The Oxford Research Encyclopedia of International Studies will be available via subscription on April 26. Visit About to learn more, meet the editorial board, or recommend to your librarian.

Dismiss
Show Summary Details

Updated on 20 Nov 2017. The previous version of this content can be found here.
Page of

PRINTED FROM the OXFORD RESEARCH ENCYCLOPEDIA, INTERNATIONAL STUDIES (internationalstudies.oxfordre.com). (c) Oxford University Press USA, 2017. All Rights Reserved. Personal use only; commercial use is strictly prohibited (for details see Privacy Policy and Legal Notice).

date: 19 April 2018

Deterrence and Crisis Bargaining

Summary and Keywords

Several strands of research on deterrence and crisis behavior were developed within different disciplinary, intellectual, and methodological traditions. Although sometimes coexisting as separate subfields, these studies share a common focus on coercive bargaining in international crises. To draw a common thread between two main subfields, strategic studies on deterrence on one hand and general literature on military crises on the other, both similarities and distinctions in their central concepts are delineated. Four general periods (“waves”) are also briefly outlined in the progression of this research area since World War II, each dominated by a distinct paradigmatic tradition. The main attention then turns to arguments about the causal conditions and mechanisms through which deterrence and crisis bargaining succeeds or fails. Since deterrence requires both capable and credible threats to work, divergent explanatory frameworks are discussed for each of these two requirements. Besides theoretical debates, there are also methodological controversies and measurement issues, which are introduced along with the major data collections that have been developed only recently in this area. In conclusion, several research paths are identified and discussed that have great promise for future advancements in the study of conflict and deterrence.

Keywords: deterrence, crisis behavior, crisis bargaining, coercive bargaining, international crises, military crises, conflict

Introduction

In this article, we survey several strands of research on deterrence and crisis behavior that were developed within different disciplinary, intellectual, and methodological traditions. Although sometimes coexisting as separate subfields, these studies share a common focus on coercive bargaining in international crises. To draw a common thread between two main subfields, strategic studies on deterrence on one hand and general literature on military crises on the other, we delineate both similarities and distinctions in their central concepts. We also briefly outline four general periods (“waves”) in the progression of this research area since World War II, each dominated by a distinct paradigmatic tradition. Our main attention then turns to arguments about the causal conditions and mechanisms through which deterrence and crisis bargaining succeeds or fails. Since deterrence requires both capable and credible threats to work, we discuss divergent explanatory frameworks for each of these two requirements. Besides theoretical debates, there are also methodological controversies and measurement issues, which are introduced along with the major data collections that have been developed only recently in this area. In conclusion, we identify and discuss several research paths that have great promise for future advancements in the study of conflict and deterrence.

Conceptual Preliminaries

Perhaps the best starting point for a survey of deterrence and crisis bargaining research is to clarify the central concept—deterrence—in order to differentiate it from related but nevertheless distinct terms. It simply refers to the persuasion of an opponent not to take an action that, it is believed, the opponent would otherwise take in the absence of the deterrent threat. Three definitional components are essential. First, deterrence is about dissuading an adversary from taking a specific action rather than compelling it into an action. The latter situation is typically identified with compellence: “The distinction is in the timing, in who has to make the first move, in whose initiative is put to the test” (Schelling, 1960, p. 195). Although in practice states often combine both strategies in coercive bargaining, this distinction is important at least for analytical purposes.

Second, deterrence is based on the deterrer’s belief that, without a threat, an adversary intends to take the prohibited action. As such, it intrinsically raises the problem of identifying deterrence success. Deterrence succeeds only if the adversary does not take the action (an observable event) but intended to do it prior to the threat—an unobservable aspect that reflects only the deterrer’s beliefs about the opponent’s intentions. This issue has both served as an obstacle in validating arguments about when deterrence works and provided an opportunity for novel arguments that address precisely this issue of unobservables. Lastly, almost all definitions of deterrence specify that the threat is intended to change the opponent’s expected payoffs, so that the costs from taking an action would outweigh any benefits as a result of the deterrer’s retaliation. These are especially critical constitutive elements in formal-theoretic stylizations of deterrence and bargaining.

The next step in conceptualizing deterrence is to delineate the types. Two classifications have turned out to be relevant for theory development. One distinguishes between direct and extended deterrence types, initially termed as primary and secondary deterrence, respectively (Snyder, 1961). Primary or direct deterrence refers to the prevention of attack on the deterrer’s home territory, whereas in extended deterrence, a state attempts to deter an attack on a third party. As the evolution of deterrence theories have shown, it is a greater problem for a deterrer to persuade the opponent that it is willing to carry out its threat when it is extended to defending a third party rather than for self-defense.

Another important classification concerns general and immediate forms of deterrence. As P. M. Morgan originally suggested,

Immediate deterrence concerns the relationship between opposing states where at least one side is seriously considering an attack while the other is mounting a threat of retaliation in order to prevent it. General deterrence relates to opponents who maintain armed forces to regulate their relationship even though neither is anywhere near mounting an attack (1983, p. 30).

This typology facilitates drawing a conceptual and theoretical link between deterrence and crisis bargaining, since the notion of immediate deterrence corresponds to the conventional understanding of militarized crises. For this reason, research on immediate deterrence is directly related to the bargaining studies of crises, and vice versa (e.g., Snyder & Diesing, 1977; Lebow, 1981; Leng, 1983, 1993; Bueno de Mesquita & Lalman, 1992).

Theoretical Foundations

General Requirements for Deterrence

From the outset deterrence scholars sought to understand the basic requirements for effective deterrence or, in other words, the factors that would give an advantage to one side over the other in crisis bargaining. There has been near consensus that the key condition for prevailing in a crisis requires that the defender’s (or deterrer’s) threat be believable, the term often used interchangeably with the notion “credible” (Zagare, 1990, p. 259). The initial formulations of “three main areas in which credibility must be established” pointed to “the areas of capability, cost, and intentions” (Kaufmann, 1956, p. 19). Reflecting continuity in the field, the conditions for credible threats were similarly restated several decades later, but were also refined to associate “costs” to “capabilities.” This resulted in two essential requirements: “(1) military capabilities sufficient to inflict substantial costs on a potential attacker, and (2) the will or intention to use those capabilities if necessary” (Huth, 1988, p. 4). In other words, a deterrer must convince the other side that it is both capable and willing to carry out its threat should its opponent decide to take the prohibited action.

A Brief Survey of Deterrence Theory Evolution

In the first-wave strategic writings (e.g., Brodie, 1946), which spanned from the dawn of the nuclear age in 1945 to roughly 1960, military capabilities, nuclear ones in particular, were seen as the key to effective deterrence. The second wave in the late 1950s and the 1960s, however, brought greater attention to the second major requirement—resolve or intentions. The main focus was on the coercive strategies that would project a deterrer’s resolve to carry out its threats even in the extreme situations of the apparent irrationality of actually doing so (Schelling, 1956, 1960, 1966; Ellsberg, 1968). Phrasing the problem as “a blackmailer’s dilemma,” Ellsberg succinctly summarized it as follows: “my problem as a blackmailer is to ensure—by actions that either change your payoffs, hence your critical risk, or that increase your expectation of punishment—that your estimate of the actual risk is greater than the critical risk. How to do this is, of course, the heart of the blackmailer’s art” (1968, p. 11). Since the second-wave deterrence theory takes a signaling and commitment approach to bargaining and war, its main tenets will be closely examined in the section on commitments and signals.

Although treated as a single wave in the original periodization (Jervis, 1979), we see two subgroups in the third-wave research (beginning in the early 1970s) as a reaction to the second-wave theory. On the one hand, the “decision-making group” (Holsti, 1972; Jervis, 1976; Lebow, 1981) argued that due to psychological, domestic political, or bureaucratic pressures, foreign threats are usually underestimated or overestimated. If leaders do not perceive the environment through the same “rational” lenses, then this undercuts the rationality assumption upon which the second-wave theory was premised. The impact of such cognitive or bureaucratic idiosyncrasies, endemic to foreign policy decision-making, consequently shed doubts on the possibility for generalizations and systematic research in the areas of bargaining and deterrence. On the other hand, the third wave is also identified with those critiques that do not question the possibility of analytic rigor and systematic empirical generalizations in this area. Rather, they target the exclusive focus on the manipulative tactics for influencing the opponent’s beliefs about one’s own resolve at the expense of understanding the nonmanipulative and more fundamental aspects of resolve such as interests (Weinstein, 1969; Maxwell, 1968).

Finally, a few influential works in the third wave combine both types of critiques, most notably Jervis (1970), George and Smoke (1974), and Snyder and Diesing (1977). Besides raising similar issues as the decision-making theorists, George and Smoke (1974), for example, additionally questioned the extent to which the risks generated in coercive bargaining can be controlled and also pointed to the importance of less manipulable commitments. Snyder and Diesing (1977) provided a plethora of new concepts and insights such as, among others, the distinction between resolve and bargaining power, or their differentiation between coercion, accommodation, and persuasion as bargaining tactics. All of these and other notions entered the conceptual and analytical currency of later research. Similarly, Jervis’s (1970) distinction between signals and indices, that is, the manipulative and less manipulative aspects of resolve, provided a general framework within which we can trace different avenues of research in the next, fourth, wave of research.

Since Jervis’s (1979) classic periodization of three deterrence waves ends with the late 1970s, we can delineate some basic directions in the most recent wave that has progressed afterwards. Note that since the first edition of this survey, Lupovici (2010) provided additional insights on the four waves in deterrence research. We would direct the reader to his somewhat different perspective on the fourth wave, which, though not incompatible with ours, also covers a similar body of literature. The fourth wave roughly started in the 1980s with a series of publications that were both theoretically and methodologically distinct from the previous waves (e.g., Huth & Russett, 1984; Huth, 1988; Morrow, 1989; Zagare, 1990; Bueno de Mesquita & Lalman, 1992). As our detailed discussion below will show, the fourth wave substantively expanded on the earlier debates and appears to have been oriented primarily toward (1) the role of private information, and (2) the issue of interests versus signals in projecting resolve.

Yet, as is the case with any periodization, there are notable exceptions that do not easily fit the classificatory mold. In this respect, in his more recent and important study, Lake (2010/2011) reverts to the third-wave theory and methodology to shed doubts on some premises of the fourth-wave research. Using the case of the Iraq War, he questions the validity of “rational behavior” assumptions and instead suggests the pivotal impact of cognitive failures and domestic politics, rather than intentional choice of strategies that take advantage of private information about one’s own resolve (see Fearon, 1995), in influencing escalation and bargaining outcomes.

Besides their substantive differences in approaching the requirements of deterrence, each wave also opted for different methodologies. The first-wave literature was characterized by traditionalist, mostly policy-oriented narratives; the second wave was primarily deductive, generally using strategic-form games; the third wave was characteristically inductive, based on comparative case studies; and the fourth wave has advanced both formal-theoretic and quantitative empirical research. These diverse methodological choices were clearly embedded in the distinct theoretical positions of each wave.

Is Theoretical Rigor Possible? Rational Deterrence Debate

Perhaps the most persistent criticism of classical (second-wave) deterrence theory comes from psychological/decision-making scholars, arguing against its deductive logic based on “abstract rationality” that ignores the barriers to a leader’s “cool-headed” calculations. Instead, the decision-making scholars place an emphasis on both cognitive (unmotivated) and affect-driven (motivated) biases that distort statesmen’s perceptions of an adversary’s commitments and signals (Holsti, 1972; Jervis, 1976; Jervis, Lebow, & Stein, 1985). In the former case, prior belief systems, for example, create cognitive shortcuts for interpreting an opponent’s signals, which in turn results in misperceptions of the adversary’s actual motivations. These misperceptions are especially accentuated in the high-stress situations endemic to crisis decision-making (Holsti, 1972; Jervis, 1976). More recently, Johnson and Tierney (2011) argued that escalatory cognitive biases, especially overconfidence, occur most potently once leaders believe that war is imminent, i.e., when crossing a “psychological Rubicon” of war. In the latter case, when motivated biases are at play, then an individual’s miscalculations are not cognitive, but rather driven by domestic vulnerabilities or other motivating factors unrelated directly to the crisis (Lebow, 1981). Instead of attributing deterrence failure to the unfavorable balance of forces or to a deterrer’s weak resolve, as classical deterrence scholars do, these studies rather ascribe it to a policymaker’s misperceptions and misjudgments.

This led to the well-known “rational deterrence” debate (Downs, 1989), triggering rebuttals on both methodological and theoretical grounds. Methodologically, the decision-making scholars rejected the formal logic of classical deterrence theory in favor of an in-depth case study approach. Yet the advantages of these comparative case studies were marred by selection bias problems resulting from an exclusive focus on situations when deterrence failed (Achen & Snidal, 1989). As a consequence, the decision-making scholars were criticized for not being able to determine whether the same cognitive or motivational biases that they attributed to deterrence failure were absent in those cases when deterrence succeeded. In addition, more recent empirical studies reexamined the findings of the third-wave comparative cases, rendering new results that were more consistent with the rational deterrence approaches stressing the pivotal role of capabilities and resolve for deterrence outcomes (Orme, 1987).

Theoretically, as Zagare (1990) pointed out, the notion of rationality implicit to rational choice models is instrumentalist and quite different from the procedural definition used by the decision-making critics. Instrumental rationality simply indicates that an individual’s preferences are connected and transitive; that is, an individual is rational if she or he can (1) compare between two choices (a and b), and (2) rank these options in a consistent manner such that, if a is preferred to b, and b is preferred to some third choice c, then a is also logically preferred to c. Misperception, on the other hand, represents an aberration from procedural rationality (i.e., well-defined objectives, a comprehensive evaluation of an exhaustive set of options, and “accurate” estimations), which has been erroneously ascribed to rational choice models by its critics.

Consequently, rather than being incompatible, rational choice approaches and those stressing psychological or domestic biases may benefit from their synergy (Zagare, 1990; Wagner, 1992). Whereas the latter can inform the former about preference formation, the former can provide more pointed expectations about behavior once factors such as “incomplete, imperfect, or even erroneous information” are incorporated into formal models. In fact, advancements in formal-modeling techniques have increasingly allowed scholars to analyze the effects of incomplete information and learning—the same factors emphasized by the decision-making scholars—on deterrence outcomes (Morrow, 1989; Wagner, 1992). Kydd (1997), for example, revisits Jervis’s (1976) spiral model from a rational choice perspective, formally stylizing it as a game of incomplete information, and shows that psychologically biased beliefs might matter only under restricted conditions. Carlson and Dacey (2014) formally stylize actors’ fear and anger in terms of their risk assessments and deduce, within the rational choice framework, their subsequent effects on initiating a deterrence crisis. Also, divergent cognitive reactions need not be irrational at all, but rather explainable as a function of whether states have “security-seeking” or “greedy” predispositions. Though skeptical views highlighting the role of cognitive-psychological factors in preventing generalizations about deterrence have occasionally reemerged (Mercer, 1996), the “rational deterrence debate” has not forestalled the analytic rigor in the deterrence research. And the main focus continued to be on the key factors of capabilities and credibility.

Power Capability

Approaches to deterrence from the standpoint of capability factors revolve around several questions: whether and how uncertainty about capabilities influences the effectiveness of deterrence; whether power parity or imbalance leads to deterrence breakdown and crisis escalation; whether nuclear weapons change the calculus of deterrence; and, specific to nuclear deterrence, whether mutual assured destruction or limited nuclear options have a greater deterrent value.

Power Balances, Power Shifts, and Uncertainty

Both balance-of-power and power transition theories of war conclude that an equal power distribution generates greater uncertainty, though they, of course, disagree about its likely consequences: the former expects caution under such conditions, whereas the latter generally expects war. Although these are not bargaining theories per se, their claims were echoed in the bargaining literature. Bargaining studies similarly show that, under conditions of a 50-50 chance of winning or losing a contest, the informational uncertainties about relative bargaining power are heightened, leading to disagreements about which side is likely to win a war (Blainey, 1988; Wittman, 2001; Reed, 2003). These divergent expectations can, in turn, lead to miscalculations as major triggers for war.

Blainey (1988) originally suggested that conflict results from the disagreement between two sides about their relative bargaining power, which is most likely under rough power parity. Wittman (2001) also formally shows that the “expectation differential” between two sides about the likely war outcome is greatest as their prewar power levels approach parity. Similarly, Reed (2003) develops an “ultimatum bargaining” model with one-sided incomplete information in which the challenger is uncertain about the defender’s capabilities. His findings are consistent with Blainey (1988) and Wittman (2001), confirming that the risk of war is greatest when an opponent’s capabilities are underestimated, which is most likely to happen under conditions of rough parity. Empirical evidence generally leans in the same direction, showing, for example, that war is less likely and immediate deterrence more likely to succeed if the balance of forces favors the defender (Huth, 1988; Fearon, 1994a).

In general, however, the observable power alone is not sufficient to predict escalation, because states hold private information about intangible capabilities, such as the quality of forces, leadership, and the like (Morrow, 1989). When observable measures of power approach parity, beliefs about the other side’s “intangible” aspects of power tend to have a greater impact on the bargaining outcome (Bueno de Mesquita, Morrow, & Zorick, 1997; Reed, 2003). Powell (1999) additionally argues that the relationship between power and bargaining is not straightforward and that what really matters is the disparity between the distribution of benefits and the distribution of power. All else equal, war occurs when these distributions are incongruent for one side, thus making it dissatisfied, and the other side does not make sufficient concessions. In their formalization of perfect (equilibrium) deterrence, Zagare and Kilgour (2000) also consider the power balance and dissatisfaction, and they provide validations for, but also further refinements of, power transition theory. Ultimately, all the cited formal-theoretic works demonstrate that the predictions from power-based theories have merits, but also show that the bargaining process must be taken into account as it can substantially alter the actors’ expectations and, consequently, crisis outcomes.

All the same, it should be noted that recent literature offered some innovative ways to think about the impact of capabilities on escalation and bargaining. Sechser (2010), for example, shows that the asymmetry of power impacts a stronger state’s ability to assess the other side’s reputation-driven resolve. Slantchev (2011) suggests that capabilities can be instrumental in transforming the bargaining game by altering each side’s preferences in anticipation of war-winning prospects. Though taking different approaches, both studies made important strides in further understanding how power itself, especially in its relational context, can shape or, alternatively, set limits to the impact of other variables, such as resolve. Nonetheless, the issue of credibility and resolve is multifaceted and, as many studies demonstrate, it is a potent factor in affecting bargaining behavior and outcomes either independently or jointly (interdependently) with capabilities. Therefore, most research is focused on the element of threat credibility in deterrence and crisis bargaining.

Conventional and Nuclear Deterrence

A witness to the unprecedented destructive power of atomic bombs in Hiroshima and Nagasaki brought on a radically new thinking about the logic of warfare in the nuclear age. Bernard Brodie, one of the first influential voices, promptly declared, “Thus far the chief purpose of our military establishment has been to win wars. From now on its chief purpose must be to avert them” (1946, p. 76). This expressed the prevalent view that nuclear weapons have fundamentally altered the nature of warfare as known throughout human history. In the prenuclear world, victory was thought to rest on a strong defense that would deny the opponent any major gains from fighting. In the nuclear age, however, the notion of nuclear victory was deemed impossible. Consequently, as seminally discussed by Snyder (1961), the old strategy for averting wars by denying an opponent any gains from fighting (deterrence by denial) was replaced by a new strategy based on the threats of inflicting intolerable costs on an attacker (deterrence by punishment). The ability and willingness to impose unacceptable damages through a nuclear retaliatory blow was considered a sufficient deterrent against an enemy’s attack.

This first wave of deterrence writings gave rise to two separate debates concerning nuclear deterrence. On the one hand, the initial reaction by the first-wave scholars such as Brodie and Morgenthau was to question the standard calculus between means and ends when considering the use of force. Morgenthau believed that the use of nuclear force was “militarily anachronistic and self-destructive” (1964, p. 35); its singular purpose was not to make its physical application effective, as was the case with traditional force, but rather to make “its physical application superfluous by deterring the prospective opponent from using it” (1964, p. 24). In other words, the balance of nuclear terror with both sides armed with a second-strike capability would be an effective deterrent against an attack by either side.

An alternative view soon emerged, notably in the writings of Albert Wohlstetter and Henry Kissinger, casting doubts on the actual effectiveness of the Mutual Assured Destruction (MAD) doctrine. The idea of limited nuclear war brought back the logic of winnable wars, though of a restricted size, as was the case in conventional fights. The arguments in what came to be known as the MAD vs. NUTs (nuclear use theories) debate, as labeled with some degree of ironic skepticism, were typically confined to the traditionalist, speculative, and mostly policy-oriented studies (for a comprehensive survey of traditionalist and policy-based writings about nuclear doctrines, see Freedman, 1989). Since it is impossible to provide empirical validation for either side, the validity of countervalue (MAD) versus counterforce (NUTs) strategies has been analyzed mainly in formal-theoretic literature, with both sides finding some support though under different contingencies (Powell, 1990; Wagner, 1991; Zagare & Kilgour, 2000).

On the other hand, it has been debated how much a state’s behavior is truly altered in the nuclear environment and whether the sheer possession of nuclear weapons has a greater deterrent value than conventional forces. Although only a few rigorous empirical studies were conducted in this respect, the findings have been mixed at best, ranging from those reporting an effective deterrent value of nuclear weapons (Bueno de Mesquita & Riker, 1982; Huth, 1990; Fearon, 1994a) to others arguing that nuclear deterrence brought little or no difference to conventional deterrence (e.g., Organski & Kugler, 1980; Huth & Russett, 1984; Huth, 1988; Geller, 1990). None of these studies, however, analyzed the effect of nuclear threats on deterrence, but rather the impact, if any, of the mere possession of nuclear capabilities on the onset and outcomes of conventional deterrence. This tradition largely remains in the post-Cold War research as well, albeit with improved methodology that uses larger data sets. In a theoretical sense, however, only a handful of studies go beyond the logic that infers crisis outcomes from the sheer fact of whether a state possesses nuclear weapons and/or which type. While Kroenig (2013) somewhat continues this old theoretical tradition in his rigorous empirical test of nuclear crises, he also suggests the importance of political stakes in impacting their dynamics and outcomes. Narang’s analysis (2013) stands out in demonstrating the importance of nuclear postures, rather than just the possession of nuclear weapons, in shaping nuclear crisis outcomes.

Threat Credibility

Schelling’s seminal approach to the “art of bargaining” was that of “not just estimating enemy intentions but influencing them” (1966, p. 35) by employing tactics such as “bridge-burning,” “competitions in risk-taking,” or “irrevocable commitments.” It shaped the course of deterrence research in two ways: one resulting from critical reactions to Schelling’s approach, the other expanding on his initial ideas of manipulative bargaining. The third wave gave rise to the criticism by highlighting the importance of alternative factors such as less manipulable “inherent resolve” and signals that combine firmness with flexibility. The fourth wave provided theoretical refinements and methodological rigor to both legacies.

Inherent Resolve

As discussed above, there were two lines of reactions to Schelling’s second-wave theory among the third-wave scholars: one highlighting the impact of cognitive-psychological factors and the other critical of Schelling’s emphasis on signals of commitment as the “alpha and omega” of the bargaining process. This latter group emphasized instead the importance of “intrinsic interests” that set limits to the need for, and potential effectiveness of, manipulative bargaining strategies (Maxwell, 1968; Weinstein, 1969; Jervis, 1970; George & Smoke, 1974; Snyder & Diesing, 1977). Maxwell’s (1968) critique particularly resonated with the charge of “Schelling’s tendency to revert to an apolitical concept of strategy” and neglect the key role of interests in influencing deterrence outcomes: “the obvious conclusion to be drawn from the failure of a state to fulfill a commitment is simply that the commitment did not represent an interest worth defending, at the level of violence and risk estimated to be necessary” (Maxwell, 1968, p. 19).

Though significant, these initial critiques were primarily focused on conceptual distinctions between manipulative and nonmanipulative aspects of resolve, as best exemplified in Jervis’s (1970) distinction between “signals” and “indices.” Similarly, Snyder and Diesing (1977, p. 190) differentiated between “inherent” and “actual” bargaining power. The former refers to “what the relative bargaining power would be if the parties had perfect information” about each other’s manifested indices of power and interests, both determining the actor’s “resolve” (i.e., inherent resolve). As such, inherent bargaining power “sets a certain limit on what the parties can achieve by manipulative tactics during the bargaining process.” On the other hand, “actual” bargaining power is “a function of perceived comparative resolve” (Snyder & Diesing, 1977, p. 190) and, as such, it can either be distorted under psychological and bureaucratic pressures (decision-making arguments) or shaped by the opponent’s signals of commitments (Schelling’s signaling theory).

Though significant in initial conceptualizations of inherent resolve, the third wave provided neither a rigorous theoretical development of this aspect of resolve as an explanatory variable nor systematic empirical analyses of the scope of its actual effect. The fourth-wave deterrence studies, mostly those primarily empirically oriented, picked up where the third-wave scholarship left off in this respect. In a series of quantitative empirical analyses of deterrence, Huth and Russett advanced the notion of “issues at stake” (1984; Huth, 1988) in terms of the strategic and economic ties between a defender and its protégé. The earliest study in their research program (Russett, 1963) provided important preliminary results showing that these ties were probably the most important predictors of deterrence outcomes. In the context of major power politics, Danilovic (2002) proposes an approach to the “inherent credibility” of threats in terms of their interests in geopolitical areas. Her findings show that, compared to relative capabilities or manipulative signals of resolve, the extent to which major power regional interests are mutually contested has the strongest impact on deterrence outcomes.

Overall, then, the notion of inherent resolve, identified in terms of military-economic interests, has been shown not only to set important limits on the bargaining range within which manipulative tactics can be effective, but also to be a critical alternative predictor for war. Its potential for theory-bridging beyond deterrence studies is further reinforced by its compatibility with the notion of issue salience in general studies of conflict (e.g., Diehl, 1992).

Commitments and Signals

Besides contextual factors such as the balance of interests and power, a state’s bargaining behavior is another critical factor influencing beliefs about its resolve. There have been several avenues of research on commitments and signals.

While the research on inherent resolve grew out of the critical reaction to Schelling’s bargaining theory, there have been several lines of research since the 1980s that have adopted and further built its premises. The emphasis here has been on the strategies aimed at influencing an opponent’s beliefs about one’s own resolve under conditions of uncertainty. Since states have private information about their true resolve, the problem of avoiding inefficient wars is reduced to the problem of convincing one’s opponent about the state’s willingness to fight (Morrow, 1989, 1999; Fearon, 1995). For this reason, states have an incentive to misrepresent their private information (i.e., bluff) to make themselves indistinguishable from those with a genuine willingness to fight. To solve the puzzle of war as a result of signaling failure under uncertainty, a number of signaling strategies have been identified.

One strand is rooted in Schelling’s approach to costly commitments that make it difficult or costly for a bargainer to retreat (1960, 1966). Strategies such as “burning bridges” and “surrendering options” were advocated precisely with this rationale in mind. More recently, Fearon (1997; see also Morrow, 1999) argues that states can credibly convey their resolve through threats or costly signals that “create some cost that the sender would be disinclined to incur or create if he or she were in fact not willing to carry out the threat” (Fearon, 1997, p. 69). Fearon further differentiates between two types of costly signals: (1) sunk cost signals which are costly for a state to send ex ante, but that have no influence on a leader’s willingness to stand firm versus fight (i.e., resolve) since they are paid regardless of whether the sender carries out the threat, and (2) tying hands signals which do influence the sender’s resolve since they are only paid ex post if she or he fails to carry out a threat. In general, Fearon’s (1997) formal analysis shows that sunk cost signals have little bearing on an opponent’s beliefs about a state’s willingness to stand firm. This is consistent with some empirical results showing that arms transfers or mobilizations, if considered as a type of sunk cost signal, have little influence on the outcomes of extended-immediate deterrence (Huth, 1988; Fearon, 1994a; Danilovic, 2002).

In contrast, tying hands signals are expected to have a much greater impact on threat credibility (Fearon, 1997), especially if understood in terms of tying a signaler’s hands domestically (Fearon, 1994b). This line of research, emphasizing the role of domestic audience costs, has burgeoned in the last two decades. These refer to the ex post domestic costs that leaders pay for reneging on their threats. Since democratic leaders are more vulnerable to domestic political penalties for policy failures, the effect of audience costs should make it unlikely for them to issue threats merely as “limited probes” or bluffs. Consequently, democratic leaders are expected to be genuinely resolute when initiating a dispute and, if their threats are thus considered more credible, they are less likely to be resisted. At the same time, the arguments differ as to why the domestic audience would be willing to punish a leader for foreign policy failure. For Fearon (1994b), it is about the nation’s “reputation” or “prestige,” while Smith (1998) argues that a leader’s competence is at stake. Schultz (2001) emphasizes the role of domestic opposition parties that can capitalize on a leader’s decision to back down in a foreign dispute. Notwithstanding these variations, systematic empirical research resulted in strong support for the domestic audience cost argument in general (Partell & Palmer, 1999; Schultz, 2001; Gelpi & Griesdorf, 2001). At the same time, some recent literature on audience costs developed in the direction of either questioning the validity of this approach (Trachtenberg, 2012) or offering modified arguments about the relevance of audience costs, as comprehensively discussed by Snyder and Borghard (2011).

Note that the idea of tying hands is similar to Schelling’s (1960) strategy of creating an irrevocable commitment by “throwing the steering wheel out the window” in the game of chicken. That is, by issuing public threats, democratic leaders make a commitment from which they cannot retreat without costs. This also raises the well-known risk–return tradeoff (Powell, 1999). If domestic audience costs are generated, it increases the possible returns from the bargaining process by signaling that the democratic state is resolute, but, precisely in doing so, it also increases the risk of war if the other side does not yield.

While it is possible to draw an analytical difference between sunk and tying-hands costs, the distinction may not be as clear in practice. Slantchev (2005), for example, shows that when leaders mobilize resources toward war, they simultaneously sink costs and tie their hands by increasing the expectations for winning a war. One of the interesting implications is that, even if leaders were not resolute at the beginning of the dispute, they can create commitments during a crisis by mobilizing their resources at a high enough rate to make fighting an attractive option. Yet, if mobilization does not deter the opponent from resisting, it increases the risk of war.

Schelling’s signaling theory, however, focuses on the international reputational costs that states pay for their irresolute behavior. Put simply, given incomplete information about their actual willingness to fight, successful bargainers need to establish their reputations for resolve. And, if events are interdependent, as Schelling argued to be the case, then a state’s acquiescence in a crisis would signal to its opponent and other international audiences (adversaries and allies alike) that it would not be willing to defend its commitments in other areas or at later points in time (Schelling, 1966). This type of reputational signaling approach has been rejected from both strands of the third-wave deterrence research on the grounds that reputations for resolve cannot be established as Schelling’s theory predicts, due to either individual psychological or domestic political factors (Jervis, 1970; Lebow, 1981; Mercer, 1996) or the greater importance of inherent resolve in conveying threat credibility (Maxwell, 1968; Snyder & Diesing, 1977; Danilovic, 2002; Press, 2004–2005).

Notwithstanding this criticism, recent research in the fourth wave has examined the empirical validity of international reputational effects and also sought to establish more precisely the types of reputations that matter. On the empirical front, several studies have found support, albeit limited, that a state’s past behavior influences its opponent’s future bargaining positions (Leng, 1983, 1993; Huth, 1988; Gelpi, 1997). One consistent theme that appears in all of these studies is that the context of the past crisis matters. Reputations from past behavior only matter when the defender is facing the same challenger (Huth, 1988; Leng, 1983, 1993; Weisiger & Yarhi-Milo, 2015), if contesting each other in the same area of their interests (Danilovic, 2002), over the same type of issue (Weisiger & Yarhi-Milo, 2015), or if the state backed down in an area of their strong interests (Clare & Danilovic, 2012).

All of these empirical studies examine how a state’s past behavior influences its opponent’s beliefs about its resolve. Some recent studies, motivated in part by the “chain store paradox” and models of entry deterrence, have begun to explore how reputational concerns drive a state’s own behavior as part of a “reputation-building” process. Sechser (2010), for example, formally models a weaker state’s reputational incentive to resist a stronger state in order to prevent recurrent challenges from the same state in the future. Clare and Danilovic (2010) argue that states are particularly concerned with the reputational consequences of irresolute past behavior if they face multiple potential challengers (strategic rivals) in the future. In this case, states have an incentive to invest in their reputations by initiating and escalating conflicts, with the primary goal of rebuilding a reputation for strong resolve and deterring future challenges from other rivals. Their approach therefore departs from other studies by introducing the possibility of proactive reputation-building in general deterrence (that is, reputational incentives for initiating disputes).

Theoretically, scholars have questioned the types of reputations that matter and sought to refine or redefine the notion of reputation itself. Guisinger and Smith (2002) show that threats are credible when reputations are attributed to individual leaders rather than states, and also when there is a mechanism for punishing leaders that damage the nation’s reputation (see also Huth, 1997). Sartori (2005; see also Guisinger & Smith, 2002) argues that reputations for honesty, rather than resolve, are more important. If this is the case, then, in some situations, not acting can be beneficial because it allows a state to maintain its reputation for honesty and reveals to an opponent that when it does resist, it is truly resolved to fight. Honest appeasement, in other words, can work by preserving the credibility of future threats.

Another important body of literature examined bargaining strategies that can project strong resolve while avoiding the perils of risky escalation. Lately better known as firm-but-flexible signals (Leng, 1983, 1993; Huth, 1988; Gelpi, 1997), these were seminally discussed in the 1970s literature on “coercive diplomacy” as an alternative to the “coercion” purely based on firmness as in Schelling’s commitment strategies (George, Hall, & Simons, 1971). Concerned with the risk of an uncontrollable spiral of events, the third-wave research pointed to the need for reducing the opponent’s costs of retreat in order to give him a face-saving option (George et al., 1971; Snyder & Diesing, 1977) and generally avoid escalatory spirals due to the tendency, heightened under psychological stress, to overestimate the enemy’s hostile intentions (Jervis, 1970; Holsti, 1972). In his influential study of cooperation, Axelrod’s (1984) computer simulations revealed that tit-for-tat reciprocity was the most effective strategy for inducing cooperation in the Prisoner’s Dilemma game. In their game-theoretic analysis, Zagare and Kilgour (2000) show that reciprocity may lead to deterrence success by establishing credibility, while preventing crises from spiraling into all-out war. Triesman’s (2004) formal stylization of the appeasement strategy suggests that even appeasing the adversary can be rational in the long run by conserving limited resources for future challenges.

The varieties of bargaining strategies were also subjected to important quantitative empirical tests. Leng (1993), for example, finds that reciprocity is most effective when the actors use a combination of threats of violence (if their opponent did the same) and conditional accommodations. The accommodative (firm-but-flexible) strategy signals that a leader is willing to reach compromise if the opponent cooperates, whereas the threats show his willingness to fight if the opponent remains obstinate. Besides Leng’s significant contributions in this area (Leng & Wheeler, 1979; Leng, 1983, 1993), Huth (1988) and Gelpi (1997) also looked at strategies of positive, negative, and mixed signals. Despite their different data sets, they all found that the “firm-but-flexible response” had a greater impact on deterrence success than “bullying” (escalatory) strategies, generally associated with Schelling’s idea of “competitions in risk-taking” (Leng, 1993). In short, cooperative moves, as long as they are mixed with firmness of response to hostile actions, do not necessarily undercut reputations for firm resolve while, at same time, avoiding unintentional escalations.

So far, we have presented a variety of approaches to commitments and signals intended to overcome the problems of bargaining failures due to informational uncertainties. Many significant contributions of the fourth-wave research have thus been in terms of how states can credibly signal their intentions, by producing ex ante sunk costs or ex post reputational costs (domestic or international, either for resolve or honesty), in order to prevail in an environment of private information.

Yet several recent studies have shown that war can also result under conditions of complete information. In this respect, the most influential argument considers war as a commitment problem (Fearon, 1995). In this context, it does not refer to a state’s commitment to carry out its threat but its willingness to uphold an agreement that has been reached. Powell (2006), for example, shows that wars result from large and rapid power shifts that make it difficult to find a settlement that all sides would prefer to war and thus credibly commit to. Specifically, any offer made by the rising state must reflect its future levels of power if it is going to make a credible commitment to uphold it. The greater the rising state’s shift in power, the less it will offer to its opponent. If this is less than the declining state could expect to gain by fighting in the current time period (i.e., when it is still somewhat stronger), then war is likely to occur. Hence, in contrast to the studies relating particular power distributions to war under uncertainty, Powell’s “commitment-based” bargaining approach predicts that even under complete information, power shifts can lead to war.

Data and Methodological Issues

Since this encyclopedia offers separate essays on data and methodological issues in the scientific study of conflict, we limit our discussion only to a few issues specific to the deterrence and crisis bargaining literature.

Notwithstanding a few methodological controversies to be presented in the next section, we should note that the research in this area has witnessed an encouraging trend in further refinements and diversification of methods. The case studies underwent a more rigorous overhaul in qualitative research design, whether as comparative (Crawford, 2003; Nincic, 2010) or single-case studies (Zagare, 2011). Extensive data sets on deterrence and compellence further facilitated the rigor in quantitative research (Huth, 1988; Danilovic, 2002; Sechser, 2011), while there has also been a rise in carefully designed experiments (e.g., Tingley, 2011; Tingley & Walter, 2011). Although the formal-theoretic work continues to take an important place in advancing our knowledge of the bargaining process, these diverse and refined strands in empirical approaches helped further both the development and testing of novel theories.

Methodological Controversies

One of the issues is ingrained in the definition of deterrence and it concerns how we identify deterrence success. Given that deterrence is based on the deterrer’s beliefs that the attacker intended to act prior to the threat, the observed absence of attack does not necessarily mean that deterrence worked. Rather, it worked only if the deterrer’s belief about the attacker’s aggressive intentions were correct, which is indeed an unobservable phenomenon (Danilovic, 2001). This general issue does not only pose a difficulty in identifying empirical cases where deterrence succeeded, but it also underscores two main debates regarding the possibilities for, and potential problems with, case identification and variable measurements.

The issue of case identification dates to the debate between Lebow and Stein (1990) and Huth and Russett (1990) about the difficulty of operationalizing deterrence due to the “elusive” aspect of intentions behind it. In their critique of Huth and Russett’s data set, Lebow and Stein questioned their method of inferring intentions from behavior and set out to reexamine the entire data set by historically documenting whether such intentions existed at all. Unsurprisingly, the resulting revised list of cases was quite deflated in comparison to the original data set. The problem with empirically establishing intentions from leaders’ statements is that these statements may not reflect their genuine intentions. This illustrates only one among many obstacles to empirically establishing the actual intentions of the involved parties. If deterrence theories are to be validated by empirical tests that lead to generalizable findings (large-N, that is), then the only viable route, though not an ideal one, seems to be through behavioral indicators as done by Huth and Russett (1990). In addition, specifically to general deterrence cases, Huth and Russett (1993) propose to identify them in terms of enduring rivalries. We can then assume ongoing adverse intentions between rivals, thus rendering general deterrence to last as long as the enduring rivalry does, which might potentially solve the selection problem in identifying such elusive cases.

This highlights the second methodological controversy, which concerns the problem of selection bias that results from looking only at the cases when general deterrence failed. This has marred a number of case studies that select on the dependent variable (Achen & Snidal, 1989). The problem is accentuated by the fact that challengers act strategically when deciding whether to issue threats. Known as the “selection effects” argument and highlighted by several fourth-wave theorists (Morrow, 1989; Fearon, 1994a), it proved useful in explaining previously unexpected or inconsistent results. For example, Fearon (1994a) suggests that, if an observably stronger state is challenged (i.e., general deterrence fails against a weaker opponent), then there is little reason to anticipate that the challenger would be deterred at the immediate deterrence stage, because of its prior unobservable strong resolve. In this case, then, the balance of forces does increase the likelihood of deterrence success, but only by “selecting out” irresolute challengers in general deterrence. Yet, if we do not observe cases of general deterrence failure, then we might erroneously infer that there is no relationship between the balance of forces and deterrence outcomes. Overall, by identifying the selection effect mechanisms, the fourth-wave scholars theoretically advanced our understanding of the dynamics of deterrence by identifying factors such as unobservable resolve or intangible aspects of capabilities. This helps to explain why deterrence might fail even if a deterrer is observably stronger, but without having to resort to psychological factors such as misperception or motivation biases.

Data Collections

Regarding data collections of deterrence cases, Russett (1963) was the first to point to the need for and possibility of quantitative studies in this area, despite the elusiveness of some aspects of deterrence as discussed above. Indeed, his was the first attempt at quantifying and systematically analyzing deterrence. It was followed by the first comprehensive effort to collect and systematize the cases of deterrence for a 100-year period (1885–1984) that resulted in the path-breaking quantitative deterrence research by Huth and Russett (1984, 1990; Huth, 1988; Petersen, 1986). For the first time, deterrence theories were tested in a systematic way. To our knowledge, there is only one other deterrence data set that is comprehensive in scope (90-year period from 1895 to 1985) and, in this instance, specific to major power conflicts (Danilovic, 2001, 2002).

Although both data sets share common operational elements for deterrence, there are some differences in operationalizations. While Huth and Russett (1984, 1990) identify deterrence outcomes in standard terms of failure (war) and success (peace), Danilovic (2001, 2002) differentiates between four outcomes: war and the defender’s peaceful acquiescence as two distinct forms of deterrence failure (the latter being obscured in a simple war-no war dichotomy of deterrence outcomes), the challenger’s acquiescence (though, for the reason discussed above, not identified as necessarily deterrence success), and compromise. These attempts at generating deterrence data for quantitative analysis were significantly augmented in the recently launched data set on compellent threats for the 1918–2001 period (Sechser, 2011). Partly due to the absence of data on compellence, this bargaining strategy has been overshadowed by research on the alternative, deterrent, strategy in crisis bargaining. This new data set will facilitate a better understanding whether and how compellence works in influencing behavior in security contests.

Despite these steps toward providing the grounds for quantitative deterrence research, the number of cases in both data sets is still far from large in comparison to traditional data sets on “militarized interstate disputes” (MID) and “international crisis behavior” (ICB) (see “Links to Digital Materials”). While some deterrence arguments might certainly be tested with either of these data sets, in their current versions they do not allow for an operationalization of some critical steps in deterrence encounters, such as the actual sequence of threats and counterthreats, or outcomes that are directly related to deterrence models. Nevertheless, the ICB data, with the detailed synopses of international crises (Brecher & Wilkenfeld, 1997), provide a valuable source for analyzing different stages of international crises: their onset, escalation, de-escalation, and outcomes. They cover the 1918–2006 period. Theirs and the MID data, covering the 1815–2001 period, offer the most comprehensive and widely used data on both contextual factors and various hostility-level aspects of militarized crises.

For a systematic study of behavioral aspects such as crisis-bargaining moves, one might want to turn to the Behavioral Correlates of War (BCOW) data set. It is designed to supplement the Correlates of War (COW) data on wars and MIDs by “obtaining more microscopic descriptions of conflict bargaining” (Leng & Singer, 1988, p. 156). It includes data on initiations, responses, and bargaining (“influence”) strategies in over 40 randomly sampled cases of international crises since 1816. Gelpi’s (1997) augmentation of the ICB data, coding bargaining strategies for 122 international crises between 1929 and 1979, now renders this data set suitable for testing arguments about bargaining behavior as well.

Future Research Directions

In the end, we would point to several recent advancements as well as areas for further improvements that, in our view, have the potential for the future accumulation of knowledge. They are not listed in any particular order of significance, as we find them all equally relevant either as novelties or as longstanding issues that have been neglected or incompletely addressed.

First, the conventional treatment of deterrence, common to all four waves, treats war as the end of the bargaining process. It has been suggested, however, that this costly-lottery premise about war can restrict our understanding of why bargaining fails and war begins. If war is seen as a continuation of the bargaining process, rather than as an end to it, then understanding why states reach settlements to end wars is directly related to the question of why they must fight in the first place (Wagner, 2000). Motivated by Blainey’s (1988) view of war as a disagreement over relative bargaining power, Wagner argues that uncertainty about power capabilities leads to bargaining breakdown. This uncertainty is reduced through intrawar bargaining until both sides’ beliefs about their relative strength converge. Filson and Werner (2002) similarly view bargaining and war as a problem of asymmetric information about capabilities and provide additional nuances to the argument about war as a continuation of the bargaining process by allowing for beliefs to be updated through battles and negotiations as the war progresses. Reiter (2009) departs from these studies by examining both informational asymmetries and commitment problems and assesses how each type of problem influences the way a war ends. He shows that wars resulting from commitment problems are much more difficult to resolve and, as a result, are more likely to end only after the complete defeat of the opponent. Wolford, Reiter, and Carrubba (2011) also include both asymmetric information and commitment problems in their formal model. When both types of bargaining problems are present, their model leads to the interesting corollary that resolving informational asymmetries through fighting can actually make a negotiated settlement to the war less, rather than more, likely.

These are only some examples of the bourgeoning literature that challenges the standard treatment of deterrence and bargaining as a strictly prewar process. While important in many respects, these studies primarily deal with pre- and intrawar bargaining as a problem of private information about power. In this respect, their results can greatly contribute to the deterrence research that examined the impact of power uncertainties on crisis bargaining (Morrow, 1989; Bueno de Mesquita et al., 1997; Wittman, 2001; Reed, 2003), but do not bear direct implications for arguments that attribute bargaining breakdown to informational uncertainties about resolve. At the same time, if modeled in terms of informational asymmetries about resolve, studies about crisis and war as interrelated stages of the same bargaining process might provide interesting solutions to some previously unresolved puzzles in this area as well.

Second, some recent formal stylizations of war under complete information have attempted to show that uncertainty is not necessarily crucial for understanding bargaining failures, the premise pervading most of the research on commitments and signals (Slantchev, 2003; Powell, 2006). In Slantchev’s (2003) analysis, complete information about each side’s abilities to inflict costs as well as to bear them sets the boundaries within which bargaining takes place both prior to and during the war. Powell’s (2006) model also hinges on complete information but, unlike Slantchev, war is not treated as a continuation of the bargaining process. Beside these rare but important formal-theoretic works on the effect of complete information, empirical studies based on inherent resolve (e.g., Russett, 1963; Huth & Rusett, 1984; Danilovic, 2002) showed that the nonmanipulative aspects of a state’s willingness to fight, such as their intrinsic interests in the dispute, set firm parameters within which bargaining strategies can be effective. And since the notion of inherent resolve by definition is based on the “as if” premise of complete information about resolve (Snyder & Diesing, 1977), these empirical findings additionally validate the need to reintroduce perfect information about resolve or capabilities to fully understand the potentials for and limits to the bargaining process.

Third, and related to the previous point, rather than compartmentalizing different requirements for effective deterrence, there is a great need for developing an integrated theoretical framework. As this survey has illustrated, the last several decades have led to many improvements in understanding the individual factors that are exogenous or endogenous to the bargaining process, but have increasingly focused on them in isolation from other factors. For example, the signaling and commitment research produced a number of useful concepts and novel ideas about the mechanisms that may make deterrence effective. Yet the ultimately indeterminate, or rather less robust, predictions regarding the precise conditions for war as a result of bargaining failure leave much to be desired. Rather than relegating this issue to war in the “error term” (see Gartzke, 1999), though this argument certainly has important merits, there can be a viable “second chance” approach: reintegrating different aspects of deterrence requirements into a single framework that specifies how their interplay produces crisis outcomes.

To illustrate, Clare’s study (2007) recently showed that, unlike generally expected by audience cost approaches, democracies do indeed renege on their threats even at the risk of being caught in a bluff. At the same time, it specified the conditions under which it is rational for their leaders to do so without necessarily rejecting the basic premises of audience cost theory, but rather by modifying it to include the interplay between ex post audience costs and ex ante inherent resolve. Hence, rather than juxtaposing the alternative notions of credibility, the study’s findings warrant exploring these factors synergistically, as they condition each other, within a single framework. The possibilities for other theoretically integrative attempts are immense and, as we see it, the value of their contributions should be judged on two essential grounds: (a) the extent to which they reduce the indeterminacy of previous predictions, and (b) the extent to which the theoretical integration is synthetic rather than simply eclectic.

Fourth, past research has mainly focused on the use of public diplomacy and costly signals as the key to conveying and manipulating threat credibility. Since private threats do not generate public audience costs, they have been seen as a form of ineffective cheap talk and largely ignored. Still, some recent formal models (Kurizaki, 2007; Leventoglu & Tarar, 2005; Trager, 2010; Ramsay, 2011) and experimental studies (Tingley & Walter, 2011) have focused on the use of private diplomacy and found that, under certain conditions, it can be as efficient as the public form. Sartori (2005) shows that, if states are concerned with maintaining a reputation for honesty, then even private diplomacy should convey information to an opponent. Kurizaki (2007) shows that, if a challenger’s public threats also generate audience costs for a defender, then the very nature of these public threats can make it more difficult for the defender to make concessions, and thus reduce the possibility of reaching a peaceful outcome. Alternatively, a private threat can allow a defender to make concessions without facing domestic ramifications. Similarly, Leventoglu and Tarar (2005) show that both sides in a bargaining situation would be better off if they did not issue their demands publicly; doing so makes it difficult for either side to make the concessions necessary for an agreement. Note that in both studies, the results require both sides to face domestic audiences, leading to the interesting implication that private diplomacy may be more efficient for reaching negotiated settlements when both states are democratic.

Fifth, the credibility problem, an already precarious issue, is further magnified in cases of extended as opposed to direct deterrence. For this reason, most studies in each research wave concentrated mostly on the deterrer’s and attacker’s beliefs and behavior in this form of deterrence. Yet the dynamics of extended deterrence are a function of the behavior of all three sides, and the role of the protégé has been neglected until fairly recently. The protégé’s behavior has been analyzed especially as a function of alliance commitments such as the protégé’s beliefs about its defender’s reliability (Smith, 1995; Zagare & Kilgour, 2003; Quackenbush, 2006). Their findings, either formal or empirical, demonstrate that bargaining outcomes are indeed affected by the protégé’s behavior and its beliefs about the defender’s reliability to protect it. Another recently raised issue concerns the impact of a protégé’s exit options, that is, alternative alliance choices to the defender, on the defender’s strategies in extended deterrence and their consequences for restraint or escalation (Crawford, 2003; Zagare, 2011). These are certainly encouraging first cuts into a promising line of research that should improve our understanding of extended deterrence. In our view, besides the cited models, the principal-agent theory could also be well suited for formally stylizing the defender–protégé relations and, as such, one of the promising research strategies in this area.

Sixth, if we are to judge according to the literature output, one of the understudied areas concerns issue-linkage options in the bargaining process. The effectiveness of issue-linkage strategies for reaching settlements within the bargaining range and avoiding inefficient wars is potentially high, as evidenced in diplomatic memoirs, biographies, and detailed narratives of successful diplomatic endings to acute crises. From a theoretical standpoint, it can also provide alternative solutions to the “issue indivisibility” problem as an obstacle to successful bargaining. Yet issue-linkage has been examined somewhat sporadically (e.g., Stein, 1980; Morgan, 1990; Morrow, 1992). Though disagreeing on certain points, such as whether the issue-linkage offer might be interpreted as a signal of weak resolve—each of these studies offers novel analytical tools. Recently, Trager (2011) has shown how formal-theoretic refinements of issue-linkage diplomacy can lead to novel expectations, whereas Poast (2013) offers methodological and measurement solutions to the challenges in quantitative testing. These should facilitate future refinements to better understand an apparently common practice in crisis diplomacy that was somewhat neglected in the scholarship.

Seventh, there is an almost urgent need and great potential for theoretical and empirical extensions of bargaining studies to nonstate (intrastate or transnational) conflicts. We disagree with those who believe that nonstate forms of conflicts are outside the scope of traditional perspectives such as bargaining theory. In fact, only a few deterrence studies are specific to the idiosyncrasies of interstate relations. It is therefore plausible to expect general models of bargaining and deterrence to be quite useful for understanding the dynamics of transnational or “postmodern warfare.”

In this respect, several recent attempts stand out in extending models of international bargaining to civil conflicts as well as terrorism. For example, Powell (2006) shows that conflicts between domestic factions can result from the same commitment problems that lead to international conflicts. Thyne (2012) argues that the difficulty with ending civil wars is related to the inability to credibly commit to negotiated settlements, and his empirical analysis shows that governments with powerful and stable executives are better suited to overcome this problem. Walter (2009) examines civil conflicts as resulting from informational problems and shows that reputations for resolve are important when governments face multiple ethnic separatist groups. This finding is consistent with the literature on reputations in international bargaining. Cunningham (2011) argues that governments will take a different bargaining approach depending on the internal cohesion of the separatist group. She finds that governments are more likely to offer concessions to internally divided groups in order to simultaneously (1) strengthen moderate factions, who could be the key to a lasting peace settlement, and (2) overcome informational asymmetries by exposing the more violent factions of the group. Findley (2013) takes the novel approach of examining the interdependent nature of bargaining at different stages of a civil war: war, negotiation, agreement, and implementation. He shows that two factors in particular—stalemates in fighting and the number of actors—can have a different effect on bargaining outcomes at the different stages of a civil war.

Bargaining and deterrence arguments have even made their way into studies of terrorism, rejecting the conventional belief that terrorists cannot be considered rational and/or are not susceptible to the same forms of coercion as states. Among formal-theoretic studies, Bapat (2006), for example, shows that the problem of negotiating with terrorists results from a credible commitment problem; that is, from the inability of terrorist organizations to credibly commit to upholding their end of an agreement. If, however, terrorists are dependent on a host (sponsor) state, then outside pressure and threats by other states can effectively force the host to ensure that the terrorist organization does not renege on its agreements. Alternatively, Bueno de Mesquita (2005) models this problem as resulting from the inability of a government to commit to its agreements. The solution to this problem is for moderates within the terrorist organization to show their willingness and ability to suppress the hard-line factions among their ranks only if the negotiating government upholds its agreements. Browne and Dickson (2010) examine the role of audience costs in secret negotiations with terrorist groups. Somewhat paradoxically, they show that public statements of resolve reduce the government’s bargaining power in negotiations but, at the same time, increase the likelihood of successful negotiations by bringing an otherwise obstinate adversary to the negotiating table.

Eighth, a final advancement that we would suggest for future research relates to a better understanding of the full array of bargaining strategies, including those offering positive inducements and not just those threatening with punitive sanctions. Since threats of punishments constitute the definitional core of deterrence as conceptualized since the second-wave research, it has been almost paradigmatic to analyze only those bargaining strategies that use threats of negative sanctions. Notwithstanding Theodore Roosevelt’s proverbial dictum, “Speak softly and carry a big stick,” the third-wave studies provide much anecdotal evidence of successful negotiations when leaders combine threats of punishments with promises of rewards. And, as Baldwin noted, “Although positive and negative sanctions are opposite in logic, they are not opposites in their behavioral consequences” (1971, p. 28).

As shown in this survey, there are only a handful of formal-theoretic and empirical studies on reciprocity strategies or those that mix threats of negative sanctions with positive inducements (e.g., Leng, 1983, 1993; Huth, 1988; Zagare & Kilgour, 2000). Nincic’s (2010) study stands out among the rare recent attempts to advance the theory of positive inducements, showing that even in the most intractable cases of bargaining with so-called “renegade” regimes, coercive measures can undercut the prospect of changes induced by positive incentives. He also finds that domestic political conditions in the target regimes are additionally critical for the effectiveness of positive and negative pressures toward their policy changes.

All these studies demonstrate that a general neglect of positive inducements in bargaining research is unjustified. There are many paths for theoretical progress in this area and, in our view, there can be great benefits from expanding the notion of promised rewards, both conceptually by pointing to their richer typology and theoretically by specifying the conditions under which these would not undercut the credibility of threats. For the latter, a research strategy is warranted that incorporates both types of signals (promises and rewards) into a single model rather than treating them as separate research programs, as exemplified in the current parallel coexistence of bargaining approaches to coercion and cooperation.

Behavioral Correlates of War. The BCOW application and data by Russell J. Leng are maintained by Inter-University Consortium for Political and Social Research (ICPSR).

Correlates of War Project: Militarized Interstate Disputes. Various data sets collected for the COW project are available for public use. The latest version (4.1) of MIDs covers the 1816–2010 period. Besides the MID data and codebook, it also provides a separate data set on the geographic locations of disputes.

Dyadic MID Dataset. Zeev Maoz’s revised version of dyadic MID data for the 1816–2001 period (version 2.0). Along with the data set and codebook, the author provides the rationale for, and detailed comments on, modifications he has made on the MID data set.

International Crisis Behavior Project. The latest release of the ICB data (version 11), originally developed by Michael Brecher and Jonathan Wilkenfeld, covers 470 international crises for the 1918–2013 period. It can be downloaded from Duke University’s Department of Political Science website. Among unique features, the website also provides online access to crisis summaries through the ICB Data Viewer.

EUGene Software Program. EUGene stands for Expected Utility Generation and Data Management, a software tool developed by D. Scott Bennett and Allan C. Stam. Primarily intended for political scientists and freely available to all, it allows a user to create, merge, and manage data from the most frequently used data sets in the field. It integrates a number of variables including those from the MID, Zeev Maoz’s Dyadic MID, and ICB data, and facilitates merging them with other data sets as well.

References

Achen, C. H., & Snidal, D. (1989). Rational deterrence theory and comparative case studies. World Politics, 41(2), 143–169.Find this resource:

Axelrod, R. (1984). The evolution of cooperation. New York: Basic Books.Find this resource:

Baldwin, D. A. (1971). The power of positive sanctions. World Politics, 24(1), 19–38.Find this resource:

Bapat, N. (2006). State bargaining with transnational terrorist groups. International Studies Quarterly, 50(1), 213–230.Find this resource:

Blainey, G. (1988). The causes of war (3d ed.). New York: Free Press.Find this resource:

Brecher, M., & Wilkenfeld, J. (1997). A study of crisis. Ann Arbor: University of Michigan Press.Find this resource:

Brodie, B. (1946). The absolute weapon. New York: Harcourt, Brace.Find this resource:

Browne, J., & Dickson, E. S. (2010). “We don’t talk to terrorists”: On the rhetoric and practice of secret negotiations. Journal of Conflict Resolution, 54(3), 379–407.Find this resource:

Bueno de Mesquita, B., & Lalman, D. (1992). War and reason: Domestic and international imperatives. New Haven, CT: Yale University Press.Find this resource:

Bueno de Mesquita, B., Morrow, J. D., & Zorick, E. (1997). Capabilities, perception, and escalation. American Political Science Review, 91(1), 15–27.Find this resource:

Bueno de Mesquita, B., & Riker, W. H. (1982). An assessment of the merits of selective nuclear proliferation. Journal of Conflict Resolution, 26(2), 283–306.Find this resource:

Bueno de Mesquita, E. (2005). Conciliation, counterterrorism, and patterns of terrorist violence. International Organization, 59(1), 145–176.Find this resource:

Carlson, L. J., & Dacey, R. (2014). The use of fear and anger to alter crisis initiation. Conflict Management and Peace Science, 31(2), 168–192.Find this resource:

Clare, J. (2007). Domestic audiences and strategic interests. Journal of Politics, 69(3), 732–745.Find this resource:

Clare, J., & Danilovic, V. (2010). Multiple audiences and reputation building in international conflicts. Journal of Conflict Resolution December, 54(6), 860–882.Find this resource:

Clare, J., & Danilovic, V. (2012). Reputation for resolve, interests, and conflict. Conflict Management and Peace Science, 29(1), 3–27.Find this resource:

Crawford, T. W. (2003). Pivotal deterrence: Third-party statecraft and the pursuit of peace. Ithaca, NY: Cornell University Press.Find this resource:

Cunningham, K. G. (2011). Divide and conquer or divide and concede: How do states respond to internally divided separatists? American Political Science Review, 105(2), 275–297.Find this resource:

Danilovic, V. (2001). Conceptual and selection bias issues in deterrence. Journal of Conflict Resolution, 45(1), 97–125.Find this resource:

Danilovic, V. (2002). When the stakes are high: Deterrence and conflict among major powers. Ann Arbor: University of Michigan Press.Find this resource:

Diehl, P. F. (1992). What are they fighting for? The importance of issues in international conflict research. Journal of Peace Research, 29(3), 333–344.Find this resource:

Downs, G. W. (1989). The rational deterrence debate: A symposium. World Politics, 41(2), 143–237.Find this resource:

Ellsberg, D. (1968). The theory and practice of blackmail. RAND transcript P-3883 of 1959 original.Find this resource:

Fearon, J. D. (1994a). Signaling versus the balance of power and interests: An empirical test of a crisis bargaining model. Journal of Conflict Resolution, 38(2), 236–269.Find this resource:

Fearon, J. D. (1994b). Domestic political audiences and the escalation of international disputes. American Political Science Review, 88(3), 577–592.Find this resource:

Fearon, J. D. (1995). Rationalist explanations for war. International Organization, 49(3), 379–414.Find this resource:

Fearon, J. D. (1997). Signaling foreign policy interests: Tying hands versus sinking costs. Journal of Conflict Resolution, 41(1), 68–90.Find this resource:

Filson, D., & Werner, S. (2002). A bargaining model of war and peace: Anticipating the onset, duration, and outcome of war. American Journal of Political Science, 46(4), 819–837.Find this resource:

Findley, M. G. (2013). Bargaining and the interdependent stages of civil war resolution. Journal of Conflict Resolution, 57(5), 901–928.Find this resource:

Freedman, L. (1989). The evolution of nuclear strategy (2d ed.). New York: St. Martin’s Press.Find this resource:

Gartzke, E. (1999). War is in the error term. International Organization, 53(3), 567–587.Find this resource:

Geller, D. S. (1990). Nuclear weapons, deterrence, and crisis escalation. Journal for Conflict Resolution, 34(2), 291–310.Find this resource:

Gelpi, C. (1997). Crime and punishment: The role of norms in crisis bargaining. American Political Science Review, 91(2), 339–360.Find this resource:

Gelpi, C., & Griesdorf, M. (2001). Winners or losers? Democracies in international crisis, 1918–94. American Political Science Review, 95(3), 633–647.Find this resource:

George, A. L., & Smoke, R. (1974). The deterrence in American foreign policy: Theory and practice. New York: Columbia University Press.Find this resource:

George, A. L., Hall, D. K., & Simons, W. E. (1971). The limits of coercive diplomacy; Laos, Cuba, Vietnam. Boston: Little, Brown.Find this resource:

Guisinger, A., & Smith, A. (2002). Honest threats: The interaction of reputation and political institutions in international crises. Journal of Conflict Resolution, 46(2), 175–200.Find this resource:

Holsti, O. R. (1972). Crisis, escalation, war. Montreal: McGill-Queen’s University Press.Find this resource:

Huth, P. (1988). Extended deterrence and the prevention of war. New Haven, CT: Yale University Press.Find this resource:

Huth, P. (1990). The extended deterrent value of nuclear weapons. Journal of Conflict Resolution, 34(2), 270–290.Find this resource:

Huth, P. (1997). Reputations and deterrence: A theoretical and empirical assessment. Security Studies, 7(1), 72–99.Find this resource:

Huth, P., & Russett, B. (1984). What makes deterrence work? Cases from 1900–1980. World Politics, 36(4), 496–526.Find this resource:

Huth, P., & Russett, B. (1990). Testing deterrence theory: Rigor makes a difference. World Politics, 42(4), 466–501.Find this resource:

Huth, P., & Russett, B. (1993). General deterrence between enduring rivals. American Political Science Review, 87(1), 61–73.Find this resource:

Jervis, R. (1970). The logic of images in international relations. Princeton, NJ: Princeton University Press.Find this resource:

Jervis, R. (1976). Perception and misperception in international politics. Princeton, NJ: Princeton University Press.Find this resource:

Jervis, R. (1979). Deterrence theory revisited. World Politics, 31(2), 289–324.Find this resource:

Jervis, R., Lebow, R. N., & Stein, J. G. (1985). Psychology and deterrence. Baltimore: Johns Hopkins University Press.Find this resource:

Johnson, D. P., & Tierney, D. (2011). The Rubicon theory of war: How the path to conflict reaches the point of no return. International Security, 36(1), 7–40.Find this resource:

Kaufmann, W. (1956). The requirements of deterrence. In W. Kaufmann (Ed.), Military policy and national security (pp. 12–38). Princeton, NJ: Princeton University Press.Find this resource:

Kroenig, M. (2013). Nuclear superiority and the balance of resolve: Explaining nuclear crisis outcomes. International Organization, 67(1), 141–171.Find this resource:

Kurizaki, S. (2007). Efficient secrecy: Public versus private threats in crisis diplomacy. American Political Science Review, 101(3), 543–558.Find this resource:

Kydd, A. (1997). Game theory and the spiral model. World Politics, 49(3), 371–400.Find this resource:

Lake, D. A. (2010/2011). Two cheers for bargaining theory: Assessing rationalist explanations for the Iraq War. International Security, 35(3), 7–52.Find this resource:

Lebow, R. N. (1981). Between peace and war: The nature of international crisis. Baltimore: Johns Hopkins University Press.Find this resource:

Lebow, R. N., & Stein, J. G. (1990). Deterrence: The elusive dependent variable. World Politics, 42(3), 336–369.Find this resource:

Leng, R. J. (1983). When will they ever learn? Coercive bargaining in recurrent crises. Journal of Conflict Resolution, 27(3), 379–419.Find this resource:

Leng, R. J. (1993). Interstate crisis behavior, 1816–1980: Realism versus reciprocity. New York: Cambridge University Press.Find this resource:

Leng, R. J., & Singer, J. D. (1988). Militarized interstate crises: The BCOW typology and its applications. International Studies Quarterly, 32(2), 115–174.Find this resource:

Leng, R. J., & Wheeler, H. G. (1979). Influence strategies, success, and war. Journal for Conflict Resolution, 23(4), 655–684.Find this resource:

Leventoglu, B., & Tarar, A. (2005). Prenegotiation public commitment in domestic and international bargaining. American Political Science Review, 99(3), 419–433.Find this resource:

Lupovici, A. (2010). The emerging fourth wave of deterrence theory: Toward a new research agenda. International Studies Quarterly, 54(3), 705–732.Find this resource:

Maxwell, S. (1968). Rationality in deterrence. Adelphi Paper 50. London: Institute for Strategic Studies.Find this resource:

Mercer, J. (1996). Reputation and international politics. Ithaca, NY: Cornell University Press.Find this resource:

Morgan, P. M. (1983). Deterrence: A conceptual analysis (2d ed.). Beverly Hills, CA: SAGE.Find this resource:

Morgan, T. C. (1990). Issue linkages in international crisis bargaining. American Journal of Political Science, 34(2), 311–333.Find this resource:

Morgenthau, H. J. (1964). The four paradoxes of nuclear strategy. American Political Science Review, 58(1), 23–35.Find this resource:

Morrow, J. D. (1989). Capabilities, uncertainty, and resolve: A limited information model of crisis bargaining. American Journal of Political Science, 33(4), 941–972.Find this resource:

Morrow, J. D. (1992). Signaling difficulties with linkage in crisis bargaining. International Studies Quarterly, 36(2), 153–172.Find this resource:

Morrow, J. D. (1999). The strategic setting of choices: Signaling, commitment, and negotiation in international politics. In D. A. Lake & R. Powell (Ed.), Strategic choice and international relations (pp. 77–114). Princeton, NJ: Princeton University Press.Find this resource:

Narang, V. (2013). What does it take to deter? Regional power nuclear postures and international conflict. Journal of Conflict Resolution, 57(3), 478–508.Find this resource:

Nincic, M. (2010). Getting what you want positive inducements in international relations. International Security, 35(1), 138–183.Find this resource:

Organski, A. F. K., & Kugler, J. (1980). The war ledger. Chicago: University of Chicago Press.Find this resource:

Orme, J. (1987). Deterrence failures: A second look. International Security, 11(4), 96–124.Find this resource:

Partell, P. J., & Palmer, G. (1999). Audience costs and interstate crises. International Studies Quarterly, 43(2), 389–405.Find this resource:

Petersen, W. J. (1986). Deterrence and compellence. International Studies Quarterly, 30(2), 269–294.Find this resource:

Poast, P. (2013). Issue linkage and international cooperation: An empirical investigation. Conflict Management and Peace Science, 30(3), 286–303.Find this resource:

Powell, R. (1990). Nuclear deterrence theory: The search for credibility. New York: Cambridge University Press.Find this resource:

Powell, R. (1999). In the shadow of power. Princeton, NJ: Princeton University Press.Find this resource:

Powell, R. (2006). War as a commitment problem. International Organization, 60(1), 169–203.Find this resource:

Press, D. G. (2004–2005). The credibility of power: Assessing threats during the appeasement crises of the 1930s. International Security, 29(3), 136–169.Find this resource:

Quackenbush, S. L. (2006). Not only whether but whom: Three-party extended deterrence. Journal of Conflict Resolution, 50(4), 562–583.Find this resource:

Ramsay, K. W. (2011). Cheap talk diplomacy, voluntary negotiations, and variable bargaining power. International Studies Quarterly, 55(4), 1003–1023.Find this resource:

Reed, W. (2003). Information, power, and war. American Political Science Review, 97(4), 633–641.Find this resource:

Reiter, D. (2009). How wars end. Princeton, NJ: Princeton University Press.Find this resource:

Russett, B. (1963). The calculus of deterrence. Journal of Conflict Resolution, 7(2), 92–109.Find this resource:

Sartori, A. E. (2005). Deterrence by diplomacy. Princeton, NJ: Princeton University Press.Find this resource:

Schelling, T. C. (1956). An essay on bargaining. American Economic Review, 46(3), 281–306.Find this resource:

Schelling, T. C. (1960). The strategy of conflict. Cambridge, MA: Harvard University Press.Find this resource:

Schelling, T. C. (1966). Arms and influence. New Haven, CT: Yale University Press.Find this resource:

Schultz, K. A. (2001). Democracy and coercive diplomacy. Cambridge, U.K.: Cambridge University Press.Find this resource:

Sechser, T. A. (2010). Goliath’s curse: Coercive threats and asymmetric power. International Organization, 64(4), 627–660.Find this resource:

Sechser, T. A. (2011). Militarized compellent threats, 1918–2001. Conflict Management and Peace Science, 28(4), 377–401.Find this resource:

Slantchev, B. (2003). The power to hurt: Costly conflict with completely informed states. American Political Science Review, 97(1), 123–133.Find this resource:

Slantchev, B. (2005). Military coercion in interstate crises. American Political Science Review, 99(4), 533–548.Find this resource:

Slantchev, B. (2011). Military threats: The costs of coercion and the price of peace. New York: Cambridge University Press.Find this resource:

Smith, A. (1995). Alliance formation and war. International Studies Quarterly, 39(4), 405–435.Find this resource:

Smith, A. (1998). International crises and domestic politics. American Political Science Review, 92(3), 623–638.Find this resource:

Snyder, G. H. (1961). Deterrence and defense. Princeton, NJ: Princeton University Press.Find this resource:

Snyder, G. H., & Diesing, P. (1977). Conflict among nations: Bargaining, decision making, and system structure in international crises. Princeton, NJ: Princeton University Press.Find this resource:

Snyder, J., & Borghard, E. D. (2011). The cost of empty threats: A penny, not a pound. American Political Science Review, 105(3), 437–456.Find this resource:

Stein, A. A. (1980). The politics of linkage. World Politics, 33(1), 62–81.Find this resource:

Thyne, C. L. (2012). Information, commitment, and intra-war bargaining: The effect of governmental constraints on civil war duration. International Studies Quarterly, 56(2), 307–321.Find this resource:

Tingley, D. H. (2011). The dark side of the future: An experimental test of commitment problems in bargaining. International Studies Quarterly, 55(2), 521–544.Find this resource:

Tingley, D. H., & Walter, B. F. (2011). Can cheap talk deter? An experimental analysis. Journal of Conflict Resolution, 55(6), 994–1018.Find this resource:

Trachtenberg, M. (2012). Audience costs: An historical analysis. Security Studies, 21(3), 3–42.Find this resource:

Trager, R. F. (2010). Diplomatic calculus in anarchy: How communication matters. American Political Science Review, 104(2), 347–368.Find this resource:

Trager, R. F. (2011). Multidimensional diplomacy. International Organization, 65(3), 469–506.Find this resource:

Triesman, D. (2004). Rational appeasement. International Organization, 58(2), 345–373.Find this resource:

Wagner, R. H. (1991). Nuclear deterrence, counterforce strategies, and the incentives to strike first. American Political Science Review, 85(3), 727–749.Find this resource:

Wagner, R. H. (1992). Rationality and misperception in deterrence theory. Journal of Theoretical Politics, 4(2), 115–141.Find this resource:

Wagner, R. H. (2000). Bargaining and war. American Journal of Political Science, 44(3), 469–484.Find this resource:

Walter, B. (2009). Reputation and civil war: Why separatist conflicts are so violent. New York: Cambridge University Press.Find this resource:

Weinstein, F. B. (1969). The concept of a commitment in international relations. Journal of Conflict Resolution, 13(1), 39–56.Find this resource:

Weisiger, A., & Yarhi-Milo, K. (2015). Revisiting reputation: How past actions matter in international politics. International Organization, 69(2), 473–495.Find this resource:

Wittman, D. (2001). War or peace? Paper presented at the Political Economy of Conflict conference, Yale University.Find this resource:

Wolford, S., Reiter, R., & Carrubba, C. (2011). Information, commitment, and war. Journal of Conflict Resolution, 55(4), 556–579.Find this resource:

Zagare, F. C. (1990). Rationality and deterrence. World Politics, 42(2), 238–260.Find this resource:

Zagare, F. C. (2011). The games of July: Explaining the Great War. Ann Arbor: University of Michigan Press.Find this resource:

Zagare, F. C., & Kilgour, D. M. (2000). Perfect deterrence. Cambridge, U.K.: Cambridge University Press.Find this resource:

Zagare, F. C., & Kilgour, D. M. (2003). Alignment patterns, crisis bargaining, and extended deterrence: A game-theoretic analysis. International Studies Quarterly, 47(4), 587–615.Find this resource: