The Myth of Entangling Alliances: Reassessing the Security Risks of U.S. Defense Pacts

For the first 165 years of its history, the United States did not form any alliances besides the one it signed with France during the Revolutionary War. Instead, U.S. leaders followed George Washington's advice to “steer clear of permanent alliance with any portion of the foreign world,”¹ a recommendation subsequently enshrined in Thomas Jefferson's inaugural pledge: “Peace, commerce, and honest friendship with all nations; entangling alliances with none.”²

Since World War II, however, U.S. leaders have dramatically reversed this policy and signed defense pacts with more than sixty countries.³ As a result, the United States is legally obligated to defend a patchwork of nations that spans five continents, contains 25 percent of the world's population, and accounts for nearly 75 percent of global economic output. To what extent does this massive commitment entangle the United States in wars it would otherwise avoid?

This is a pivotal question for U.S. foreign policy. For sixty-five years, the United States has maintained a global network of alliances, and President Barack Obama reaffirmed U.S. commitments to defend allies during recent crises between North and South Korea,⁴ Iran and Israel,⁵ Russia and Ukraine,⁶ and China on the one hand and Japan and the Philippines on the other.⁷ A growing number of prominent scholars, however, argue that such commitments are dangerous and should be abandoned. In this view, U.S. alliances are not valuable assets but rather “transmission belts for war”⁸ that “risk roping the United States into conflicts over strategically marginal territory”⁹ and “ensnaring the United States in wars that it otherwise need not fight against nuclear-armed adversaries.”¹⁰ One study even compares the U.S. alliance network to the tangled web of European security commitments that helped catalyze World War I.¹¹ According to these scholars, the United States should reduce its entanglement risk by revising,¹² scaling back¹³ or scrapping altogether its alliance network¹⁴—measures that if implemented would constitute the biggest shift in U.S. grand strategy in two generations.

Despite this deep policy interest in entanglement, however, there are few rigorous studies on the subject. Most studies on alliances examine issues only partially related to entanglement, such as the design of alliance agreements,¹⁵ the reliability of allies in wartime,¹⁶ the causes of alliance formation,¹⁷ and the effect of alliances on the likelihood of international conflict.¹⁸ The few studies that focus on entanglement engage primarily in theory building, rather than theory testing, and analyze, at most, a handful of cases.¹⁹ As a result, scholars' hunches about the security risks of the United States' alliances remain largely just that—hunches.

To address this shortcoming, this article tests two competing perspectives on entanglement. The first, which I call “entanglement theory,” holds that alliances drag states into wars by placing their reputations at risk, socializing their leaders into adopting allied interests and norms, and provoking adversaries and emboldening allies.²⁰ The other perspective, which I call “freedom of action theory,” maintains that great powers can avoid entanglement by inserting loopholes into alliance agreements, sidestepping costly commitments, maintaining a diversified alliance portfolio that generates offsetting demands from different allies, and using explicit alliance commitments to deter adversaries and dissuade allies from initiating or escalating conflicts.²¹

I test these competing perspectives by analyzing the extent of U.S. entanglement in all postwar U.S. military conflicts. Specifically, I examine every militarized interstate dispute (MID) in which the United States participated from 1948, the year it signed its first standing alliance, to 2010 and ask a basic question: To what extent was U.S. involvement driven by formal alliance commitments?

Admittedly, this is a limited ambition, as I evaluate only the frequency of past cases of U.S. entanglement, not the current level of U.S. entanglement risk. Moreover, I do not account for other potential costs of maintaining alliances, such as higher levels of military spending or poorer diplomatic relations with non-allied countries. Nevertheless, entanglement is arguably the most important cost of alliances, and providing a systematic account of past cases is an important step in gauging the United States' current level of entanglement risk.

The results of this analysis strongly support freedom of action theory. Over a sixty-two-year period in which the United States maintained more than sixty alliances, I find only five ostensible episodes of U.S. entanglement—the 1954 and 1995–96 Taiwan Strait crises, the Vietnam War, and the interventions in Bosnia and Kosovo in the 1990s. Even these cases are far from clear-cut, because in each there were other important drivers of U.S. involvement; U.S. policymakers carefully limited support for allies; allies restrained the United States from escalating its involvement; the United States deterred adversaries and allies from escalating the conflict; or all of the above.

To be sure, the United States has intervened on the side of allies on numerous occasions. In most cases, however, U.S. actions were driven by an alignment of interests between the United States and its allies, not by alliance obligations. In fact, in many cases, U.S. policymakers were the main advocates of military action and cajoled reluctant allies to join the fight.

In addition, there are many cases in which alliances restrained the United States, or in which the United States restrained its allies or sidestepped costly commitments. I examine only cases of military conflict and therefore cannot evaluate fully the prevalence of such cases of nonconflict. Even within my biased sample, however, there are at least four cases in which alliances prevented U.S. escalation, and another seven cases in which the United States reneged on security commitments, restrained an ally from attacking a third party, or both

At worst, therefore, alliances have had a mixed effect on U.S. involvement in military conflicts—some alliances at some times have encouraged U.S. military involvement; others have discouraged it; and some have simply been ignored by U.S. policymakers. The only way to build a powerful case for entanglement theory is to commit serious methodological errors. For example, one could spin correlation as causation by characterizing cases in which the United States backed allies for self-interested reasons as cases of entanglement. One also could select on the dependent variable by highlighting evidence of entanglement while ignoring instances in which the United States shirked alliance commitments or in which alliances restrained the United States or prevented conflicts from breaking out in the first place. Such practices are common in the existing literature, but they are flawed and feed an exaggerated fear of entangling alliances in American society.

The finding that U.S. entanglement is rare has important implications for international relations scholarship and U.S. foreign policy. For scholars, it casts doubt on classic theories of imperial overstretch in which great powers exhaust their resources by accumulating allies that free ride on their protection and embroil them in military quagmires.²² The U.S. experience instead suggests that great powers can dictate the terms of their security commitments and that allies often help their great power protectors avoid strategic overextension.

For policy, the rarity of U.S. entanglement suggests that the United States' current grand strategy of deep engagement, which is centered on a network of standing alliances, does not preclude, and may even facilitate, U.S. military restraint. Since 1945 the United States has been, by some measures, the most militarily active state in the world. The most egregious cases of U.S. overreach, however, have stemmed not from entangling alliances, but from the penchant of American leaders to define national interests expansively, to overestimate the magnitude of foreign threats, and to underestimate the costs of military intervention. Scrapping alliances will not correct these bad habits. In fact, disengaging from alliances may unleash the United States to intervene recklessly abroad while leaving it without partners to share the burden when those interventions go awry.

A better route to a disciplined defense policy would be to bolster domestic constraints on the president's ability to send U.S. forces into battle while retaining allies, who can provide U.S. leaders with a second opinion on the merits of potential interventions; political cover when U.S. leaders decide not to intervene; and troops, supplies, intelligence, and money when the United States does intervene. In short, military restraint is best pursued through allied engagement, burden sharing, and domestic reform—not retrenchment. Americans tend to blame their military misadventures on foreign allies, but the blame for these debacles, and the responsibility for avoiding future fiascos, lies at home.

The rest of this article proceeds as follows. First, I define entanglement. Second, I discuss entanglement theory by explaining how alliances drag states into military conflicts. Third, I present freedom of action theory by discussing how great powers can avoid entanglement. Fourth, I examine the empirical record, summarizing the findings from all of the MIDs in the sample (188 in total) and then discussing the 5 ostensible cases of entanglement individually. Finally, I discuss the implications of the results for U.S. foreign policy.

Entanglement occurs when a state is dragged into a military conflict by one, or more, of its alliances. The essence of entanglement, then, is that loyalty trumps self-interest: a state is driven by moral, legal, or reputational concerns to uphold an alliance commitment without regard to, and often at the expense of, its national interests.

Many studies call this phenomenon “entrapment,” following Glenn Synder's pioneering work on alliance politics.²³ Tongfi Kim, however, persuasively argues that the term “entrapment” should be used only when an ally instigates a conflict with a third party in order to drag its alliance partners into the dispute.²⁴ For example, some scholars warn that Taiwan could entrap the United States in a war with China by declaring independence.²⁵ The term “entrapment” is thus a subset of a broader phenomenon that Kim calls “entanglement,” which encapsulates a variety of situations in which a formal alliance encourages a state to participate in or escalate a conflict in a manner not dictated by its national interests.

There are four general types of entanglement. First, there are cases of direct defense in which a state is compelled by an alliance obligation to back an ally in a fight. This form of entanglement includes cases of entrapment as well as cases in which a state defends an ally that is attacked despite refraining from provocative behavior. For example, if Taiwan made no moves toward independence, but China attacked it anyway and the United States defended Taiwan to uphold the Taiwan Relations Act, such a case would constitute entanglement but not entrapment.

A second form of entanglement involves cases in which a state fights to maintain its credibility as a security guarantor even when its allies are not involved in the conflict. In such cases, a state fights not to defend its allies, but to reassure them of its resolve to use force on behalf of others. For example, some scholars argue that the United States intervened in the Bosnian War in 1995 to demonstrate its commitment to NATO.²⁶

A third form of entanglement involves cases of co-optation in which the existence of an alliance causes a state to expand the scope of its interests. For example, some scholars claim that the formation of the U.S.-Japan alliance caused U.S. policymakers to develop an interest in securing Southeast Asian resources and markets for Japan and ultimately spurred U.S. military involvement in Vietnam.²⁷

Finally, there are cases in which an alliance makes a state a target of another country. For example, some scholars claim that America's de facto alliance with Israel places the United States in the cross-hairs of some Middle Eastern countries and thereby increases the likelihood that U.S. forces will be attacked.²⁸

The common theme among these various forms of entanglement is that, in each, an alliance drags a state into a military conflict against its national interests. To spot instances of entanglement, therefore, analysts first must identify a state's national interests and then distinguish actions intended to serve those interests from those intended to uphold an alliance. In other words, because entanglement is defined as a form of suboptimal behavior, identifying cases of entanglement requires a clear standard of optimal behavior, which the entanglement literature generally defines as the strict pursuit of predefined national interests.

The concept of entanglement also relies on a distinction between alignments and alliances. An alignment is a confluence of interests among states. An alliance, on the other hand, is a formal agreement among states to cooperate militarily. Whereas alignments emerge and dissolve as states' interests change, alliances often outlive the conditions that initially spawned them and persist in the face of clashing interests. For example, the United States and Pakistan remain bound together by an alliance even though their interests are decidedly unaligned on numerous issues.²⁹ By contrast, Saudi Arabia and the United States are not formal allies, but they share an alignment based primarily on the exchange of weapons for oil. This distinction between alliances and alignments is central to the concept of entanglement, because entanglement occurs only when a state fights to uphold a formal alliance commitment and not to capitalize on an alignment of interests.

Entanglement is thus distinct from chain-ganging, which occurs when a state backs an ally because the defeat or defection of that ally would threaten its own vital interests.³⁰ In cases of chain-ganging, states support allies out of self-interest, and formal alliance obligations are causally insignificant. In cases of entanglement, by contrast, alliances are of overriding significance; they drag states into conflicts against their interests.

Alliances can entangle states into conflicts by placing their reputations at risk, socializing leaders into adopting allied interests and norms, provoking adversaries, and emboldening allies.³¹ Below, I discuss each of these mechanisms.

According to entanglement theory, a state that violates one of its alliance agreements risks undermining the credibility of all of its international commitments, just as defaulting on a loan can destroy a person's credit score.³² Consequently, “alliance commitments do not merely ‘reveal’ existing interests, they create new ones, by incurring reputation costs.”³³ Such reputational concerns can drive states into wars over trivial interests in peripheral places. As Jack Levy explains:

The alliance comes to be perceived as an end in itself, transcending the more concrete national security interests for which it was initially conceived. Political decision makers come to believe that support for one's allies, regardless of its consequences, is essential for their national prestige, and that the failure to provide support would ultimately result in their diplomatic isolation in a hostile and threatening world. This symbolic significance of an alliance commitment may also become linked with public opinion (for example, “national honor”) and the domestic security of elites, thus further increasing the importance of alliance solidarity. It is for these reasons that the policies and precipitous actions of secondary states often drag their Great Power protectors into war.³⁴

This argument reflects a tenet of liberal institutionalist theory that holds that international institutions can take on “a life of their own” and shape state behavior by placing states' reputations for compliance on the line.³⁵

Alliances also can shape states' military policies through processes of socialization. Allies typically engage in regular meetings, military exercises, and displays of solidarity, and allied leaders often make public appearances and support interest groups in each other's societies. Such social interactions may, over time, encourage states to adopt each other's interests and internalize alliance norms, including the norm of collective defense.³⁶ As Glenn Snyder explains, “Alliances introduce a sense of obligation” and a “moral convention that promises should be kept. These normative elements are enhanced by the solemnity of ceremonies” and may cause allies to “redefine their ‘selves’ so that they embrace at least some of the partner's interests as their own” and “have a joint interest in keeping [the alliance] alive” that “may very well overcome a divergence of self-regarding interests.”³⁷

Alliances may provoke adversaries and thereby foment conflicts that ultimately suck in states against their interests. The logic of this argument stems from the spiral model of conflict, which holds that international relations are often characterized by vicious cycles of fear and aggression.³⁸ The anarchical nature of the international system forces states to make worst-case assumptions about the motives of others. As a result, steps that one state takes to enhance its security, such as forming or bolstering an alliance, may “turn tomorrow's potential adversary into today's certain one”³⁹ and “trigger a spiral of insecurity that heightens the prospect of war between alliance members and non-signatories.”⁴⁰

Firm security commitments also may spark conflicts by emboldening allies to act more aggressively than they otherwise would.⁴¹ As Barry Posen explains, “U.S. security guarantees also encourage plucky allies to challenge more powerful states, confident that Washington will save them in the end—a classic case of moral hazard. This phenomenon has caused the United States to incur political costs, antagonizing powers great and small for no gain and encouraging them to seek opportunities to provoke the United States in return.”⁴² Such emboldenment not only increases the likelihood of wars against non-allies, but also exacerbates “internal security dilemmas” among allies that can spiral into wars, such as those between Greece and Turkey⁴³

According to entanglement theory, alliances rope states into conflicts by placing states' reputations on the line, by socializing leaders into adopting allied interests and norms, and by provoking adversaries and emboldening allies. Entanglement theory also suggests that there is a linear relationship between a state's number of alliance commitments and the frequency with which the state will suffer from entanglement. As Ja Ian Chong and Todd Hall argue, “The greater the number of commitments the United States assumes, the higher the number of conceivable flash points capable of drawing both it and its partners into a larger confrontation. Moreover, the multiplication of commitments also raises the stakes: the United States may see itself as having to take a firm stance in a conflict scenario on a matter of peripheral interest out of concern for its reputation among allies, friends, and adversaries.”⁴⁴

If entanglement theory is correct, what trends should scholars expect to see in the empirical record? Unfortunately, the writings cited above do not make precise predictions about how much entanglement one should observe in the case of the United States. But given that the United States has maintained more than sixty alliances for more than sixty years, it seems reasonable to expect to find at least a few clear-cut cases—that is, those in which the existence of an alliance functioned as a necessary condition, if not a necessary and sufficient condition, for U.S. involvement or escalation in a costly conflict. It also seems reasonable to expect to find within many other cases evidence of the entanglement dynamics discussed above—for example, U.S. leaders expressing concerns about the United States' reputation as an ally, leaders expanding their conception of U.S. national interests to include those of allies, or alliances sparking conflicts by provoking adversaries or emboldening allies. Failure to find such evidence would not mean that U.S. alliances are cost-free commitments, but it would cast doubt on the claim advanced by advocates of retrenchment that “America's alliances are transmission belts for war that ensure that the U.S. would be embroiled in Eurasian wars.”⁴⁵

The entanglement mechanism described above assumes that if a U.S. ally becomes involved in a conflict, the United States “will be automatically swept up in the fighting—regardless either of its degree of interest in the conflict or the costs and risks of involvement.”⁴⁶ Yet a significant body of scholarship suggests that great powers can avoid entanglement by inserting loopholes into alliance agreements, sidestepping costly commitments, maintaining a diversified portfolio of alliances, and using alliances to deter adversaries and restrain allies from initiating or escalating conflicts. I discuss each of these methods below.

In discussions of entanglement, it is often assumed that states are unable or unwilling to limit the scope of their alliance commitments. In essence, alliances are characterized as “blank checks” that can be cashed under any circumstance.⁴⁷

Numerous studies, however, show that alliances “are rarely blanket commitments of support,” but instead provide members with escape clauses or wiggle room regarding the conditions under which they must assist each other and the type of assistance they must provide.⁴⁸ These barriers to entanglement are especially common in alliances between major and minor powers (so-called asymmetric alliances) because strong states can limit their commitments to weaker allies while demanding unconditional loyalty in return.⁴⁹ In short, most alliance treaties are vague, so the specific obligations of each ally are left open to negotiation and are usually determined by relative power, with weaker allies giving up more of their autonomy than stronger allies.⁵⁰

Not surprisingly, all U.S. alliances are vague and contingent commitments. The North Atlantic Treaty requires only that each member “provide assistance” and “take action as it deems necessary … in accordance with their respective constitutional processes” if an ally is attacked. In other words, intervention is not automatic; it is contingent on the authorization of domestic veto players. Similarly, the Organization of American States (OAS) charter obliges members “to provide for common action on the part of member states in the event of aggression” without defining “aggression” or specifying what actions must be taken. Furthermore, the OAS charter states: “No State shall be required to use armed force without its consent.” Finally, all of the United States' bilateral defense pacts merely ask each side to “act to meet the common danger in accordance with its constitutional process.” In sum, U.S. alliance agreements contain loopholes that allow the United States to maintain its freedom of action. Case studies suggest that the United States often exploits these loopholes while preventing its weaker allies from doing the same.⁵¹

Because the international system lacks a central authority, states must provide for their own protection and jealously guard their military assets. When conflicts erupt, therefore, states will be tempted to “remain on the sidelines if vital interests are not threatened” or at least to avoid incurring significant costs fighting on behalf of others.⁵² In particular, states are likely to sidestep costly alliance commitments by providing minimal support for allies (e.g., providing air support or supplies but not ground forces, providing diplomatic support but not military support, etc.) or by abandoning allies altogether.

Sidestepping may not be a viable option for weak states that depend on a single great power for protection. But for the United States, a superpower with many allies, sidestepping is eminently feasible.⁵³ During the Cold War, the United States accounted for roughly 75 percent of allied defense spending, so the loss of any ally could be “accommodated without disastrously distorting, or even much affecting, the balance between America and Russia.”⁵⁴ The United States could thus “dissociate” itself from costly allied conflicts without suffering long-term costs.⁵⁵ Today, as the world's only superpower, the United States has even less need for staunch allies and, consequently, is more insulated from entanglement.⁵⁶ Although the United States may wage bloody wars to demonstrate U.S. resolve to adversaries and neutral countries, it is unlikely to incur major costs to display loyalty to allies that depend on U.S. protection and patronage for their survival.

As noted earlier, entanglement theory tends to assume that the more alliances a state joins, the more likely it will fall prey to entanglement. By contrast, freedom of action theory holds that great powers can mitigate their entanglement risk by maintaining a large and diversified portfolio of alliances.

With multiple allies scattered across many regions, the United States will likely face conflicting demands from different allies and will need to restrain its military activities in support of one ally so that it can satisfy others. Moreover, in most conflicts, only a few allies will be directly threatened and demand U.S. intervention. The vast majority of allies, by contrast, will favor U.S. restraint, because their security would suffer if the United States sunk resources into a peripheral region or escalated a faraway conflict into a global war—what British leaders have called “annihilation without representation.”⁵⁷

For this reason, Dominic Tierney explains, most allies can be expected to “pressure Washington to moderate its policy” and “provide political cover for U.S. compromise by sharing responsibility for concessions.”⁵⁸ In fact, Tierney suggests that the United States has been far more militaristic than its allies during the postwar era and that allies often have tried to restrain the United States from acting on its “escalatory bias.”⁵⁹ If this is correct, then a large alliance network may, on balance, help keep the United States out of costly wars, even if particular allies occasionally try to drag it into conflicts.

Entanglement theory assumes that alliances do not reduce conflict sufficiently to offset the increased liability inherent in security commitments. In fact, as discussed earlier, a key tenet of the theory is that alliances ignite spirals of hostility by provoking non-allies and emboldening allies.

On the other hand, the deterrence model of conflict, which holds that the best way to prevent war is to stand firm against enemies,⁶⁰ suggests that alliances prevent conflict by committing members to form a united front against aggression.⁶¹ Even when deterrence fails, alliances can facilitate peaceful settlement by encouraging belligerents to make smaller demands of each other; in other words, “the shadow of the alliance influences the bargaining stances of both challengers and targets” and expands the range of potential settlements.⁶² Alliances also “serve as vehicles to facilitate communication and transparency”⁶³ and as “tools of power management and control” that enable allies to resolve intra-allied conflicts peacefully and to restrain each other from engaging in reckless behavior.⁶⁴

As the most powerful member of the world's largest alliance network, the United States is especially well positioned to reap these advantages of alliances and practice “dual deterrence,” simultaneously deterring aggressors and restraining allies⁶⁵ Adversaries of the United States or one of its allies must consider the possibility that they will face a superpower backed by a posse of nearly seventy states if they engage in aggression. Meanwhile, U.S. allies know that what they do will be watched by an overwhelmingly powerful actor, so they are less likely to jockey with one another for military supremacy. Allies also know that their security depends on continued U.S. protection. As a result, the United States can run what Victor Cha calls a “powerplay”—that is, use its leverage over weaker allies to prevent them from initiating conflicts that threaten U.S. interests.⁶⁶

Freedom of action theory holds that great powers typically limit their alliance commitments to weaker states, fulfill them only when doing so does not threaten their vital interests, and use them to deter adversaries and allies from initiating or escalating conflicts. Freedom of action theory also maintains that a large and diversified alliance network can reduce the likelihood of entanglement by generating conflicting demands from different allies, thereby limiting the ability of any individual ally to goad their great power protector into conflicts against its interests.

If this perspective is correct, then one should not observe any clear-cut cases of U.S. entanglement, though entanglement dynamics (e.g., allies trying to persuade the United States to intervene in their conflicts, U.S. officials expressing concerns about U.S. reputation as a security guarantor, or alliances provoking adversaries or emboldening allies) may be present in some cases. Freedom of action theory does not predict that entanglement dynamics will be totally absent from U.S. foreign policy, but it does predict that such dynamics will not play a significant (i.e., a necessary or sufficient) role in explaining U.S. involvement in costly military conflicts. This prediction implies that, for every U.S. conflict, one should be able to show that the United States intervened or was attacked while in deliberate pursuit of national interests; that the United States successfully limited its military risks, actions, and costs; or both.

To analyze the prevalence of U.S. entanglement, I employ a method that, for lack of a better term, I call “active searching.” This method involves looking for evidence of a phenomenon across the universe of potential positive findings and has been used in studies on preemptive war and audience costs.⁶⁷

Specifically, I examine every MID (188 in total) in which the United States participated from 1948 to 2010.⁶⁸ This sample encompasses all possible cases of U.S. entanglement during that period, because the United States could only have been entangled in militarized disputes that it actually participated in. Of course, this sample is biased because it excludes cases of peace and, therefore, may omit many cases of non-entanglement. Consequently, it is suitable only for estimating the total number of cases of U.S. entanglement, not for assessing the number of cases of entanglement relative to the number of non-cases or for analyzing factors that make entanglement more or less likely.

The United States formed formal alliances with sixty-six countries from 1948 to 2014 (table 1).⁶⁹ Thirty-three are located in the Western Hemisphere and are allied to the United States through the OAS; twenty-seven are located in Europe and allied via NATO; five are located in Asia and tied to the United States through a “hub and spokes” system of bilateral defense pacts; and the Australia, New Zealand, United States Security Treaty (ANZUS) is a defense pact among those three countries. All of these alliances were established in the early years of the Cold War, although the OAS and NATO have added new members several times since. New Zealand withdrew from ANZUS in 1986, and Cuba was suspended from the OAS in 1962 and readmitted in 2009.

I make two additions to this list. First, I extend the dates of the U.S.-Taiwan alliance from 1954 to the present because the Taiwan Relations Act arguably constitutes “the functional equivalent” of a defense pact.⁷⁰ Second, I count Israel as a “de facto” ally because U.S. leaders, beginning with President John F. Kennedy in 1962, have pledged to defend Israel from attack.⁷¹

As noted earlier, to identify instances of entanglement, analysts must distinguish military actions intended to serve national interests from those driven by alliance concerns. This aim, in turn, requires analysts to start by identifying a state's national interests.

Scholars generally agree that the United States has pursued three broad interests throughout the postwar era: security, prosperity, and the projection of U.S. values.⁷² To advance each of these interests, U.S. policymakers have sought several more specific objectives (table 2). To maintain security, policymakers have sought to prevent foreign attacks against U.S. citizens and territory, contain the rise of great power rivals, and prevent the spread of weapons of mass destruction.⁷³ To ensure prosperity, U.S. officials have sought to maintain U.S. access to resources, transit routes, and markets abroad.⁷⁴ To project American values, U.S. leaders have promoted the international spread of democracy and capitalism;⁷⁵ and in the 1990s, preventing genocide arguably became a national interest, as it became “a central and insistent preoccupation in U.S. discourse” and “a permanent program requiring special doctrines.”⁷⁶

U.S. leaders have pursued these interests in diverse ways and with varying degrees of vigor, ranging from speeches to the use of force. I do not attempt to explain these variations. Instead, I aim only to identify militarized disputes in which alliances caused U.S. leaders to deviate from the strict pursuit of these interests.

To do so, I employ historical process tracing, scouring primary and secondary sources for evidence of entanglement in each of the 188 MIDs. In an online appendix, I provide a narrative for each MID and sort the cases using two coding schemes.⁷⁷ First, I code the effect of alliance politics on the level of U.S. military involvement in each conflict using the following scheme:

1. Alliances reduced the level of U.S. involvement.

2. Alliances are unnecessary to explain the observed level of U.S. involvement, and there is no evidence that they played a role in U.S. decisionmaking.

3. Alliances are unnecessary to explain the observed level of U.S. involvement, but alliance concerns featured in U.S. decisionmaking.

4. Alliances are necessary to explain the observed level of U.S. involvement.

Second, I code the overall level of U.S. military involvement in each dispute using the following scheme:

1. U.S. forces were uninvolved.

2. U.S. forces were placed at a higher risk of attack but did not take any actions of their own.

3. U.S. forces engaged in nonmilitary operations (e.g., resupply operations and transport for other countries' militaries).

4. U.S. forces engaged in military operations (e.g., shows of force, blockades, combat).

From this initial analysis, 5 conflicts (comprising 18 MIDs) stood out in terms of the influence of alliances on U.S. decisionmaking and the extent of U.S. military involvement:⁷⁸ the 1954–55 and 1995–96 Taiwan Strait crises; U.S. interventions in Indochina culminating in the Vietnam War; and the interventions in Bosnia and Kosovo in the 1990s. It is worth noting that other studies have flagged these cases as the most egregious cases of U.S. entanglement as well.⁷⁹ Later, I discuss each of these cases in detail.

As I explain further in the appendix, the empirical record does not contain other apparent cases of entanglement. In 9 MIDs—the 1958 Taiwan Strait crisis, the 1973 Arab-Israeli War, a 1976 decision to cave to Israeli demands to suspend arms sales to Saudi Arabia and Iran, military base negotiations with Greece in the 1980s, a 1999 Chinese missile test, the 2001 decision to sell arms to Taiwan, and crises between North and South Korea in 1999, 2004, and 2010—alliance concerns featured in U.S. decisionmaking, but the scope of U.S. military involvement and the risks to U.S. forces were limited.

Nineteen other MIDs involved Cold War–era military exercises or skirmishes on or near the borders between East and West Germany or North and South Korea. Although U.S. forces were deployed on those borders to defend allies from attack, U.S. policymakers perceived a clear national interest in holding the line there against communist forces. U.S. involvement was therefore driven by U.S. alignments with West Germany and South Korea, rather than by alliance obligations.

In 131 other MIDs, I could find no evidence that alliance obligations played any role in U.S. decisionmaking. Instead, American forces intervened, or were attacked, while in deliberate pursuit of predefined national interests: U.S. efforts to contain the Soviet Union, China, or both accounted for 58 of these MIDs; the 1991 and 2003 wars with Iraq account for another 20 MIDs; counter-terrorism operations, mostly conducted in and around Afghanistan, account for 14 MIDs; 11 other MIDs involved clashes in the 1980s with forces from Muammar al-Qaddafi's Libya, which sponsored terrorist attacks against U.S. citizens; 6 MIDs involved U.S. counterproliferation activities against North Korea and Iran; and 5 disputes stemmed from the 1979 Iranian Revolution and subsequent Iranian threats to close the Strait of Hormuz. In addition, U.S. forces intervened seven times in the Western Hemisphere variously to protect American citizens after coups, to restore democracy, or to stem drug trafficking into the United States. The United States was involved in 5 fishing disputes with various countries. In 1974 U.S. Customs officials arrested an American draft evader living in Canada. Later Liberian troops fired on the U.S. embassy in 1998 after a Liberian opposition leader took refuge there. There were also 4 bilateral disputes (2 with Libya in the 1980s, 1 with Canada in 1975, and 1 with the Soviet Union in 1965) on which I could not find any information.

In 4 other MIDs—U.S. escalation during the Korean War, the 1958 and 1961 Berlin crises, and the 1961 Laotian crisis—there is strong evidence that coalitions of allies helped dissuade U.S. leaders from escalating U.S. military involvement. In 7 other MIDs—the 1954 Nicaragua–Costa Rica border dispute, the 1956 Suez crisis, the 1962 Taiwan Strait crisis, the 1967 Six-Day War, the aftermath of the USS Pueblo Incident, and the 1983–84 Lebanon peacekeeping mission—the United States reneged on security commitments to an ally, restrained an ally from attacking a third party, or sided against an ally.

In sum, there are five cases (comprising 18 MIDs) that stand out as potential cases of U.S. entanglement. In the following sections, I examine each of these five cases in detail. I divide my discussion of each case into two parts. First, I discuss the effect of alliance politics on U.S. decisionmaking. Second, I discuss the role of U.S. national interests in U.S. decisionmaking. As will be shown, even in these five cases, the effect of alliance politics on U.S. decisionmaking is mixed, and the overall level of entanglement is limited.

In the mid-1950s, the United States faced possible entanglement when its de facto ally, the Nationalist Chinese government on Taiwan, confronted the People's Republic of China (PRC) over several offshore islands that Nationalist forces had occupied after fleeing the mainland during the Chinese civil war. On September 3, 1954, China began shelling Jinmen (known as “Quemoy” at the time), a small Nationalist-held island located one mile off the Chinese coast near the port of Xiamen. Two months later, China began bombing Nationalist garrisons on the Dachen Islands, a nine-square-mile archipelago located thirty miles off the coast of China's Zhejiang Province. And in January 1955, PRC troops seized Yijiangshan, an island just north of the Dachens. Throughout this period, PRC forces massed on the shores opposite Taiwan.

Although alliance politics precipitated the conflict—China began shelling Jinmen to deter the United States from allying with the Nationalists, a development that Chinese leaders mistakenly believed was imminent⁸⁰—the U.S. decision to respond by threatening China with military retaliation and by signing a formal alliance with the Nationalists was driven by well-established U.S. security interests. Nationalist attempts to entrap the United States in a larger conflict ultimately failed because the United States, under pressure from other allies, limited its commitments to the Nationalists while deterring the PRC from escalating its attacks.

The United States had not yet signed a defense pact with the Nationalists when the conflict began in September 1954, and the treaty signed by both sides in November did not cover the offshore islands and contained secret notes that established a U.S. veto over Nationalist military activity. As Secretary of State John Foster Dulles wrote to President Dwight Eisenhower: “[The treaty] stakes out unqualifiedly our interest in [Taiwan] and the [Penghus] and does so on a basis which will not enable the Chinese Nationalists to involve us in a war with Communist China.”⁸¹

Avoiding entanglement, however, turned out to be more difficult than Dulles predicted, because the Nationalists deployed tens of thousands of their soldiers on Jinmen and Mazu (known as “Matsu” at the time), a small archipelago located just north of Jinmen and ten miles off the Chinese coast. The Eisenhower administration believed that these troops were vital to Taiwan's security, even if the islands they were based on were not, and therefore attempted to compel the Nationalists to “redeploy and consolidate” these forces on Taiwan.⁸² The Nationalists, however, refused to budge without a firm U.S. security guarantee.

In January 1955, therefore, Eisenhower reluctantly traded a secret pledge to defend Jinmen and Mazu for a Nationalist withdrawal from the Dachens. The Nationalists then pocketed this American pledge, dug in their forces on Jinmen and Mazu, and rebuffed subsequent requests from exasperated U.S. officials to withdraw them, even turning down a U.S. offer to blockade China's coast opposite Taiwan in exchange. This uneasy situation persisted until late April, when China proposed peace talks and the crisis subsided.

Some scholars believe that if China had invaded the offshore islands, the United States would have upheld its secret pledge to defend them.⁸³ In this view, the United States escaped serious entanglement only because the PRC refrained from escalation. This claim, however, is questionable given that U.S. officials refused to publicize the pledge,⁸⁴ and then retracted it altogether during a meeting with Taiwan's premier, Chiang Kai-shek, in April 1955.⁸⁵

Even if the United States had maintained this pledge, moreover, other allies might have restrained the United States from intervening on Taiwan's behalf. As numerous declassified documents reveal, U.S. officials worried that defending the Nationalists would rupture relations with European allies, most of whom were horrified that the United States was risking war with China over tiny islands.⁸⁶ In fact, there is clear evidence that U.S. officials decided against attacking China in March 1955 because they feared such actions would disrupt the allied unity necessary to bring West Germany into NATO.⁸⁷ In addition, there is strong evidence that British leaders compelled Eisenhower to refrain from committing publicly to defend Jinmen and Mazu by threatening to withdraw British support for UN neutralization of the Taiwan Strait.⁸⁸

Finally, even if Eisenhower truly intended to defend Jinmen and Mazu, declassified PRC documents reveal that China refrained from invading them in large part because Chinese leaders feared that the United States would bomb the Chinese mainland in retaliation.⁸⁹ In short, successful deterrence precluded further U.S. entanglement.

The preponderance of evidence suggests that U.S. involvement in the conflict was driven not by alliance concerns but rather by two preexisting policies, both of which emerged in the wake of the Korean War and reflected core U.S. security interests.

The first was the United States' “policy of pressure” against China.⁹⁰ China's intervention in the Korean War in October 1950 convinced U.S. policymakers that the PRC was a Soviet proxy—“a colonial Russian government, a Slavic Manchukuo … driven by foreign masters,” in Dean Rusk's words⁹¹—and that the two communist powers were bent on world conquest.⁹² Based on this assessment, the Truman and Eisenhower administrations implemented a China policy that went “beyond containment” and aimed to subvert the PRC and rupture the Sino-Soviet alliance.⁹³ Taiwan was deemed “an essential weapon” in implementing this policy,⁹⁴ because it provided U.S. forces with a staging point and manpower for raids, blockades, and, if necessary, full-scale attacks against China.⁹⁵ Conversely, U.S. officials feared that if the PRC took Taiwan, it would allow the Soviet Union to build military bases on the island.⁹⁶

Thus, as China attacked one Nationalist-held island after another in 1954– 55, a consensus emerged among U.S officials that China was testing American resolve, that failure to respond would spur further Chinese expansion, and that such expansion would rupture the U.S. defense perimeter in Asia if it spread to Taiwan.⁹⁷ “Communist probing will go on,” Secretary of State Dulles explained at a pivotal National Security Council (NSC) meeting, “until the United States decides to ‘shoot off a gun’ in the area.”⁹⁸ In sum, the decisive impetus for U.S. involvement, Robert Accinelli writes, “came not from any sentimental attachment to Chiang Kai-Shek or the Nationalists … but rather from the pursuit of American national security interests, which were enmeshed with images of an aggressive Communist China linked to the Soviet Union.”⁹⁹

The second policy guiding U.S. behavior was Eisenhower's “New Look” strategy. After the Korean War, Eisenhower decided to cut U.S. military costs by substituting nuclear coercion for conventional forces. By mid-1954, however, this policy faced declining support at home because of a deadly nuclear test near Japan and eroding credibility abroad resulting from U.S. inaction in Indochina. Eisenhower's desire to stem these trends and “ratify the New Look” encouraged him to issue veiled nuclear threats during the Taiwan crisis and “disinclined [his administration] to look for a peaceful resolution.”¹⁰⁰

In sum, the extent of U.S. entanglement in this case was ultimately limited by the prominent role of national interests in U.S. decisionmaking; the steps Eisenhower took, under pressure from allies, to limit U.S. liability for Nationalist security; and the conflict-dampening effects of U.S. security guarantees.

For two decades, the United States poured military forces into Indochina. During that time, nearly 60,000 Americans and roughly 2 million Vietnamese, Laotians, and Cambodians lost their lives.¹⁰¹ Alliance concerns featured in U.S. policy debates and may have delayed U.S. withdrawal from Vietnam. In general, however, the effect of alliance politics on U.S. decisionmaking was mixed. U.S. intervention and escalation were driven primarily by a perceived national interest in containing communism and Chinese and Soviet influence in Indochina. This interest emerged prior to, and existed independent of, the formation of the United States' Cold War alliance network. Moreover, claims that alliance credibility concerns drove U.S. involvement are belied by the fact that the United States waited to intervene in Vietnam until after France, a U.S. ally, had been defeated there and then steadily escalated its involvement in the face of explicit allied opposition.

The United States never formally allied with South Vietnam and, in fact, cut aid and backed coups there when South Vietnamese leaders failed to commit fully to the struggle against the Viet Minh, the communist organization that led the independence movement against French colonialism, governed North Vietnam, and sought to unify all of Vietnam under its rule.¹⁰² Declassified documents reveal that top U.S. officials favored escalating U.S. military involvement in Vietnam “government or no government”—that is, with or without a viable and willing South Vietnamese regime.¹⁰³

In 1955 the United States created the Southeast Asian Treaty Organization (SEATO) to rally anticommunist forces in the region and provide a rationale for U.S. intervention. France and Britain, however, vetoed SEATO involvement in Vietnam, and South Vietnam itself was barred entry to the organization by the Geneva accords. As a result, SEATO “wasted away as little more than a hollow pact” and did not shape U.S. policy in the conflict.¹⁰⁴

Other alliances, however, may have entangled the United States in Vietnam. First, some scholars trace the United States' initial military involvement to European alliances. According to this “Europe made us do it” hypothesis, France entrapped the United States by making French support for the European Defense Community (EDC) contingent on U.S. support for French forces in Indochina.¹⁰⁵

This argument, however, is difficult to sustain given that the United States, often under pressure from other allies, repeatedly rebuffed French requests for U.S. military intervention. Most notable, the Eisenhower administration rejected French pleas for military assistance during the climactic battle of Dien Bien Phu, which occurred just months before the French parliamentary vote on the EDC and in which Viet Minh forces surrounded a French garrison of 14,000 troops and shelled it for fifty-five days, killing roughly 2,000 French soldiers and catalyzing France's ultimate withdrawal from Indochina.¹⁰⁶

Another problem with attributing U.S. intervention in Vietnam to French arm-twisting is that, as Fredrik Logevall points out, France's war in Indochina drained resources from European defense. Thus, “had senior U.S. officials cared only or mostly about the EDC and NATO, they would have insisted that France rid herself of the Indochina burden by promising complete and total independence for Vietnam.”¹⁰⁷ Instead, U.S. officials “pushed [the French] to fight and then left them to die,”¹⁰⁸ perhaps the opposite of what one would have expected U.S. officials to do if they were concerned primarily with gaining French support for European defense or maintaining U.S. credibility as a reliable security guarantor. Tellingly, after France decided to withdraw from Indochina, U.S. Secretary of State Dulles described Dien Bien Phu as a “blessing in disguise” because “now we enter Vietnam without the taint of French colonialism.”¹⁰⁹

A second potential source of entanglement stemmed from the U.S.-Japan alliance. Some scholars claim that the United States “crossed the most crucial threshold on the road to the Vietnam War” when it guaranteed Japan's security in the 1950s, a goal that, in turn, required the United States to safeguard Japanese access to Southeast Asian markets and resources.¹¹⁰

Yet, although the United States certainly sought resources and markets for Japan, it procured them primarily by trading with and investing in Japan, not by invading Vietnam. In fact, in 1957 the United States scuttled a Japanese proposal to establish a Southeast Asian Development Fund to invigorate the region's economies.¹¹¹ The Eisenhower administration then upgraded the U.S.-Japan alliance and oriented the Japanese economy around the American market.¹¹² As a result, U.S.-Japanese trade was three times greater than Japan's trade with all of Southeast Asia combined throughout the 1950s and 1960s.¹¹³ By the time President Lyndon Johnson decided to escalate U.S. involvement in Vietnam, Japan ranked among the wealthiest countries in the world and was clearly not dependent on Southeast Asian resources or markets.

Third, many studies claim that concerns about maintaining U.S. credibility as a security guarantor contributed to Johnson's decision to escalate U.S. military involvement in Vietnam and to Johnson's, and then Richard Nixon's, decisions to delay U.S. withdrawal.¹¹⁴ The documents most commonly cited in support of this argument are drafts of a memo written in late 1964 and early 1965 by John McNaughton, the assistant secretary of defense for international security affairs, in which he asserted that “70 percent” of the U.S. war effort was “to avoid a humiliating U.S. defeat (to our reputation as a guarantor)” and that it was “essential” for the United States to persist in Vietnam, like a “good doctor” striving to save a patient, to “avoid harmful appearances that will affect the judgments by, and provide pretexts to, other nations regarding how the U.S. will behave in future cases.”¹¹⁵ Although McNaughton's memo provides the clearest articulation of U.S. alliance credibility concerns, such concerns also show up in at least a dozen other declassified documents and memoirs.¹¹⁶

As Francis Gavin has pointed out, however, “the credibility argument has long been simultaneously the most convincing and most suspect” because U.S. policymakers were painfully aware, certainly by late 1964 if not earlier, that most allies staunchly opposed U.S. involvement in Vietnam.¹¹⁷ Allies worried not only that the United States would sap its strength by intervening, but also that American leaders would pressure them to contribute to the struggle.

On numerous occasions, therefore, allies tried to restrain the United States.¹¹⁸ These attempts ultimately failed, of course, but they did make clear to American officials that allied confidence in the United States did not depend on, and might even be undermined by, U.S. escalation in Vietnam.¹¹⁹ As Undersecretary of State George Ball wrote in an internal memo, “The principal anxiety of our NATO allies is that we have become too preoccupied with an area which seems to them an irrelevance and may be tempted to neglect our NATO responsibilities. … By and large, therefore, they would … regard a compromise solution in South Viet-Nam more as new evidence of American maturity and judgment than of American loss of face.”¹²⁰

Even staunch advocates of U.S. escalation agreed with this assessment. For example, in October 1964 William Bundy, the assistant secretary of state for East Asian and Pacific affairs, drafted a memo in which he refuted various arguments for U.S. withdrawal but conceded that “none [of America's allies] are likely to feel that [U.S. abandonment of South Vietnam] really proves that we cannot be counted on in their individually different situations, or that they themselves cannot handle these situations with our help as needed.”¹²¹ A month later, a major NSC report reached the same conclusion: “Within NATO (except for Greece and Turkey to some degree), the loss of South Vietnam probably would not shake the faith and resolve to face the threat of Communist aggression or confidence in us for major help. … In other areas of the world, either the nature of the Communist threat or the degree of U.S. commitment or both are so radically different than in Southeast Asia that it is difficult to asses the impact.”¹²² The NSC report then asserted, “We must maintain, particularly to our key NATO allies, the picture of a nation that is strong and at the same time wise in the exercise of its power. … European allies … could become seriously concerned if we get ourselves involved in a major conflict that degraded our ability to defend Europe.”¹²³

Of course, Johnson ultimately involved the United States in a major conflict in Vietnam. It is not clear, however, that alliance concerns drove his decisionmaking. To begin, several of Johnson's closest advisers subsequently rejected McNaughton's assertion that 70 percent of U.S. war aims was to maintain credibility as a security guarantor; Robert McNamara, Johnson's secretary of defense, stated that McNaughton lacked sufficient access to the inner decisionmaking circle to make such an assessment; and William Bundy decried the importance historians have attached to McNaughton's memo, saying that “it isn't serious evidence” and proves only that bureaucrats “could write a memorandum at the end of the day to themselves.”¹²⁴ More important, as Logevall's detailed study of Johnson's decisionmaking concludes, Johnson was “less inclined than his foreign-policy aides to see the war in terms of its international implications, and thus he likely did not frame the issue in the manner that McNaughton had.”¹²⁵ It is possible that Johnson, and later Nixon, eventually became concerned about allied credibility and delayed U.S. withdrawal as a result. But this claim is difficult to square with the steady stream of assessments, made by top officials in both administrations, that the United States was straining its relationships with key allies by fighting on in Vietnam.¹²⁶

Further casting doubt on the argument that alliances entangled the United States in the Vietnam War is the mass of evidence showing that U.S. policymakers, from the late 1940s to early 1970s, perceived a clear national interest in containing the spread of communism and Chinese and Soviet influence in Vietnam.¹²⁷ In 1949 President Harry Truman's administration concluded that Ho Chi Minh was “an outright commie”¹²⁸ and that “Southeast Asia [was] the target for a coordinated offensive directed by the Kremlin.”¹²⁹ Then, in 1950 China and the Soviet Union recognized Ho's government and the Korean War erupted. By the time Truman left office, it was an article of faith among U.S. officials that the communist powers were on the move, that Vietnam was their next target, and that a communist victory there would tip the global balance of power against the United States.¹³⁰

This perceived vital interest in containing communism in Vietnam was passed down to successive U.S. administrations. Eisenhower officials deemed it their “top priority in foreign policy” and therefore exerted great effort in building a sovereign “South Vietnam” on the bones of the French colonial regime.¹³¹ Under Kennedy, the preservation of South Vietnam was “simply a given, assumed and unquestioned,”¹³² and officials “undertook no wide-ranging discussion … of the struggle's importance to American national security.”¹³³ Instead, policy debates “centered on how to save Vietnam, not whether to save it,”¹³⁴ and Kennedy, whose Cold War credentials had been marred by the Bay of Pigs fiasco, nonintervention in Laos, the Vienna summit, and the erection of the Berlin Wall, ultimately doubled down on Eisenhower's policies and increased from 600 to 16,000 the number of U.S. military “advisers” in Vietnam.¹³⁵

Similarly, many studies and government documents suggest that Johnson's decision to bomb North Vietnam and deploy troops in the South was driven by a “genuine fear … that giving the Communists a free hand in Vietnam … would lead to a Soviet-American and/or Sino-American confrontation”¹³⁶ and spark communist “wars of national liberation” all over the world.¹³⁷ It should be noted that other studies question whether Cold War imperatives actually drove Johnson's decisionmaking. These studies, however, do not link Johnson's decision to alliance politics, but rather to elements of Johnson's personality—his proclivity to view “the war as a test of his own manliness,”¹³⁸ his fear that retreat would tarnish his legacy and undermine his domestic agenda,¹³⁹ his demand for internal consensus,¹⁴⁰ and his membership in the “G.I. generation”¹⁴¹ and related tendency to equate retrenchment with Munich-style appeasement.¹⁴²

A similar blend of geopolitical interests and presidential ego is typically used to explain Nixon's dogged pursuit of “peace with honor.”¹⁴³ A long-time proponent of the domino theory, Nixon believed that “if we fail to end the war in a way that will not be an American defeat, and in a way that will deny the aggressor his goal, the hawks in Communist nations will push for even more and broader aggression.”¹⁴⁴ To maintain domestic support for waging the Cold War and a reputation for resolve vis-à-vis the Soviet Union and China, Nixon sought to commit the United States to “permanent war” in Vietnam,¹⁴⁵ or at least to ensure a “decent interval” between the United States' withdrawal and South Vietnam's collapse.¹⁴⁶ To that end, his administration bolstered South Vietnam's military, invaded Cambodia, and bombed and blockaded North Vietnam. “I'm the only President… who'd have the guts to do what we're doing,” he declared.¹⁴⁷

After the breakup of Yugoslavia in 1991, the multiethnic Socialist Republic of Bosnia and Herzogovina held a referendum on independence. Bosnian Serbs, however, boycotted this referendum, established their own republic, and mobilized forces backed by the Serbian government of Slobodan Milošević to secure Serb populations within Bosnia and Herzegovina. Fighting, mostly between Bosnian Serbs on the one hand and Bosnian Croats and Muslims on the other, erupted in April 1992. In response, European countries imposed an economic embargo on Serbia and Montenegro, deployed 9,000 peacekeepers in Bosnia, and crafted several failed peace accords. They also resisted U.S. demands to authorize air strikes until the summer of 1995, when the United States led NATO in the largest military operation in its history to date.

Alliance concerns catalyzed the U.S. decision to intervene, but national interests also played important roles, and for two years, from mid-1993 to mid-1995, alliance politics actually restrained the United States from striking. Moreover, U.S. military involvement was ultimately limited—allies provided all of the ground forces and most of the postconflict peacekeeping forces and did not differ dramatically from the intervention plan that President Bill Clinton's administration had been pushing on U.S. allies since coming to office.

Alliance politics initially restrained the United States from intervening in Bosnia. On May 1, 1993, the Clinton administration began pushing allies to lift the arms embargo and authorize air strikes against Serb forces.¹⁴⁸ NATO allies, however, opposed this so-called lift-and-strike option because it risked escalating the conflict and jeopardizing the safety of their peacekeepers.¹⁴⁹ France and Britain, in particular, threatened to pull their troops out of Bosnia if the United States implemented “lift-and-strike” and effectively gave the United States a choice: stay out of the conflict or contribute troops to the UN peacekeeping mission.¹⁵⁰ Clinton chose the former, explaining later: “I didn't want to divide the NATO alliance by unilaterally bombing Serb military positions … and I didn't want to send American troops there.”¹⁵¹ Declassified documents show that other American officials shared Clinton's concern that unilateral military action would rupture NATO.¹⁵²

Nevertheless, alliance obligations ultimately played a necessary role in compelling U.S. intervention. In December 1994, Clinton pledged to send U.S. ground troops to evacuate UN peacekeepers if their presence became unsustainable, a pledge he hoped would persuade European allies to keep their forces in Bosnia and, thereby, allow U.S. forces to stay out.¹⁵³ This plan backfired in mid-1995 when Bosnian Serb forces overran several UN “safe areas,” took hundreds of peacekeepers hostage, and slaughtered thousands of civilians. At that point, some allies threatened to withdraw their peacekeepers, and the French president, Jacques Chirac, publicly challenged the United States to intervene. The Clinton administration determined that NATO's credibility was at stake and that bombing Serb forces was preferable to breaking a promise to allies or sending 25,000 U.S. soldiers to cover a UN retreat.¹⁵⁴ Operation Deliberate Force commenced in August 1995.

Although alliance concerns contributed to the U.S. decision to intervene, national interests also played a necessary role. First, humanitarian interests were critical to keep the conflict on Clinton's agenda. During the 1992 presidential campaign, Clinton advocated air strikes and criticized the George H.W. Bush administration for “turning its back on violations of basic human rights.”¹⁵⁵ Once elected, Clinton faced relentless pressure to live up to this campaign rhetoric from members of Congress,¹⁵⁶ from humanitarian activists within his administration,¹⁵⁷ and from the media, which beamed home horrific images and published numerous articles comparing the conflict to the Holocaust and U.S. inaction to the appeasement of Adolph Hitler.¹⁵⁸ These domestic actors may not have been able to force the administration to intervene, but they raised the domestic political costs of doing nothing.

Second, security interests encouraged the administration to play more than a token role in addressing the conflict. In the early 1990s, U.S. policymakers sought to preserve American primacy by, among other things, smothering European attempts to forge independent security policies.¹⁵⁹ This hegemonic impulse is clearly evident in U.S. decisionmaking on Bosnia. For example, at a principals meeting on February 5, 1993, Clinton declared that a “failure to [act] would be to give up American leadership.”¹⁶⁰ When a senior official countered that “we could say this is a European problem and they should take responsibility for enforcing a settlement,” Clinton cut him off, explaining: “We can't do that without giving up our whole position in the world.”¹⁶¹ Senior officials reiterated this rationale in numerous meetings and memos on Bosnia,¹⁶² and a major study on the issue concludes that U.S. intervention was driven primarily by a perceived need to prevent the emergence of an independent European center of power.¹⁶³

Given these interests, it is perhaps not surprising that the Clinton administration's early plans for Bosnia called for U.S. military action and were similar to what Operation Deliberate Force ultimately entailed—European ground operations, U.S. air strikes and, in the event of a peace settlement, U.S. contributions to a multinational peacekeeping force.¹⁶⁴ In 1995 the United States provided 45 percent of coalition aircraft and flew 65 percent of the sorties, but did not deploy ground forces until after the signing of the Dayton accords, instead relying on NATO and Croat troops to conquer and hold territory.¹⁶⁵ The United States contributed 18,400 of the 54,000 postconflict peacekeepers, but cut its forces by roughly 50 percent each year until 2004, when the European Union took over peacekeeping operations entirely.¹⁶⁶

On May 22, 1995, the Clinton administration approved a visa for Taiwanese President Lee Teng-hui to visit the United States to attend his graduate school reunion at Cornell University. The visit set off a dispute between the United States and China that climaxed in March 1996, when China conducted military exercises and missile tests near Taiwan, and the United States responded by deploying two carrier battle groups to waters east of Taiwan. On March 23, President Lee was reelected and China ceased its military exercises.

The available evidence suggests that alliance concerns are necessary and perhaps even sufficient to explain U.S. actions, but the extent of U.S. military involvement was limited to a show of force, which was made only after China and the United States had privately reassured each other that they would not escalate the conflict.

Most studies conclude that the United States deployed the carrier battle groups to convey U.S. resolve to the PRC and to maintain a reputation for loyalty to its security partners.¹⁶⁷ The key decision was made on March 9, 1996, in a meeting of the foreign policy principals, and “everyone at the meeting agreed that the United States should counteract China with a show of force, one that would reassure both Taiwan and American allies in Asia.”¹⁶⁸ Ten days later, the House and Senate, which had been bombarding the Clinton administration with calls to defend Taiwan's threatened democracy, passed resolutions declaring the “sense of Congress” to be that the United States, “in accordance with the Taiwan Relations Act… should assist in defending [Taiwan] against invasion, missile attack, or blockade by the People's Republic of China.”¹⁶⁹

Some scholars claim that the U.S. show of support for Taiwan caused China to back down.¹⁷⁰ If true, then one could interpret the outcome of the crisis as a case of successful extended deterrence. Yet, other sources cast doubt on this view by showing that China never planned to escalate hostilities absent additional Taiwanese moves toward independence. Indeed, China's actions during the crisis were “carefully nonprovocative.”¹⁷¹ The People's Liberation Army did not fill the Taiwan Strait with air superiority fighters or an armada of warships; its ships and planes did not approach or lock their radar on U.S. ships; and Chinese diplomats privately assured American officials that China would not attack Taiwan.¹⁷² Beijing also announced the location of its missile firings in advance to ensure that foreign vessels would not be in the area, and Taiwanese spies revealed that the missiles China launched did not carry warheads.¹⁷³ China's caution suggests that the risk of major conflict was low and that the U.S. show of force did not place U.S. forces at a high risk of attack. China's cautious behavior, however, also casts doubt on the claim that the U.S. display of force was critical to prevent China from escalating.

There is little evidence that U.S. policymakers believed that vital U.S. interests were at stake in the crisis. Although the United States retained an interest in containing rival powers, it is not clear that the Clinton administration intervened in the 1995–96 crisis to advance this interest, though this claim is difficult to assess without access to the internal records. Some U.S. policymakers backed Taiwan because it is a democracy, but most studies suggest that defending democracy was secondary to the goal of maintaining alliance credibility.¹⁷⁴

The lack of clear national interests in the conflict helps explain why American officials took pains to limit U.S. military involvement. China conducted missile tests and military exercises in July, August, and November 1995, but the Clinton administration did not respond militarily until March 1996. Meanwhile U.S. officials, including President Clinton, privately told Chinese leaders that the United States did not support Taiwanese independence or UN membership and that future visits to the United States by Taiwanese leaders would be rare. Administration officials later explained that these steps were taken to defuse the conflict and forestall U.S. military involvement.¹⁷⁵

After the breakup of Yugoslavia in 1991, ethnic Albanians in Kosovo declared independence from Serbia. In response, Serbia deployed military forces in Kosovo and, at times, forcibly expelled ethnic Albanians from the territory. Throughout the 1990s, clashes occurred between Serb forces and the Kosovo Liberation Army (KLA), an ethnic Albanian paramilitary group. In 1998 and early 1999, several high-profile massacres of Albanian civilians took place and galvanized international intervention. On March 24, 1999, NATO began a seventy-eight-day bombing campaign against Yugoslavia to stop Serb forces from attacking or expelling Albanian civilians.

Alliance politics featured in U.S. decisionmaking, but it is not clear that they were a decisive cause of U.S. intervention. As in the Bosnian case, humanitarian concerns shaped U.S. decisionmaking, and allies initially tried to restrain the United States from attacking. Unlike in Bosnia, however, there were no allied forces on the ground prior to the decision to intervene. Consequently, the main mechanism of U.S. entanglement in Bosnia—an American commitment to evacuate allied peacekeepers—was not present in Kosovo. Concerns about NATO's credibility may have encouraged some initially reluctant U.S. officials to endorse air strikes, but there is strong evidence that humanitarian concerns, plus assurances that allies would handle most of the postconflict peacekeeping, were more important.

Before January 1999, the Clinton administration was divided on whether to intervene militarily in Kosovo. Supreme Allied Commander Gen. Wesley Clark and senior State Department officials, led by Secretary of State Madeleine Albright, advocated threatening Yugoslavia with air strikes, but National Security Adviser Samuel Berger and top Pentagon officials, including Secretary of Defense William Cohen, rejected this option, fearing that issuing military threats would embroil the United States in a Vietnam-style quagmire.¹⁷⁶

Concerns over NATO's credibility may have played a necessary role in breaking this internal deadlock. In October 1998, NATO threatened air strikes against Yugoslavia if it failed to scale back Serb security operations in Kosovo. When Serb forces launched an offensive in the winter of 1998/99 in retaliation for attacks by the KLA, Albright and other State Department officials argued that the United States had to retaliate to uphold NATO's credibility. In subsequent interviews, some officials claimed that this argument helped persuade reluctant Pentagon officials to endorse military intervention.¹⁷⁷

On the other hand, Pentagon officials may have been less concerned with NATO's credibility than with garnering allied contributions to the anticipated postconflict peacekeeping mission. As Stefano Recchia's detailed study of U.S. decisionmaking concludes, “What finally persuaded the Pentagon and especially the JCS to agree to Albright's proposal, however, was their understanding that NATO was now locked in and the United States would not be left with a heavy postwar stabilization burden. After the latest Contact Group meeting on January 22, the Joint Chiefs were confident that the Western Europeans fully supported the new strategy and could be relied upon to generate the majority of troops for Kosovo's long-term stabilization.”¹⁷⁸ In other words, alliances encouraged U.S. military action by making it less costly. It is questionable whether this enabling mechanism truly constitutes entanglement.

Moreover, NATO allies restrained the United States from striking on several occasions.¹⁷⁹ Even after a high-profile Serb massacre of Albanian civilians at Racak in January 1999, European NATO members refused to endorse the use of force without a final attempt at diplomatic negotiations.¹⁸⁰ For that reason, an extensive study on the conflict concludes, “allied hesitancy” was the “main obstacle” to military action and functioned as a “useful corrective to the inclination of some Clinton administration officials to lash out violently against Milošević.”¹⁸¹

Humanitarian interests—specifically, a desire to prevent mass slaughter as had occurred in Bosnia—played a decisive role in driving senior U.S. officials to advocate intervention.¹⁸² As Alex Bellamy argues, there was a “direct correlation” between the amount of violence in Kosovo and top leaders' commitment to intervention, a trend suggesting that humanitarian concerns fundamentally affected the U.S. decisionmaking process.¹⁸³ For example, when Serb forces massacred dozens of women and children in Gornji Obrinje in September 1998, Berger became more amenable to military action, telling colleagues that the Serbs had breached the “atrocities threshold.”¹⁸⁴ Similarly, the Racak massacre “transformed the West's Balkan policy as singular events seldom do”¹⁸⁵ and “altered the dynamics” of the NSC.¹⁸⁶ Indeed, just hours before the massacre the NSC had decided against military intervention; but when news of the killings reached Washington, the NSC reconvened and endorsed Albright's proposal threatening NATO air strikes.¹⁸⁷

American concerns about entangling alliances are as old as the Republic itself. During the post-World War II era, however, there have been only five ostensible episodes of U.S. entanglement, and even these cases are questionable. The case in which alliance obligations had the largest impact on U.S. decisionmaking (the 1995–96 Taiwan Strait crisis) entailed minimal military action, and the case that entailed the most military action (the Vietnam War) contained only a marginal role for alliance politics in U.S. decisionmaking. In the other three cases (the 1954–55 Taiwan Strait crisis and the wars in Bosnia and Kosovo), both the effect of alliance obligations on U.S. policy and the costs suffered by U.S. forces were moderate. And beyond these cases, entanglement was virtually nonexistent in U.S. foreign policy.

Against this limited evidence of entanglement are numerous cases in which alliances restrained the United States. Allies dissuaded the United States from escalating the Korean War and crises in Laos and Berlin, and struggled in vain to prevent the United States from entering or escalating other conflicts, the 2003 Iraq War being only the latest major example. Indeed, instances of alliance-induced restraint are evident even within the five cases of entanglement discussed above: in the 1954–55 Taiwan Strait crisis, concerns about European alliances discouraged U.S. policymakers from bombing the Chinese mainland and publicly committing to defend Jinmen and Mazu; in the Vietnam War, allies impeded U.S. entry into the war and then repeatedly implored the United States to get out; and in Bosnia and Kosovo, U.S. allies initially restrained the United States from lashing out violently and then provided all of the NATO ground forces and most of the postconflict peacekeepers for the eventual operations.

There also are several cases in which the United States sidestepped inconvenient alliance commitments, restrained an ally from attacking a third party, or openly sided against an ally—and this list could probably be expanded by looking within other cases, including the five ostensible cases of entanglement. As explained earlier, the United States blatantly retracted a pledge to Taiwan to defend Jinmen and Mazu in 1955, refused to save the French at Dien Bien Phu in 1954, delegated ground operations and most of the postconflict peacekeeping in Bosnia and Kosovo to allies, and waited for eight months and the receipt of private security assurances before responding militarily to China's provocative behavior near Taiwan in 1995–96.

In sum, the empirical record shows that the risk of entanglement is real but manageable and that, for better or worse, U.S. security policy lies firmly in the hands of U.S. leaders and is shaped primarily by those leaders' perceptions of the nation's core interests. When the United States has overreached militarily, the main cause has not been entangling alliances but rather what Richard Betts calls “self-entrapment”—the tendency of U.S. leaders to define national interests expansively, to exaggerate the magnitude of foreign threats, and to underestimate the costs of military intervention.¹⁸⁸ Developing a disciplined defense policy therefore will require the emergence of prudent leadership, the development (or resurrection) of guidelines governing the use of force,¹⁸⁹ the establishment of domestic institutional constraints on the president's authority to send U.S. forces into battle, or some combination of these.¹⁹⁰ Scrapping alliances, by contrast, would simply unleash the United States to act on its interventionist impulses while leaving it isolated diplomatically and militarily.

To be sure, certain alliances may need to be revised or dissolved as circumstances change. Past performance is no guarantee of future results, and U.S. entanglement risk may shift over time. For example, China's development of antiaccess/area-denial capabilities may substantially increase the risks to the United States of maintaining alliance commitments in East Asia.¹⁹¹ Conversely, U.S. allies may be able to use similar capabilities to defend themselves and thereby allow the United States to maintain alliance commitments while limiting risks to U.S. forces.¹⁹² This study does not account for such emerging trends and, therefore, cannot rule out the possibility that the U.S. alliance network will need to be revised in the future.

What this study does suggest, however, is that such revisions should be modest. The historical record shows that allies often help keep U.S. troops at home not only by bearing some of the burden for U.S. wars, but also by encouraging the United States to stay out of wars altogether. Large-scale retrenchment would sacrifice these and other benefits of alliances while doing little to compel U.S. leaders to define national interests modestly or choose military interventions selectively. How to accomplish those goals will continue to be the subject of debate, but those debates will be more productive if they focus on domestic culprits rather than foreign friends.

For comments on earlier drafts, the author thanks Brendan Rittenhouse Green, Jonathan Markowitz, Daryl Press, Andrew Radin, Stefano Recchia, Sebastian Rosato, Jeffrey Taliaferro, Stephen Walt, William Wohlforth, and the anonymous reviewers.