Deterrence is both a theory in international relations and a strategy of conflict management. Although not called deterrence, threat-based strategies that attempt to manipulate the cost-calculus of other actors have long been practiced. There is ample evidence of their use by all the ancient empires. My treatment of deterrence begins with the advent of nuclear weapons, which put a premium on policymakers to find ways of preventing catastrophically destructive wars yet exploiting any strategic nuclear advantage for political gain. I describe early approaches to deterrence, their application in practice and the critique of them that subsequently developed. Drawing on works that made use of Soviet, American, Chinese and Israeli archives, and interviews with officials from these countries and Egypt, I offer an overall assessment of the consequences of deterrence during the Cold War. I conclude with a brief discussion of post-Cold War deterrence and promising areas for research.
The Deterrence Debate
Scholars and policymakers became interested in deterrence following the development of the atom bomb. Early publications on the subject (Brodie, 1947) recognize that a war between states armed with atomic weapons could be so destructive as to negate Carl von Clausewitz's (1976: 75-89) classic description of war as a continuation of politics by other means. In 1949, the problem of deterrence gained a new urgency as the Cold War was well underway and the Soviet Union, in defiance of all American expectations, exploded its first nuclear device in October of that year. In the 1950s, often referred to as the golden age of deterrence, Bernard William Kaufmann (1954), Henry Kissinger (1957) and Bernard Brodie (1959) among others developed a general approach to nuclear deterrence that stressed the necessity but difficulty of imparting credibility to threats likely to constitute national suicide. The 1960s witnessed an impressive theoretical treatment by Thomas Schelling (1964) that analyzed deterrence in terms of bargaining theory, drawn from microeconomics, and elaborated a set of bargaining tactics based on tacit signals. Deterrence played a central role in the American strategy in Indochina during the Johnson and Nixon administrations. Deployment of forces, the character of the engagements they sought and the level and choice of targets for bombing were never intended to defeat the Viet Cong or North Vietnam but to compel them to end the war and accept the independence of South Vietnam. The Indochina intervention ended in disaster and helped to spawn a series of critiques of the theory and strategy of deterrence in the 1970s. The end of the Cold War and with it, an opening in the archives of the participants, brought another wave of reassessment. No consensus has emerged, but the issues have been clarified and enriched by much new evidence. The contemporary debate is far more international than it was during the Cold War, in part because there are more nuclear powers. Studies of deterrence by Indian, Pakistani and Chinese scholars and military think tanks have supplemented those of the U.S., U.K. and Israel. In the West, the focus of deterrence has turned away from restraining large state actors with nuclear weapons to smaller, so-called "rogue" states thought to be trying to acquire such weapons. Since 9.11, there has also been a debate about the applicability of deterrence to the problem of terrorism.
Deterrence is an attempt to influence another actor's assessment of its interests. It seeks to prevent an undesired behavior by convincing the party who may be contemplating such an action that its cost will exceed any possible gain. Deterrence presupposes that decisions are made in response to some kind of rational cost-benefit calculus, that this calculus can be successfully manipulated from the outside, and that the best way to do so is to increase the cost side of the ledger. Compellence, a sister strategy, uses the same tactics to attempt to convince another party to carry out some action it otherwise would not.
We must distinguish theories of deterrence from the strategy of deterrence. The former address the logical postulates of deterrence and the political and psychological assumptions on which they are based. Another important distinction is between general and immediate deterrence (Morgan, 1983). General deterrence is based on the existing power relationship and attempts to prevent an adversary from seriously considering any kind of military challenge because of its expected adverse consequences. Immediate deterrence is specific; it attempts to forestall an anticipated challenge to a well-defined and publicized commitment. Immediate deterrence is practiced when general deterrence is thought to be failing. It is almost impossible to know when general deterrence succeeds because non-action by a target state can be the result of many reasons, including any lack of intention to use force. Because cases of immediate deterrence success and failure are somewhat easier to identify, most research has sought to explain their outcomes. Analyses of immediate deterrence that ignore its relationship to general deterrence offer a biased assessment of its success rate and an incomplete picture of the conditions and processes that account for its outcome. Empirical studies of immediate deterrence are surrounded by considerable controversy in the absence of compelling evidence about the intentions and calculations of the leaders of target states (Huth and Russett, 1984, 1988; Lebow and Stein, 1990). There turn out to be striking differences among leaders on opposing about who was practicing deterrence. In many so-called deterrence encounters (Garthoff, 1989;Lebow and Stein, 1990) both sides considered themselves the deterrer. This is often due to different interpretations of the status quo. In the Cuban missile crisis (Lebow and Stein, 1994), Khrushchev understood the secret Soviet missile deployment in Cuba to be part and parcel of his attempt to deter an American invasion of Cuba. Kennedy and his advisors interpreted the deployment as a radical and underhanded effort to upset the strategic status quo.
The Golden Age
The theory of deterrence developed as an intended guide for the strategy of deterrence. The first wave of theorists wrote from the late 1940 through the mid-1960s. The early literature Kaufmann,1954; Brodie, 1959; Schelling 1966) begun with the assumption of fully rational actors and was largely deductive in nature. It stressed the importance of defining commitments, communicating them to adversaries, developing capability to defend them and imparting credibility to these commitments. It explored various tactics leaders could exploit toward this end, concentrating on the problem of credibility. This was recognized as the core problem when deterrence was practiced against another nuclear adversary – and the implementation of the threats in question could entail national suicide (Jervis, 1979). Thomas Schelling (1963) went so far as to suggest that it was rational for a leader to develop a reputation for being irrational so his threats might be believed. Richard Nixon indicates that he took this advice to heart in his dealings with both the Soviet Union and North Vietnam (Kimball: 76-86).
As a practicing economist, Schelling might have been expected to privilege material capabilities in his analysis. In Arms and Influence he makes a ritual genuflection in this direction on the opening page when he observes that with enough military force a country may not need to bargain. His narrative soon makes clear that military capability is decisive in only the most asymmetrical relationships, and only then when the more powerful party has little or nothing to lose from the failure to reach an accommodation. When the power balance is not so lop sided, or when both sides would lose from non-settlement, it is necessary to bargain. Bargaining outcomes do not necessarily reflect a balance of interests or military capabilities. Three other influences are important.
First is context, which for Schelling describes the stakes, the range of possible outcomes, the salience of those outcomes, and the ability of bargainers to commit to those outcomes. In straight forward commercial bargaining, contextual considerations may not play a decisive role. In bargaining about price, there will be a range of intervals between the opening bids of buyer and seller. If there is no established market price for the commodity, no particular outcome will have special salience. Either side can try to gain an advantage by committing itself to its preferred outcomes. Strategic bargaining between states is frequently characterized by sharp discontinuities in context. There may be a small number of possible outcomes, and the canons of international practice, recognized boundaries, prominent terrain features, or the simplicity of all or nothing distinctions can make one solution more salient than others. Salient solutions are easier to communicate and commit to, especially when the bargaining is tacit.
The second consideration is skill. Threats to use force lack credibility if they are costly to carry out. To circumvent this difficulty, clever leaders can feign madness, develop a reputation for heartlessness, or put themselves into a position from which they cannot retreat. Other tactics can be used to discredit adversarial commitments or minimize the cost of backing away from one's own.
The third, and arguably most important, determinant of outcome is willingness to suffer. Paraphrasing Carl Von Clausewitz, Schelling describes war as a contest of wills. Until the mid-twentieth century, force was used to bend or break an adversary's will by defeating its army and holding its population and territory hostage. Air power and nuclear weapons revolutionized warfare by allowing states to treat one another's territory, economic resources, and population as hostages from the outset of any dispute. War is no longer a contest of strength but a contest of nerve and risk-taking, of pain and endurance. For purposes of bargaining, the ability to absorb pain counts just as much as the capability to inflict it.
Schelling does not say so, but it follows from his formulation that the capacity to absorb suffering varies just as much as the capacity to deliver it. Clausewitz recognized this variation. Increases in both capabilities, he argued, made possible the nation in arms and the revolutionary character of the Napoleonic Wars (Clausewitz: 585-94). By convincing peoples that they had a stake in the outcome of the wars, first the French, and then their adversaries, were able to field large armies, extract the resources necessary to arm and maintain them, and elicit the extraordinary level of personal sacrifice necessary to sustain the struggle.
The Clausewitz-Schelling emphasis on pain has wider implications for bargaining. The ability to suffer physical, economic, moral or any other loss is an important source of bargaining power and can sometimes negate an adversary's power to punish. Realist approaches to bargaining tend to neglect this dimension of power and focus instead on the power to hurt and how it can be transformed into credible threats. Schelling also ignores the pain absorption side of the power-pain equation -- when analyzing compellence in Vietnam, an oversight that led to his misplaced optimism that Hanoi could be coerced into doing what Washington wanted.
The power to punish derives only in part from material capabilities. Leaders must also have the will and freedom to use their power. Schelling observes that Genghis Khan was effective because he was not inhibited by the usual mercies. Modern civilization has generated expectations and norms that severely constrain the power to punish. The American bombing campaign in Vietnam, in many people's judgment the very antithesis of civilized behavior, paradoxically demonstrates this truth. The air and ground war aroused enormous opposition at home, in large part because of its barbarity, and public opinion ultimately compelled a halt to the bombing and withdrawal of American forces from Indochina. The bombing exceeded World War II in total tonnage but was also more restricted. The United States refrained from indiscriminate bombing of civilians and made no effort to destroy North Vietnam's elaborate system of dikes. The use of nuclear weapons was not even considered. Restraint was a response to ethical and domestic political imperatives. Similar constraints limited American firepower in Iraq in the Gulf War of 1990-91, and enabled the Republican Guard and Saddam Hussein to escape destruction.
From the beginning, deterrence theory and strategy has spawned critiques. The most interesting are those that evaluate deterrence strategy in light of empirical evidence from historical cases. The work of Milburn (1959), Snyder and Diesing (1977), George and Smoke (1974), and Lebow (1981) and Jervis, Lebow and Stein (1984) is representative. George and Smoke recognized that challenges short of full scale attacks – what they called “probes” – were difficult to deter and might be instituted by adversaries to test a state’s resolve. They and Milburn attempted to put deterrence into a broader context and argued that it might be made a more efficacious strategy if threats of punishment were accompanied by promises of rewards for acceptable behavior.
Janice Gross Stein and Richard Ned Lebow (Lebow 1981, Jervis, Lebow and Stein 1984, Lebow and Stein 1987) developed an extensive critique of immediate deterrence with three interlocking components: political, psychological and operational. The political component concerns the motivation behind foreign policy challenges. Deterrence is unabashedly a theory of “opportunity.” Adversaries are assumed to seek opportunities to make gains and pounce when they find them. Case studies of historical conflicts point to an alternative explanation for challenges, including resorts to force, which Lebow and Stein term a theory of “need.” Strategic vulnerabilities and domestic political needs can push leaders into acting aggressively. Khrushchev’s Cuban missile deployment, to cite one instance, was motivated by his perceived need to protect Cuba, offset American strategic superiority and anger at Kennedy for deploying missiles in Turkey – making him look weak in the eyes of hardliners (Lebow and Stein, 1994: 19-66). When leaders become desperate, they may resort to force even when the military balance is unfavorable and there are no grounds for doubting adversarial resolve. Deterrence may be an inappropriate and provocative strategy in these circumstances.
The psychological component is also related to the motivation behind deterrence challenges. To the extent that policymakers believe in the necessity of challenging commitments of their adversaries they become predisposed to see their objectives as attainable. When this happens, motivated bias can be pronounced and take the form of distorted threat assessments and insensitivity to warnings that the policies to which our leaders are committed are likely to end in disaster. Policymakers can convince themselves, despite evidence to the contrary, that they can challenge an important adversarial commitment without provoking war. Because they know the extent to which they are powerless to back down, they expect their adversaries to accommodate them by doing so. To continue with our Cuban missile crisis example, Khrushchev brushed aside advice of top political and diplomatic advisors who warned him that the missiles would be discovered before they were operational and would provoke a serious crisis with the U.S. He sought refuge instead in promises of marginal military officials with little knowledge of Cuba or American intelligence capabilities (Lebow and Stein, 1994: 67-93).
The practical component highlights the distorting effects of cognitive biases and heuristics, political and cultural barriers to empathy, and the differing cognitive contexts the deterrer and would-be challengers are apt to use to frame and interpret signals. Problems of this kind are not unique to deterrence; they are embedded in the very structure of international relations. They nevertheless constitute particularly severe impediments to deterrence because of a deterrer’s need to understand the world as it appears to the leaders of a would-be challenger in order to manipulate effectively its cost-benefit calculus. Failure to do this in the desired direction can make the proscribed behavior more attractive to a challenger. In the case of Cuba, Kennedy’s deployment of Jupiter missiles in Turkey and his warnings that under some circumstances the U.S. would not hesitate to strike first given its strategic nuclear advantage were intended to moderate Khrushchev, but instead convinced him of the even greater costs to the Soviet Union of remaining passive in the face of these American threats. Kennedy in turn had made these threats because of Khrushchev’s browbeating of him at the Vienna summit and threats to the Western position in Berlin (Lebow and Stein, 1994: 19-50). The missile crisis was, in effect, the product of a series of escalating threats and actions by both sides, each attempting, unsuccessfully, to deter the other.
For many years, empirical research on deterrence, whether qualitative or quantitative, drew primarily on cases of immediate, conventional deterrence. Beginning in the late 1980's evidence on Soviet and Chinese foreign policy began to become available and it became possible for the first time to reconstruct critical Soviet-American and Sino-American deterrence encounters, and to make some observations about the role of general deterrence in these relationships. Research on the Cuban missile crisis, the Soviet-American crisis arising out of the 1973 Middle East War, and the two Taiwan Straits crises of 1954 and 1958 tend to confirm the findings of critics of conventional deterrence. So does research on general nuclear deterrence. Zhang (1992), Hough (1994), and Lebow and Stein (1994) find that deterrers do worry about their reputations and the credibility of commitments, but that the targets of deterrence rarely question their adversary’s resolve. For this reason efforts to communicate resolve were often perceived as gratuitously aggressive behavior and sometimes provoked the kind of challenges they were designed to prevent. In doing so, the strategy of deterrence helped to provoke the Cuban missile and Taiwan Straits crises and to prolong the Soviet-American and Sino-American conflicts.
Based on the study of Soviet-American relations in the Khrushchev and Brezhnev eras, Lebow and Stein offer the following conclusions about the role of general nuclear deterrence: 1) Leaders who try to exploit real or imagined nuclear advantages for political gain are not likely to succeed; 2) Credible nuclear threats are very difficult to make; 3) Nuclear threats are fraught with risk; 4) Strategic buildups are more likely to provoke than to restrain adversaries because of their impact on the domestic balance of political power in the target state; 5) Nuclear deterrence is robust when leaders on both sides fear war and are aware of each other's fears.
Despite considerable evidence from both superpowers opinion about the role deterrence played in the end of the Cold War is deeply divided. Conservatives credit Ronald Reagan's arms build and Strategic Defense Initiative (Star Wars) for bringing the Soviet Union to its knees (Matlock, 1995). They maintain that Gorbachev and his advisors became convinced that they could not compete with the United States and ought to negotiate the best deal they could before Soviet power declined even further (Davis and Wohlforth). Liberals and most Soviet policymakers and scholars attribute the end of the Cold War to “New Thinking” and the political transformation it brought about within the Soviet leadership. Gorbachev, they contend, considered the Cold War dangerous and a waste of resources and sought to end it to bring the Soviet back into Europe, facilitate political reform at home and free resources for domestic development (Brown, 1996; English, 2000).
The big question for scholars may not be whether deterrence helped to prevent World War II but why leaders and lesser officials in both superpowers and so many scholars thought it was so necessary, and how, until the advent of Gorbachev, they repeatedly confirmed this belief tautologically. Such behavior has not stopped with the end of Cold War. Reputable scholars routinely claim that nuclear weapons have kept the peace between India and China and that American intervention in Iraq brought about an about-face in Libyan foreign policy. With the possible exceptions of Israel and conservative British defense analysts, the United States appears to stand alone in the faith it places in deterrence and the credit it gives it for preventing war. What theorists say about deterrence may tell us more about their ideological assumptions and their country’s strategic culture than it does about the nature and efficacy of threat-based strategies.
One of the more interesting characteristics of post-Cold War deterrence and compellence is that it has been more effective against those who would stop violence than against its perpetrators. In Somalia, the U.S. withdrew its forces after losing eighteen Rangers. In Rwanda, genocidal Hutus deterred Western intervention by killing ten Belgian soldiers. In Bosnia, compellence clearly failed against Milosovic, who continued his policy of ethnic cleansing of Albanians in Bosnia despite Western threats. Pushed by Western public opinion, NATO finally screwed up its courage to intervene, but then failed to go after known war criminals because of the vulnerability of its lightly armed forces, whose primary mission was the distribution of aid (Freedman:124-25). There is an important lesson here, and one that has been consistently ignored by theorists of threat-based strategies. It has to do with the ability to inflict pain versus the willingness to suffer it. As we observed, Schelling and American policymakers ignored the latter in Indochina, concentrating only on how much damage they could inflict on North Vietnam and the Viet Cong. The U.S. lost the war because its Vietnamese opponents were willing to accept far more suffering than the American people. This phenomenon is equally pronounced today. Self-deterrence, in effect, prevented intervention in Rwanda and stalled it for a long time in the former Yugoslavia. It did not have this effect in Afghanistan and Iraq, where the Bush administration grossly underestimated its costs and duration.
Two concluding observations are in order. The first grows out of the record of deterrence and compellence during the Cold War and its aftermath. These conflicts suggest that he political and psychological dynamics governing cost estimates and relative willingness to bear the costs of military action remain the critical consideration for leaders contemplating the use of threat-based strategies and their probability of success or failure. Much important research can be done in this connection, especially in conflicts that pit highly developed industrial powers, with a low tolerance for loss of life, against weaker, less developed, more traditional countries where honor remains important and death in combat or by suicide missions is more acceptable.
The second concerns the general efficacy of deterrence as a strategy. Its many drawbacks do not mean that it should be discarded. Rather, scholars and statesmen must recognize the limits and inherent unpredictability of deterrence and make greater use of other strategies of conflict prevention and management.