To Disclose or Deceive? Sharing Secret Information between Aligned States

“Truth for friends and lies for enemies,” a phrase coined by Charles Horton Cooley many years ago, nicely captures how international relations scholars typically think about what information is communicated in world politics.¹ The literature on alliances has long focused on the conditions that give rise to and help maintain military cooperation. Notwithstanding a recent resurgence on the topic of secrecy in international politics more generally, the dynamics of deception—to include lying and concealing information—between cooperating states have received much less attention.² As John Mearsheimer noted in Why Leaders Lie, one reason for that might be that “because allies can help a state deal with a formidable rival, there are strong incentives for countries to have good relations with their allies and to build a modicum of trust with them, which is hardly served by lying.” Indeed, states caught lying to their friends are likely to pay a stiff price for their deception, which will “undermine trust and damage the partnership” and “ultimately hurt the country that told the lie.”³

Yet, alongside military cooperation, deception between aligned states sometimes occurs—even in between states that enjoy a long-term relationship, when their strategic interests are aligned, and when the states are particularly able to uncover lies or deception by their counterpart. For example, U.S. allies such as Israel, South Korea, and Taiwan have repeatedly deceived the United States about their respective efforts to build nuclear programs.⁴ Allies have also concealed from each other treaties signed with third parties, as was the case with Bismarckian Germany's secret alliance network.⁵ Moreover, it is an open secret that allies continually spy on each other and deny doing so when caught. One of the most famous instances is former U.S. Navy intelligence analyst Jonathan Pollard, who was caught spying for Israel in 1985. More recently, the National Security Agency was accused of tapping phone calls involving German Chancellor Angela Merkel, with international outrage somewhat lessened by claims that German intelligence had also been eavesdropping on U.S. secretaries of state.⁶

The most pronounced, consequential, and intriguing cases are those that involve a state deceiving an aligned state—that is, a foreign country with which it has formal or extensive informal ties in the security realm—about its intentions to use offensive military force against a threatening third party. Given the ability of the aligned state to provide military, economic, and diplomatic cover, the choice to deceive or conceal information is especially important to explore.

The dynamics of deception between aligned states become ever more crucial to understand, for both theoretical and policy reasons, when put in a larger framework of strategic choices about how much and what information to share with a trusted state. Current scholarship has been silent on the conditions that lead states to adopt one kind of information-sharing strategy over another. But what explains this variation? For example, why do states sometimes choose to share information about their intentions to use force, yet other times lie about their plans? When do they choose to conceal such plans, and when will they be transparent?

In this article, we seek to answer these questions by developing and testing a new theory that explains how states choose an information-sharing strategy when thinking about using military force. Specifically, we offer a parsimonious theory about the conditions under which a state chooses among four different information-sharing strategies: collusion, compartmentalization, concealment, and lying.

Our point of departure is that a state contemplating a limited offensive use of force against another state must decide whether and how much information about its secret plans to share with a state with which it is aligned. We argue that three main considerations govern decisions to fully share, compartmentalize, conceal, or lie: the state's assessment of whether it needs its partner's capabilities to succeed at the military mission, the state's perception of whether the partner will be willing to support the state in the requested role, and the state's expectation of costs it would incur if caught deceiving its partner state. We examine how these perceptual variables explain a state's choice of which information-sharing strategy to adopt.

The article makes three main contributions to the literature on information sharing and military cooperation between allies. First, it reconsiders how scholars typically understand the dynamics between aligned states regarding the use of force. We thus challenge the conventional wisdom that deception in the military sphere is characteristic of the interactions between adversaries but not aligned states. Our theory specifies important causal variables and scope conditions that explain when such dynamics are likely to be more or less salient.

Second, our theory contributes to a growing literature on secrecy in international security. Whereas most studies about secrecy focus on when states choose to act covertly versus overtly, and the consequences of such actions, we consider a state's decisions about whether and when to share secret information with an allied state about its intention to use force.⁷ As such, we look within the covert sphere to show how information-sharing dynamics affect not only the likelihood, but also the type of covert action we should observe.

Third, not only does the choice of an information-sharing strategy reveal the extent of states' intentions to cooperate and on what issues, but it also acts as an important mediating factor connecting pre-operational planning to the form of military force that will be employed and the likely blowback that could follow from these operations. Such intentions thus carry significant implications for international dynamics more broadly. For example, a choice among the four information-sharing strategies significantly affects whether states use military force bilaterally or unilaterally; it shapes the degree to which states are willing to exercise military restraint in their use of force; and it constitutes an important condition under which the use of force can lead either to tighter cooperation or to a lack of trust. Put differently, the choice of information-sharing strategy is an independent variable shaping these broader dynamics and outcomes, but one that current scholarship has yet to study.

The rest of the article proceeds as follows. First, we discuss the extant literature on aligned states and how secrecy and deception have been undertheorized with regard to allies. We then develop a theory that explains the variation in information-sharing strategies—that is, collusion, compartmentalization, concealment, and lying—that states choose to employ when contemplating the use of force. Next, we test our theory by examining four case studies, each corresponding to one of the information-sharing strategies we consider. The case studies include Israel, Britain, and France's decision to use force against Egypt during the Suez Crisis (illustrating the logic of collusion between France and Israel, as well as concealment vis-à-vis the United States), Israel's bombing of Syria's al Kibar reactor in 2007 (as a case of compartmentalization), and Israel's internal deliberations from 2010 to 2012 about whether to attack Iran's nuclear facilities (as a case of lying).⁸ In addition to offering variation in both the explanatory and dependent variables, the cases allow us to control for a host of potential confounding variables, including the extent of Israel's dependence on each of its main allies. We conclude with a discussion of the implications of our theoretical framework for states' cooperative or manipulative behaviors.

Our theory on information-sharing strategies and military cooperation draws from earlier theories related to alliance restraint, that is, the diplomatic effort of one state to influence its ally not to proceed with a proposed or existing military policy.⁹ The alliance restraint literature assumes that the state knows the military plans of its allies and thus can intervene in the implementation of those plans. Our article, instead, theorizes about whether a state will share information with an aligned state that may interfere or retaliate in response to that information. Most literature on alliance restraint and inter-alliance control focuses on formal alliances, whereas our theory extends to informal alignment relationships. Works discussing informal reliance relationships, such as military coalitions or allies of convenience, suggest temporary allegiance against some urgent national security threat.¹⁰ Our article shows that even these informal dealings can have predictable effects across repeated interactions, despite changes over time, in leadership, or in the relevant adversary.

We make a point to differentiate between alignments and alliances, a distinction often neglected in the literature. Stephen Walt's traditional definition tends to conflate these terms: “An alliance (or alignment) is a formal (or informal) commitment for security cooperation between two or more states, intended to augment each member's power, security, and/or influence.”¹¹ In actuality, however, an alliance is but one subset of the broader concept of alignment. Alignment is a “set of mutual expectations between two or more states that they will have each other's support in disputes or wars with particular other states.“¹² Patterns of alignment preexist formal alliances and are driven by common interests, or conflicts, between states.¹³ Therefore, as Glenn Snyder notes, there is a ”vast area of political interaction lying between the relatively static interests and conflicts in the system and the formation and management of formal alliances. In this area much turns on nuances—friendly gestures, concessions, symbolic demonstrations, and so on—well short of alliance or even entente.“¹⁴ Alignments do not exist apart from larger signals of support between states, which may take the form of a public alliance or private military aid.¹⁵

Scholars understand how reassurances function within alliances, but the function of secrecy in alliance maintenance is understudied.¹⁶ Our article adds to the growing, but still small, literature on secrecy and the relationships between allies. Most work on secret alliances examines the effect of their existence being kept secret, either for offensive advantage or as a means to create strategic uncertainty about states' capabilities.¹⁷

Studies of deception within alliances are also limited, with only a few examining when allies withhold information about themselves from partners or when allies share false information about others.¹⁸ For example, allies may bluff regarding their own intentions or capabilities. They may also use interpersonal relationships to more easily and effectively dissemble.¹⁹ Other works theorize that alliance deception may occur as the result of external, intense threats.²⁰ Mearsheimer suggests that lying can sometimes be a rational response to an anarchic system, in which states “have powerful incentives to sometimes act in ruthless and deceitful ways to ensure their survival.“ He predicts that leaders are more likely to lie to rivals than to allies, because states caught in a lie are likely to irreparably damage their alliance.²¹ Yet, leaders do, in fact, sometimes lie to their friends, and our theory adds scope conditions to when leaders will or will not choose to lie.

Although scholars have some understanding about why states will lie to each other, we know less about the conditions under which aligned states choose lying versus concealment to manage or maintain their relationships with other states. Furthermore, the extant literature has rarely explored the choices that aligned states make between strategies of cooperation and deception. Our theory fills this gap. In so doing, it contributes to scholarship on the strategic use of secret information and carries implications for the study of alignment behavior, alliance restraint and cooperation, and international relations more generally.

In considering whether to use military force, decisionmakers must weigh the likely benefits against strategic, domestic, political, economic, and myriad other risks. Once they have decided to use military force against a target, they face another dilemma: Should the state share information about its intentions and plans with an aligned state? Or, should it keep this information secret from its partner? Or, should it feed this state intentionally false information?

We consider two states: an Initiator, which wants to use military force against a target state, and a Partner, a state aligned with the Initiator that has the potential to assist it. Our study examines the Initiator's choice of whether to share information about potential uses of military force that (1) occur beyond the Initiator's borders, (2) directly or indirectly implicate the security interests of the Partner, and (3) the Initiator does not want publicized to the target state.

Before we present our theory, we should be clear about the situations and the conditions to which it pertains. First, it applies to limited offensive uses of military force that are kept secret until they begin, in order to maintain operational surprise or to avoid pressure from domestic or international actors. In some cases, the operation will remain secret after it takes place, so that the state can maintain plausible deniability, control the risk of escalation, or reduce the chance that it is seen as acting hypocritically. Such operations may include preventive and preemptive strikes, as well as other types of offensive military missions such as cyberattacks with kinetic effects, covert rescue missions, targeted killing missions, and surprise offensive operations to seize territory.²²

Second, the theory applies to situations where the planned use of force has implications for the Partner's security interests, writ large. Given their offensive nature, these operations could result in blowback between the partner states.²³ Although our theory focuses on when and how secret information is shared between partner states, it could also apply to formal alliances, especially to those whose terms are vague about whether a state is obligated to disclose its military plans to its allies. Even these states, however, may dispute when a disclosure is required for any given action. Moreover, while formal alliances may raise the costs of deception through their design—and intelligence-sharing mechanisms built into some of these arrangement could make it harder for the Initiator to conceal or lie about its military plans—they cannot prevent Initiators from attempting to deceive the Partner.

Third, our theory is not concerned with the general balance of power, and so the Initiator and the Partner may or may not be asymmetrical in terms of their military capabilities. Instead, we are interested in the contextualized balance of power: Does the Initiator, in a particular context and against a particular target, have the military capabilities to succeed in its mission? For instance, the Partner may have unique capabilities that make it especially useful for a certain mission but may otherwise not have superior capabilities. In this way, our theory does require some degree of dependency, but we consider the extent of this reliance relationship in calculating the costs of deceiving that Partner, as discussed below.

Finally, our interest is neither in whether the Initiator was correct in its assessment of the Partner's support or capabilities, nor in the deception costs it would face, nor in whether the mission was successful. Rather, we focus on whether and how these subjective assessments contribute to the information-sharing strategy the Initiator adopts.

In this section, we detail four types of information-sharing strategies an Initiator could adopt when contemplating the offensive use of military force. We identify two cooperative strategies: collusion and compartmentalization. We also include two deceptive strategies: concealment and lying (see table 1).

These information-sharing strategies involve varying degrees of truth-telling. At one extreme is collusion, in which the Initiator fully shares with its Partner its secret military plans to use force against a third party, including the nature, timing, and scope of the attack. This choice creates the potential for significant benefits and risks for the Initiator. For example, the Partner can potentially provide the Initiator with material benefits, such as assistance in planning or executing the attack or in providing cover, intelligence, and logistical support. These benefits need not be tangible, however. The Partner may provide an important source of legitimacy once the Initiator has employed force and must justify its actions before the international community. One example is the years of close cooperation between U.S. and Pakistani intelligence agencies to capture Khalid Sheikh Mohammed in 2003.²⁴

At the same time, the Initiator must also consider potential risks of sharing information with its Partner, including possibly disruptive (intentional or unintentional) leaks by the Partner that could jeopardize the success of the military mission by eliminating the element of surprise or the possibility of plausible deniability. Additionally, once privy to the information, the Partner could seek to impose certain demands for changes to the Initiator's plans, thereby reducing the Initiator's autonomy over the mission. The Partner could further coerce the Initiator by engaging in strong-arm tactics, for example, imposing economic sanctions or threatening to help the target state. The Partner could also choose to sabotage the mission and reduce the ability of the Initiator to achieve success. This type of coercion can be seen in Britain's rejection of military intervention in Iran as a result of U.S. opposition. In 1951, the United States threatened to withhold its assistance to the British if a dispute between the British and the Iranians drew Soviet involvement.²⁵

A strategy of compartmentalization, in which the Initiator provides only limited information to its Partner, mitigates some of the risks associated with collusion. This “need-to-know basis” approach ensures that the Partner is not blindsided by a military action that may adversely affect its foreign policy aims. For example, the Initiator could choose to selectively share its intentions, without operational details, to prevent unintended leaks. A Salvadoran army battalion trained at the U.S. Army School of the Americas during Ronald Reagan's administration was later implicated in human rights abuses during El Salvador's “dirty war,” but the Salvadoran government had not shared operational details of the military's efforts with U.S. military advisers or U.S. diplomats.²⁶ On the other hand, the Initiator can choose to share operational details with its Partner in an effort to appear forthcoming, without revealing its full foreign policy objectives associated with the mission. Still, compartmentalization could invite further scrutiny from the Partner. Thus, while compartmentalization reduces some of the risks of sabotage and loss of autonomy inherent in sharing secret information with the Partner, it does not eliminate them.

We also identify two deceptive strategies. The first involves the Initiator fully concealing information from the Partner. Concealment refers to situations wherein the Initiator does not share information and leaves the Partner completely in the dark about its intentions. Concealment benefits the Initiator, in that it is likely to prevent the Partner from sabotaging or restraining the Initiator, and it reduces the risk of pre-operational leaks. Concealment, however, involves additional types of costs compared with collusion and compartmentalization. Complete concealment could risk political or diplomatic retaliation from the Partner once the secret operation is revealed, especially if the Initiator's actions interfere with the Partner's foreign policy. Indeed, if the Partner believes that it should have been informed of the operation, it might impose “deception costs” on the Initiator. Deception costs refer to the expected reputational, material, diplomatic, or political costs the Initiator could anticipate, before the mission, that it would face as a result of being caught lying to (or concealing information from) its Partner.

The second deceptive strategy—lying—refers to situations wherein the Initiator chooses to communicate an untrue statement to the Partner with the intent to deceive or to give a false or misleading impression of its plans or intentions. Lying can involve some (but not all) of the information the Initiator shares with its Partner. For example, the Initiator can offer disinformation about the goal of the intended operation, but true information about the timing. We consider lying a means to ensure the Partner's support of the Initiator's military mission and willingness to play the role requested by the Initiator. The lie might be exposed before or after the mission is carried out, but the key is that the Initiator uses this type of information-sharing strategy during the lead-up to the mission.

Lying may involve two types of disinformation. The first type involves false private information about the Initiator's intentions or capabilities, such as misrepresenting its true military strengths or goals of the mission.²⁷ The second involves the Initiator's misrepresentation of intelligence about other states' capabilities or intentions. For example, Britain used propaganda in an attempt to shift U.S. attitudes away from isolationism in the lead-up to World War II. It spread disinformation that “ranged all the way from publicizing misleading information about allied strategy to undermining the prestige of an individual Nazi by encouraging salacious gossip about his private life.”²⁸

In the next section, we explain the conditions that shape an Initiator's decision of which information-sharing strategy to adopt vis-à-vis its Partner.

In this section, we argue that an Initiator's choice of information-sharing strategy depends on (1) the extent to which the Initiator believes that the Partner's capabilities are required for the military mission to succeed; (2) the Initiator's perceived likelihood that the Partner will support the Initiator in the form the Initiator requests; and (3) the Initiator's assessment of the deception costs it would likely incur if the Initiator deceived its Partner about its military plans. Importantly, our theory assumes that, all else being equal, an Initiator would prefer to adopt cooperative strategies with its Partner over deceptive ones, as long as the Partner is not expected to intentionally or inadvertently expose the military mission prematurely and will cooperate in the manner intended by the Initiator.

The first issue an Initiator must address is whether it needs its Partner's help to accomplish its military mission. The answer can vary depending on the type of mission being contemplated and the context-specific capabilities of the Initiator. Such assistance could include access to military or technological capabilities, intelligence, or manpower, or political, military, or diplomatic cover if the mission does not succeed as planned. If such help is deemed necessary, all else being equal, the Initiator is more likely to share information with the Partner if it chooses to go ahead with the mission.

Second, the Initiator must assess the Partner's likely reaction to the Initiator's willingness to cooperate in the form the Initiator desires. The Initiator is considering sharing information with the Partner in the hope that the Partner will play a certain type of role in the military mission. This role can be minimal or expansive, and can assume various forms, such as ground forces, air support, or intelligence, or even providing funds or political support to the Partner, but going no further. If the Initiator assesses that the Partner would be willing to play the requested role, all other things being equal, the Initiator should be more likely to share information. On the other hand, if the Initiator believes that its Partner will be utterly unwilling to play the requested role, it is likely to be deterred from sharing information out of concern that sharing could result in sabotage, premature exposure, or undesired loss of autonomy over the mission.

In such cases, the Initiator cannot risk the possibility that the Partner may sabotage its planning and thus is more likely to opt for a deceptive information-sharing strategy. In other cases, the Initiator might assess that the Partner is unlikely to object to participation and would not sabotage the mission, but the Partner might be unlikely to play the role the Initiator intends for it to play. For instance, the Initiator may want the Partner to take a small role and remain in the shadows, while the Partner may want to be an active participant. In the 1990–91 Gulf War, for example, Israel sought to retaliate against Iraqi missile attacks, but the United States feared that such action could destroy the coalition fighting against Iraq.²⁹ Alternatively, the Initiator may want the Partner to play a large role, but the Partner prefers to act in a supporting role, such as providing intelligence but not ground support. In such cases, the Initiator will have to make a determination whether it can persuade the Partner to play the desired role and, if so, engage in a cooperative information-sharing strategy, or whether it is unlikely to budge, leading the Initiator to opt for deception, all else being equal.

To gauge the Partner's likely reaction, the Initiator might use trial balloons in the form of anonymous leaks to the media or informal conversations between unofficial envoys to gauge the reaction of the Partner.³⁰ Trial balloons are a key strategy designed to minimize the Initiator's uncertainty with regard to the Partner's capabilities, intentions, or deception costs to be imposed. Trial balloons are not always ideal, however; they may be too vague and solicit no reaction, or they may raise alarms and cause the Partner to respond out of concern that the Initiator is about to act. Moreover, trial balloons may reveal the Initiator's intentions and invite scrutiny or slide into deception and provide information that ends up not being quite true. Trial balloons may also be prone to miscommunication, leading to misunderstandings between what constitutes a green or yellow light. Judgments about the Partner's likely reaction are also inherently difficult to make when the Partner is not a unitary actor, or when there may be no clear intentions ex ante.

Third, the Initiator needs to assess the expected deception costs it would incur if it chose to conceal or lie. Deception costs are those perceived costs that the Initiator estimates before the military operation and that would be paid after the operation begins or becomes public. Here, we focus on deception costs imposed by the Partner, but theoretically they could also be imposed by an international organization, other allies, or even the Initiator's own domestic public, as long as they are imposed on the Initiator because it concealed or lied to the Partner about the use of force.

Because assessing likely deception costs is inherently difficult, the Initiator may look to several sources to determine whether and what types of costs might be incurred. Past interactions between the Initiator and the Partner over similar issues could give the Initiator an indication as to whether the Partner is likely to impose material costs—such as decreased military aid or a pause in diplomatic relations—or other costs in response to deception. Similarly, the Initiator may consider how the Partner has reacted to deceptive acts by other states. Additionally, the Initiator might engage in trial balloons to assess potential deception costs, but this strategy is subject to the caveats and risks mentioned above. Finally, the Initiator could try to assess not only what costs the leader of the Partner state would like to impose, but also whether it will be able to impose them: Is the Partner likely to face significant domestic political or strategic constraints if it tried to impose deception costs? If the answer is yes, then the Initiator might assess the actual deception costs to be low.

Given the subjective nature of these assessments, different Initiators might use different metrics to assess likely deception costs. With this caveat in mind, we postulate that, all else being equal, deception costs might be perceived to be high when there are repeated interactions between Initiator and Partner states with a high level of dependence. Also, some types of formalized alliances may have more institutionalized mechanisms for sharing privileged information, stipulating high deception costs in the event the rules are not followed. Finally, if the mission fails or otherwise goes wrong, creating an increased risk of the Partner becoming entrapped or embarrassed, the Partner might be forced to share some of the costs from failure. In such cases, we anticipate that the Partner would wish to impose higher deception costs.

Figure 1 illustrates how the independent variables—as perceived by the Initiator—combine in a stepped theory to result in four ideal types of information-sharing strategies. When the Initiator determines that the Partner offers crucial capabilities for the success of the mission, the Initiator considers whether the Partner is likely to provide support in the way the Initiator desires. If so, then the Initiator will be more likely to engage in a cooperative strategy, because cooperation with a Partner is preferable to deception (as long as the Partner is not likely to expose the mission prematurely when informed about it). Which cooperative strategy, either collusion or compartmentalization, will depend on the Initiator's desire for autonomy, the risk of leaks, and other operational considerations, such as the particular capabilities the Initiator requires from the Partner. But because the Initiator needs the Partner's capabilities for the mission to succeed and expects the Partner to play the desired role, the Initiator, all else being equal, might be more likely to engage in collusion and fully share information with the Partner to both increase the likelihood of a successful mission and induce a vote of confidence in the Partner.

On the other hand, if the involvement of the Partner is necessary to the success of the mission, but the Partner is unlikely to agree to play the desired role, then the Initiator will have to consider whether it will incur deception costs if it deceives the Partner. If these deception costs are high, the Partner's perception that the Initiator is intentionally lying could destroy more of the trust between the two than if the Initiator pursued concealment or aborted the plan altogether. If the Initiator anticipates high deception costs, the likely strategy is indeterminate, because the Initiator's next steps depend on contextual factors. Specifically, the Initiator will likely first attempt to persuade the Partner to cooperate. Persuasion involves truthful arguments about capabilities, intent, or other aspects of the Initiator's plan. Persuasion can also involve reevaluating current information, providing new intelligence, or linking to other issues important to the Partner. Ultimately, though, if persuasion fails, the Initiator's insufficient capabilities to carry out the mission by itself could lead the Initiator to abort the mission (or delay it until more favorable conditions emerge), which would render moot the question of what information-sharing strategy to pursue. An example of an aborted mission occurred during the Gulf War, when Israel chose not to retaliate against Iraq for its Scud missile attacks following a phone call from Secretary of Defense Dick Cheney emphasizing that Israel risked losing U.S. patronage if it engaged in military action.³¹

When the Initiator needs its Partner's capabilities and assesses the Partner will be unwilling to play its desired role in the mission but is likely to impose only low costs for deception, the Initiator is more likely to choose to lie to the Partner. The goal of using disinformation is to shift the Partner's attitudes about its role in the mission toward what the Initiator desires. Anticipating low deception costs might make the idea of lying appealing to the Initiator. Nevertheless, given that these are ex post costs that the Initiator needs to anticipate ex ante, there is room for miscalculation and blowback.

As the decision tree in figure 1 shows, the Initiator may not need its Partner's help to successfully execute the military mission. In such circumstances, if the Partner is willing to assume the Initiator's desired role, the Initiator is likely to adopt one of the cooperative strategies (i.e., collusion or compartmentalization). Given the risk of unintended leaks inherent in collusion, the desire for more autonomy over the use of force, and the secondary role the Partner is expected to play, the Initiator is likely to engage in compartmentalization rather than collusion, withholding some details about the operation from the Partner.

Finally, when the Partner's support is unnecessary for the success of the military mission, but the Partner is unlikely to provide the desired support, the Initiator might be tempted to conceal all information from the Partner prior to embarking on the mission. Concealment is more likely when the Initiator estimates the deception costs to be relatively low. Concealment might be difficult, however, especially when the Partner has good intelligence capabilities; when the Initiator is a democracy, in which leaks are common; and when the operation is so big or complex that maintaining operational secrecy until it is completed might be difficult. If the deception costs are anticipated to be high, however, the Initiator's likely strategy is again hard to predict. The Initiator will again be faced with the decision of whether to persuade the Partner to join, or abort the mission if persuasion fails or is too risky as it could result in exposure by the Partner. Given the Initiator's ability to carry the mission by itself, the anticipated high deception costs must be weighed against the security cost of aborting the mission if the Partner is unpersuadable.

We consider an information-sharing campaign to begin with the Initiator's consideration of the use of force against the target state and to include all internal deliberations up to the final choice of an information-sharing strategy prior to the execution (or aborting) of the military mission.

The Initiator's information-sharing strategy may change during this period in response to new information about a number of factors: (1) as more information becomes available about the specific capabilities needed to complete the mission, the Initiator might change its assessment regarding whether the mission can be accomplished without help from the Partner; (2) in light of new intelligence or information obtained from trial balloons, the Initiator might reassess whether the Partner will agree to play the role desired by the Initiator; and (3) if the Initiator engages in a deceptive strategy, it may reevaluate the deception costs it may occur. New information can move the Initiator up/down or left/right, leading to a shift in strategy (see figure 1).

As long as deliberations over what information-sharing strategy to adopt are internal and not revealed or implemented, the Initiator might be at liberty to shift between strategies. But once the Initiator engages the Partner with either cooperation or deception, path dependencies could act as barriers to shifting strategies. For example, if the Initiator implements a collusion strategy, it will be more difficult for the Initiator to shift to concealment with the same Partner. Similarly, if the mission is exposed or revealed prematurely, the information-sharing strategies available to the Initiator might be considerably reduced.

We test our theory against four historical case studies (see table 2).³² To gain explanatory leverage, in most of our cases we hold constant the identity of the Initiator, Israel, as the actor contemplating the use of military force to eliminate or reduce a threat to its security, and its Partner, the United States. In so doing, we consider the U.S.-Israel relationship over time. We also examine the information-sharing strategy at work in France joining with Israel in the attack against Egypt in 1956, in which case France serves as the Initiator. In each case, we show how our variables shape the behavior of the Initiator and process trace its decisionmaking regarding whether and what information it should share with its Partner.

Our cases offer variation in the theory's independent variables and in the outcomes we seek to explain—the extent of Israel's information sharing with its main allies. The Suez Crisis features France colluding with Israel in the invasion of Egypt and actively concealing its plans from a disapproving United States. Although cases of collusion may be the least interesting to study as a phenomenon, because typically allies do cooperate in secret, this case is important because it highlights two information-sharing strategies to different allies at the same time. It also allows us to use primary documents to highlight the dilemmas involved in making such decisions. In cases of deceptive strategies (which are less intuitive and more surprising), we consider the role of contextualized alternative explanations for the Initiator's decision to conceal or lie.

The repeated interactions between Israel and the United States allow us to show that the dynamics we hypothesize are at work, despite variation in the leaders in power in both countries, the strategic environment of the Cold War, or the type of force being contemplated. One potential limitation of studying the U.S.-Israel dyad is that it is asymmetric. This asymmetry implies that, given Israel's dependence for security on the United States, Israel is more likely to need to share information with the United States before taking on military missions. We argue, however, that despite Israel's reliance on the United States for its security, Israel has attained significant independent military capabilities to carry out complex missions without explicit U.S. assistance or authorization, especially dating from the early 1970s.³³ As such, the U.S.-Israel relationship is not one in which Israel consistently needs its Partner's capabilities. Nor is it one where consistently high deception costs would dissuade any challenges to the Partner's position. One might object to our focus on Israel on the grounds that the U.S.-Israel relationship is special, in that the United States would provide unconditional support to Israel. We show empirically that this assumption is incorrect, especially because Israel's policy directly conflicted with that of the United States in several of these cases.

In two of our four cases, we hold constant not only the identities of the two actors, but also the threat of one of the Initiator's enemies developing an operational nuclear program. Despite these commonalities, we see extremely different information-sharing strategies adopted by Israel vis-à-vis the United States. Thus, the similarities across these cases and clear variation in the outcome of interest provide us with strong cases to test the strength of our causal claims.

The Suez Crisis demonstrates two information-sharing strategies: collusion among France, Britain, and Israel in the period from July 1956 until the invasion of Egypt in October 1956, and concealment from the United States of both the extent of that cooperation and the upcoming invasion, which we discuss in the next section.

The Suez Crisis originated with the rise to power of Col. Gamal Abdel Nasser, who played a role in the overthrow of Egypt's king in 1952 and became its second president from 1954 to 1970. Nasser sought to promote a pan-Arab movement of political unity by denouncing Israel and the West. During this time, the United States, Britain, and France sought to maintain the territorial status quo resulting from the 1948 Arab-Israeli War through the 1950 Tripartite Agreement. In September 1955, Nasser announced an arms deal with Czechoslovakia, in which Soviet weapons would give Egypt qualitative and quantitative superiority in combat aircraft over Israel. Tensions escalated when Nasser began to block traffic through the Straits of Tiran, cutting off the Israeli port of Eilat to the Red Sea. In response, the United States broke off talks about funding the Aswan Dam across the Nile River. Nasser retaliated by nationalizing the Suez Canal on July 27, 1956.

With Britain's approval, France began engaging in full information sharing with Israel about its plans to use force against Egypt in response to closure of the canal. We focus on the information-sharing strategy in which France—which was the true Initiator of the fully formed strike as codified in the secret Protocol of Sèvres—engaged with Israel. France was concerned with eliminating Nasser's influence and Egypt's sponsorship of the rebels fighting against France in the Algerian War.³⁴ The French attitude toward Nasser was “one of unrelenting hostility and fear.”³⁵ As Philip Zelikow and Ernest May note, “By the middle of September it is Paris that becomes the main driver of the moves toward war.”³⁶ Britain agreed with France's decision to share information with Israel, but it was the last party to sign off on the joint military plans.³⁷

In what follows, we show that, consistent with our theory, France needed Israel to play a key role in its attack against Egypt, and it had good reason to suspect that Israel would agree to take part. France chose to collude with Israel (together with Britain), and an agreement between Israel and France (and Britain) emerged. Secret conversations continued between the three sides to finalize their operations, which was codified in the secret Protocol of Sèvres. Consistent with those plans, on October 29, Israeli Air Force P-51 Mustangs launched a series of attacks on Egyptian positions in the Sinai.

Even before Nasser's announcement to nationalize the Suez Canal, France's armed forces had undertaken studies of the possibility of an intervention in Egypt. The memoirs of Gen. Paul Ely, chief of staff of the French armed forces, discuss studies considering interventions by various combinations of France, Israel, Britain, and the United States. France was unable to act alone because it was unable to counter Egyptian air strength with no air base near Egypt and no aircraft carrier for jets. Operations involving France, Israel, and Britain would involve delays of a month or more. Any U.S. involvement was considered unlikely.³⁸ Indeed, while France needed British capabilities, especially access to Britain's bases in Cyprus to land its forces, it did not need Israel's capabilities to succeed in its military plans against Nasser. Instead, France saw Israel's inclusion in the Anglo-French plan as an important means of justifying the use of force against Egypt.

On July 28, the day after Nasser's nationalization of the Suez Canal, French Prime Minister Guy Mollet and British Prime Minister Anthony Eden discussed that the impending threat to Israel “might be the pretext for military action against Abd al-Nasir that both nations had sought.”³⁹ Indeed, France began consulting with the Israelis almost immediately following the nationalization of the canal. During this time, Israel sought to remedy the shifting balance of power in the region by purchasing arms from the West. Rebuffed by the United States and Britain, in late September 1955, Israel was able to secure arms from France, including aircraft, tanks, and anti-tank guns. Mystère jet interceptors were to be delivered to the Israelis by April 1956.⁴⁰ Collusion between Israel and France intensified during the summer of 1956, as the two countries significantly expanded the scope of their intelligence sharing, primarily with regard to Egypt.⁴¹ During this period, the French believed that a lack of parliamentary and public support was causing the British leadership to waver in its commitment to the plan against Egypt.⁴²

Against this background, on July 27, 1956, French Defense Minister Maurice Bourgès-Maunoury met with Shimon Peres, director-general of Israel's Defense Ministry, at the French minister's request. There, Bourgès-Maunoury asked Peres, “Would Israel be prepared to take part in a tripartite military operation in which Israel's specific role would be to cross the Sinai?” Peres replied, “without hesitation” that “under certain circumstances I assume that we would be so prepared.”⁴³ On August 1, Peres was told that “the English and the French had decided in principle on a joint military action to conquer the Canal.”⁴⁴

Prime Minister David Ben-Gurion immediately made it clear that Israel would cooperate fully with France. Furthermore, he instructed Peres to share with France any information they requested: “If we hide anything that is vital for them, we will be betraying their trust in us.”⁴⁵ In mid-September, Ben-Gurion reiterated to the French that he was interested in cooperation, but that further discussions should follow with regard to the role Israel's military would play. Indeed, Ben-Gurion had become convinced that the Israeli alternative to such participation was a campaign in which Israel would stand alone against Egypt.⁴⁶ Moshe Dayan, chief of the General Staff of the Israel Defense Forces (IDF), recorded in his diary, “If indeed Britain and France capture the Suez Canal and restore its international status by force of arms, the political implications for us will be of the highest importance.” Ben-Gurion told Dayan and Peres that “in principle, we are willing to cooperate.”⁴⁷

At the same time, Ben-Gurion balked at pursuing a secret mission in which Israel would be the aggressor. To ease Ben-Gurion's concerns, Dayan told the prime minister that “the only contribution that Israel was in a position to make to the Anglo-French operation was to provide the pretext. Everything else could easily be managed without Israeli help. That ability to supply the pretext was Israel's historic opportunity.”⁴⁸ He added, “It would be easy now to extinguish this tiny flame of [French] readiness to go to war against Nasser, but it will be impossible to rekindle [it].“⁴⁹ Ben-Gurion followed Dayan's advice.

On September 30, the French and Israelis met in secret in France to discuss France's commitment of air and naval support for Israel. The meetings represented the first in-depth discussions of French-Israeli military operations against Egypt. French Foreign Minister Christian Pineau questioned whether the Israelis would be interested in joining France in a military operation if the British dropped out in response to domestic opposition, to which the Israelis responded in the negative.⁵⁰

With Israel's cooperation secured, the French and the Israelis held a secret meeting on October 14, 1956, to discuss Israel's exact role in the mission. Gen. Maurice Challe, deputy chief of staff of the French Forces, proposed “that Israel should be invited to attack Egypt across the Sinai Peninsula and that France and Britain … should then order ‘both sides’ to withdraw their forces from the Suez Canal, in order to permit an Anglo-French force to intervene and occupy the Canal on the pretext of saving it from damage by fighting.”⁵¹ Ben-Gurion was angered that this scenario was a means to satisfy British political concerns while exposing Israel to military and political dangers and the risk of being branded as the aggressor. Ben-Gurion agreed to meet with Mollet and Eden in Paris.⁵²

Indeed, during that time, Israel was extremely suspicious of the British in general, and of Eden in particular. Following Jordanian infiltrators harming several Israeli citizens, an Israeli reprisal raid against Jordan resulted in the loss of seventy to ninety Jordanian soldiers and police and several Israeli officers. The raid led the British to contemplate military intervention on behalf of Jordan, before the fighting there ended.⁵³ A private source warned Ben-Gurion that the British were preparing to take military action against Israel. As a result, Ben-Gurion had little faith in British assurances regarding cooperation in Suez, and “he feared that Britain might turn its back on Israel or even turn against it.”⁵⁴

The French were well aware of this mistrust and tried to offer Israel assurances and additional weapons shipments to allay its fears. While Israel was eager to participate in the Suez mission, it felt that there was little it could do to shape the exact role it would play; as a small state in need of weapons, Israel felt pressure to stay in the good graces of its French benefactor.⁵⁵ Ben-Gurion conceded that there was an unarticulated expectation that, to guarantee the flow of arms, Israel would abide by France's desire to go to war.⁵⁶ Israel's main goal in any cooperative relationship was to be seen as an equal partner.⁵⁷

On October 22, Ben-Gurion, Dayan, and Peres traveled to Sèvres, in the suburbs of Paris, to negotiate an agreement. On the second day of the meeting, Dayan proposed an IDF paratroop drop, accompanied by a mechanized brigade, that would constitute “a real act of war” and provide the pretext for intervention by Britain and France.⁵⁸ Plans were finalized at Sèvres for Israel to mobilize on October 26. French Mystère squadrons would arrive in Israel via Cyprus throughout October 27 and 28. French warships would reach positions off the Israeli coast by October 29, the date Ben-Gurion set for the invasion of Egypt. The Israelis would attack in the Sinai with the objective of reaching the canal as fast as possible. The British would surprise the Egyptian air force while grounded on its airfields and continue bombing as a warning to other Arab states. Landings by the English and the French would protect the canal against both sides. Israel would accept France and Britain's ultimatum to cease fire; Israeli troops would then have twelve hours to withdraw to a point ten miles west of the canal.⁵⁹ This secret agreement was codified in the Protocol of Sèvres and was signed by all the parties.⁶⁰

The Suez case fits our theory's predictions as to when we should observe concealment. First, as discussed above, France and Britain—both members of the North American Treaty Organization (NATO)—had sufficient military capabilities, especially when combined with Israel, to retake the Suez Canal. Second, by early September 1956, the three parties had concluded that the United States was unwilling to act against Nasser for a variety of domestic and strategic reasons, opting for continued diplomacy instead.⁶¹ Third, the French, British, and Israelis anticipated that they would incur low costs for their deceptive behavior against the United States for several reasons. The French believed that the United States would condone France, joined by Israel and Britain, in attacking Egypt, because France and Britain were two major NATO allies.⁶² Pineau convinced Israel's foreign minister, Golda Meir, that acting without the United States would not result in U.S. military intervention against Israel, France, and Britain. Furthermore, economic sanctions against Israel would be unlikely, because President Dwight Eisenhower was seeking to secure the Jewish vote in the November election.⁶³ Ben-Gurion believed that “the United States would not get involved so close to a presidential election and in opposition to its European allies. He expected that if America stayed out, so would Russia.”⁶⁴ At the same time, the Israelis were uncertain about how the United States would react, and whether it would break diplomatic ties with Israel.⁶⁵ In retrospect, the three parties clearly miscalculated the fury of U.S. leaders when the collusion was eventually exposed.

On August 1, Secretary of State John Foster Dulles told Pineau, “It is very important that this [the British-French-Israeli] operation not end in a gain for [Nasser]. The United States Government does not believe that military action would be justified at the present moment.”⁶⁶ The next day, Dulles, Pineau, and their British counterpart, Selwyn Lloyd, agreed to pursue international management of the canal through a multilateral conference. As Charles Cogan describes, “If Nasser rejected the conference plan, the French understood that the Americans would not act jointly with England and France but would accept Anglo-French action. The French thought of this as American ‘moral support,‘ enough to keep the Russians at bay.“⁶⁷ French intelligence reported that the Eisenhower administration was adamant that no action toward Egypt would be taken before the November 6 president election. After that, the United States would leave the initiative to Britain and France. Dulles declined Lloyd's requests to share military plans with their U.S. counterparts. As Jill Kastner describes, ”British officials assumed this was intended to save face should awkward questions arise during the American election campaign.“⁶⁸

Around the same time, the United States began to suspect that Britain and France were engaging in coordination that did not include it, but U.S. leaders were unsure of Israel's involvement. At a National Security Council meeting on August 30, the chairman of the Joint Chiefs of Staff stated that “we do know in a general way what the British and French intend to do and the character of the forces they are mobilizing for possible use against Egypt.”⁶⁹ The details about such an attack were unknown, however. A National Intelligence Estimate (NIE) from September 5 stated, “We consider it highly unlikely that the Israeli government would take advantage of a British-French military operation against Egypt to launch unprovoked major attacks on the Egyptian forces in Sinai. … In particular, the Israeli government almost certainly recognizes that the Western powers could not let it get away with such an attack if the West wished to preserve any standing with the other Arab states.”⁷⁰

The United States was unable to gather clear intelligence on the French, British, and Israeli plans. U-2 flights revealed the presence of sixty French Mystère jets in Israel, when France had reported the transfer of only twenty-four aircraft.⁷¹ Eisenhower remarked that the Mystères seemed “to have a rabbit-like capacity for multiplication.”⁷² With this, Eisenhower made clear that he knew that France was arming Israel in contravention of the 1950 Tripartite Agreement. The only additional information available to the Central Intelligence Agency (CIA) in October 1956 was that French military officials had recently met with their British counterparts.

Eisenhower continued to receive routine cables from U.S. embassies and CIA reports concerning meetings between French and British military planners, but he received little information from his Atlantic partners.⁷³ Steven Spiegel notes that as France, Britain, and Israel were concluding their collusion at Sèvres, Eisenhower and Dulles “had only tidbits of information: tension on the Jordanian-Israeli frontier, continuing Israeli mobilization, and Franco-British buildup in the Mediterranean, a suspicious termination of regular, high-level communications with Washington by Paris and London, a sizeable growth in Israeli-French diplomatic radio traffic, a large increase in French Mystère pursuit planes for Israel beyond the number reported to Washington.”⁷⁴

The British, French, and Israelis engaged in operational secrecy against the United States as the attack date approached. France and Britain intentionally created the “drying up of normal channels of information” with the United States, and Dulles reported a “loss of contact with their French and British colleagues” and “almost complete blackout of information from French and British with us regarding Middle East matters.”⁷⁵ For example, Eden had Britain's ambassador to the United States, Roger Makins, recalled on October 11. Until November 8, Britain had no ambassador in Washington. Cole Kingseed argues that this “withdrawal marked the beginning of a deliberate attempt by Eden to mask his true intentions from the American president.” During this time, “Eden and Mollet deliberately lied to Eisenhower to conceal the extent of their collusion with Israel.”⁷⁶ Between October 16 and the launching of the Israeli attack on Egypt on October 29, effective communication among Washington, Paris, and London on Suez almost ceased. As the State Department post-mortem describes, “Working level contacts between NEA [Near East and Africa, meaning the Bureau of Near Eastern, South Asian, and African Affairs at the Department of State] and the British and French Embassies in Washington virtually stopped. … Both at home and abroad the extent and kind of Anglo-French commitment was a closely held secret. British and French Foreign officials, except for a very few, were kept in the dark. No British or French ambassador in the Near East or Washington appears to have been informed.“⁷⁷

One might wonder whether France, Britain, and Israel concealed information from the United States out of concern for a superpower conflict if the United States were apprised of the plans against Egypt. Although there is some evidence to show concern about Soviet involvement in the Suez Crisis, the Initiators were primarily concerned with the United States' reaction. For example, the French ambassador to Moscow confirmed “real concern” on the part of the Soviet Union concerning Anglo-French military preparations in September 1956. The ambassador also assessed, however, “the generally acknowledged opinion—which I personally share—is that in case of military operations, the USSR will not intervene directly as long as the United States itself abstains.” But, he warned, “Any Anglo-French decision to intervene militarily should take into account these formidable risks and weigh them against the grave consequences that a success of Nasser would have vis-à-vis the British and French positions in Africa and Asia.”⁷⁸

U.S. intelligence began to suspect Israeli intentions to move against Egypt, as made clear in an October 28 report to the Intelligence Advisory Committee, consisting of the chiefs of the national intelligence agencies, that “new evidence of heavy Israeli mobilization on a scale which would permit Israel to,” among other things, “penetrate Egypt to the Suez Canal and hold parts of Sinai for a considerable time.”⁷⁹ On the same day, Ben-Gurion told the U.S. ambassador to Israel that Israel had mobilized only “a few units” strictly as a “defensive precaution.”⁸⁰ That day, Israel's ambassador to the United States, Abba Eban, met with U.S. Assistant Secretary of State for Near Eastern Affairs William Rountree and assured him of Israel's “defensive posture.” The meeting was interrupted with the delivery of a note stating there had been “a massive eruption of Israeli forces around the Egyptian boundary and a parachute drop deep into Sinai.“ As Eban relates, Rountree said, ”I expect you'll want to get back to your embassy to find out what is happening in your country.“⁸¹ On the morning of October 29, Dulles cabled the U.S. ambassador in Paris to say that ”bits of evidence are accumulating which indicate that the French Government, perhaps with British knowledge, is concerting closely with [the] Israelis to provoke action which would lead to Israeli war against Egypt with possible participation by French and British.“ News of the Israeli attack was received about four hours later, at approximately 3:30 p.m.⁸²

When news of the first attacks reached the White House, Eisenhower exclaimed: “What does Anthony think he's doing? Why is he doing this to me?” He then told Dulles, “We have to stop them—fast.” Robert Murphy, the deputy undersecretary of state, responded: “Washington simply was not informed of what the British and French were doing. We didn't know they intended going so far.”⁸³

Israel's 2007 attack on Syria's nuclear reactor in al Kibar is a case of a state employing a compartmentalization information-sharing strategy. Consistent with our theoretical expectations, Israel did not require U.S. capabilities to succeed in its mission. Instead, Israel wanted the United States to lead the operation to send a political message that it could both prevent Syrian nuclear proliferation and deter Iran's nuclear program. Although the United States was unwilling to be the face of the operation, it was willing to provide diplomatic cover to Israel as it prepared its military mission. The United States supported maintaining the covertness of the operation and signaled to Israel that it would not tell the Israelis to abort the mission. Throughout this process, Israel shared information and intelligence with the United States that preparations for the strike were almost complete. However, the United States was not informed of the exact timing of the attack prior to it being carried out. Thus, Israel provided the United States with limited information on a need-to-know basis confirming that an attack would take place, but without bringing it into the operational fold.

israel's information-sharing dilemma. Both Israel and the United States were aware of Syria's desire to obtain nuclear weapons after Syrian President Hafez al-Assad sought to buy nuclear research reactors from Argentina and Russia in the 1990s. As a result of U.S. pressure, however, the deals collapsed. But in 2006, Israel obtained intelligence regarding possible renewed Syrian attempts to obtain a nuclear weapons capability. In March 2007, operatives from Mossad, Israel's intelligence agency, recovered from the home of the head of Syria's Atomic Energy Commission photographs taken from inside the al Kibar reactor, “indicating that it was a top-secret plutonium nuclear reactor.”⁸⁴

Prime Minister Ehud Olmert decided to inform the United States of the intelligence regarding the Syrian reactor that Mossad had amassed. In April 2007, Mossad's chief, Meir Dagan, met with Vice President Dick Cheney, National Security Adviser Stephen Hadley, and Elliott Abrams, deputy national security adviser handling the Middle East.⁸⁵ Although the United States had satellite reconnaissance, Dagan shared something it did not have: photos from inside the al Kibar reactor, including one photo featuring the head of the Syrian Atomic Energy Commission with a North Korean nuclear official.⁸⁶ After President George W. Bush was subsequently shown the photos, he spoke with Olmert on the phone. Olmert said, “George, I'm asking you to bomb the compound.” Bush replied that he would examine the intelligence and respond.⁸⁷

Why did the Israelis choose to engage in a compartmentalization information-sharing strategy instead of concealment? Given the Bush administration's hawkishness and many of its members' strong support of Israel, the Israelis assumed that some in the administration would favor an attack against the Syrian reactor and would see it as an opportunity to regain U.S. credibility as a force against nuclear proliferation in the region, after the failure of U.S. troops to find evidence of weapons of mass destruction in Iraq. Moreover, the Israelis thought the United States might have an interest in exposing and terminating Pyongyang's cooperation with Damascus.⁸⁸ Based on these factors, the Israelis took a calculated risk to share intelligence on the Syrian reactor with the United States, believing that they could gain intelligence on North Korea's involvement and whet U.S. interest in counterproliferation.

a step back by the united states. The Bush administration entered a heated internal policy debate, with several options having been considered. A purely diplomatic approach failed to gain much traction. A covert operation involving a small team that would destroy critical equipment was deemed to be too risky by the CIA and the Department of Defense. An air bombing raid risked the United States being seen as engaging in a second preemptive war in the Middle East. The Bush administration hesitated giving Israel approval for a direct military option or ordering a U.S. strike out of fear of escalation into the wider Middle East. Lastly, the administration considered an international effort that would publicize the existence of the reactor and dismantle the Syrian nuclear project under the auspices of the International Atomic Energy Agency (IAEA).⁸⁹

Olmert met with Bush administration officials in Washington on June 19, arguing that a U.S. strike would “kill two birds with one stone,” and also serve to dissuade Iran from pursuing its own nuclear program. Olmert told Bush that Israel would attack al Kibar regardless, but that Israel wanted to provide the administration the information and incentives it would need to join the effort.⁹⁰ Bush explained to Olmert that the administration first would need to inform Congress that Israel was the source of intelligence about the reactor, potentially hindering U.S. action. He further explained, “I cannot justify an attack on a sovereign nation unless my intelligence agencies stand up and say it's a weapons program.” At that point, U.S. intelligence could not confirm this was the case. Bush offered to send Secretary of State Condoleezza Rice to the region, but Olmert was not interested, fearing that pursuing a diplomatic route would enable Syria to stall until the reactor came online. In his memoir, Bush wrote that “the prime minister was disappointed” at the U.S. preference for diplomacy, adding that Olmert told him, “This is something that hits at the very serious nerves of this country.”⁹¹

At no point, however, did Bush suggest that the United States would block Israeli action. “Olmert said he did not ask Bush for a green light, but Bush did not give Olmert a red light,” an Israeli general stated. “Olmert saw it as green.” Indeed, as Abrams recalls, when Olmert was making the case to the president why Israel could not pursue diplomatic means and was planning on attacking on its own if need be, “Bush was listening calmly, hanging up, and admiringly saying, ‘That guy has guts’” (or, reportedly, another anatomical variant of the same thought). From that point on, as CIA Director Michael Hayden noted, the United States needed to “step back,” because “we cannot enable someone to do something (even in their own national interest) that we are not authorized to do ourselves. We had no authority to bomb. We couldn't help anyone else do it, either.” Thus, the Bush administration stepped back and waited.⁹²

israel's unilateral action in communicating with washington. On September 1, an Olmert aide communicated to the White House that Israel's preparations for the strike on al Kibar were almost complete. Israel also shared information about the strike with Britain's foreign intelligence service, MI6. It did not, however, inform Washington of “the precise timing so that both nations could claim, with technical accuracy, that the Americans had not known about the attack in advance.”⁹³ On September 5, Israel's F-15s and F-16s launched at midnight and destroyed the reactor.⁹⁴ In this case, Israel retained operational secrecy, although the United States (and Britain) had received prior notice that the attack was imminent.

In the days following the strike, the Israeli government asked the Bush administration not to reveal what it knew. President Bush promised that the administration would remain “buttoned up” and suggested that “we let some time go by and then reveal the operation as a way to isolate the Syrian regime.”⁹⁵ To avoid retaliation and escalation, Israel opted for a targeted strike that would allow President Bashar al-Assad to deny the existence of a nuclear program in Syria. The Israeli and U.S. governments did not announce the secret raid for seven months.⁹⁶

Iran revealed a secretly constructed uranium enrichment facility in 2002. The United States announced in mid-2005 that Iran had a secret weapons program; in response, Iran suspended cooperation with the IAEA. However, a 2007 NIE backtracked to conclude, “We judge with high confidence that in fall 2003, Tehran halted its nuclear weapons program.” In a first draft, months earlier, the authors wrote that Iran still had a nuclear weapons program. This conclusion was revised after the CIA received personal communications from suspected weapons specialists that convinced them the Iranian government had terminated the weapons program in late 2003. This NIE was highly controversial, in that it did not support the narrative of the Bush administration regarding the dangerousness of Iran.⁹⁷ After the imposition of further sanctions by the UN, Iran agreed to recommence negotiations in October 2009. When Iran's counterproposals were rejected by the United States and the European Union, in June 2010 the Security Council imposed a complete arms embargo on Iran and banned it from any activities related to ballistic missiles. In January 2011, renewed talks made no progress. But in late 2011 and early 2012, the IAEA confirmed that it had evidence of Iran conducting ongoing weapons research. The confirmation sparked rumors that Israel might launch a unilateral attack on Iran.⁹⁸

Some of Israel's efforts to sway the United States to assist in preventing Iran's proliferation are consistent with an information-sharing strategy of lying. Specifically, Israel used the costly signal of a military buildup together with warnings of its intentions to attempt to sway the U.S. intelligence community and military establishment to believe that it was indeed making credible movements toward unilateral action against Iran. Importantly, we do not argue that the sole purpose of the buildup was to bluff; in fact, the buildup had two purposes—to deter Iran, and to convince the United States that it was willing to act alone against the Iranians. Deep divisions within Israel on the appropriateness of military force against Iran were kept within the debates of top political and military figures, and not shared with the United States. Thus, we find deception in that top Israeli leaders likely did not intend a unilateral attack.⁹⁹ Importantly, while Israel engaged in a campaign to shift U.S. policy that included disinformation, there were other elements in that campaign that did not feature lying, such as engaging in domestic political manipulation by communicating directly to Congress.

Consistent with our theoretical expectations, Israel resorted to a lying information-sharing strategy because it lacked the military capabilities necessary to ensure effective destruction of Iran's nuclear infrastructure. U.S. support could also add legitimacy if Israel tried to justify that a strike against Iran was done in self-defense. Also, Israel understood that the United States was going to oppose using force given its investment in a diplomatic resolution to hinder Iran's proliferation efforts. To push U.S. policy in a favorable direction, Israel exaggerated its willingness to attack Iran unilaterally. Israeli leaders believed that this exaggeration was likely to result in low deception costs. Israel's true intentions were difficult to assess, by design, given the secrecy in the upper echelons of the Israeli government and engagement by Israeli leaders in extreme rhetoric when discussing the threat posed by Iran in the past, with little retribution. Israel's efforts ultimately failed to convince the United States to cooperate in a military strike against Iran. To date, Israel has refrained from attacking Iran's nuclear reactors unilaterally.

israel's dilemma and misinformation about intentions. As noted above, one option to slow Iran's progress toward obtaining a nuclear weapons capability was a preventive attack on its nuclear facilities. Israel quickly realized, however, that unlike with the Osiraq or al Kibar attacks, this time it needed U.S. military support to complete the mission. Israel possessed penetrating munitions, but only the United States had the military capabilities to hinder Iran's nuclear success. The United States had high-precision heavy munitions, stealth air attack capabilities, and “bunker busters” designed to destroy underground facilities, such as the uranium enrichment facility at Natanz. Israel's bombs, on the other hand, would only have the limited ability to damage entrances. Israel may also have lacked the intelligence to determine if an attack had destroyed all aspects of Iran's decentralized nuclear program.¹⁰⁰

By late 2011, the Iran nuclear issue had gained urgency. Israeli officials had come to understand that the United States was not interested in using military force. Senior Israeli administration members communicated to Minister of Defense Ehud Barak that a military strike was extremely unlikely to happen in a Barack Obama administration. Barak contacted U.S. Secretary of Defense Leon Panetta in late 2011 to ask if a joint military exercise, scheduled to take place in Israel in April 2012, could be delayed. This request was intended to signal that Israel was contemplating military action. U.S. defense officials and former officials communicated to the press that they feared an Israeli attack possible between April and June.¹⁰¹

Barak also held a series of meetings with Panetta, National Security Adviser Tom Donilon, Secretary of State Hillary Clinton, and President Obama. He shared only the outlines of Israel's plans with Panetta. In response, Panetta urged him to “think twice, three times” before ordering a military strike. At one point, he even asked, “If you do decide to attack the Iranian facilities, when will we know?” Barak said the United States would not get more than a few hours' notice, but that he “recognize[d] our responsibility not to leave the Americans in the dark.” Obama, in a conversation with Barak, highlighted that domestic support for an attack was weak: “We hear that even people high up in your military, in military intelligence and the Mossad, are against it.”¹⁰²

Seeking to persuade U.S. decisionmakers to reconsider their attitude toward a military strike against Iran, Prime Minister Benjamin Netanyahu addressed a joint session of Congress on May 24, 2011. In an effort to convince the Obama administration that Iran's nuclear program required military intervention, Netanyahu emphasized the urgency of addressing Iran's proliferation.¹⁰³ Netanyahu calculated that the U.S. debate could be swung in favor of military strikes on Iran by persuading the United States that the time line of Iran's nuclear program posed an imminent threat. The Israeli prime minister, however, had a reputation for exaggerating the imminence of this nuclear threat, having claimed on many occasions that Iran was only six months away from obtaining a nuclear bomb.¹⁰⁴ In fact, intelligence from Mossad contradicted Netanyahu, concluding that Iran was “not performing the activity necessary to produce weapons.”¹⁰⁵ Netanyahu's use of disinformation was indeed known among U.S. diplomats. As former Secretary of State Rex Tillerson once observed, “In dealing with Bibi, it's always useful to carry a healthy amount of skepticism in your discussions with him.”¹⁰⁶

deceptive signals. With diplomatic pressure on the United States having failed, Israeli decisionmakers decided to pursue a lying information-sharing strategy to obtain the much-needed U.S. military support for a strike against Iran. The key act of deception occurred in 2012, when Israel, in an effort to underscore the credibility of its threat to attack Iran, launched a massive military buildup. In describing Israel's “war scare” strategy, Daniel Sobelman writes, “To enhance its leverage over the United States, Israel led its principal strategic ally to infer that a unilateral Israeli attack on Iran could be imminent.” Unlike its attacks on Osiraq and al Kibar, “Israel deliberately created the impression of an impending unilateral attack and then harnessed this perception in a deliberate effort to limit Obama's flexibility, influence U.S. policy, and alter Iran's strategic calculus.”¹⁰⁷

Israel understood that military expenditures represented a costly signal and could serve as a last-ditch effort to make its intentions to attack appear more credible. The Netanyahu government was spending billions of dollars on a military buildup, as well as consolidating military cooperation with Azerbaijan near Iran's northern border. This buildup included preparations for “an all-out air attack, supported by commando forces, in the heart of Iran.” In September 2012, Netanyahu notified the IDF and Mossad that they should prepare for a full-scale attack on Iran to occur in thirty days.¹⁰⁸ Barak believed that U.S. radar systems and electronic intercepts had detected the volume and nature of air force exercises Israel had conducted over recent months, and so the United States understood that Israel was preparing for military action.¹⁰⁹

Alarmed, the Obama administration shifted its assessment as more high-level U.S. officials began to assess the Israeli threat as credible.¹¹⁰ One official, Gary Samore, the Obama administration's White House coordinator for arms control and weapons of mass destruction, said that “with the exception of Vice President Joseph Biden, who thought it was all a big bluff,” senior members of the administration, including Obama and Panetta, felt “pressured” by Israel's signals.¹¹¹

Some observers might argue that Israel may have chosen to engage in deception because of the notoriously bad relationship between Obama and Netanyahu.¹¹² Perhaps this is one of the reasons why Israel's actions did not convince the Obama administration to abandon its diplomatic efforts in favor of military actions against Iran's proliferation. At the same time, Israel's strategy was not just about changing the personal attitudes of President Obama, but about targeting the United States as a whole, as its alleged deceptive efforts extended to Congress and the U.S. intelligence community through the tangible military buildup.

By the summer of 2012, however, the idea of an attack against Iran seemed to have lost steam. Barak writes: “I finally realized that an Israeli strike would not be possible. This wasn't because I doubted the damage it could still do to Iran's nuclear efforts, but because of the damage it would do to our ties with the United States. No matter how we might explain our attack, with the joint exercises about to begin, it would come across as a deliberate attempt to implicate our most important ally in a potential conflict with Iran, against the explicit wishes and policy of President Obama and the US government.”¹¹³ Barak balked at the suggestion of one of Netanyahu's allies to strike immediately before the 2012 U.S. presidential election and force U.S. support. Israeli threats to attack Iran dissipated around Barak's meetings with senior U.S. officials in September 2012. That same month, Netanyahu stated in a speech at the UN that there were still months left to stop Iran, hinting that an imminent attack was unnecessary.¹¹⁴

A year later, there were suggestions—including one from former Israeli Prime Minister Ehud Olmert—that Netanyahu had spent billions to make his threats look more credible to Washington rather than to prepare Israel's military for an attack.¹¹⁵ Journalist Ronen Bergman interviewed Ehud Barak, who approved of the air strike, but did not believe that Netanyahu really intended to attack. Others assess that Netanyahu “only wanted to make Obama believe that he intended to attack, in order to force Obama's hand, to steer him to the conclusion that America would inevitably get embroiled in the war anyway, so it would be better for the United States to carry out the attack itself, in order to control the timing.”¹¹⁶ Israel's national security adviser at the time, Yaakov Amidror, said that he “truly believed they were not bluffing.” Mossad Director Tamir Pardo claims to have had doubts the entire time, saying, “A deception at this level requires that no more than one or two people be in the loop.” However, “when the prime minister ‘tells me to commence the countdown, you realize that he is not playing games with you. These things [entering a state of preparedness] have enormous implications. It's not something he is allowed to do only as a drill.‘”¹¹⁷ The opacity surrounding Israel's intentions remains because the decision was kept between Netanyahu and Barak and not shared with Israel's security chiefs, to enhance the credibility of the threatened assault.

In this article, we offered a typology of information-sharing strategies, developed a theory that explains this variation, and tested the theory's predictions against several illustrative case studies. Our findings highlight unique and understudied dynamics in world politics. We showed that the choice of information-sharing strategies plays an important role in shaping not only whether Initiators ultimately choose to use force or abort a particular military mission, but also whether that use of force is unilateral or multilateral. Moreover, the information-sharing strategies the Initiator chooses to adopt can shape subsequent interactions between the aligned states, leading to more trust and further cooperation, or the imposition of deception costs and alliance discord.

One set of implications of our argument involves the capacity of the Partner to restrain the Initiator. Information-sharing strategies have important implications for both aligned and allied states because they affect how states manage and alter the behavior of friendly states. All else being equal, the Partner has the least restraint potential when the Initiator chooses to conceal its plans from the Partner. Without information about the impending plans to attack, there is little the Partner can do to dissuade the Initiator. Although it is hard to restrain the Initiator if the use of military force is a surprise, the Initiator also faces an increased risk of blowback if it engages in concealment. This may take the form of future reluctance to assist in similar military missions or in other aspects of the reliance relationship. The Partner also has the lowest restraint potential when lying, given that a Partner cannot act without relevant true information. The Partner has the highest restraint potential when the Initiator engages in collusion. This is because the Partner is privy to the greatest amount of information about the plans of attack, and its involvement is likely necessary for the operation to succeed; thus, the Partner has many points of leverage to restrain the Initiator. The Partner enjoys a medium level of restraint potential when it provides information on a need-to-know basis. The Partner is less able to restrain the Initiator because the Initiator likely has the capabilities to use force on its own, and because it probably has withheld some important aspect of the planned mission that affects the Partner's restraint potential.

Our findings also shed light on how states assess whether to use force against nuclear proliferators. We uncovered significant variation in the information-sharing strategies that states use in deciding whether to pursue a preventive strike. Specifically, assessments of the Partner's reaction to a preventive strike and its capabilities to contribute to such an attack play a crucial, and understudied, role in whether states will attack proliferators and what form such an attack will take. The patterns identified in this article also explain other cases in which Israel operates against nuclear proliferators, such as Israel's attack on Iraq's Osiraq reactor in 1981 and the Stuxnet cyberattack on Iran's Natanz nuclear enrichment facility, discovered in 2010.¹¹⁸

Since the United States' withdrawal from the international agreement on Iran's nuclear program, the Joint Comprehensive Plan of Action, discussions about an Israeli attack on Iranian nuclear facilities have resurfaced. Our theory and findings suggest that the United States' capabilities remain a necessary condition for a successful attack, and, thus, perceptions about U.S. support of such an attack are likely to continue to shape Israeli decisionmakers' planning on whether to attack the Iranian facilities. If U.S. decisionmakers gave Israel the green light (or if Israel infers such a green light), a joint Israel-U.S. operation would be likely. If the United States were reluctant to approve such an attack, however, Israel might opt to deceive it once again. The increase in cooperation between Israel and Persian Gulf states following the August 2020 Abraham Accords—which emerged in large part from joint concern of a nuclear-armed Iran—might lead to collusion between those countries over the use of force against Iran's nuclear facilities. It is unclear what role the United States would play in such an attack, but the possibility of another Suez scenario (or a joint disinformation campaign) in the event that the United States were unsupportive cannot be ruled out.

Finally, our theory and findings raise important directions for future research. Because deception and secrecy among allies is an understudied phenomenon, considering the various forms of deception and their probability of success would be instructive. Moreover, scholars could examine when states are effective in using deception—with regard to their own intentions and capabilities, or through the use of fabricated intelligence about a third state—to convince other states to support their foreign policy. Finally, future studies could further expand the scope of this study by examining the choice of information-sharing strategies between adversaries. Given a different incentive structure, the puzzling cases between adversaries or competitors would be ones of collusion and compartmentalization, as opposed to deception.

The authors thank the anonymous reviewers, the participants in the workshop on New Research on Secrecy in International Relations, as well as Austin Carson, Alexander Lanoszka, and Danielle Lupton, for their comments on an earlier draft of this article.