%20-%20https://sketch.com%20--%3e%3ctitle%3e3ECBBF50-FD08-412A-943D-2CFC5E025A16%3c/title%3e%3cdesc%3eCreated%20with%20sketchtool.%3c/desc%3e%3cg%20id='Stylesheet'%20stroke='none'%20stroke-width='1'%20fill='none'%20fill-rule='evenodd'%3e%3cg%20transform='translate(-926.000000,%20-668.000000)'%20id='BUTTONS'%3e%3cg%20transform='translate(100.000000,%20597.000000)'%3e%3cg%20id='BTN-/-RIGHTS'%20transform='translate(816.000000,%2065.000000)'%3e%3cg%20id='ICON-/-Rights'%20transform='translate(10.000000,%206.000000)'%3e%3cpath%20d='M13.5752841,0%20L1.34943182,0%20C0.59375,0%200,0.586666667%200,1.33333333%20L0,22.6666667%20C0,23.4133333%200.620738636,24%201.34943182,24%20L17.6505682,24%20C18.40625,24%2019,23.3866667%2019,22.6666667%20L19,5.36%20L13.5752841,0%20Z%20M13.5752841,1.89333333%20L17.0838068,5.36%20L13.5752841,5.36%20L13.5752841,1.89333333%20Z%20M1.34943182,22.6666667%20L1.34943182,1.33333333%20L12.1988636,1.33333333%20L12.1988636,5.36%20C12.1988636,6.10666667%2012.7926136,6.69333333%2013.5482955,6.69333333%20L17.6235795,6.69333333%20L17.6505682,22.6666667%20L1.34943182,22.6666667%20Z'%20id='Shape'%20fill='%23595959'%20fill-rule='nonzero'%3e%3c/path%3e%3ccircle%20id='Oval'%20stroke='%23595959'%20stroke-width='1.5'%20fill='%23595959'%20cx='9.5'%20cy='13.5355339'%20r='5.5'%3e%3c/circle%3e%3cpolyline%20id='Path-2'%20stroke='%23FFFFFF'%20stroke-width='1.5'%20transform='translate(9.500000,%2013.035534)%20rotate(45.000000)%20translate(-9.500000,%20-13.035534)%20'%20points='8%2016.0355339%2011%2016.0355339%2011%2010.0355339'%3e%3c/polyline%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
James Fearon's seminal (Reference Fearon1994) article on the domestic political costs of getting caught bluffing in an international crisis has spurred over 400 references in scholarly journals to “audience costs.” The main claim of this literature is that democracies whose leaders threaten force in international crises tend to get their way because the high domestic political costs of backing down make those threats credible in the eyes of their adversaries.
This argument might be viewed in a benign light as part of a broader claim that the peace among democracies rests on their ability to credibly signal their interests and commitments in international bargaining (Fearon Reference Fearon1994, 577; Lipson Reference Lipson2003). However, a less benign implication is that democracies (and perhaps those authoritarian leaders who are accountable to domestic elites) can successfully bully nonaccountable states by using belligerent foreign policy rhetoric to tie their own hands (Weeks Reference Weeks2008). This elaborates on Thomas Schelling's (Reference Schelling1960, Reference Schelling1966) analyses of the tactical advantages of throwing out the steering wheel in teenage hot rodders’ games of chicken, a tactic that Schelling (Reference Schelling1960, 28) and Fearon (Reference Fearon1994, 587) argue democracies are especially well suited to exploit. Viewed in this light, the anticipation of audience costs could lock opponents into commitments to fight over stakes that would otherwise not warrant going to war (Fearon Reference Fearon1994, 583).
Notwithstanding the widespread popularity and weighty implications of these conjectures, empirical tests of domestic audience costs theory in a real-world international setting have been inconclusive. One problem is that audience costs theorists have argued that rational leaders will not make bluffing threats in circumstances where they might have to pay heavy audience costs, so audience costs are likely to come into play only in cases where they cannot be observed (Schultz Reference Schultz2001b). Another is the difficulty in statistical studies of distinguishing the effects of hands-tying threats from other possible informational advantages of democracy, such as greater transparency or better policy analysis. To overcome such inference problems, Kenneth Schultz has suggested historical process tracing of crises where the anticipation of potential domestic audience costs mechanisms may have shaped tactics and outcomes (Reference Schultz2001b, 53). Beyond a few brief narratives, however, there has heretofore been little effort to examine crisis processes in this way.
Based on a review of post-1945 crisis histories, we find that domestic audience costs mechanisms rarely play a significant role, let alone the “crucial” one that Fearon (Reference Fearon1994, 577) claims, for four main reasons. First, leaders see unambiguously committing threats (ones that commit a previously uncommitted actor) as imprudent. They almost always seek to retain significant flexibility, rather than lock in. Second, domestic audiences care more about the overall substantive consequences of a leader's policy than about consistency between the leader's words and deeds. We found no case where a public holding dovish policy views punished a leader for the inconsistency of failing to carry out a public threat. Third, domestic audiences’ concern about their countries’ reputations for resolve and national honor is largely independent of whether a leader has issued an explicit threat. And fourth, authoritarian targets of democratic threats do not perceive audience costs dynamics in the same way that audience costs theorists do.
On the relatively rare occasions when domestic audience costs do come into play, they tend to emerge in situations in which public opinion is already inclined to favor a hard-line diplomatic stance, and where the public has pressed the leader to make the threat in the first place.Footnote 1 In that situation, it is difficult to disentangle the domestic costs of getting caught bluffing from the costs of defying a public with substantively hawkish preferences. Schultz is right that leaders avoid reckless bluffs, but the reason is that empty threats make for substantively imprudent policy, not just that leaders fear being caught in an inconsistency. Domestic audience costs, defined as the domestic political price a leader pays simply for failing to carry out a public threat, are a minor, derivative consideration.
We first review and criticize four assumptions of audience costs theory that must hold true for domestic audience costs to have a significant effect on crisis outcomes. We do not claim that this critique constitutes an alternative theory of crisis bargaining. Rather, we make two less ambitious points: (1) even staying within the logic of rationalist bargaining theory, audience costs theory depends on implausible simplifying assumptions that ignore or minimize powerful incentives shaping the choices of prudent leaders and publics, and (2) cognitive limitations of the threatened authoritarian leaders are likely to further erode domestic audience costs effects. Following this theoretical critique, we then briefly review previous efforts to test domestic audience costs theory quantitatively, experimentally, and qualitatively. Finally, we explain how we use post-1945 crises to address inference problems in this literature, present our case selection criteria, and assess the causal impact of audience costs in several crises, especially the Suez crisis, the Iran hostage crisis, the Sino-Indian War, and the Cuban missile crisis.
AUDIENCE COSTS THEORY
“Domestic audience costs” can be pithily defined as “the domestic price that a leader would pay for making foreign threats and then backing down” (Tomz Reference Tomz2007, 821).Footnote 2 Audience costs theorists have argued that the domestic public would feel that such a leader would have a severely compromised reputation for making credible commitments, and thus should be removed from power (Guisinger and Smith Reference Guisinger and Smith2002; Schultz Reference Schultz2001b, 34), or that voters might assume that leaders who back down after making a threat are generally incompetent in foreign affairs (Smith Reference Smith1998).
Some scholars have used the term with a much broader meaning, including any public dissatisfaction with a leader who fails to protect the nation's honor, whether related to an empty threat or not, or an authoritarian regime that permits its domestic public to express hawkish opinions (Weiss Reference Weiss2006). We do not doubt that leaders might pay costs if they pursue weak policies in the face of a hawkish public, but this has little to do with the core logic of audience costs theory. We use the term in the narrow, analytically precise sense of Fearon and Tomz.
Fearon clarifies the meaning of the term by distinguishing between “two sorts of costs that leaders face for backing down in a crisis.” “First, there is the domestic and international price for conceding the issues at stake, which is the same regardless of when concessions are made or after how much escalation. Second, there are whatever additional costs are generated in the course of the crisis itself.” These additional costs include, he says, the physical costs of the crisis, the risk that the crisis will escalate to accidental or preemptive war, and “international and domestic audience costs” (Reference Fearon1994, 579–80). This last category consists of “signaling costs” in the eyes of these audiences when a leader backs down after making “a show of force” or “public demands or threats” (Reference Fearon1994, 580–81).
Fearon makes expansive claims for the “centrality of domestic audience costs” in shaping crisis interaction (Reference Fearon1994, 578), arguing that the character of crises as “public events taking place in front of domestic political audiences . . . is crucial to understanding why they occur and how they unfold” (Reference Fearon1994, 577). Because his model assumes that public information about capabilities and the opponents’ respective evaluations of the stakes is already taken into account in leaders’ decisions to embark on a crisis, he claims that additional audience costs incurred during the crisis, especially domestic costs, “may be primary” in determining the tactics and outcome of the crisis bargaining (Reference Fearon1994, 578, 581).
Fearon says that domestic audience costs affect the bargaining in two ways. First, the willingness of leaders to make public threats that expose them to potential audience costs credibly signals their high resolve to their opponents. Second, leaders who are accountable to domestic constituencies may be able to use public threats to engage reputational stakes that “lock in” a sufficient motive for fighting, even if their precrisis motivation fell short of this level (Reference Fearon1994, 583). If it becomes clear during the course of crisis escalation that one of the sides has become “locked in” in this way, that side prevails in the crisis bargaining. However, if both sides become locked in before the crisis interaction clarifies which side has the greater resolve, then war can occur as a result of these “additional” domestic audience costs.
From these assumptions, Fearon derives a number of hypotheses about bargaining tactics and outcomes. Regime types with greater accountability to a domestic audience, especially democracies, are more likely to prevail in crisis bargaining. Democracies and other accountable regimes are less likely to engage in limited probes that require making threatening moves with the possibility of backing away in the face of resistance. Crises involving democracies (and other regimes accountable to a domestic audience) should see fewer escalatory steps than crises between authoritarian states, because credible communication is easier to accomplish. Because democracies have stronger incentives to escalate in crises when facing authoritarians, the chance of war should be greater in such cases, Fearon contends.
PROBLEMATIC ASSUMPTIONS OF AUDIENCE COSTS THEORY
This audience costs framework and the hypotheses deduced from it depend on a few key assumptions, some of them largely implicit in Fearon's initial article. Here we emphasize four that we find problematic in ways that severely limit the power of the theory: (1) leaders seek lock-in, not flexibility, (2) publics care a great deal about consistency between threats and deeds, independent of their preferences on policy substance, (3) domestic audience costs are substantial, independent of other reputational considerations, and (4) targets of threats understand domestic audience costs mechanisms. We rely on these four assumptions as a benchmark to assess the empirical validity of audience costs theory in our case studies, using the assumptions to identify the causal mechanisms predicted by audience costs theory.
Assumption 1: Leaders seek lock-in, not flexibility. For audience costs to exert “crucial” effects on crisis outcomes, threats must be reasonably clear and compelling. If leaders simply warn that unspecified adverse consequences might somehow follow at an unspecified time if the opponent takes some ill-defined action, the public will be unable to pin down the leader for inconsistency between word and deed.
In fact, historically grounded comparative research suggests that international crises have rarely hinged on bridge-burning ultimatums. When Glenn Snyder and Paul Diesing (Reference Snyder and Diesing1977, 213–14) studied the use of threats in twentieth-century crises, they found no instances of “committing threats,” that is, of a threat that “creates a commitment for the threatener that he did not have before, via the engagement of additional values (honor, resolve, reputation, etc.) that would be lost if the threat were not fulfilled.” Rather than locking into a rigid stance, leaders typically issued threats that retained a degree of ambiguity regarding the time, place, and nature of the response, as well as the specific actions that would trigger it. When leaders issued unambiguous ultimatums, it was almost always because they were indeed resolved to fight and sometimes because they sought to use the crisis diplomacy to justify attacking (Jervis Reference Jervis1979, 303; Lebow Reference Lebow1981, 29–37; Snyder and Diesing Reference Snyder and Diesing1977, 215). Warnings of unspecified “grave consequences” were far more common than threats specifying explicit military sanctions, which risk engaging the honor not only of the threatener but also of the target (Snyder and Diesing Reference Snyder and Diesing1977, 220). When leaders did make threats, they left loopholes, as in the Eisenhower–Dulles doctrine of responding to aggression “at places and with means of our own choosing” (Gaddis Reference Gaddis1982, 121, 147; Snyder and Diesing Reference Snyder and Diesing1977, 214, 216–18, 221). Pure bluffs were extremely rare and couched in ambiguity.Footnote 3
This preference for ambiguity stems from the very nature of the role of threats in crisis bargaining, as explained by Snyder and Diesing. Crises involving the risk of war are interactions under uncertainty in which the parties test whether they can achieve their most valued goals at an affordable price. In such interactions, threats play the dual role of increasing the threatener's bargaining leverage by incurring potential reputational costs for backing down and also of gathering information about the other side's willingness to run risks. The goal of enhancing credibility militates in favor of unambiguous threats, whereas the goals of prudent information gathering and avoiding unnecessary wars militate in favor of ambiguity. Leaders balance these contradictory incentives in using threats in a crisis. If leaders were to follow Schelling's strategy and lock into positions before learning more about the opponent's preferences and resolve, unnecessary wars would be far more frequent.
Because leaders normally do hedge their threats, they can usually escape the full ignominy of being caught in a bluff, especially if the public otherwise agrees with their policies. Threateners who want to escape from a commitment may be able to reduce the reputational costs of backing down by arguing that the opponent's action was not of the type covered by the threat (e.g., they are not offensive missiles), that the opponent's concessions on other issues obviated the need to act (they gave us part of Cameroon in exchange for Morocco), that the threat actually had been carried out through low-cost actions (we boycotted their Olympics), or that counteraction would be taken in the future or through indirect means (we are increasing our defense budget) (Snyder and Diesing Reference Snyder and Diesing1977, 216). Especially if the opponent collaborates in helping the bluffer save face, they can claim that the setback is minor or limited (Hitler: “This is my last demand”), that they are bowing to the opinion of neutral parties (Eden deferring to Eisenhower over Suez), that the change was rule-governed and not under duress (there was a plebiscite in the Saar), or that the concession is not a precedent because the situation is sui generis (Snyder and Diesing Reference Snyder and Diesing1977, 202–3). Thus, the combination of ambiguity in the original commitment and creativity in justifying an escape from the commitment may reduce audience costs substantially—and if this loophole is foreseen, may circumscribe the bargaining advantage of undertaking the ambiguous commitment.
In a follow-up article on signaling, Fearon himself admits that “leaders often make statements or take actions that have ambiguous but potentially threatening implications concerning future performance and then back down later on” (Reference Fearon1997, 83). Fearon interprets retreat following ambiguous threats as a form of bluffing that “does not occur in equilibrium in the signaling game” (Reference Fearon1997, 83). This is “inconsistent with the model in an interesting way,” he says, in that it points to other considerations that weigh on the decision maker: e.g., the disadvantage of challenging the other side's reputation, a concern to avoid upsetting dovish audiences, or the danger that hawkish allies might seize upon the threat as a blank check allowing them to provoke a fight. Regardless of the motivation for hedging threats, our point is that the impact of audience costs diminishes when threats are sufficiently vague that leaders cannot be held fully accountable for failures to carry them out.
Assumption 2: Publics care a great deal about consistency between threats and deeds, independent of their preferences on policy substance. Audience costs theorists invoke various justifications for the assumption that domestic audiences care enough about empty threats to try to remove or punish the leader. These reasons include the leader's tainted credibility or general incompetence. Fearon himself proceeds from the assumption that “shared expectations” will converge on a “historical norm” that “seems to have domestic political audiences punishing a leader who concedes after having deployed troops more than one who concedes outright” (Fearon Reference Fearon1994, 581).
If such a norm exists, why should it be decisive in crisis deliberations? If, as Fearon argues, all public information about capabilities and interests, including the preferences of the leader's own constituents, has already been factored into the bargaining calculus when the parties decide to engage in the crisis, such a norm might seem to loom large in crisis deliberations. But even within the logic of Fearon's own framework, private information about capabilities and interests can affect crisis interactions. Indeed, audience costs come into play in his framework because of the uncertainty about relative bargaining power created by that private information. Because of private information, or because the public might not have fixed preferences at the outset of a crisis, the leaders’ strategic assumptions and preferences may diverge unpredictably from those of the public to which they are accountable. The public may turn out to be more hawkish or more dovish than expected, and the public may care more about the substantive policy stakes than expected, relative to their concern about the alleged norm against empty threats.
A more basic problem, however, is the highly restrictive conditions under which it is plausible that the public would care more about a discrepancy between a leader's word and deed than about the substance of the leader's policy. For purposes of theory-building, one might entertain the fiction that the direction and intensity of the public's ideologies, beliefs, misperception, partisanship, cultural biases, parochial economic interests, and all other sources of hawkish or dovish preferences are somehow “taken into account” in the decision to start the crisis. But if they are, that could only mean that the leader has decided that those factors are sufficiently supportive of making the threat in the first place. Indeed, this comes close to the argument made by Schultz, who stresses the need for support by the opposition party to make a democratic leader's threats credible.Footnote 4 In other words, publics are expected to punish leaders who back away from threats only if they agree with the threats on substantive grounds. This claim diverges from Fearon's original audience costs theory. It also directly opposes Tomz's research design, which assumes that domestic audience costs and the public's policy preferences are experimentally separable.
Similarly, other scholars have theorized that accountable leaders’ threats are seen as especially credible only when their states already hold the upper hand in a crisis because of their greater power or stake in the outcome (Clare Reference Clare2007). This seems plausible, but this, too, is a long retreat from Fearon's original conjecture that such bargaining advantages have been factored out at the moment the crisis starts, and it vastly diminishes the scope of audience costs.
Assumption 3. Domestic audience costs are substantial, independent of other reputational considerations. Fearon is not simply arguing that publics care about national honor. It may be true that leaders who fail to confront opponents that encroach on their state's interests are likely to incur a cost to their personal reputation and that of their state, but this can be true whether or not they have made an explicit commitment to resist the encroachment. As Schelling (Reference Schelling1966, 118) says, “unlike those sociable games it takes two to play, with chicken it takes two not to play. If you are publicly invited to play chicken and say you would rather not, you have just played.” This is a broader reputational concern that should not be confused with domestic audience costs theory, which deals only with the tactical use of threats in crisis to affect bargaining leverage.Footnote 5 We find plenty of evidence that domestic audiences care about national honor, but little evidence that they care independently about empty threats.
Assumption 4. The targets of threats understand domestic audience costs mechanisms and expect them to be decisive. A final necessary assumption is that the states that are the targets of attempts to manipulate audience costs in crises understand how this mechanism works (Schultz Reference Schultz2001a, 18). At a minimum, an audience costs account needs to show that the target looked for information about the attitude of the threatener's audience, assessed the likelihood that the threatener would be punished for backing down, and assessed whether the threatener felt vulnerable to such pressures. More specifically, the target should believe that the likelihood of punishment hinges on inconsistency between the accountable leaders’ threats and deeds, rather than on substantive policy differences between leaders and their constituencies.
This assumption runs counter to the well-established findings of cognitive psychologists and international relations scholars who have applied their insights to diplomatic interactions. They show that people tend to see the other side as more unitary than it really is, and that people view the opponent as acting out of dispositional inclinations rather than under compulsion from situational constraints (Jervis Reference Jervis1976, 35–48, 319–42). Both of these tendencies limit the ability of targets to perceive domestic audience costs dynamics in democracies that threaten them. So do more prosaic factors such as perceptual self-absorption and lack of knowledge about the threatening state (Jervis Reference Jervis1976, 343–55). Even remaining within a rationalist framework, there is no reason to assume that authoritarian targets will expect potential domestic audience costs to trump other factors bearing on the credibility of threats, such as the policy preferences of the opposition party (as in Schultz's theory) or private information about the military balance and the attitude of allies (as in other aspects of Fearon's own theory).
Audience costs theory depends on simplifying assumptions that underplay incentives that decisively shape the choices of prudent leaders and publics. In addition, cognitive biases, ideologies, and parochial outlooks are likely to further erode domestic audience costs effects in international crises.
PREVIOUS TESTS OF AUDIENCE COSTS THEORY
Empirical tests of domestic audience costs theory have so far consisted of quantitative analyses of international disputes, survey-based experiments, and a handful of illustrative case studies. These testing efforts are at best inconclusive because of causal inference problems, validity issues, and mixed empirical results.
One might think that the most direct and obvious way to evaluate domestic audience costs theory would be to check whether accountable leaders who back down in crises suffer substantial domestic political punishment. One such test suggests that domestic audience costs effects do not increase the punishment of democratic leaders (Gelpi and Grieco Reference Gelpi and Grieco2000). Audience costs theorists, however, have argued that tests of this kind do not take into account selection effects: Accountable leaders tend to make only those threats that will succeed, so audience costs may be a potent, but unobservable source of leverage in crisis bargaining (Schultz Reference Schultz2001b, 33). Although it is true that potential selection effects need to be taken into account in any testing strategy, this argument goes too far. First, the logic of domestic audience costs theory depends on the leader taking a risk of punishment, and if there is no risk, there is no effect. Second, leaders sometimes do fail to carry out their threats in a crisis, and the consequences of those empty threats can in fact be observed.
To avoid inference problems caused by alleged selection effects, the indirect strategy adopted by a number of statistical studies is to test whether states with high potential audience costs, especially democracies (or other accountable regimes), tend to prevail in international coercive bargaining over low-audience-costs states, such as authoritarian (or “personalistic”) regimes. The typical problem confounding this strategy, however, is the difficulty of isolating the effects of domestic audience costs from other informational mechanisms that might explain superior bargaining outcomes for democracies. For example, better policy analysis may allow democratic initiators to succeed in international crisis bargaining because they are better than others at picking the targets of their threats. Alternatively, the greater transparency of the democratic political process may make their threats more credible (Schultz Reference Schultz1999, 245). Schultz's statistical analysis of Cold War and pre-Cold War crisis behavior (Reference Schultz2001a) shows that democratic leaders are generally loath to issue irrevocably committing threats unless the opposition party agrees with the substance of the policy. Thus, democratic leaders’ threats are likely to succeed not because the target realizes that the democratic leader's signaling credibility is at stake, but because the target realizes that the democracy is united behind a policy that makes sense in terms of the country's power and interests.
Other statistical tests are plagued by validity questions and mixed empirical results. For example, Alexander Downes and Todd Sechser's (Reference Downes and Sechser2010) review of prominent statistical studies (Eyerman and Hart Reference Eyerman and Hart1996; Gelpi and Griesdorf Reference Gelpi and Griesdorf2001; Partell and Palmer Reference Partell and Palmer1999; Schultz Reference Schultz2001a) shows that the off-the-shelf databases these studies use are swamped by a large proportion of cases in which leaders make no threats of the kind that would be relevant to audience costs theory.
To surmount the selection and mechanism problems in observational studies of historical crisis outcomes, Tomz (Reference Tomz2007, 830–31) studies domestic audience costs through experimentally manipulated surveys and finds that domestic audience costs are high even when respondents’ substantive policy views are taken into account. He also offers direct evidence of causal mechanisms when he quotes the reasons for which respondents in his survey said they disapproved of hypothetical leaders who failed to follow through on threats (Reference Tomz2007, 835). As predicted by the theory, respondents tended to remark on the leader's inconsistency, tainted credibility, and general incompetence. We look for similar evidence in our historical cases. Notwithstanding the rigorous design of Tomz's experiment, the relevance of his findings to the real world of politics is in doubt (Barabas and Jerit Reference Barabas and Jerit2010). In particular, we think that the survey format makes the hypothetical leader's failure to carry out the threat salient to the respondent in a way that is not necessarily true of real-world news coverage.
Matthew Levendusky and Michael Horowitz (Reference Levendusky and Horowitz2009) mimic Tomz's experimental approach by using surveys to gauge responses of U.S. audiences to hypothetical scenarios in which a U.S. president makes a threat and does not follow through. They find that the most important factor affecting an audience's decision to punish is the ability of the president to control the information the audience receives in the early stages of a foreign policy crisis. This is especially the case when the president can point to “new information” that made backing down a prudent policy decision. When the president is able to shape the public's perception of the crisis, “audience costs shrink dramatically” (Reference Levendusky and Horowitz2009, 3). Thus, views of substantive policy costs trump concerns about failure to carry out a threat.
In addition to looking for correlational evidence, scholars have also suggested testing audience costs theory by looking directly at causal mechanisms in historical crises. In such case studies, audience costs theory needs to show that the risk of domestic punishment comes not from substantive policy disagreement between the public and the leader but from public dismay at the leader's empty threat. Empirically, this turns out to be quite difficult to demonstrate, because audience costs situations rarely arise when the public favors a soft-line policy before the threat is made.
Schultz (Reference Schultz2001b, 41), for example, offers some colorful quotations showing that observers thought that the British government had locked itself into an uncompromising negotiating position by publishing its hard-line diplomatic correspondence during its 1898 showdown with France over Fashoda. However, he also notes that the British public was immediately “incensed” upon learning of French encroachments on Britain's position on the Nile and demanded strong action, whereas Prime Minister “Salisbury himself was inclined to be more flexible.” Cabinet hard-liners, especially Joseph Chamberlain, added to the pressure on Salisbury to stand firm. Only then did Salisbury signal to the French that his hands were tied by publishing the diplomatic correspondence (Sanderson Reference Sanderson1965, 320; Schultz Reference Schultz2001a, 182–88). Thus, hawkish public and cabinet policy preferences, anchored in Britain's massive naval superiority and its perceived stake in protecting its colonial lifeline through the Suez Canal, were firmly established before Salisbury's committing move. Hard-line domestic preferences explain the bridge-burning commitment, not vice versa, so the audience costs mechanism is at best redundant.
In short, problems of causal inference, external validity, and mixed empirical results have bedeviled previous attempts to test domestic audience costs theory. Process tracing of causal mechanisms in historical crises, nearly absent among previous testing strategies, can help overcome these problems if employed in a methodologically self-aware manner.
LOOKING FOR EVIDENCE OF DOMESTIC AUDIENCE COSTS PROCESSES
In historical process tracing, a “smoking gun” test in favor of domestic audience costs theory would be one in which the public was opposed to military action before a threat was issued and then explicitly punished the leader for not following through with a threat with which it disagreed. However, we could not find a clear-cut case of this kind. Consequently, to distinguish the effects of domestic policy disagreements from domestic audience costs, we need to fall back on tracing causal mechanisms in other kinds of cases. These include cases that lead to the outcomes predicted by audience costs theory, in which a democracy prevails over an authoritarian state, as well as cases with outcomes that are not predicted by audience costs theory.
When a democratic leader issues a public threat, the target can concede or resist; the threatener can then stand firm or back down; and the target can subsequently stand firm or back down (Schultz Reference Schultz2001b, Figure 1, 37). According to domestic audience costs logic, as the threatener's ability to generate high audience costs increases, the probability of the target resisting the threatener decreases, and the probability of the threatener standing firm increases; as the target's ability to generate high audience costs increases, the probability of the threatener standing firm decreases, and the probability of the target standing firm increases (Reference Schultz2001b, 40). For Schultz, these results create selection problems wherein “the higher the [audience] costs, the less likely they are observed” (Reference Schultz2001b, 40). However, these selection problems do not prevent us from carrying out process tracing to see whether this logic is at work in cases where leaders may have exposed themselves to potential audience costs.
Our case selection includes examples that conform to the outcome predicted by audience costs logic (the authoritarian target of the democratic threat backs down), as well as examples of deviations from that trajectory. In the cases that reflect behavior anticipated by audience costs theory, we look for evidence that the target backed down because of audience costs reasoning, rather than for other reasons. For example, the threat might have been credible and compelling because the democracy enjoyed overwhelming military power and had substantial interests engaged in the dispute, or because the domestic public was hawkish even before the threat was issued. Later we discuss the Cuban missile crisis as a case that seems superficially to lie on the outcome path predicted by audience costs theory, but in which key causal mechanisms deviate from the theory's expectations and support other explanations.
There are three ways in which a deviation from the audience costs trajectory can occur, with each deviation increasingly problematic for audience costs theory:
1. The nondemocratic target stands firm in response to the democratic challenger's threat (our example is the Sino-Indian War).
2. The democratic threatener backs down when the nondemocratic target stands firm and is punished by the domestic audience (the Iran hostage crisis).
3. The democratic threatener backs down when the nondemocratic target stands firm and is not punished by the domestic audience (the Suez crisis).
Why might these unpredicted trajectories occur? In the first of these scenarios, the prospect of war might fail to deter authoritarians who think the threat to fight is credible, but believe that the democracy is too weak militarily or diplomatically to prevail in the fight, or who think that the democracy's stake in the struggle is small and so it will quit after sustaining a few military setbacks. Another possibility is that the authoritarian thinks the democratic leader is bluffing and will not fight. The authoritarian might reach this conclusion for any of several reasons: The democracy's underlying power and interests seem insufficient to warrant carrying out the threat; public opinion in the democracy is too dovish to support carrying out the threat; the democratic leader has made a threat that is too imprecise or weak to generate sufficient audience costs; the democratic leader seems firmly in power and is therefore impervious to audience costs; or the authoritarian leader does not understand how audience costs work. If any or all of these reasons were found to be prevalent in a significant proportion of cases, they would cast doubt on audience costs theory's claim to be a powerful, general explanation for democratic successes in crisis bargaining.
Also unpredicted by audience costs theory would be the democratic leader's failure to carry out the committing threat if the authoritarian state failed to comply, despite the risk that he might then be punished. Why might this occur? One reason could be that a lame duck leader—or one who is domestically invulnerable—does not care if punishment follows being caught in a bluff. This does not undermine the core logic of audience costs theory, but if such situations are common, it would circumscribe its relevance. A second possibility is that the leader erroneously thinks he can escape severe punishment because the threat was vague, leaving loopholes that might permit the evasion of accountability. This undermines audience costs theory if it happens frequently enough. A third possibility is that the leader intended to carry out the threat, but new information came to light that carrying out the threat would be much costlier than anticipated. This is a substantive policy consideration that erodes the power of audience costs theory. Even if the public does punish the bluffer, an audience costs explanation needs to show that the cause was the bluffing per se rather than the policy preferences of the public. For example, a hawkish public that wanted to use force even before the threat was made might punish the leader for diverging from its preferred policy, whereas a dovish public might punish the leader for having made what it considered an unwise threat in the first place.
Finally, a failure to punish the democratic leader's called bluff would be even further off the path predicted by audience costs theory. This could happen because the audience considered the bluff a smart, low-cost tactic (Walt Reference Walt1999); because substantive policy considerations outweighed the public's dismay at getting caught in an empty threat; because the domestic audience did not care about the issue at stake; or because the leader's threat was vague. Any of these, if common enough, would raise doubts about the potency of audience costs theory.Footnote 6
Any evaluation of a failure to punish requires a counterfactual inference: Even if the leader's standing did not decline after an empty threat, would it have been higher had he not failed to follow through on the threat? The counterfactual needs to assess whether other favorable trends, such as economic improvements unrelated to the crisis, buoyed the leader's status and thus masked the effect of the punishment. We consider these factors when assessing the leader's approval rating.
One particular escape route from punishment, an appeal to changed circumstances, requires further discussion. A leader who wants to back away from a threat could explain to his public that the threat was appropriate when it was made, but changed circumstances had made carrying it out imprudent or unnecessary. Perhaps the opponent had become too strong to resist in the disputed location, or perhaps the opponent's ruling coalition had switched from one with unlimited aims to one with limited aims, such that the concession would have no broader implications. Or perhaps the threat was issued based on erroneous information. For example, the Reagan Administration backed away from a threat to Nicaragua after realizing that the Soviets were not actually sending MiG fighter aircraft there.
Audience costs theory, focusing on inconsistency between word and deed, would presumably expect the public to take a dim view of appeals to changed circumstances, feeling that the leader should not have undertaken unequivocal commitments in such fluid circumstances. In contrast, a perspective emphasizing substantive policy preferences would expect the public to breathe a sigh of relief to learn that the leader was not going to follow through with what had become an imprudent commitment (Slantchev Reference Slantchev2006, 449). If leaders can easily escape from audience costs by claiming changed circumstances, this does not necessarily refute the theory, but it seriously limits its scope.
SCOPE CONDITIONS
Audience costs theorists have tended to make narrow scope claims with respect to the range of threats and commitments that the theory addresses, and so do we.Footnote 7 Audience costs literature generally focuses on specific threats or displays of force undertaken during a crisis interaction, rather than public commitments made well in advance of the crisis. The existence of the crisis shows prima facie that previous commitments were insufficient to deter the challenge, so new, more credible threats are needed to prevail in crisis bargaining.
Explicit verbal threats and implicit threats arising from shows of military force are both seen as creating potential audience costs. We will consider both. Some studies also portray limited use of force as generating even larger potential audience costs if the leader later withdraws from the fight without going all out to win it (Tomz Reference Tomz2007, 828). Although we see the logic in this, we also worry about stretching the concept of audience costs to apply to any conflict involving sunk costs. One problem with this extension is that any use of force beyond the merely symbolic is likely to yield information about the probable success of using military power to defeat the adversary. In this situation, it might be difficult to distinguish an empty threat from Bayesian updating leading to withdrawal from a disadvantageous fight. For that reason, we will examine cases that involve uses of force that are too limited to provide significant information about the likely outcome of more extensive fighting.
Limited uses of force that only partially carry out a verbal threat raise an important question of interpretation. Such actions might be seen as mitigating audience costs on the grounds that the leader made some effort to fulfill the threat, even if the actions did not achieve the stated purpose of the threat. However, we follow Fearon and Tomz in turning that view on its head. Displays of force that fail to compel compliance simply add to the leader's reputational exposure, according to the logic of the theory.
CASE SELECTION
We used Brecher and Wilkenfeld (Reference Brecher and Wilkenfeld2000) as our primary source to identify cases of democratic states that issued threats to authoritarian states in the period after 1945, for which public opinion data are often available.Footnote 8 To facilitate our case selection process, we coded all of the Brecher and Wilkenfeld post-1945 crises involving the United States according to the clarity of the threat, including whether the threat was accompanied by a show or use of force, and the outcome of the interaction. A table showing our coding of these cases is available in a supplemental online Appendix (at http://www.journals.cambridge.org/psr2011013).Footnote 9 We similarly surveyed the Brecher and Wilkenfeld cases involving other democracies in a qualitative manner. We double-checked with the Militarized Interstate Disputes dataset (Ghosn, Palmer, and Bremer Reference Ghosn, Palmer and Bremer2004) to make sure we had not missed any crises.
In choosing cases for more detailed research, we looked in particular for cases that should be easy for domestic audience costs theory to explain, according to two main criteria. First, we looked for cases that involved a clear ultimatum, especially in combination with a show of force. In these cases, there can be little doubt that the audience costs mechanism should have been engaged. The Suez crisis of 1956 is the clearest case by this criterion. Second, we looked for cases that on the surface exhibited domestic political processes that seemed promising for an audience costs narrative. Specifically, we looked for any cases in which a concern about national credibility or honor was voiced in public discourse, domestic public opinion pressed the leader to stand firm, the prudence of the leader's threat became a subject for domestic political discussion, decision makers privately worried that they would be punished for backing away from their threats, the leader was punished for showing weakness in the crisis, or scholarly accounts argued that the leader became trapped into standing firm as a result of his or her own threatening rhetoric.
Based on these criteria, we provide narratives here for the most plausible cases in favor of audience costs theory. In particular, we use these narratives to assess whether audience costs mechanisms were in fact significant once we take the public's substantive policy preferences into account. The Sino-Indian war of 1962 is the strongest case for audience costs theory by this criterion. The Cuban missile crisis and the Iran hostage crisis are ambiguous and worthy of detailed discussion. Our four main cases also happen to illustrate the four trajectories of threat and response discussed earlier, including the path predicted by audience costs theory and three deviant paths. However, the cases were chosen because they are the strongest cases for audience costs theory, not because they are illustrative of these various trajectories. We also briefly mention a few other crises to illustrate the broader range of cases that follow the four trajectories.
One possible source of bias in our pool of cases is the prominent role of the United States, an uncommonly powerful actor that tends to get its way when it makes threats and tends to carry out the threat when it does not. The United States was involved in 54 crises during the period we surveyed, more than twice the number of the next most crisis-prone democracy. U.S. power advantages might make it harder to demonstrate the decisive effect of audience costs on crisis outcomes, because power considerations and audience costs would often point in the same direction. Moreover, the United States's overwhelming capability might make it unnecessary to use audience costs tactics, because its military threats automatically carry considerable credibility. But many cases remain untainted by these problems. The Soviet Union constituted an imposing foe to the United States in several crises. In addition, the domain of cases also includes numerous weaker democracies that did not always get their way. Finally, these post-1945 cases should be assessed against the backdrop of our summary of the findings of Snyder and Diesing's study, which included many pre-1945 crises in which the democracies were not militarily dominant.
For each case, we address issues that correspond to the four assumptions of domestic audience costs theory that we questioned earlier: (1) whether threats were unambiguous enough to generate potential audience costs for the leader, (2) whether any domestic punishment or discussion of possible punishment of the leader was based on unfulfilled threats or on policy disagreements, (3) whether the public's reputational concerns were related to unfulfilled threats, and (4) whether the foreign target of the threat assessed and understood the potential for audience costs. In evaluating domestic punishment, we ask whether the leader or the leader's party suffered a setback in elections or public opinion surveys. To assess the reasons for punishment, we look at opinion surveys, legislative speeches, and commentary in major newspapers, systematically covering at least some of these sources during the period before, during, and after the crisis. We take into account the possibility that rising approval on other issues might mask punishment for empty threats.
WEAK FOLLOW-THROUGH ON A THREAT—NO AUDIENCE COSTS PAID: SUEZ
Audience costs theorists have not offered examples of called bluffs leading to domestic punishment of democratic leaders. Such cases, if they exist at all, are scarce in part because rational leaders generally avoid blatant bluffing. The Suez crisis, however, offers a rare opportunity to observe what happens when democratic leaders fail to follow through effectively on an unambiguous ultimatum that is defied by an authoritarian target. British and French leaders were not punished by domestic opinion, largely because the average citizen was in general agreement with their policies except for the ultimatum. Even critics of those policies failed to emphasize the discrepancy between the threat and weak subsequent actions.
Clarity and Implementation of Threat
In the 1956 Suez crisis, the leaders of Britain and France issued an ultimatum to Egyptian President Gamal Abdel Nasser to reverse his seizure of the canal, were defied by Egypt, briefly undertook inconclusive military operations in the canal area, and quickly withdrew. The crisis was triggered when the United States and Great Britain reneged on their commitment to financially assist Nasser in building the Aswan Dam following large Soviet arms deliveries to Egypt. In response, on July 26, 1956, Nasser nationalized the Suez Canal, which had been leased to the French- and British-owned Suez Canal Corporation (Brecher and Wilkenfeld Reference Brecher and Wilkenfeld2000, 275–76). Conservative Prime Minister Anthony Eden and the British cabinet, concerned about the security of oil supplies and trade, decided to restore Western control of the canal by force, if necessary, hoping to use the invasion as an opportunity to depose Nasser (Fursenko and Naftali Reference Fursenko and Naftali2006, chap. 4; Neustadt Reference Neustadt1970, 12). In France, the Socialist government of Prime Minister Guy Mollet, fearing that Nasser's bold stroke would encourage the Algerian national liberation movement, likewise decided on a forceful response, secretly inviting Israel to invade the Egyptian Sinai to provide Britain and France with a pretext to seize the canal. When multilateral negotiations failed to resolve the issue, the British, French, and Israelis pressed ahead with their secret plan, despite warnings from President Eisenhower and Secretary of State Dulles that the United States would not support the use of force against Nasser (Neustadt Reference Neustadt1970, 12–13).
On October 29, 1956, the Israelis invaded the Sinai Peninsula, prompting an Anglo-French ultimatum calling on the Israelis and Egyptians to order their forces to withdraw ten miles from the Suez Canal within 12 hours and demanding that Egypt allow British and French forces to temporarily occupy positions guarding the canal. In the case of a failure to comply, British and French forces would intervene in “whatever strength may be necessary to secure compliance” (Times of London, October 31, 1956). On October 31 the British and French forces bombed Egyptian airfields and after several days established only a small ground presence at the northern end of the canal (Coles Reference Coles2006, 100–18; Epstein, Reference Epstein1960, 201–24; Neustadt Reference Neustadt1970, 24–25). Egypt appealed for help to the Soviets, who threatened missile attacks against London, Paris, and Tel Aviv. Eisenhower sent Secretary of the Treasury George Humphrey to insist that the British immediately halt operations in Egypt, or else the United States would bar British access to U.S. dollars from the International Monetary Fund and deny British credit from the Export–Import Bank at a time when the pound sterling was already in jeopardy (Neustadt Reference Neustadt1970, 26). Following the U.S. threat, the British and French immediately decided to declare a cease-fire and began withdrawing their forces on November 6, 1956 (Brecher and Wilkenfeld Reference Brecher and Wilkenfeld2000, 276). A United Nations Peacekeeping Force was sent to Egypt to supervise the withdrawal.
Domestic Punishment and Criticism
Eden, Mollet, and their governments suffered little domestic punishment for failing to follow through on their threat. Their public approval ratings remained roughly constant during and after the crisis, and public surveys showed majority agreement with their policies, except that the British public opposed an ultimatum and a military seizure of the canal. After the crisis, Mollet remained in office. Eden stepped down for compelling health reasons in January 1957, replaced by his cabinet colleague Harold Macmillan (Pearson Reference Pearson2003, 173). In the next general election in 1959, the Conservative party increased the size of their majority despite ineffectual attempts by Labour and the Liberals to revive the Suez issue (Beloff Reference Beloff and Roger Louis1989, 333; Epstein Reference Epstein1964, 170–71; Fry Reference Fry2005, 154; Kunz Reference Kunz1991, 142).
Criticism of Eden focused mainly on his poor judgment in putting Britain in an untenable position in the first place—not on failing to persist in this mistake—and on lying to Parliament about collusion with Israel's attack on Egypt (Carlton Reference Carlton1988, 92–93). The dovish Labour opposition criticized Eden for the substance of his policy, not for failing to carry out his threat. A hawkish minority within the Conservative party did criticize the failure to follow through with military action, but on specific policy grounds, not because it would tarnish the credibility of British threats in general, and it remained supportive of the Conservative government.
Before the ultimatum to Nasser, British public opinion largely supported the British government's handling of the crisis, which included various economic and political sanctions as well as participation in international conferences to resolve the crisis. In August 1956, 59% of British respondents indicated that they approved of the government's overall handling of the Suez problem.Footnote 10 An overwhelming majority of respondents (ranging from 65 to 93%) approved various steps taken by the British government, including freezing Egyptian sterling credit, recalling reserve forces, strengthening forces in the Mediterranean, cutting off British arms shipments to Egypt, and calling for an international conference. However, when asked what action the British government should take against Egypt if the conferences failed, only 33% of respondents favored military action, 47% favored economic and political action, and 20% did not know. Of the 33% who urged military action, an overwhelming majority (27%) supported military action even without American support. Public skepticism about the use of force continued in September: 65% of British respondents said France and Britain should not have taken military action against Egypt immediately when Nasser seized the canal. Asked whether Britain should “give Egypt an ultimatum that unless she agrees to our proposals we will send in troops to occupy the Canal,” 49% of respondents disapproved, whereas 34% approved.
The British public disapproved of carrying out the October 30 ultimatum with military force when Nasser failed to comply with its terms. On November 1–2, 44% of respondents thought Britain was wrong to take military action in Egypt, whereas 37% approved. Following the announcement of a cease-fire, British public opinion expressed approval of Eden and his policies. On November 10–11, 53% said they “agree with the way Eden has handled the Middle-East situation” (Epstein Reference Epstein1964, 143). With the safety of hindsight in December 1956, after the cease-fire but before Eden's resignation, 49% of respondents retrospectively indicated they thought Britain had been “right to take military action in Egypt,” whereas 36% thought Britain had been “wrong.” Nonetheless, 53% said it was right to accept the cease-fire, whereas only 34% said the military operation should have continued. The Guardian noted with irony that “the Government gained credit for having stopped a war” (November 15, 1956).
Policy preferences, not general concerns about threat credibility, motivated critics on the left and the right. In the House of Lords, some expressed the opinion that the ultimatum “was a tragic mistake, almost incomprehensible, because at a time when unity was wanted it had split the nation more bitterly and more completely than it had been split before.” The context made clear, however, that the objection was not to the discrepancy between word and deed, but to the policy of force (“Peers’ Views on Suez,” The Times, Wednesday, December 12, 1956).
On the eve of a December 6 confidence vote, the Labour Party criticized the prudence and legality of the Conservative government's policy of force, not its failure to follow through on the threat (“Suez Withdrawal Debate,” The Times, Saturday, December 1, 1956). The government easily survived the vote, though 15 Tories abstained. These were a hard core of hawkish backbenchers who had favored hard-line policies from the beginning of the crisis (“Back-Benchers Fear of Weakening,” The Times, Friday, November 30, 1956). Some hawkish Conservatives may have gently pushed Eden to resign (Beloff Reference Beloff and Roger Louis1989, 320–22; Pearson Reference Pearson2003, 167–68; Rothwell Reference Rothwell1992, 236–37), but most Conservatives believed that Eden had adequately stood up for Britain's interests (Thorpe Reference Thorpe2003, 542–44).
Schultz's excellent short case study of the Suez crisis correctly argues that Eden's threats lacked credibility in part because of Labour's vocal opposition to them (216–22). This supports his theory, which emphasizes the centrality of substantive policy views. It counts against Fearon's audience costs theory, because Labour's sharp criticism started when the threat was issued, not when it was abandoned.
Some might ask whether the British leaders were not punished because they did partially carry out their ultimatum. This is unconvincing on grounds of both theory and facts. Tomz (Reference Tomz2007, 828) correctly argues that an inadequate, token use of force increases the leader's exposure to audience costs. Indeed, none of the British argued that the half-hearted military operation had successfully carried out the threat. A memorandum issued by Macmillan before Eden's resignation captured the mood well: “Suez had been a gamble which failed. Now we must go on as if it never happened” (Pearson Reference Pearson2003, 168).
Even more than the British government, the French government was doubly immune from domestic audience costs. In July 1956, before Nasser nationalized the canal, 29% of French respondents were satisfied with Mollet as President of the Council, a figure that rose to 39% by March 1957. Mollet was not punished for failing to follow through effectively on the ultimatum because he was covered on both flanks. The majority of the French public was disinclined to use force in the first place, and the minority who wanted to use force could not blame Mollet, who would have been willing to continue military operations in Egypt had it not been for the unilateral decision of the British cabinet to withdraw (Neustadt Reference Neustadt1970, 28).Footnote 11
Throughout the crisis, French public opinion was on the whole more supportive of political and diplomatic action than of the use of military force. When asked during the London Conference of September 1956 what would be the best solution to the crisis, a plurality of respondents (41%) preferred to “entrust to an international authority both the management of the Canal and control of the navigation.” Only 28% believed the solution should be forcibly imposed. When asked what to do if Egypt did not accept the decisions taken at the London Conference, 49% of French respondents expressed support for economic and political measures, whereas 20% supported military action. Three-fourths of the latter group indicated that they would support military action even if the United States did not support France and Britain. Notwithstanding the limited public support for the use of force, the French parliament was more supportive of the ultimatum and military action (Times of London, Wednesday, October 3, 1956). After the withdrawal, Mollet's government easily prevailed in a vote of confidence, 325 to 210, in the National Assembly.
Reputational Concerns
The British public and media commentators viewed the nationalization of the canal as a challenge to Britain's national honor and prestige weeks before Eden's ultimatum (Kunz Reference Kunz1991, 77–78). The Manchester Guardian opined that “It is the further weakening of Western prestige in the Middle East rather than the disposition of the canal . . . which is the more serious consequence of these recent transactions” (July 28, 1956). An August 1956 survey found that 23% of respondents thought that past British weakness was to blame for the canal crisis. Insofar as reputation mattered, it took the form of substantive concern about the trend of national power and policy, not formalistic concern over consistency between words and deeds. The British public was concerned about national reputation regardless of Eden's ultimatum.
Reputational costs also played a role in the French public's reaction to the crisis. These international reputational concerns focused on the weakness of French policy in facing Arab challenges in general, not on the domestic audience costs of failing to carry out threats. Explicitly drawing an analogy to “the Munich syndrome,” the French press and politicians argued that weakness on the Suez issue would undermine France's credibility in dealing with rebellions in North Africa. Mollet himself made public comparisons between Nasser and Hitler (Vaisse Reference Vaisse, Roger Louis and Owen1989, 134–35). The public shared this view. In September 1956, 58% of respondents indicated that the Suez crisis had a “very important” impact on the situation in Algeria, whereas only 7% believed it had no effect. Fifty percent of respondents believed the crisis also had an important effect on the relationship between France and Tunisia and Morocco. However, the public's other policy concerns outweighed these reputational concerns. Although the public worried about the reputational consequences of retreats, they saw military intervention as unwise.
Target's Assessment of Domestic Audience Costs
Schultz argues convincingly that the Labour Party's opposition to the use of force undermined the credibility of the Anglo-French ultimatum in Nasser's eyes (Reference Schultz2001a, 220). Nasser's adviser Mohammed Heikal has written vividly of Nasser's reaction to films of a mass antiwar rally in Trafalgar Square, which led him to cancel plans to go into hiding and fight a guerrilla war (Reference Heikal1986, 195). This supports Schultz's theory about the crucial role of opposition support in making democratic threats credible, and it contradicts Fearon's audience costs theory, because Eden's accountability to the public made his threats less credible. There is no evidence that Nasser took audience costs considerations into account when responding to the Anglo-French ultimatum.
In short, the Suez Canal crisis represents a particularly clear case calling into question the efficacy of audience costs. The British and French publics did not punish Eden and Mollet despite their failure to follow through effectively on their very clear ultimatum. Rather, because both publics were reluctant to use force, they preferred their leaders’ policies of backing away from the ultimatum. Criticism of the withdrawal came only from the faction that was already hawkish before the ultimatum was issued.
WEAK FOLLOW-THROUGH ON A THREAT—DOMESTIC POLICY COSTS PAID: IRAN HOSTAGE CRISIS
In the Iran hostage crisis, President Jimmy Carter threatened to use military force unless Iranian students released U.S. diplomats in Teheran, and Iran failed to comply. The United States initiated an aborted attempt to use military force to rescue the hostages, but Carter made no subsequent attempts to make good on his threat. Carter was punished by his domestic audience, though, for the weakness and incompetence of his policy in general, not because of a discrepancy between threats and actions. The public turned hawkish before Carter did and pushed Carter into making threats that he was reluctant to implement. Notwithstanding Carter's poor standing in approval surveys and defeat by the more hawkish Ronald Reagan in the 1980 presidential election, we find little evidence that this punishment conforms to the logic of audience costs theory.
Clarity and Implementation of Threat
Following the seizure of the U.S. embassy in Tehran by Iranian students on November 4, 1979, the Carter administration initially rejected the use of force to compel the release of the hostages (Saunders Reference Saunders, Christopher and Kreisberg1985, 73). The New York Times quoted an administration official on November 5 as saying, “Since we don't have the Shadow or Superman in our employ, any force would be crazy.” However, Carter himself warned of “grave consequences” if the hostages were harmed or put on trial (Los Angeles Times, Nov. 24) and sent an aircraft carrier to the Arabian Sea. “U.S. honor is at stake in Iran and is more important than the lives of the 49 Americans hostages held there,” Carter said in a White House briefing (Los Angeles Times, Nov. 24).Footnote 12 “Other actions which I might decide to take would come in the future after those peaceful means have been exhausted,” Carter said ambiguously. “This nation will never yield to blackmail.” (New York Times, November 28, 1979).
Domestic public opinion was initially supportive of Carter's approach, but turned hawkish well before he issued threats of military action to compel release of the hostages in April 1980.Footnote 13 In November, 71% of respondents agreed that Carter “is doing all he can to free the Americans being held hostage in the U.S. Embassy in Iran,” (NBC News/Associated Press survey, November 27–28) and in December, 76% of respondents approved of the way Carter was handling the crisis (Gallup, December 7–10). At this stage, the public overwhelmingly disapproved of military solutions to the crisis: 70% of respondents opposed a U.S. policy of bombing the main Iranian oil terminal if Iran put the U.S. hostages on trial, and 65% disapproved of a rescue attempt (ABC News/Louis Harris and Associates, November 26–29). Carter's overall approval rating increased from 32% just before the hostage taking to 61% in early December (New York Times, December 10).
Nonetheless, criticism from hawks was gradually building. In December 1979, 48% thought Carter had been “tough enough” on Iran, but 47% said he had not been tough enough (Chicago Tribune, December 24). Beginning in January 1980, public opinion became increasingly hawkish and began to shift against Carter: 58% of respondents indicated that Carter should have taken stronger actions to end the crisis (CBS News/New York Times, January 9–13). In March, 49% believed Carter's handling of the crisis had been a failure, whereas only 33% believed it had been a success (ABC News/Louis Harris and Associates, March 5–8); 60% believed stronger action should have been taken to free the hostages (CBS News/New York Times, March 12–15); 45% agreed that “it is time to do whatever is necessary . . . even if it involves military action,” compared to 44% who disagreed (NBC News/Associated Press, March 20–21).
This increased public support for a military solution helped to push Carter toward making further public threats. By April 1980, says Harold Saunders (140), “a feeling pervaded the administration that the patience of the American people was running out and that firm and decisive steps had to be taken.” On April 7, Carter cut off diplomatic relations with Iran, issued a full embargo, froze Iranian assets in the United States, cancelled Iranian visas, and expelled Iranian diplomats. The public, however, did not believe those additional moves would be sufficient to end the crisis: 51% believed those actions were not tough enough (Newsweek and Gallup Organization, April 9–10); 77% believed “it was a big mistake not to give Iran an ultimatum in the first 72 hours after the hostages were seized”; 51% approved of a military strike against Iran if the hostages were put on trial, “even if it might endanger the lives of the hostages” (ABC News/Louis Harris and Associates, April 8).
On April 17, Carter announced a new set of economic sanctions and issued a public threat to the Iranians: “If this additional set of sanctions that I've described to you today, and the concerted action of our allies, is not successful, then the only next step available that I can see would be some sort of military action, which is the prerogative and the right of the United States under these circumstances” (New York Times, April 17, 1980). Carter added, “The Authorities in Iran should realize that the availability of peaceful measures—like the patience of the American people—is running out” (Chicago Tribune, April 18, 1980). Notwithstanding potential audience costs, these threats failed, and on April 24 the rescue mission was launched.
It is unclear whether the rescue attempt should be seen as fulfilling Carter's vague threat to take military action. In a nationally televised address on April 25, Carter said, “the mission on which they were embarked was a humanitarian mission. It was not directed against Iran” (New York Times, April 26, 1980). This suggests Carter was distinguishing between the rescue attempt and a military response, which never occurred. Carter had previously said that mining Iranian ports seemed to be the most viable military option (Sick Reference Sick, Christopher and Kreisberg1985, 145–46). Even if one asserts that the “humanitarian” mission was in fact a military response, the fact that Carter halted the mission before it could fulfill its stated goal exposed Carter to potential audience costs for having failed to follow through on a threat.
Domestic Punishment and Criticism
In the immediate aftermath of the aborted rescue attempt, the hawkish public supported the policy but criticized Carter's incompetence in carrying it out: 65% believed the United States should have attempted the rescue mission earlier (CBS News, April 26–27); 70% of American believed the “U.S. was right to have tried the rescue attempt and should try again”; 62% rated Carter negatively “on his handling of the hostages situation”; and 72% rated Carter negatively “on his ability to ‘get things done’” (Boston Globe, April 27).
Carter's approval ratings continued to fall. In June 1980, 74% of respondents rated Carter negatively on his handling of the crisis (ABC News/Louis Harris and Associates, June 5–9). In October, 38% of respondents believed Reagan would do more to get the hostages out of Iran, compared to the 29% who believed Carter would do more (Time and Yankelovich, Skelly, and White, October 14–16). In December 1980, 56% of respondents agreed with the statement, “Right after the hostages were seized, the U.S. should have given Iran an ultimatum to give back the hostages by a certain day, and then taken direct military action against Iran, even if this might have endangered the lives of the hostages” (ABC News/Louis Harris and Associates, December 3–6). In January 1981, 71% of respondents agreed that “Iran released United States hostages this week because they were afraid of dealing with Ronald Reagan as President” (NBC News/Associated Press, January 21–23).
In the November 1980 Presidential election, 8% of voters said their choice had been affected primarily by the hostage issue (Roper, November 15–23). Whatever punishment Carter received as a result of his Iran policy needs to be considered in the context of his low general approval rating before the crisis, continuing displeasure with his stewardship of the economy, and the probably mixed effect of the Soviets’ December 1979 invasion of Afghanistan, which at least briefly boosted his reputation for the perceived toughness of his response (New York Times, January 16, 1980).
Reputational Concerns
There is no question that Carter's domestic audience punished him for his handling of the crisis, but there is very limited evidence that this had to do with any mismatch between Carter's words and deeds. Rather, the public disapproved of Carter because they had more hawkish policy preferences and because they were dismayed by the reputational costs of failing to redress the hostage taking, irrespective of Carter's statements.
During the course of the crisis, the public did become concerned with the consequences for U.S. national honor, but because of the inherently humiliating situation, not because of unfulfilled threats per se (Saunders Reference Saunders, Christopher and Kreisberg1985, 87). At the beginning of the crisis, the public was evenly divided about the impact of the crisis on national honor: 26% believed the crisis caused the world to have less respect for the United States, 30% believed it caused the world to have more respect for the United States, and 38% believed there was no change (Newsweek and Gallup, December 5–6, 1979). As the crisis progressed the answers became significantly more negative. In early April 1980, before Carter's ultimatum, 71% agreed that “Up to now, the U.S. has been at the mercy of the Ayatollah, making us look weak and helpless” (ABC News/Louis Harris and Associates, April 8). In December 1980, 70% continued to agree with that statement (ABC News/Louis Harris and Associates, December 3–6). On October 29, 1980 the Wall Street Journal reported that Reagan's strongest question to Carter in the candidates’ debate had been, “Is America respected throughout the world?”
These were concerns about reputational costs, but not audience costs per se. We found no direct complaints about discrepancies between Carter's threats toward Iran and his actions. Polls did not ask that precise question. Carter's political opponents did claim, however, that his foreign policy in general was “inconsistent.” Complaining that Carter had not invested in sufficient naval might to back up his diplomacy against Iran, Republican National Committee chairman Bill Brock said that Carter had been “scabbard-rattling in the last couple of weeks, but without anything in the scabbard.” George H.W. Bush likewise hoped that Carter was not “bluffing” beyond his military means (New York Times, January 20 and 30). But these were points made in policy debates about defense budgets, and not evidence of audience costs in Fearon's sense.
Moreover, when respondents were forced to choose between specific policy costs and diffuse reputational costs, they tended to care more about the former. Asked in December 1980 whether it was “more important . . . to preserve the lives of the 52 hostages or to preserve the honor of the United States,” 52% said the lives of the hostages (Louis Harris, December 3–6). This preference order was reversed only when the question was posed in a way that created a presumption that no policy had any hope of securing the hostages (ABC News/Louis Harris and Associates, April 8).
Target's Assessment of Domestic Audience Costs
Although Carter's public threats failed to compel the Iranians to release the hostages, earlier threats, including private ones, of “extremely grave” consequences if the hostages were harmed or put on trial do seem to have been effective. To back up those deterrent threats, Carter sent the USS Kitty Hawk to the Persian Gulf as part of the largest fleet gathered in the Indian Ocean since World War II. In response, the Iranian regime toned down its rhetoric. Khomeini refrained from threatening to execute or try the hostages except in the event of a U.S. military attack (Sick Reference Sick, Christopher and Kreisberg1985, 147–48). National Security Council staffer Gary Sick says, “On balance, there is reason to believe that the U.S. threat of retaliation was received in Tehran, that it was considered credible, and that it was heeded by those in authority” (Sick Reference Sick, Christopher and Kreisberg1985, 149). Comments by Ali Akbar Hashemi Rafsanjani and others in the regime illustrate that the Iranians were aware of the electoral cycle in the United States and the effect of the hostage crisis on Carter's bid for reelection (Sick Reference Sick1991, 86, 164). Thus, threats deterring action against the hostages were credible, but threats designed to compel their release were unpersuasive. What was decisive in this case was the action demanded, not the publicity of the threat.
Overall, it appears that the U.S. public cared first and foremost about substantive policy costs—getting the hostages home safely—and then about reputational costs irrespective of whether threats were made—preserving the honor of the United States and punishing Iran. There is little evidence that the public cared at all about—or was even aware of—audience costs as Fearon and Tomz define them.
THREAT FAILS AND IS CARRIED OUT: SINO-INDIAN WAR
It is fairly common that a target fails to comply with a democracy's threat and the democracy carries it out. A number of post-Cold War cases fit this pattern: the failure of Saddam Hussein to withdraw from Kuwait in 1991; his failure to fully comply with U.S. demands for maximally intrusive weapons inspections and later for his abdication in 2003; the failure of the Taliban to hand over al-Qaeda leaders responsible for the September 11, 2001, terror attack; and Serbian President Slobodan Milosevic's refusal to accept highly invasive U.S. peace terms for Kosovo in the Rambouillet talks. All of these were ambitious attempts at compellence (warnings demanding a change in the status quo), not deterrence. The failure to comply arguably stemmed not from a lack of U.S. credibility but from inability to comply in the Taliban case, the prohibitively high cost of compliance for Saddam in 2003 and Milosevic at Rambouillet, and the target's preference in each case to fight and lose rather than give up without a fight. In no case did U.S. leaders use committing threats to tie their hands to carry out policies that they would otherwise have been reluctant to implement. Audience costs are largely irrelevant to these cases.
Among Cold War crises, the Sino-Indian war of 1962 comes closest to illustrating the causal mechanisms of audience costs theory in the case of a democratic leader carrying out a threat when the target fails to comply. The threat consisted of an unyielding negotiating position on defining the remote, mountainous Sino-Indian border, combined with shows of force. In the most prominent account of this crisis in international relations scholarship, by Richard Ned Lebow, the case appears to provide a straightforward example of the audience costs mechanism: During the crisis Indian Prime Minister Jawaharlal Nehru went public with Sino-Indian diplomatic correspondence that aroused domestic antipathy toward China, locking him into a policy of confrontation and eventually war. In fact, the story diverges from a straightforward audience costs account in decisive respects.
Clarity and Implementation of Threat
The western part of the Sino-Indian border in the trackless mountains was ill defined in the 1950s because of ambiguous historical precedents and the lack of a continuous presence on the part of officials of either state. Tibetan separatism from China heightened the significance of this boundary. To solidify access to western Tibet and to control cross-border movements of Tibetans, the Chinese began road-building in the disputed territory. Discovering the roadwork in 1958, India demanded Chinese withdrawal in private diplomatic correspondence. Minor armed clashes occurred as India made forward troop deployments in 1959. With each side believing that it would prevail in the event of an armed confrontation, both remained intransigent.
Nehru's allegedly hands-tying public threat came when India released the first of the diplomatic correspondence in September 1959. Nehru's presentation of the issue in Parliament portrayed China as a “bully” usurping India's rightful territory (Maxwell Reference Maxwell1970, 121) and repeatedly alluded to the possibility of war. At the same time, Nehru left himself some room for bargaining, waffling on the question of ambiguities in the delineation of the boundary (Maxwell Reference Maxwell1970, 117–19). This was the first that the Indian public learned of the dispute. Subsequent White Papers were released in the lead-up to the war (Hoffmann Reference Hoffmann1990, 68). Forward Indian deployments encroached on China's strategic road, adding a show of force to the threat implied in the diplomatic dispute. With neither side willing to back down, the two armies jockeyed for position in the disputed territory, and on October 20, 1962, the Chinese attacked, expelled the Indian forces, and then withdrew unilaterally on November 21.
Domestic Punishment and Criticism
Lebow argues that Nehru “apparently found anti-Chinese sentiment useful in strengthening his hand against Peking and did his best to encourage it.” Later in the crisis, Lebow claims, Nehru could not moderate his policy because he “became the prisoner of the crude nationalist passions he had helped to arouse.” Thus, Nehru's release of the diplomatic documents “was directly responsible for arousing the popular passions which ultimately compelled him to pursue his challenge of China to the point of war” (Reference Lebow1981, 184, 188). More recently, Steven Hoffman has likewise argued that public sentiment made it “imperative for the Nehru government to adopt only those policies that could conceivably meet with approval from an emotionally aroused parliament and press” (Reference Hoffmann1990, 68, 70–74). Assessing a possible territorial swap, Nehru is alleged to have confided to a colleague, “If I give them that I shall no longer be Prime Minister of India” (Maxwell Reference Maxwell1970, 161). In short, this sounds like a domestic audience lock-in by means of an implicit, but committing, threat.
However, the reason that the Indian political class and attentive public came to oppose concessions was not concern about Nehru's credibility but the substantive view that China was an aggressor that had to be stopped from encroaching on Indian territory (Maxwell Reference Maxwell1970, 103). Although China's motive in seeking a more favorable delineation of the border was to protect a strategic road needed to suppress Tibetan rebels, Indian opinion agreed with Nehru that China was acting from aggressive, not defensive, motives. The Indian public's furor when it learned about Chinese moves, says Hoffman, reflected a longstanding, widespread view that “throughout its long history China had shown a tendency to be ‘aggressive’” (Reference Hoffmann1990, 53). In other words, the Indian public was hawkish because of its concern about substantive policy costs, not because Nehru had put his credibility on the line.
This calculation of policy costs by both Nehru and the Indian public was strongly influenced by their unwarranted optimism that India would prevail in a showdown with China. Because Nehru and his military had an inflated estimate of India's prospects if matters came to a fight, he discounted the likelihood of China taking confrontational military steps against India (Guruswamy and Singh Reference Guruswamy and Singh2009, 74; Vertzberger Reference Vertzberger1984, 129). “I have no fear of China, great, and powerful as that country is,” Nehru said in October 1959 (Vertzberger Reference Vertzberger1984, 123). In a speech to Parliament in August 1961, Nehru declared that the balance of power was changing in India's favor (128). Failing to appreciate the unpreparedness of the Indian army and exaggerating Chinese internal weaknesses as a result of the Great Leap Forward, Nehru stuck with this assessment until events decisively refuted it on October 20, 1962 (123–25). Because public opinion heard no alternative strategic analysis, they assumed that India would win the war and thus saw no reason that Nehru should give in to an aggressor (Lebow Reference Lebow1981, 190–92; Maxwell Reference Maxwell1970, 132, 178, 240–42). Because the Indian public got its strategic information from India's governing elites, Nehru and Indian opinion had the same substantive preferences and the same assessment of the bargaining power of the two sides. This is a case of misperception and manipulation of apparent policy costs, not a case of audience concerns about inconsistency between word and deed.
Moreover, it is doubtful that audience costs concerns led the democratic leader to undertake actions that he would otherwise have been reluctant to carry out. Nehru had hawkish preferences throughout the crisis, regardless of public opinion. His shift toward more moderate language at one stage of the crisis reflected a desire not to compromise but to seek a negotiated settlement almost entirely on Indian terms (Maxwell Reference Maxwell1970, 118–19, 131–32). Nehru freely undertook the crucial decision at the end of 1960 for forward military deployments that threatened China's lifeline to Tibet. “The Government was not being spurred to implement the forward policy by popular pressure,” says Neville Maxwell, a premier historian of the crisis and Lebow's main source. By that time, “public interest in the boundary dispute had simmered down” (205).
According to Maxwell, Nehru's decision to publish his uncompromising diplomatic correspondence with China “worked only to push him in directions he had chosen himself . . . so the controls were locked—but in the positions in which Nehru had set them” (134). Nehru was using publicity not to lock in what Snyder and Diesing would call a committing threat, but simply to underscore what they would call a warning. Moreover, the lock-in mechanism depended on Nehru convincing the Indian public that an uncompromising policy was correct, not mainly on his putting his own credibility on the line. This is not an audience costs case in the narrow sense in which Fearon and his successors have defined it.
Reputational Concerns
Hoffman (Reference Hoffmann1990, 48–49) argues that the Indian public's mindset was hypersensitive about affronts to Indian sovereignty. If so, this attitude might have inclined them to punish leaders whose reputation for standing up to Chinese challenges had been tarnished. Indeed, in the aftermath of the diplomatic revelations of 1959, Defense Minister Krishna Menon was attacked for allegedly neglecting the army and advocating the earlier policy of friendship with China (Lebow Reference Lebow1981, 187). However, this criticism had nothing to do with backing away from a threat. Rather, it rebuked Menon for not being tough enough on China in the first place.
The reputation of Nehru, the 70-year-old founder of independent India and leader of a one-party-dominated democracy, was not much in doubt. Maxwell (Reference Maxwell1970, 113) says that “Nehru's dominance was by no means absolute,” but opposition to him was “latent” and “muted.” Notwithstanding Nehru's disingenuous remark about losing power if he compromised, he seems far too entrenched a leader to have gained much bargaining leverage from potential audience costs. This limits the potential scope of whatever audience costs dynamics might have been operating during the crisis.
Target's Assessment of Domestic Audience Costs
Nehru's threats failed to deter the authoritarian Chinese opponent for several possible reasons, any of which would call into question an audience costs account. Some scholars argue that the Chinese leadership failed to appreciate how Indian public opinion might act as a constraint on government policy. For example, according to research based on transcripts of a 1959 Chinese Politburo meeting on the Tibetan rebellion, “because the Chinese leader [Mao] did not appreciate India's democratic and pluralist political system, he regarded all the criticism from India (including from the Indian media) as indications of the Indian government's support for the rebels” (Guruswamy and Singh Reference Guruswamy and Singh2009, 78; Jian Reference Jian2006, 54–101). One People's Republic of China military historian, however, says that Mao believed that Nehru was carrying out his bellicose strategy in part to divert the attention of the Indian public from its domestic problems.Footnote 14 If so, this would indicate the salience of Indian domestic politics in Chinese thinking, but not an audience costs mechanism. Alternatively, it seems plausible that the Chinese did expect India to attack, but believed that Nehru had grossly overestimated India's diplomatic and military advantages, and thus did not fear a fight (Christensen Reference Christensen, Ross and Johnston2006, 64).
THREAT SUCCEEDS: CUBAN MISSILE CRISIS
A number of post-1945 international crises have proceeded on the path expected by audience costs theory, in the sense that an authoritarian state complied with a democratic state's threat. But in fact these cases had little to do with the causal mechanism stipulated in the strictly stated version of audience costs theory. Often the military balance favored the democratic power, so the target backed down. Domestic politics frequently did not matter very much, and when it did, substantive policy views were more salient than concerns over consistency between the leader's talk and action.
Several cases superficially fit Tomz's experimental finding that audience costs are greater in cases of shows of force than in cases of merely verbal threats. To mention some examples, verbal threats to the Communists to withdraw from Trieste in 1945 achieved nothing until military forces were deployed; President Kennedy's verbal warning about the Pathet Lao's encroachment on Thailand had little effect until troop deployments led to a negotiated settlement; complaints to the North Koreans following the deadly incident over the pruning of a poplar tree in the DMZ in 1976 accomplished nothing until the United States made a show of force. But domestic hands-tying had nothing to do with these incidents. Rather, the powerful U.S. military became directly engaged against a weaker adversary in circumstances that strongly implied that there was little room or time left to avoid an immediate clash that the opponent could hardly want (Brecher and Wilkenfeld Reference Brecher and Wilkenfeld2000).
In some cases of successful threats, domestic political considerations cut against an audience costs account. For example, President Clinton threatened invasion to bring order to Haiti in 1994, deploying naval forces and readying a landing party to intimidate Haitians who were resisting international intervention. These Haitian foes, recalling Clinton's abandonment of the humanitarian intervention in Somalia in the face of light casualties, gathered on the dock and shouted “Mogadishu! Mogadishu!” Nonetheless, their leaders capitulated, but hardly because audience costs tied Clinton's hands. Only 31% of Americans supported the intervention, and even after it had succeeded, only 41% did (Brecher and Wilkenfeld Reference Brecher and Wilkenfeld2000). This doubly undermines the audience costs argument, first because opinion was inclined to punish carrying out the threat, and second because opinion was driven more by views on substantive policy costs than by concerns about consistency between word and deed.
Though superficially on the expected path, the Taiwan straits crisis of 1954–55, when Communist China shelled the Kuomintang-held coastal islands of Quemoy and Matsu, turns audience costs theory on its head. Instead of using threats to lock in deterrence by raising domestic costs of withdrawal, President Eisenhower publically hedged his threat in order to bring doves into a domestic consensus behind his policy. Only when he decided that mollifying domestic critics was less important than sending a stronger deterrent signal did he switch to a more threatening stance, saying that the United States would use tactical atomic weapons to defend the offshore islands “just as you would use a bullet or anything else” (Chang Reference Chang1988, 108). When Eisenhower's nuclear threats provoked howls of domestic U.S. and international protest, the President once again changed his tack. Telling his national security team in March 1955 that war over Quemoy and Matsu would be “undesirable” because the allies would not support it and U.S. opinion would be divided, he dispatched Admiral Radford to Taipei to propose that Chiang Kai-shek pull his troops out of the offshore islands (Taylor Reference Taylor2009, 478–82). The mainland Chinese eventually backed down despite U.S. public protests characterizing the Pentagon's war plans as reckless (George and Smoke Reference George and Smoke1974, 291), contrary to the expectations of audience costs theory.
The Cuban missile crisis looks superficially to be on the path expected by audience costs theorists in that Kennedy made threats and Khrushchev backed down. Moreover, Kennedy did indeed worry about being perceived as failing to carry out a threat and suffering domestic punishment for that reason. Nonetheless, audience costs theory is a poor guide to understanding the dynamics of the crisis and its outcome. Kennedy's threat to respond to any emplacement of offensive missiles in Cuba was ambiguous, leaving him room to maneuver and escape punishment. He issued the threat not to tie his own hands, but because he was already under pressure from a hawkish public and opposition. Kennedy's reputational concerns focused mainly on how he would appear to the Soviets and to U.S. allies, not to domestic critics. Finally, Khrushchev seems to have been completely tone-deaf to the audience costs mechanism.
Clarity and Implementation of Threat
Until September 1962, Kennedy made no clear threat to deter the Soviet decision to ship missiles, bombers, and nuclear warheads to Cuba because his intelligence service saw this contingency as highly improbable. Only after Republican Senator Kenneth Keating publicized rumors of significant Soviet missile deployments did Kennedy warn on September 4 that “the gravest issues would arise” if offensive ground-to-ground missiles were deployed in Cuba.Footnote 15 Just 72 hours later Kennedy learned of photographs indicating that the Soviets had already done exactly that (Fursenko and Naftali Reference Fursenko and Naftali1997, 206, 227). After 13 days of deliberation, Kennedy went public on October 22 with the compellent demand that the Soviet Union withdraw offensive strategic forces from Cuba. He announced that a naval “quarantine” would be a first step toward enforcing U.S. demands. In Snyder and Diesing's terms, these were neither full-fledged committing threats nor warning threats, because they were ambiguous about the actions the United States would take. Rather, they were “problematic threats,” because the threatener (not to mention the target) was unsure what he would do in the face of noncompliance. As it turned out, Khrushchev removed the missiles and warheads, and Kennedy gave him a public but not formal promise not to invade Cuba and a secret promise to remove obsolete U.S. missiles from Turkey.
Domestic Punishment and Criticism
U.S. officials worried about their domestic political vulnerability in light of the hawkish attitudes of the U.S. public.Footnote 16 Criticism from domestic hawks goaded President Kennedy to issue his September 4 threat, and top officials later discussed whether this threat had created potential audience costs if Kennedy failed to carry it out sufficiently. This case supports our conjecture that audience costs are most likely to arise as a second-order complication when public opinion already has hawkish preferences and pushes the democratic leader toward making a committing threat.
U.S. public opinion was quite hawkish throughout the crisis, even before Kennedy made any public threats. In the wake of the failure to follow through with the Bay of Pigs invasion, Cuba was the only foreign policy issue for which Kennedy received negative public opinion ratings at the beginning of 1962. Even before the revelations about the Soviet missiles, 70% of the increasingly hawkish U.S. public wanted tougher action against Castro's Cuba—not war, but measures to “starve them out.” Overall Kennedy's approval rating was 60% before the crisis, rose to 76% immediately afterward, but by March 1963 fell back to 66% in part because of continuing Soviet troop presence in Cuba (Freedman Reference Freedman2000, 161, 225, 249; Weisbrod Reference Weisbrod2001). In this climate, Kennedy hardly needed to manufacture additional audience costs to convey the possibility that he might be punished for taking too weak a line on offensive missiles in Cuba.
Concerns about the potential domestic cost of taking too soft a line against the missile deployments pervade the ExComm discussions. Early in the crisis on October 16, Assistant Secretary of State Edward Martin opined that “you've got to move immediately, or . . . you're going to have a lot of instability in this country,” to which President Kennedy responded, “Oh, I understand that.” At the same session, Defense Secretary Robert McNamara said, “I don't believe it's primarily a military problem. It's primarily a domestic political problem” (Stern Reference Stern2003, 82, 93). At the October 18 meeting, Treasury Secretary Douglas Dillon jotted to speechwriter Theodore Sorensen, “Have you considered the very real possibility that if we allow Cuba to complete installation and operational readiness of missile bases, the next House of Representatives is likely to have a Republican majority?” (George Reference George2003, 115, 125, 204).
Reputational Concerns
Kennedy's concern about his reputation for resolve in the wake of the Bay of Pigs fiasco and his harrowing meeting with the blustering Khrushchev in Vienna had both international and domestic dimensions. His televised statement announcing the blockade, like his early book Why England Slept, invoked the appeasement of Hitler: “The 1930s taught us a lesson: Aggressive conduct, if allowed to grow unchecked and unchallenged, ultimately leads to war” (Fursenko and Naftali Reference Fursenko and Naftali1997, 246). This echoed the October 16 ExComm discussions in which Attorney General Robert F. Kennedy worried that “a year from now” the Cubans could blackmail other Latin American countries, and the President observed that a failure to act would make Cuba appear “coequal with us,” or as Dillon put it, as if “we're scared of the Cubans.” Martin stressed “the psychological factor of our having taken it,” meaning that the United States could not appear to “have sat back and let ‘em do it to us, that is more important than the direct threat” (Stern Reference Stern2003, 81–82).
Kennedy's recorded tête-à-tête with his brother Robert on the day after his televised announcement revealed that this foreign policy danger interacted with his domestic concerns. The President told him, “It looks like it's gonna be real mean, doesn't it? But on the other hand, there's really no choice. If they get this mean on this one—Jesus Christ! What are they gonna fuck up next?” The Attorney General responded, “No, there wasn't any choice. I mean you woulda had a . . . you woulda been impeached.” “Well, that's what I think,” the President replied, “I woulda been impeached” (Stern Reference Stern2003, 204).
Thus, the President and his advisers worried about the reputational foreign policy costs of seeming weak and about the risk of domestic punishment by a hawkish public, whether or not they had made explicit statements of commitment. In addition, they also worried about audience costs stemming from unfulfilled threats. Historian Sheldon Stern provides atmospheric commentary that helps shed light on the following JFK quotation from the October 16 ExComm meeting, regretting his September 4 public threat:
“Last month,” he began, in an almost jocular tone as several advisers chuckled in the background, “I said we weren't going to [accept offensive missiles in Cuba] and last month I should have said we're . . . well, that we don't care. But when we said we're not going to, and then they go ahead and do it, and then we do nothing,” he suddenly became very somber, “then . . . I would think our . . . risks increase.” (Stern Reference Stern2003, 82)
But was Kennedy referring to domestic audience costs of his failed deterrent threat, or foreign ones? On the sole occasion when an ExComm participant used the word audience, it was in a strictly international context. On October 27, National Security Adviser McGeorge Bundy said “there are two different audiences here, Mr. President,” the Soviets and the NATO allies; “if we sound as if we wanted to make this trade to our NATO people and all those who are tied to us by alliance, we are in real trouble” (Stern Reference Stern2003, 316).
McNamara, however, invoked audience costs in a clearly domestic context on October 16, but argued that domestic audience costs could be escaped without having to attack Cuba: “This is a domestic political problem. In our announcement, we didn't say we'd go in and not that we'd kill them. We said we'd act. Well, how will we act?”Footnote 17 Implying that Kennedy's September 4 statement had been ambiguous about means, McNamara suggested that the threat could be carried out by intensive surveillance, naval blockade, and a more explicit deterrent threat against the launch of the missiles that were already in place (Stern Reference Stern2003, 91–92).
Kennedy's critics, however, warned him that reputational costs could not be escaped so easily, at least in the international arena. Senator Richard Russell, meeting with the President just before his televised speech, reminded Kennedy that “you have warned these people [the Soviets] time and again . . . as to what would happen if there was an offensive capability created in Cuba. They can't say they're not on notice.” We should “assemble . . . an adequate force and clean out that situation,” or else “time's gonna come, Mr. President, when we're gonna have to take this gamble in Berlin, in Korea, in Washington, D.C. and Winder, Georgia, . . . for the nuclear war.” “The more that we temporize, the more surely he [Khrushchev] is to convince himself that we are afraid to . . . really fight” (Stern Reference Stern2003, 165).
In all of these examples, the risk of audience costs was directly bound up with the substantive argument that failure to resist aggression would lead to more aggression in the future. Domestic audience costs from a possible failure to carry out a threat were a concern only because a sizeable, powerful domestic constituency was already worried about international reputational costs before the threat was issued. McGeorge Bundy later claimed that Kennedy's September 4 warning was issued to defuse hawkish pressures at home: “We did it because of the requirements of domestic politics, not because we seriously believed the Soviets would do anything as crazy from our standpoint as placement of Soviet nuclear weapons in Cuba” (Weisbrod Reference Weisbrod2001, 93). In short, hawkish domestic attitudes about reputational costs cannot in this case be “held constant” to test the separate impact of domestic audience costs because these underlying hawkish preferences set up the conditions giving rise to audience costs. No preexisting international reputational concerns, no domestic audience costs.
Though concerned about the consequences of looking weak, the Kennedys were far from locked in by international reputational costs or by domestic audience costs. In fact, they saw political dangers in behaving too belligerently as well as too passively. Robert Kennedy, wondering how it would look to reject the Soviet proposal to trade for the Turkish missiles, notes that the man in the street could easily think that the Soviet offer “is very reasonable . . . And we just turned it down, now suddenly we drop the bomb on Cuba” (Stern Reference Stern2003, 320). In response to McNamara's harping on Kennedy's domestic political problem, Kennedy adopted a “lecturing” tone, says Stern, pointing out that an invasion of Cuba would strain the NATO alliance, because the allies see the United States as “slightly demented on this subject” and do not want to risk a major war over Cuba (100). Likewise, Kennedy later argued that the allies would accept a trade of the Turkish missiles once they realized that the alternative would be Soviet retaliation against them for a U.S. attack on Cuba (324).
The Kennedys frequently demonstrated an awareness of the importance of publicity as a factor constraining their own as well as Soviet options, but they sought to manage this problem tactically and not let it drive their overall strategy. Managing publicity problems emerges as a tactical issue repeatedly. The President argued that Khrushchev's proposal on the Turkish missiles could not simply be ignored because it was public (Stern Reference Stern2003, 422). Kennedy was willing to trade away the Turkish missiles, but was determined to keep the deal private to avoid seeming an unreliable protector in allies’ eyes. Publicity of a pledge not to invade Cuba was less worrisome, because forswearing risky aggression had no reputational disadvantages for the audience of U.S. allies, though Kennedy still did not want to release a letter containing the pledge until he was sure the deal would be accepted (334). In another example earlier in the crisis, Robert Kennedy worried that Soviet Foreign Minister Gromyko would announce the missile deployment just before the President's television appearance in a way that would lock the Soviets into a commitment, and RFK suggested a quick preemptive statement (148). In all these ways, the Kennedys adjusted their tactics to considerations of publicity, trying to contain domestic audience costs and especially international reputational costs, but they did not use audience costs to try to lock into an irrevocable bargaining position.
Target's Assessment of Domestic Audience Costs
Although Khrushchev made numerous remarks about the effect of U.S. domestic politics on Kennedy's policy toward the missile issue, he does not seem to have had a coherent understanding of audience costs and was not deterred because of them. Insofar as Khrushchev took Kennedy's audience problems into account in designing his strategy, it seems that Khrushchev expected Kennedy to keep quiet about any missile intelligence until after the November 1962 Congressional elections, by which time the missiles would be operational and thus too risky to attack (Taubman Reference Taubman2003, 543, 555). More generally, Khrushchev held the somewhat self-contradictory view that, on the one hand, the Soviets needed to bargain from strength with Kennedy, who was captive to the “reactionaries” around him, but on the other hand that Kennedy lacked the resolve to prevent the Soviets from achieving such a position of strength in Cuba (552). Khrushchev failed to consult any experts on the United States (Stern Reference Stern2003, 112). Gromyko told him well before Kennedy's September 4 warning that the missile deployment would “cause a political explosion in the U.S. I am absolutely certain of that.” But Khrushchev told him he had no intention of changing his policy (Taubman Reference Taubman2003, 544).
Khrushchev tried to banish U.S. domestic audience costs and reputational costs through sheer obstinacy. At Vienna, Kennedy warned Khrushchev not to demand concessions that would weaken Kennedy domestically in light of his narrow margin of victory over Nixon and the strength of the hawkish Congressional opposition. Reassuringly, Khrushchev told him not to worry about a withdrawal from Berlin, because “U.S. prestige will not be involved, and everybody will understand this” (Taubman Reference Taubman2003, 499).
Sometimes it seemed that Khrushchev simply paid no attention to the way his tactics heightened Kennedy's reputational and audience costs. According to Khrushchev's foreign policy assistant Oleg Troyanovsky, “It never occurred to anyone that publicizing the Turkish aspect of the deal would create additional difficulties for the White House” (Taubman Reference Taubman2003, 570). As the ambassador to the United States, Anatoly Dobrynin, said later, “Frankly, I don't have the impression that everything was thought through” (552). Superficially, it might appear that the Cuban missile crisis developed more or less on the path expected by audience costs theory, but it would be hard to argue that Khrushchev was compelled to withdraw the missiles from Cuba because he was sensitive to or impressed by Kennedy's audience costs.
CONCLUSION: AUDIENCE COSTS IN PERSPECTIVE
Domestic audience costs, defined narrowly by Tomz (Reference Tomz2007, 821) as “the domestic price that a leader would pay for making foreign threats and then backing down,” seem to have at most a very small effect on crisis behavior. This is true for several reasons. First, unambiguous committing threats (ones that commit a previously uncommitted actor) are rare in crises at the brink of war, because leaders correctly see them as imprudent. The more ambiguous threats that such leaders actually make contain escape clauses that allow them to mitigate audience costs if they decide to back down. Second, domestic audiences understandably care more about policy substance than about consistency between a leader's words and deeds. Where these criteria are in conflict, punishment is more likely to be doled out for an unpopular policy than for a failure to carry out a threat. Third, domestic audiences do care about their country's reputation for resolve and national honor, but they care about meeting such challenges independent of whether the leader has issued an explicit threat. Fourth, authoritarian states probably do try to understand the domestic political context of democratic states’ crisis threats, but narrowly defined audience costs are only a small part of that domestic context, and there is little evidence that authoritarian targets of democratic threats perceive audience costs dynamics in the same way that audience costs theorists do.
We have commented on some of the most prominent post-1945 international crises between a democracy and an authoritarian state, and we have built on studies of all twentieth-century crises. We surveyed comprehensive databases and chose what we thought would be the easiest cases for audience costs theory, in which threats were fairly clear and the domestic political audience of the democratic threatener was actively engaged in the crisis. We looked at cases that were both on and off the path predicted by the theory.
The fact that we found so little evidence of audience costs mechanisms in these cases suggests that the significance of the theory needs to be reassessed. We accept the logic that leaders might sometimes face some domestic cost for inconsistency between their threats and actions, but the significance of this claim must be very heavily qualified in light of a more comprehensive understanding of crisis bargaining and the actual track record of twentieth-century crisis behavior. This conclusion is important not only for theory but also for policy. Future leaders of democracies should not come away from their political science classes having gained the impression that democracies can safely get their way in a crisis by publically committing themselves to fight for otherwise unpersuasive objectives.
'/%3e%3cpath%20d='M27.19%208.879L19.12.81C18.605.295%2017.905%200%2017.205%200H2.726C1.99%200%201.326.295.811.81.295%201.327%200%201.99%200%202.727v29.548C0%2033.784%201.216%2035%202.726%2035h22.548C26.784%2035%2028%2033.784%2028%2032.274v-21.48c0-.7-.295-1.4-.81-1.915zm-1.548.663h-5.968c-.663%200-1.216-.553-1.216-1.216V2.395l7.184%207.147zm-.368%2023.874H2.726c-.626%200-1.142-.516-1.142-1.142V2.726c0-.294.11-.59.332-.81.22-.221.516-.332.81-.332h14.148v6.742c0%201.548%201.252%202.8%202.8%202.8h6.742v21.148c0%20.626-.516%201.142-1.142%201.142z'%20transform='translate(4)'/%3e%3c/g%3e%3c/svg%3e)
'/%3e%3cpath%20d='M17%207.971L15.711%206.677%2010.912%2011.495%2010.912%200%209.088%200%209.088%2011.495%204.289%206.677%203%207.971%2010%2015z'%20transform='translate(8%208)'/%3e%3c/g%3e%3c/g%3e%3c/svg%3e)
Comments
No Comments have been published for this article.