BOOK REVIEW: The Limited Power of Nuclear Threats

May 2017

Reviewed by George Perkovich

Nuclear Weapons and Coercive Diplomacy
By Todd S. Sechser and Matthew Fuhrmann
Cambridge University Press, 2017, 333 pp.

As Russian President Vladimir Putin continues to flaunt Russia’s nuclear weapons, U.S. President Donald Trump tweets that he will “greatly strengthen and expand [U.S.] nuclear capability,” and Kim Jong Un makes bellicose nuclear threats, an old question re-emerges: do nuclear weapons enable bullies to get their way? The answers to that question have policy ramifications. The dangers of nuclear proliferation and the policy responses to it should be assessed differently if nuclear weapons do not significantly augment a possessor’s coercive power. So too in deciding the quantitative and qualitative requirements of a nuclear arsenal, policymakers should assess whether superiority adds to ­effective coercive power.

In Nuclear Weapons and Coercive Diplomacy, political scientists Todd S. Sechser and Matthew Fuhrmann use statistical analysis and historical case studies to answer whether and under what conditions nuclear weapons “provide countries with advantages in international bargaining.” Coercion or blackmail is the type of bargaining they have in mind, as when a state seeks to compel an adversary to relinquish territory, stop committing massive human rights abuses, remove military forces, release hostages, or otherwise change the status quo.

The late Nobel laureate economist Thomas Schelling, who used game theory to model strategic interactions, famously distinguished coercion, which he called “compellence,” from deterrence. A deterrent motivates someone not to take an action that might be contemplated. Coercion seeks to change a current behavior or position. Sechser and Fuhrmann think that nuclear weapons have been and can be effective deterrents against offensive threats. (With characteristic fairness, they acknowledge that even this value of nuclear weapons is disputed by some scholars and activists.)

Although deterrence remains vital, current policy and scholarly debates about nuclear weapons tend explicitly or implicitly to revolve around coercion. It is said that if Iran acquires nuclear weapons, it will be able to impose its will on its neighbors; that Russia is using nuclear threats to coerce its neighbors and NATO; that the United States needs to augment its nuclear arsenal to restore its international power; and so on. In these situations and others, deterrence also could be at play. A state could intervene in neighboring states and create new facts on the ground, as Russia did in Ukraine, because victims would be deterred from resisting for fear of triggering escalation that could lead to nuclear use. Sechser and Fuhrmann assess these arguments as part of their overall project.

They begin with statistical analysis, a de rigeur move for political scientists. To analyze the role of nuclear weapons in “crisis bargaining,” they composed a data set of 210 militarized so-called compellent threats issued from 1918 to 2001. A compellent threat occurs when a “challenger” state demands that a “target” state change the status quo, with the demand backed by the threat of military force. Threats to use force are transmitted verbally or communicated tacitly through militarized actions such as troop maneuvers, force deployments, or exercises. Sechser and Fuhrmann analyzed whether and when compellent threats succeeded, meaning the targets complied with a challenger’s demands without the challenger having to use military force. They also ran statistical regressions to assess “partial compellence success,” where only some of the challenger’s demands were met.

By the Numbers

The results contradict conventional wisdom. Of the 210 compellent threats in the data set, nuclear-weapon states succeeded just 10 times. Indeed, nuclear-armed states, whether they made nuclear threats or not, were less successful at coercion than states without nuclear weapons: 20 percent of the time versus 32 percent. Further, having a nuclear monopoly over the state one is trying to coerce provided statistically no advantage. To the contrary, nuclear monopolists prevailed just 16 percent of the time, compared to the 33 percent success rate of states without nuclear weapons.

Examples of failures include the U.S. demand in 1964 that Vietnam stop supporting the Viet Cong, China’s demand in 1979 that Vietnam end its occupation of Cambodia, the United Kingdom’s demand in 1982 that Argentina withdraw from the Falkland Islands, and U.S. demands on Iraq in 1990 to withdraw from Kuwait and, in 1998, to readmit weapons inspectors.

The 10 cases of successful compellent threats by nuclear-armed states include some that make a reader laugh at statistical analysis: restoration of Jean-Bertrand Aristide to power in Haiti in 1994 and conduct of elections in the Dominican Republic in 1961. Moreover, three of the 10 successful cases involved the same event: the compulsion of Serbia to withdraw heavy artillery from Sarajevo. This event is counted three times ­because three countries (the United States, France, and the UK) made the threat. Perhaps the most famous perceived success was by the United States in motivating the Soviet Union to withdraw its missiles from Cuba in 1962.

In chapter 4, the authors ran a different set of regressions, analyzing 348 territorial disputes between 1919 and 1995. These disputes may or may not have involved the making of explicit coercive demands. The question is whether ­nuclear-armed states extracted concessions in negotiations over territorial disputes at a higher rate than non-nuclear-armed states and whether they succeeded more in gaining disputed territory after using force. The analysis shows that possessors of nuclear weapons do not win territorial concessions at a higher rate than non-nuclear-armed states (35 ­percent versus 36 percent).

What about success in using military forces to gain territory? Twenty-three military confrontations occurred within seven different territorial disputes from 1945 to 1995. Seventy percent of the time, use of force did not result in major territorial gains. For example, China could not alter the territorial status quo with Vietnam. Pakistan could not gain territory in Kashmir from India either in 1999 or 2001–2002. This analysis “throws cold water on the argument that nuclear weapons are useful for redrawing the map. If nuclear weapons do anything in these disputes, they preserve the status quo.”

Sechser and Fuhrmann offer three broad reasons “why nuclear coercion is uniquely difficult compared to other types of coercion—even against non-nuclear adversaries” or, stated another way, why “it is exceedingly difficult to make coercive nuclear threats believable.”

First, to be credible, a threatened ­action must be able to accomplish the given objective if the target does not comply. A nuclear attack is not a viable way to seize territory, end human rights abuses, or make a government change leaders or extradite terrorists. Moreover, in many cases, a nuclear-armed state has other military capabilities to accomplish such objectives. If the target is unmoved by the more credible threat that conventional capabilities will be exercised, it is unlikely to be moved by nuclear threats.

“Second, the political, military, and economic costs of launching a nuclear ­attack are prohibitive, particularly if a challenger’s political objectives are offensive in nature,” they write. Other influential actors in the international system are likely to react harshly against a state that breaks the nuclear taboo for any reason other than national self-defense against an existential threat. Political opprobrium from allies, adversaries, and neutral countries is an obvious cost. Economic sanctions and military action also could be mobilized against the nuclear attacker. Moreover, the use of nuclear weapons for coercion could stimulate other states to seek nuclear weapons to deter such offensive exertions in the future. Finally, depending on the state and the circumstance, domestic opposition could be mobilized against a government that initiated use of nuclear weapons for coercive purposes.

“Third, coercion by its very nature generally is a lower-stakes enterprise for coercers, in contrast to deterrence,” ­according to the authors. Psychological and behavioral-economic studies demonstrate clearly that people and nations are more highly motivated to avoid losses of what they have than to take risks for new gains. Coercers seek new gains through changing the status quo; their targets are trying to retain what they have. Defenders seeking to deter are generally more resolved than aggressors seeking to coerce.

Unlike many of their scholarly competitors, Sechser and Fuhrmann ­acknowledge the potential inadequacies of statistical analysis. Statistical studies account for general trends, but “they may not explain the most important historical cases.” Including crises that “did not have an overt nuclear component, including disputes between nonnuclear states, to understand the efficacy of nuclear threats,” may be misleading or irrelevant. The small number of overt nuclear ­coercion threats may make statistical findings flawed. The coding of threats and responses may be disputable. Finally, correlation does not mean causation; maybe something other than nuclear weapons was decisive in the cases analyzed.

Sechser and Fuhrmann respond ­effectively to each of these points. The few, most dangerous nuclear crises on which scholars often focus may themselves be more anomalous than others. Including and examining nonnuclear crises provide a useful basis for comparison, much as studies of medicinal drugs wisely compare the effects of placebos with the effects on subjects given the drug. Including disputes that did not involve overt nuclear threats helps assess the argument that the possession of nuclear weapons creates a background effect even when specific nuclear threats are not made. Regarding the coding of cases, Sechser and Fuhrmann have made their data set openly available and invited others to review it, recode cases, and run their own regressions.

Historical Case Studies

To further test their results and the interpretations they derive from them, the authors complement their statistical analysis with historical case studies. In chapter 5, they use primary and secondary sources to analyze nine cases between 1946 and 2016 when nuclear coercion failed. These include the canonical Soviet ultimatums over Berlin in 1958 and 1962, the U.S. threat against North Korea to release the captured USS Pueblo electronic surveillance ship in 1968, the Nixon-Kissinger “madman nuclear alert” in 1969 to coerce the Soviet Union and Vietnam to adopt more forthcoming positions in the Paris negotiations with South Vietnam and the United States, and North Korean leader Kim Jong Un’s 2013 threat to conduct “pre-emptive nuclear attack to destroy the strongholds of aggression” if UN Security Council sanctions were not lifted and U.S.–South Korean military exercises were not curtailed.

The conduct and outcome of these episodes confirm the authors’ central arguments why nuclear coercion is difficult to achieve.

With similar rigor, in chapter 6, Sechser and Furhmann re-examine 10 cases that are widely believed to have been successes for nuclear coercion. These include the U.S. nuclear threats against China and North Korea from 1950 to 1953, the U.S. nuclear threats against China during the Taiwan Strait crisis in 1954–1955, the Cuban missile crisis, and the Sino-Soviet conflict over disputed islands in the Ussuri River in 1969.

The authors conclude that none of these were “unequivocal” successes. In each case, there is “some doubt—often considerable doubt—about whether nuclear weapons provided states with coercive leverage.” Moreover, the putative successes all occurred early in the Cold War. “There is scant evidence that nuclear blackmail has worked since the collapse of the Soviet Union,” Sechser and Fuhrmann conclude.

In these historical analyses, as in their statistical analyses, the authors display great methodological care. They recognize that governmental records and memoirs may reflect more or less than the truth. Memoirists, of course, are inclined to make themselves look good and their adversaries not so much. Early in the nuclear age, memoirists tended to exaggerate the impact nuclear weapons had on events. For example, U.S. President Harry Truman, in his memoirs, suggested that his threat to drop the bomb compelled Soviet leaders to withdraw their troops from Iran in 1946. Yet, Truman never made such a threat. As the taboo against initiating use of nuclear weapons has strengthened, memoirists tend to downplay or omit nuclear threats they signaled when in office. Classification requirements no doubt precluded then-Secretary of State Henry Kissinger from mentioning what he and President Richard Nixon did during the so-called madman nuclear alert in 1969, but it is difficult to imagine he was eager to discuss the episode. Governmental records are not immune from artifice either.

From their findings, Sechser and Fuhrmann construct a theory of “nuclear skepticism.” They juxtapose this to what they call “the nuclear coercionist school,” which has two variants. “Absolutists” such as diplomats John Foster Dulles and John Bolton and scholars Victor Asal and Kyle Beardsley have presented views that nuclear powers have coercive advantages regardless of their opponents’ capabilities. “Relativists” such as scholars Robert Pape and Matthew Kroenig tend to argue that nuclear superiority, however defined, is key to coercive victories. Sechser and Fuhrmann provide what might be labeled “nuclear realism.” Their work is simply more realistic than that of scholars and policy practitioners who assert that nuclear weapons have exceptional coercive value.

Nuclear Myths

Sechser and Fuhrmann close by unpacking 10 “myths” about the political effects of nuclear weapons. Two of these are particularly important today.

The first is that “nuclear-armed countries can use low-level violence with relative impunity, knowing that their targets will hesitate to retaliate due to the possibility of nuclear escalation.” An act of coercion creates a new status quo that is then preserved by nuclear deterrence. The authors tested this proposition in their statistical analyses by assessing whether states were more likely to issue coercive threats, make military challenges to the territorial status quo, escalate conflicts over territory, or resolve territorial disputes in their favor using military force after they had acquired nuclear weapons than before they had. They found scant evidence for any of this, although they acknowledge that Russia’s recent actions in Ukraine may be a troubling exception. They soundly argue that Pakistan has not succeeded in coercing territorial or diplomatic concessions from India, although Pakistan’s nuclear arsenal deters India from conducting major punitive military operations in the Pakistani heartland.

The second is that “nuclear superiority matters.” Although “the notion that more power equals greater influence is intuitive,” they write, “it is not supported by real-world events.” The case studies show that leaders “rarely behave as if nuclear superiority provides them with advantages in coercive bargaining.” Indeed, if anything, “nuclear-inferior states are often the biggest risk-takers in serious nuclear crises,” as suggested by the Soviet Union in the Berlin crises and North Korea in 2013 and perhaps now. Statistically, they conclude that “there is no relationship between nuclear superiority and coercive victories.” Even when superior states did prevail in crises, examination indicates that the nuclear balance was not the cause. For example, in the Berlin crises the key factor was that the United States was seeking to deter a change in the status quo and the Soviet Union to change it.

Going further, it is interesting to consider how Sechser and Fuhrmann’s work might apply to three questions that may bedevil nuclear scholars and policymakers in the coming years. The United States since the end of the Cold War has appeared to pursue regime change or significant behavior modification to varying degrees in countries such as Iraq, Iran, North Korea, and even Russia and China. This coercive objective has been pursued under nuclear shadows, either while trying to prevent nuclear proliferation or to contest exertions of adversary regimes in situations where conflict could ensue with nuclear dimensions. Sechser and Fuhrmann’s analysis suggests that such coercive exertions are unlikely to succeed and, unfortunately, may motivate adversaries to seek or deploy nuclear weapons to deter them. More work would be welcome to clarify whether and how states’ perceived attempts to change others’ regimes affect the conduct and requirements of nuclear deterrence and the ­prospects of nuclear proliferation.

Nuclear Deterrence

The second question is whether and how this analysis adds insight to the challenge of providing extended nuclear deterrence. This is a paramount objective of the United States and its allies today. Are coercive threats against U.S. allies more or less likely to succeed than coercive threats issued directly against the United States, that is, are providers of extended deterrence more susceptible to coercion than if they were threatened directly? Would clearer U.S. nuclear superiority over Russia and China, however defined, improve the situation? The authors, based on this work, might say that the answer is “not to any significant degree.” A related question is whether such analysis would help reassure U.S. allies.

Writing for the War on the Rocks website in March, Yale University nuclear scholar Paul Bracken acknowledges that “narrow nuclear blackmail (‘Give up Kashmir immediately or we’ll attack Mumbai with atomic weapons’) is not very likely anymore. But, the opportunities for blackmail in a nuclear context are greatly increased today,” meaning attempts by nuclear-armed states to compel someone without issuing a specific threat to use nuclear weapons. Bracken emphasizes the possibility of such blackmail being directed not only against enemies but also against allies—Washington pressing Seoul “not to take Pyongyang” in a conflict. Bracken’s essay shows the salience of Sechser and Fuhrmann’s book, but their rich analysis seems to puncture his ­assumptions and assertions.

Unfortunately, it is difficult to imagine that high-level policymakers in any country will read this book or any book like it. Even if they did read and draw the proper cautionary conclusions, every nuclear crisis is new. Knowledge of the historical record and probabilities that may be derived from it will not prevent decision-makers from concluding that a new crisis is different. What rarely worked in the past might well work now. This possibility heightens the importance of reinforcing the taboo against first use of nuclear weapons because, of any single variable Sechser and Fuhrmann cite in explaining the difficulty of nuclear coercion, the most important is the sense of the international cost that would be imposed on the actors who use nuclear weapons first.