiding catastrophic nuclear escalation.  Second, the Treaty Banning Nuclear Weapons Tests in the Atmosphere, in Outer Space and Under Water (a.k.a. the Partial Test Ban Treaty [PTBT] or Limited Test Ban Treaty [LTBT]) prohibited the kind of aboveground testing that had been producing such problematic nuclear fallout. 

These agreements represented the first steps down the nuclear arms control road, and over the following decades these two agreements would be followed by many more.  Notably, however, these pioneering efforts didn’t address nuclear armaments per se.  Instead, they aimed to help the United States and the Soviet Union do better in crisis management in order to avoid nuclear escalation (i.e., a form of risk reduction), and also to more safely contain the externalities of nuclear competition (in that case, test fallout). 

These early efforts at arms control, moreover, occurred at a time in which the superpowers were engaged in intense competitive nuclear build-ups, and neither side was interested in agreed force limits until the early 1970s.  But there is precedent, in other words, for at least some modest – but constructive – arms control engagement even during a period of significant competition and arms racing.

It is helpful here to conceptualize arms control not as a single, discrete “thing,” but rather a sort of “toolkit” that contains a range of different potential implements.  You can break these tools down into a number of conceptual categories – a number of types of arms control – based on a number of different historical precedents.

By way of a typology, I would suggest at least the following list of types of thing one might try to do with arms control, each with an illustrative example from historical practice:

• Crisis communications (Hotline Agreement of 1963);

• Limitation on  externalities of peacetime competition [e.g., fallout] (LTBT of 1963);

• Prohibit nuclear weapons deployment in certain potential battlespace domains (Outer Space Treaty of 1967);

• Superpower cooperation on rules restricting the development of capabilities by third parties (Treaty on the Non-Proliferation of Nuclear Weapons (NPT) of 1968);

• Numerical constraints upon strategic delivery systems (Strategic Arms Limitation Treaty [SALT I] of 1972);

• Limit possession of specified types of strategically-relevant capability (Anti-Ballistic Missile (ABM) Treaty of 1972);

• Provide “best practices” and limit forms of conduct deemed liable to create nuclear escalation risks (Incidents at Sea Agreement of 1972);

• Prohibit types of offensive arms deemed destabilizing [e.g., high-capacity multiple-warhead missiles] (second Strategic Arms Limitation Treaty (SALT II) [signed but never ratified]);

• Multilateral communications protocols for data exchange, launch notifications, and other transparency and crisis communication applications (Agreement on the Establishment of Nuclear Risk Reduction Centers in 1987);

• Prohibit a class of delivery systems (INF Treaty of 1987);

• Protocols for verified dismantlement of delivery systems (INF Treaty of 1987);

• Limits on and transparency for deployments of conventional forces the clash of which could lead to nuclear escalation (Conventional Forces in Europe (CFE) Treaty of 1990);

• Data exchanges  (Strategic Arms Reduction Treaty [START] of 1991)

• On-site inspections to help verify and increase confidence in numerical limits (START); and

• Permissive technical monitoring [e.g., photographic aerial overflights] (Open Skies Treaty of 1992).

As this list suggests, the arms control “repertoire” provided by this list of historical precedents is a fairly broad one.  So I’d like to make two points about them.

The first point is about when and the circumstances under which such measures were adopted.  Many of the more iconic steps we associate with arms control today (e.g., numerical reductions and on-site inspections) are ones that – in their original instantiation, at least – were adopted in a strategic context of waning tensions, rather than increasing ones. 

But others (e.g., crisis communications, externality management, and arguably also domain restrictions) were first possible, historically, even in an environment of considerable geopolitical tension and in which the superpowers were still building up their nuclear arsenals at a rapid pace.  Still other important tools (e.g., numerical limits, behavioral regulation, and type limits) became possible in a sort of historical era of “intermediate” tensions – specifically, the period of détentein the early 1970s, when confrontation between the superpowers remained intense and arms racing continued, but also a time in which relations had improved considerably from how they had been before.

The second point is that not all of these different tools necessarily have anything to do with numbers – and thus need not necessarily be pushed off the table of available options even in an era in which we may need increasing numbers.  Together, these two points offer two at least potential points of hope for the future of arms control even in this challenging time of increasing tensions and a growing need for the United States to increase nuclear weapons numbers.

IV.          Some Cautions

But let’s be careful.  To say that arms control remains to some extent conceivable is not the same thing as predicting it will occur. 

At the moment, we seem rather short of good-faith counterparties.  China today, for instance, seems completely uninterested in any type of arms control – even things like crisis communication improvements and transparency and confidence-building measures (TCBMs).

As for Russia, the jury is out in various ways.  It is out on whether or not Moscow would be willing to accept any future arms control agreements.  It is also out on whether we would be able to trust the Russian Federation in any future arms control deal, given its track record of repeated arms control violations, nuclear saber-rattling, and the employment of nuclear weapons threats to facilitate conventional wars of aggression and annexation.  And the jury is also out on whether it is even possible to define Russo-American force limits that would make sense in light of China’s current sprint toward nuclear weapons parity or even superiority. 

(Part of me is intrigued by the idea of having a third-party “escalation clause” in some future Russo-American deal pursuant to which both parties could lawfully augment their treaty-limited forces if a third-party such as China exceeded some specified threshold of nuclear capability.  But that’s a discussion for another day.)

Frankly, given the almost existential flavor of the rhetoric between Moscow and Washington since the advent of Putin’s full-scale war of aggression and war crimes in Ukraine, I’m skeptical that it would be possible for the parties to negotiate anything new at this point anyway.  We shall see.

V.            Conclusion

For present purposes, my point is merely that we need to do two things.  First, we need to give upon the enthusiasms for reductions and for disarmament that have preoccupied Western arms control for the last generation.  Those things are simply not possible now, at least for some while.  (Indeed, in current circumstances, a U.S. policy agenda of desperately doubling down on trying to achieve them would probably be not only fruitless but also dangerous – a gift to our strategic adversaries and a curse upon Allies who need us.)

Second, we need to open our mind to the broader range of what arms control can potentially include beyond the reduction agenda that China and Russia have worked so hard and so successfully to take off the strategic table.  Frankly, it is far from clear that there is any hope for any arms control in the near future.  But if there is, my guess is that we will need to start with modest measures that don’t look much like post-Cold War arms control, and which are inspired more by non-numerical, pre-1970s precedents than by those thereafter. 

I look forward to our discussion.  Thanks!

—Christopher Ford