How Likely Is an Irrational US Presidential Order to Use Nuclear Weapons?

By March 5, 2017

BESA Center Perspectives Paper No. 417, March 5, 2017

EXECUTIVE SUMMARY: There is no scientifically valid way to predict US presidential irrationality on expressly nuclear commands. Nor is there any analytically acceptable way for an American president to foretell the nuclear outcomes of any considered policy decision. There is another, less frequently considered prospect that is just as impossible to predict: that President Trump could refuse to follow through on certain deterrent threats, thereby allowing assorted Russian aggressions – including those that might involve Israel.

Back in 1976, I was a young professor of political science researching a new book on nuclear strategy and nuclear war. Because I was especially interested in the distant but still conceivable prospect of an irrational American president with nuclear decisional authority, I wrote to General Maxwell Taylor, a former Chairman of the Joint Chiefs of Staff, for some informed clarification. On 14 March 1976, in a handwritten letter to me, General Taylor advised: “….the best protection is not to elect one.”

This was not a disingenuous response. On the contrary: it revealed, generically, that the so-called “two-man rule” of redundant nuclear safeguards simply does not apply at the highest level of American political decision-making. However odd or ironic, this utterly unrestricted or ultimate form of presidential authority still obtains today. While it is increasingly under discussion among concerned persons in the just-beginning Trump era, salient apprehensions have generally been expressed surreptitiously, in more-or-less decorous whispers.

It is unlikely that such core apprehensions will remain suppressed, even in the near term. Over the next several months, President Trump could, for example, be severely tested by Pyongyang, and may have to make time-urgent decisions concerning North Korea’s expanding nuclearization. How is he likely to respond to such tests?

From a purely analytic or scientific standpoint, there is no reliable way to answer such a vital question. More precisely, to credibly assess whether Mr. Trump can be expected to react rationally or irrationally, we would first have to assume the existence of pertinent and recoverable past experience. This is because all scientifically meaningful assessments of probability need to be based upon a determinable frequency of past events.

This is a requirement for all excursions into probabilistic outcomes, whether in areas of international relations, or in actuarial judgments of individual client life expectancies.

Is it true that there have been no pertinent past events regarding national nuclear decision-making? The only conceivable case, it seems, would be the 1962 Cuban missile crisis. During that uniquely uncertain episode, President John F. Kennedy ordered his “quarantine” of Cuba (a euphemism for “blockade,” always a possible casus bellum) with an apparent awareness of the corresponding risks.

According to Theodore Sorensen, his biographer, JFK seemingly believed that even his intentionally softened escalatory response would carry heavy odds of an ensuing nuclear war with the USSR: more precisely, with an expected probability of “between one out of three and even.” Although we now know, based upon what we have learned about ascertaining probabilities and the discernible frequency of past events, that any such estimate was necessarily without any scientific foundation, what matters is that at the time, JFK believed in those ominously high odds. Was JFK’s “quarantine” a genuine instance of nuclear decisional irrationality, one that turned out well only by happenstance? Could it have ended, as then Secretary of Defense Robert McNamara had feared, in some sort of “Armageddon”? Or was it rather an example of what might today be most aptly termed the “rationality of pretended irrationality?”

The answer is important. If the latter, JFK was playing a more or less carefully calculated strategic game, much like the game of Chicken once played in cars by teenagers. In Chicken, the objective of each player is twofold: not to be a chicken, but also not to be dead. JFK plausibly sought similar objectives, although, of course, at a far wider collective level, and for immeasurably higher stakes.

President Trump’s upcoming nuclear-related policy judgments can never be based upon scientifically garnered probabilities, an incapacity that has nothing whatever to do with his own personal strengths or weaknesses. We must further inquire about a related but still discrete presidential decisional problem. This issue references the corollary matter of President Trump being unable to calculate the probable outcomes of any particular nuclear decision that he himself should ultimately make. Again, this expected inability has nothing to do with any personal intellectual deficit on his part; to be sure, the very same analytic problem would confront any American president. It has to do rather with the indisputable absence of any pertinent past events.

Let’s assume, for example, that President Trump should seek out an “expert” probability assessment or prediction concerning North Korean escalation to nuclear weapons (in the near term, such an escalation could more or less plausibly affect Japan, US forces in the region, and/or certain targets in Alaska or Hawaii). There would be no relevant history to evaluate. In any such scenario, there would be, by definition, no available opportunity to render a scientifically meaningful assessment of applicable probabilities, or “odds.”

On the core issue of prospective presidential irrationality regarding nuclear weapons, it is at least conceivable that such consequential missteps could become less likely over time, on the “common sense” assumption that experience would correlate favorably with prudential caution. But that conclusion would merely offer a “man-in-the-street” explanation. At best, it would represent a naively contrived extrapolation from earlier historical eras –  one, moreover, wherein the main argument could only have made sense in the pre-nuclear past.

In any event, during a nuclear crisis involving the US, President Trump would somehow have to strike an optimal balance between the unavoidably competitive search for “escalation dominance,” and the corresponding need to avoid being locked into any increasingly desperate sequence of move and countermove. In any such sequence, the inherently self-limiting pattern could lead inexorably to a nuclear exchange or even full-blown war.

Fortuitously, perhaps, President Kennedy revealed in 1962 that strategic risk-taking can be advantageous up to a point – but figuring out exactly where that point should be is by no means an easy task. Indeed, various well-documented histories of the Cuban Missile Crisis all seem to agree that the superpowers came close to a very different, and genuinely calamitous, conclusion. An important lesson to be learned is that a great deal rests on chance when there is a lack of pertinent historical examples.

Nuclear strategy is a game that sane national leaders must learn to play, but never with any reassuringly plausible guarantees of probable outcomes. The only way this can change is if, in the years ahead, examples should begin to accumulate of nuclear escalations and outcomes. Naturally, this is not something one ought to wish for. Instead, it would be better to continue to have to concede our incapacity to reliably “figure the odds” in any nuclear crisis engagement, and also to augment this “concession” with a renewed dedication to stable nuclear deterrence.

We can’t yet determine how likely it is that America’s new president would ever give an irrational order to use American nuclear weapons, but we can still advise Mr. Trump that unimaginable nuclear dangers lurk not only in sudden “bolt from the blue” enemy attacks, but also in unanticipated and uncontrolled forms of nuclear escalation. As far as pretending irrationality is concerned – a tactic that may or may not have figured in the Cuban Missile Crisis, depending upon one’s own interpretation of JFK’s 1962 strategic calculations – it could rapidly become a double-edged sword.

One final observation. Expressed worries about an irrational US presidential order to use nuclear weapons notwithstanding, it is increasingly plausible that Mr. Trump could do precisely the opposite. He might steadfastly refuse to make good on a variety of America’s core deterrent threats, and/or give the impression of such an intended refusal to one or more prospective aggressors. If, for example, Moscow should believe it can undertake further aggression in Ukraine and Syria with impunity – that is, without any fear of consequential reprisals from Washington – Putin might well decide to proceed.

The consequences of any such Russian calculations in Syria could even impact the safety and security of Israel.

View PDF

Louis René Beres is Emeritus Professor of International Law at Purdue and the author of twelve books and several hundred articles on nuclear strategy and nuclear war. He is a frequent contributor to Harvard National Security Journal (Harvard Law School), The Bulletin of the Atomic Scientists, The Atlantic, The Jerusalem Post, US News & World Report, and Israel Defense. His newest book is Surviving Amid Chaos: Israel’s Nuclear Strategy (Rowman & Littlefield, 2016).           

BESA Center Perspectives Papers are published through the generosity of the Greg Rosshandler Family

Prof. Louis René Beres

Louis René Beres (Ph.D., Princeton, 1971) is Professor of Political Science and International Law at Purdue. The author of many books and scholarly articles in the field, he contributes regularly to several major world newspapers and magazines. Dr. Beres was Chair of Project Daniel (Israel, 2003).