Chapter 1 A Theory of Inadvertent Expansion

Expansion: Intentional and Inadvertent

The broad outcome this book seeks to explain, or dependent variable, is “expansion,” which I define narrowly as the coercive acquisition of foreign territory that is intended to be long term or permanent for the expanding state. Expansion, as used here, includes the gain of territory through armed conquest or political annexation. This definition excludes largely voluntary transfers of territory, such as when states engage in treaty-based territorial exchange or the adjustment of shared borders.1 It also excludes purchases or leases of territory that do not include an important coercive component, such as when the United States purchased Alaska from Russia in 1867, as well as various forms of “domestic expansion,” such as territorial expropriation, ethnic cleansing, counterinsurgency, and other pacification campaigns.2 Finally, it excludes forcible acquisitions of territory that are intended to be temporary or provisional in nature, such as military occupations or, in its modern guise, “neotrusteeship.”3

The definition of expansion this book employs differs from others that are more broadly conceptualized. Fareed Zakaria, for instance, understands the term to include not only imperialism but also “an activist foreign policy that ranges from attention to international events to increases in diplomatic legations to participation in great power diplomacy.”4 The problem with such an inclusive definition is its conceptual clarity and precision: it becomes so broad as to lose much of its meaning.5 A more circumscribed definition, such as the one employed here, mitigates this problem.

Expansion is related to, but distinct from, a number of concepts in international relations scholarship with which it is often conflated. For instance, expansion is not identical with “imperialism,” since expansion can also consist of the extension of a state’s national borders, and, it only covers the initial acquisition of foreign territory, not its ongoing administration.6 Expansion is also not the same thing as “war,” but it often includes the use of force and regularly precipitates or follows from the occurrence of war.7 Expansion is not identical to “power-maximization,” as this concept refers not merely to extended territorial control but also to a variety of other policies intended to increase the relative economic and military capabilities of a state.8 Finally, expansion is distinct from the somewhat opaque concept of “rise,” which most often refers to the secular increase in economic and military capabilities of one state relative to others.9

Most international relations scholarship tends to treat expansion as if it were a singular phenomenon: the largely strategic and leader-directed extension of the state’s territorial domain. But this is not always the case. For instance, Colin Elman draws a distinction between what he refers to as “manual and automatic expansion,” the former consisting of “conscious bids for hegemony” and, the latter, “incremental, localized attempts to expand, with a view to immediate opportunities . . . [not] with the deliberate purpose of becoming the dominant state in the international system.”10 Similarly, Randall Schweller argues that “we must distinguish between aggressive and nonaggressive expansion and between safe and risky expansion. All states . . . can be expected to engage in nonaggressive expansion as well as cheap, relatively safe expansion on the periphery.”11 These are welcome distinctions, to be sure. But because the concepts are not thoroughly developed, we are not given much of an indication of when to expect one form over the other.12 And, more importantly for our purposes, both forms remain largely leader centric.

The primary distinction I draw is between what I refer to as “intentional expansion” and “inadvertent expansion.” Intentional expansion is territorial expansion that is planned and ordered by the central leadership of a state or empire and is executed in accordance with those plans. Thorough planning for expansion by state leaders will include where it will take place, when it will take place, how much territory to acquire, and how it will be acquired. Intentional expansion is exemplified in many cases of territorial expansion that are well known in the historical record and in international relations scholarship, such as the German conquest of Poland in 1939 and Japan’s conquest of much of Southeast Asia in 1941–42.

In contrast, inadvertent expansion is territorial expansion that is planned and executed by actors on the periphery of a state or empire without the authorization of central leadership in the capital. A case of territorial expansion is inadvertent if the peripheral actor who carries it out lacks prior orders, permission, or authority from their superiors in the capital or, in some cases, acts against direct orders to the contrary. Inadvertent expansion is exemplified by the case that opened this book—the British conquest of Sind in 1843—as well as those that populate later chapters (among many others).

It is important to note that this idea of inadvertent expansion is not entirely new, particularly in the history of empire. In summarizing this literature, Michael Doyle refers to a whole category of theories as “peri-centric,” which root the sources of imperialism in the periphery.13 In the history of the British Empire, this includes foundational works by John Gallagher and Ronald Robinson on the “Imperialism of Free Trade” and informal empire, John Galbraith on the “turbulent frontier” and the “man on the spot,” and David Fieldhouse on “colonial sub-imperialism.”14 A number of historians have also taken a more interactionist view of the relationship between center and periphery in producing imperialism, such as Robinson, John Darwin, and Ronald Hyam.15

In international relations, specifically, only Paul MacDonald has sought to articulate a concept similar to inadvertent expansion, which he calls “autonomous expansion.” MacDonald defines autonomous expansion as the “political expansion of overseas empires without prior authorization or approval from officials in the metropole.”16 This is virtually equivalent to inadvertent expansion, but I do not adopt MacDonald’s terminology for two reasons. First, the term “autonomous” has a number of meanings, including self-governing, independent, and operating without human control—none of which exactly capture what occurs in cases of inadvertent expansion.17 And having multiple meanings associated with the term can hinder conceptual clarity, leading readers to ask, “autonomous in what sense?” Second, if the term “autonomous” applies to any party in these cases, it is to the actors on the periphery, as they are the ones acting autonomously, without direction from the center. Yet what is just as interesting about inadvertent expansion is how little control those in power, the leaders in the capital, often have over the process. In contrast, the term “inadvertent” has fewer meanings, being defined as “not resulting from or achieved through deliberate planning” and geared toward the actors in the center, which I would argue thus makes it a more fitting term for the phenomenon.18

Principals, Agents, and Peripheral Expansion

Inadvertent expansion is a regular feature of the modern history of great power politics, occurring dozens of times over the course of more than a century. However, despite the variety of actors involved, regions in which it has occurred, temporal contexts, and territories acquired, instances of inadvertent expansion at a basic level tend to follow a common two-step sequence: “unauthorized peripheral expansion” and “subsequent central authorization.”

In the first step of unauthorized peripheral expansion, actors on the geographic periphery of a state or empire plan, execute, and, in many cases, fully carry out an instance of territorial expansion without the authorization of the leaders at home in the capital.19 These actors on the periphery consist of a variety of state and nonstate actors including members of the national military, colonial armies, imperial officials, members of private-chartered companies, merchants, and explorers. The leaders in the capital whose authorization would normally be required to engage in territorial conquest or annexation include the head of state or government—such as the president, prime minister, autocrat, or monarch—and the key members of his or her cabinet primarily responsible for national or imperial security policy, including the defense, foreign, and colonial ministers.

Unauthorized peripheral expansion can take the form of a military or colonial army engaging in armed conquest, such as when French colonial forces invaded German Cameroon in the opening months of World War I, without orders from Paris.20 It can also take the form of diplomatic or colonial officials annexing territory, such as when Sir Stamford Raffles acquired Singapore on behalf of the British Empire in 1819, without prior authorization from London.21 It can also consist of members of private-chartered companies and other economic actors seizing territory for economic gain, such as the German trader Adolph Lüderitz annexing territory without government authorization in what would become Germany’s first overseas colony of South-West Africa (now Namibia).22 Despite their many differences, what these forms of peripheral expansion share in common is that they were not specifically ordered or authorized by leaders at home in the capital.

The first question that needs addressing, then, is under what conditions peripheral actors are more likely to claim foreign territory without authorization. In short, what explains unauthorized peripheral expansion? The best way to understand unauthorized peripheral expansion is as a particular manifestation of what is known as a “principal-agent problem.”23 Whenever one actor, known as the “principal,” delegates some decision-making authority to another actor, known as the “agent,” the possibility for principal-agent problems arises. In the theory of inadvertent expansion, the principals are leaders in the capital and the agents, peripheral actors.24 In almost all cases of unauthorized peripheral expansion, these agents are either state actors—such as members of the military, diplomatic, or colonial apparatus—or nonstate actors who are operating under government orders and support. In these cases, the delegation of authority is direct and fairly explicit.25 Yet every act of delegation entails potential losses in terms of agency or control on the part of the principal, and this is where problems can arise.

There are two jointly necessary conditions for principal-agent problems to occur. The first of these involves information asymmetries in favor of the agent. Agents will typically possess, or have greater access to, information that is, or would be, of value to the principal. This can be due to the agent’s proximity to the task at hand or expertise in a given issue area or the significant costs associated with monitoring the agent’s behavior. In cases of inadvertent expansion, these information asymmetries can be particularly severe. Many cases involve relatively few people communicating over vast distances, often using nineteenth-century transportation and communications technology. For instance, in 1850, when the explorer Gennadii Nevelskoi unilaterally annexed the Amur Basin to the Russian Empire, it took at least five months to get a reply to a message sent from Saint Petersburg.26 Similarly, in 1852, when Commodore George Lambert independently sparked the Second Anglo-Burmese War, leading to the British conquest of lower Burma, it took approximately four months to get a reply to a message sent from London.27 These conditions greatly exacerbated information asymmetries in favor of peripheral agents and hampered central control by leaders in the capital. As the British colonial secretary Michael Hicks Beach told a colleague, shortly before his peripheral agent independently conquered Zululand in 1879, “I cannot really control him without a telegraph.”28

However, the information asymmetries common in cases of inadvertent expansion are not simply a matter of vast distances and slow communications. Even distance-demolishing technologies such as the telegraph would not solve all problems in this regard—though, as I show in chapter 2, they make a profound difference. First, even as telegraph cables were rapidly connected to coastal regions throughout the world in the late nineteenth century, communication with the interior of many regions remained difficult to establish. Second, leaders in the capital still relied upon their peripheral agents to send information from the frontier, allowing these agents a great deal of artistic license in how they dealt with the facts.29 As one historian of the Russian Empire in Central Asia notes, the “real power which local officers wielded . . . [lay] in their monopoly of information. It was extremely difficult to resist the arguments for action which these officers pressed when the only information available came from them too.”30 And third, even when specific orders could get through to them in time, peripheral agents without representatives from the capital peering over their shoulders always had the option of turning a blind eye. As the famed counterinsurgent of French Algeria Thomas Robert Bugeaud once declared, “In the face of the enemy, one must never accept any precise instructions or plans imposed from above. . . . One should burn instructions so as to avoid the temptation of reading them.”31

The second necessary condition for principal-agent problems to occur is a divergence of preferences between principal and agent. In delegating, the principal wants the agent to act exactly as the principal would if it were making the decisions itself, while agents will be tempted to make their own decisions, at least some of the time. There are many reasons why preferences can diverge in delegated-agency relations, including agent self-interest, barriers to communication, varied perceptions, differing risk orientations, and divergent structural conditions facing principal and agent. In cases of inadvertent expansion, the most important driver of preference divergence is the differing positions and associated perceptions of the leaders in the capital and their peripheral agents. And the most important issue on which preferences diverge is on national or imperial security policy, with peripheral agents often wanting to act quickly and aggressively and leaders in the capital often preferring to act slowly and cautiously.

These differences can be neatly summed up with the ideas of the “view from the capital” and the “view from the frontier.” In cases of inadvertent expansion, the view from the capital is that which is typically held by leaders in the capital. Their responsibilities are broad, being concerned with the interests of the state or empire as a whole. Their areas of concern are often global, including not only the territory at issue but also adjacent territories, other important regional actors, relationships with the other great powers, and, of course, domestic politics at home. And the leaders’ distance from the frontier means that their sense of urgency to act in response to any given contingency there is generally lower than that of the peripheral agents. From the perspective of the capital, the daily fluctuations of threats and opportunities present on the periphery are seen as distant and abstract.

The view from the frontier, in contrast, is that which is typically held by peripheral agents in cases of inadvertent expansion. Their responsibilities are relatively narrow, being concerned, at most, with a portion of the state or empire. Their areas of concern are often local, being essentially limited to the territory at issue and, perhaps, some surrounding areas. And their proximity to the frontier means that their sense of urgency to act in any given contingency is often much higher than that of the leaders back in the capital. The threats and opportunities these peripheral agents regularly face—though often mild when considering the security of the state or empire as a whole—are, for them, immediate and very real. Thus, viewing a given problem from the capital or from the frontier can lead to highly divergent perceptions and to corresponding differences in policy preferences.

This difference between the view from the capital and the view from the frontier is well articulated in a letter from the governor-general of India, Lord Ellenborough, to Queen Victoria, justifying the recent conquest of Sind. “Your Majesty will not have failed to observe,” Ellenborough wrote in October 1843,

how very different a position the British Government stands in Europe from that which it is placed in India. In Europe peace is maintained by the balance of power amongst several states. In India all balance has been overthrown by our preponderance, and to exist we must continue to be supreme. The necessity of our position may often render necessary here measures wholly unsuited to the state of things which prevails in Europe. The least appearance of weakness or of hesitation would lead to a general combination of all against a foreign, and necessarily an unpopular, Government.32

Thus, a combination of information asymmetries and preference divergence will tend to produce principal-agent problems. The most important among these, for our purposes, is what is known as “moral hazard.”33 Given that the agent is acting at the behest of the principal, and given that it is generally the principal who is responsible for the outcomes in question, the agent may feel emboldened to make riskier, less responsible choices than it otherwise would, potentially creating significant problems for the principal. When it comes to inadvertent expansion, the most important manifestation of moral hazard is unauthorized peripheral expansion. Since these peripheral agents are representatives of a powerful state back home—or, at the very least, claim to be acting in its name—they will often have a well-founded expectation that they will be supported by leaders in the capital, regardless of whether their actions were authorized or not.

Now, at least in theory, leaders in the capital should not be totally powerless in the face of these usurpations of authority from the periphery. According to principal-agent theory, there are a number of ways principals can mitigate the risk of wayward agents.34 However, each of these is complicated by the challenging context of globe-spanning empires. The first possible solution is screening: leaders can presumably screen and select peripheral actors who are more likely to be obedient. The problem here, however, is of “multiple agents.”35 There are a small number of principals but, in the case of major empires, a vast number of agency positions to be filled. The idea that each and every agent can be adequately screened—and that each and every position will have multiple eligible potential candidates—is simply not realistic. And this is especially true for the rugged, distant, and often unprestigious posts on the far-flung edges of many nineteenth- and early twentieth-century states and empires.36 A second potential solution is sanction: leaders can threaten punishment for disobedient agents and promise reward for those who follow their orders. The problem here, as I discuss later, is political. Elites in the capital and the broader public often rally in favor of these wayward peripheral actors, largely insulating them from significant punishment. A third possible solution is monitoring: leaders can more closely monitor their agents and thereby directly exert control over their behavior. But as noted, with vast distances and mostly nineteenth-century transportation and communication technology, monitoring is prohibitively costly and, therefore, effectively impossible.

Unauthorized peripheral expansion, then, is the first step in the process of inadvertent expansion. But what motivates peripheral agents to engage in unauthorized peripheral expansion in the first place? In some cases, the territory in question has intrinsic strategic or economic value, and an irresistible window of opportunity presents itself.37 In others, the agents are genuinely threatened by actors on the other side of the frontier and expand for defensive reasons.38 Peripheral agents may also have more bureaucratic or parochial interests, such as promotion within the military or advancement in their organizations.39 Finally, some may simply want the glory and prestige that is so often historically associated with territorial conquest.40 The strength and relevance of these motivations vary widely—both across peripheries and across time—as personnel and circumstances change.

Whatever their individual motivations may be, the crucial permissive factor enabling unauthorized peripheral expansion is the degree of control leaders in the capital have over these peripheral agents. When leaders in the capital have only limited control over their peripheral agents, unauthorized peripheral expansion is more likely. In contrast, when leaders in the capital have significant control over their peripheral agents, unauthorized peripheral expansion is much less likely. In line with principal-agent theory, “control” results from the ability to monitor the agents’ behavior, often coming in the form of advanced transportation and communications technology. However, there are other ways in which central control can break down. For instance, in cases of weak civilian oversight of the military, control by leaders in the capital over their peripheral agents can be severely hampered.41

In sum, unauthorized peripheral expansion is most likely when control by leaders in the capital over their agents on the periphery is relatively low. It is a principal-agent problem driven by a combination of information asymmetries favoring peripheral agents, a divergence of preferences between the capital and the periphery, and moral hazard on the part of peripheral agents. This is the first step in the process of inadvertent expansion.

Peripheral Pulls

The second step of inadvertent expansion consists of subsequent central authorization.42 After claiming territory without authorization, the actors on the periphery present the leaders in the capital with their acquisitions as a fait accompli.43 In some cases, they directly contact their metropolitan bosses to notify them of their gains. In other cases, news filters back to the capital more indirectly through other channels. In either case, knowledge of the acquisition among leaders in the capital initiates a deliberative process on whether to reject or accept the fait accompli. If leaders in the capital decide to reject the fait accompli, then they effectively cut these peripheral actors loose, instructing them to return the acquired territories or fend for themselves. In contrast, if the leaders decide to accept the fait accompli, as they have in dozens of cases in the modern era, the territory then becomes part of the state or empire, thereby completing the process of inadvertent expansion.

The key question that needs addressing for this second step, then, is why leaders in the capital accept these faits accomplis in some cases, but not in others. More simply, what explains subsequent central authorization? Before getting into what explains variation in leader decision-making, it is important to note that, for leaders facing these faits accomplis, there are important constraints on their freedom of choice. Across numerous cases of inadvertent expansion, leaders appear to reluctantly come up with arguments for retaining these territories ex post such that it is difficult to imagine they would marshal in favor of their acquisition ex ante. The primary reason for this is that, by engaging in unauthorized expansion, the peripheral agents change “the facts on the ground,” altering the strategic calculus of the leaders in the capital. And for three central reasons, these changed facts place constraints on these leaders, which can make returning the territory to its previous occupants surprisingly difficult.

First, unauthorized peripheral expansion will often sink the costs of territorial acquisition for leaders in the capital.44 Sunk costs are costs that have already been incurred and, therefore, cannot be recovered.45 In cases of inadvertent expansion, by the time leaders in the capital become aware of what is transpiring on the periphery, their agents have already used state or imperial resources to partially or wholly conquer the territory or negotiate its annexation. If we think of the costs of territorial expansion as being a combination of current “acquisition costs” and future “governance costs,” unauthorized peripheral expansion often sinks a good portion, if not all, of the first half of the equation. Once expended, these costs—often paid in both treasure and blood—are not recoverable, and this will often encourage authorization of the territorial fait accompli.46 For many leaders, the prospect of giving up newly acquired territories will appear as a missed opportunity and, given the irrecoverable costs incurred, a wasted one. The existence of these sunk costs tends to strengthen the arguments of proexpansion hawks in the capital and to weaken the arguments of cost- or risk-conscious doves, pushing the leaders toward acceptance.

Second, unauthorized peripheral expansion also tends to generate domestic political pressure on leaders in the capital to support their own agents and nationals, regardless of “who started it” or how unscrupulous they may have been.47 This sometimes comes in the form of political, military, and imperial elites in and around the capital putting pressure on the leadership to look after their subordinates or nationals. These effects can also be powerfully felt from the opposition in government. In cases where the leadership is opposed to an acquisition, but powerful political opponents are supportive, leaders will often keep their opposition quiet out of a fear of looking unpatriotic or being accused of “siding with the enemy.”48 In still other cases, the state or empire’s domestic public at large becomes aware of events on the periphery, rallies in favor of their conationals, and puts pressure on the government to support them as well. These kinds of pressures, both elite level and public, narrow the options of leaders and make ordering a prompt withdrawal from the territory difficult.

To be clear, this kind of pressure does not always materialize. In certain cases, the public is largely disengaged from events on the frontier. In others, some may even be opposed to the acceptance of the fait accompli. Yet, in most cases, both political elites and the public pressure leaders to support their conationals, and even the expectation of this kind of pressure among leaders can be enough to make them think twice about rejecting the fait accompli. It is simply very politically difficult for leaders to deny their own citizens aid in times of need. As David Landes puts it, “Imperialism was in large measure built on the fait accompli . . . with the state almost always ready to pull its nationals’ chestnuts out of the fire.”49

Third, unauthorized peripheral expansion engages the prestige, honor, and reputation of leaders in the capital, and the state or empire as a whole, in ways they simply were not engaged before.50 This can make relinquishing the territory and retrenching appear prohibitively difficult from the perspective of leaders in the capital.51 It is not that leaders see territorial expansion as invariably prestige or status enhancing—though, in some cases they do. Rather, it is that they often fear that backing down and rejecting the territorial fait accompli is status diminishing. These concerns are particularly potent in the colonial context, where great powers often have other territorial holdings in the area and feel that their continued control rests heavily upon the maintenance of a reputation for resolve in the face of a challenge.52 As Czar Nicholas I of Russia remarked upon learning of the unauthorized annexation of the Amur Basin, “Where once the Russian flag has been raised, never shall it be lowered.”53

Thus, for reasons of sunk costs, domestic political pressure, and the engagement of national honor, leaders in the capital will often feel powerfully bound to accept their peripheral agents’ territorial faits accomplis. Peripheral agents often know this and will use these very same forces to manipulate leaders in the capital, further constraining them toward acceptance. For instance, they can help gin up domestic political support in favor of their cause through means such as waging press campaigns, engaging in pamphleteering, sending allied individuals and organizations on speaking tours, and directly lobbying members of the great power’s government. Some of their strategies are incredibly shrewd. In the process of his unauthorized acquisition of what would become Rhodesia in 1890, for example, Cecil Rhodes made a point of recruiting young men from influential families for his expeditions, so that the government in London would feel extra pressure to intervene on their behalf if needed.54 Peripheral agents also make regular reference to concerns of national honor and prestige in their communications with the capital to make these concerns salient. These communications include, as examples, arguments that backing down would weaken the great power’s authority in the region and elsewhere, that it would only invite attack and encourage dissension, that the locals “respect only force,” that the other great powers “are watching,” and, more generally, that there is a need to support the great power’s important political and social mission and to defend the “honor of the flag.”55 And the knowledge among peripheral agents that these incentives exist, and of their ability to manipulate them, only feeds the moral hazard that leads them to engage in unauthorized peripheral expansion in the first place.

Unauthorized peripheral expansion can also generate incentives for leaders in the capital to reject the territorial fait accompli, largely out of a desire to rein in rogue peripheral agents and to reassert central control.56 However, this incentive often pales in comparison with the stronger domestic political pressures and honor concerns that leaders feel bound by, particularly in cases of successful peripheral expansion, when sunk cost concerns will also be powerfully felt. Once the important segments of the elite and public become aware of what has transpired on the frontier and rally to the peripheral agents’ cause, it becomes very politically difficult for leaders in the capital to censure or punish their agents in any way. This, as noted earlier, explains why peripheral agents who engage in unauthorized peripheral expansion are so often not only not punished but rewarded for their insubordination. But, perhaps more importantly, these are often deeply embarrassing experiences for leaders in the capital, threatening to make them look weak, incompetent, and out of control in front of their political opponents, their populace, and the world at large. A quick and tidy acceptance—along with an effort to cover the inadvertence up—is often a more appealing alternative than engaging in a long, drawn out, and very public process of rejection and retrenchment.57

Geopolitical Risk and Central Authorization

With so many incentives pushing leaders in the capital toward acceptance of the fait accompli, it would seem that there is little reason then for them to reject it. Nonetheless, there are a number of instances in which they decided to rebuff their peripheral agents and to refrain from subsequently authorizing the acquisition. The historical record of inadvertent expansion shows a number of different motivations for these decisions. When the British colonial secretary Lord Glenelg rejected his agents’ claims of the Transkei region of South Africa in 1835, it was a combination of the perceived worthlessness of the territory and popular revulsion at the inhumane treatment of the locals that led to its relinquishment.58 When French leaders rejected their peripheral agent Joseph Galliéni after his unilateral annexation of Senegal in 1881, it was due to the unacceptable terms of the annexation treaty he had negotiated.59 And when the United States refused to accept Hawaii in 1893 after a coup orchestrated, in part, by its own local consular official, the primary reason was that it occurred in the context of a leadership transition in which the incoming president, Grover Cleveland, was utterly uninterested in the territory.60 Thus, like any important decision in world politics, leaders are influenced by a variety of factors.

However, what stands out as the most important factor weighing on the minds of leaders in the capital when deciding whether to accept or reject a fait accompli is the amount of geopolitical risk associated with doing so. The prospect of significant geopolitical repercussions is often enough to discourage leaders from authorizing their peripheral agents’ deeds. And the absence of these potential risks will usually clear the way for the constraints imposed by sunk costs, domestic politics, or national honor to take hold and for leaders to accept the fait accompli and engage in inadvertent expansion. Geopolitical risk is understood here as an objective condition rather than a purely perceptual variable. This is not to say that leaders do not periodically get it wrong, either over- or underestimating the geopolitical risk of a given circumstance. It is also not to say that objectively risky circumstances do not occasionally fail to materialize, because they definitely do. Yet the theory assumes that most leaders will accurately assess the geopolitical risk associated with territorial expansion most of the time. While this is not the sole motivation driving decisions to accept or reject agents’ unauthorized claims, the quantitative evidence and the historical case studies this book examines show that it is the most important determinant of this decision.

There are three key kinds of geopolitical risk that give leaders pause and lead them to rethink the option of accepting the fait accompli. The first is the prospect of crippling international economic isolation. If leaders think that authorizing peripheral expansion will lead to international trade sanctions, substantial foreign economic disinvestment, or being cut off from other forms of international economic exchange, they will be far less likely to accept the fait accompli. The second is the prospect of armed conflict with a local regional power. The possibility of becoming embroiled in a war with a distant regional power—even one that is significantly inferior to the great power in question—will often give leaders pause. The third, and perhaps the gravest, risk is the prospect of encroaching on the interests of another great power to the point of risking a major crisis or even war. Leaders will be highly reluctant to accept the offerings of their peripheral agents when the consequences could be this grave. In contrast, the absence of geopolitical risk clears the way for sunk costs, domestic political pressure, and national honor concerns to take hold, and leaders under these conditions will be significantly more likely to accept the fait accompli.

A skeptical reader might point out, however, that the leaders of powerful states have regularly risked economic isolation and war for territorial gain.61 We need look no further than the United States’ annexation of Texas in 1845, which sparked the Mexican-American War, as well as the expansion of Japan and Germany in the late 1930s, to see that leaders can be highly risk acceptant when pursuing territorial gains. What makes inadvertent expansion different? The key difference between inadvertent expansion and intentional expansion in this regard is, quite simply, planning and preparation by the state’s leaders. Territorial expansion is among the least certain and highest-risk foreign policy initiatives a state can undertake. While it can promise the acquisition of strategic territory and valuable resources, it also risks overburdening the state with obligations, getting bogged down in a long and costly pacification campaign, and impinging upon the interests of other great powers. Thus, when it comes to intentional expansion, leaders will generally carefully consider where to expand, when to do so, how much territory to take, and how to do so. And leaders will time and conduct their campaigns in ways that aim to minimize adverse consequences or, at the very least, to prepare themselves to weather them. In contrast, inadvertent expansion is thrust upon leaders in the capital unwittingly. They have had no control over what territory was taken, when it was taken, how much was taken, and the mode by which it was taken. And they have also had little opportunity to try to avoid adverse consequences or to dig in and bear them. Given these differences, leaders will usually be hard pressed to face significant risks for policies that they have had no control over.

It is worth highlighting that the two forces present at this second stage of subsequent central authorization push in opposite directions. On the one hand, leaders will often feel constrained to accept the territory due to sunk acquisition costs, domestic political pressure, and honor concerns. On the other hand, leaders may face severe geopolitical risk associated with doing so, discouraging them from accepting the fait accompli. The confluence of these countervailing forces can periodically put leaders in a deeply unenviable position. On the one hand, they may face a nationalistic public rallying for acceptance, an opposition in government urging them to defend the national honor, and, in extreme cases, the risk of a coup or even assassination for appearing to look weak. On the other hand, leaders may face the risk of crippling economic isolation, conflict with a regional power thousands of kilometers from home, or a territorial dispute with another great power, possibly escalating to all-out war.

This veritable minefield of decision-making for leaders in the capital is what I refer to as the “dilemma of inadvertent expansion.” Thankfully for the leaders involved, these situations are relatively rare. In most cases, leaders in the capital are presented with territorial faits accomplis by their peripheral agents with few significant risks associated with acceptance, and acceptance is the result. In other cases, metropolitan leaders do face substantial risks associated with acceptance, yet the public is not fully engaged or there is minimal interest at the elite level, giving leaders a safe off-ramp to reject the fait accompli. However, in some cases, these two potent forces come together, requiring leaders to weigh the costs and risks associated with making one decision or the other. Even in these difficult circumstances, considerations of geopolitical risk tend to predominate. But they are, nonetheless, remarkably challenging positions for leaders to find themselves in—with their political positions, and even their lives, at stake.

To sum up the theory in its entirety: inadvertent expansion, the outcome or dependent variable this book seeks to explain, occurs in two basic steps—unauthorized peripheral expansion and subsequent central authorization. The most important explanatory variable for unauthorized peripheral expansion is the degree of control leaders in the capital have over their agents on the periphery, with less control being associated with an increased likelihood of unauthorized peripheral expansion. Once this occurs, the very act of unauthorized peripheral expansion tends to constrain leaders in the capital toward acceptance of the territory being offered due to sunk territorial acquisition costs, the emergence of domestic political pressure, and the engagement of national honor. Yet, at the second step of subsequent central authorization, leaders in the capital also need to consider the degree of geopolitical risk associated with acquiring the territory, with lower perceived risk being associated with an increased likelihood of authorization. Thus, we should expect to observe most inadvertent expansion when leaders in the capital both have relatively low control over their peripheral agents and perceive few geopolitical risks associated with expansion into a given territory. Figure 1.1 illustrates the theory.

Research Design and Methods

Defining the relevant counterfactual of inadvertent expansion is not entirely straightforward.62 If the outcome of interest is inadvertent expansion, is the relevant counterfactual intentional expansion, with the most important variation being between forms of expansion? Is it nonexpansion, with the most important variation being whether peripheral expansion occurs or not? Or is it failed inadvertent expansion—essentially, unauthorized peripheral expansion with leaders in the capital rejecting the fait accompli at the subsequent central authorization stage? A further complication is that some of these counterfactuals, particularly failed inadvertent expansion, are not easily observable in the empirical record. For most historians and in most histories, as I discuss in chapter 2, these are essentially “non-events,” and are, therefore, less likely to appear in the historical record. Thus, given the subject matter of this book, defining and observing the relevant counterfactual poses serious challenges.

This book settles on examining two counterfactuals. The first, which I consider in chapter 2, is variation between forms of expansion: intentional and inadvertent. This helps deal with the problem of the low observability of failed inadvertent expansion and illustrates what is unique and distinctive about inadvertent expansion within the broader universe of cases of territorial expansion by the great powers. The second counterfactual, which I examine in chapters 3–7, is variation between successful inadvertent expansion and failed inadvertent expansion—that is, variation between cases where leaders accept or reject the territorial fait accompli. This is much closer to the important variation explained by the theory and clarifies what explains leaders’ choices when faced with similar inadvertent expansion opportunities.

This is a mixed-methods study of inadvertent expansion, making use of both quantitative and qualitative methods. The quantitative component, located in chapter 2, presents and analyzes data on all instances of great power territorial expansion from 1816 to 2014. Each territorial expansion observation is coded as to whether it is intentional or inadvertent, and this variation provides the primary basis for comparison. The purpose of the quantitative component is threefold: first, to demonstrate the generality of the phenomenon of inadvertent expansion by presenting the universe of post-1815 cases; second, to present basic descriptive statistics regarding inadvertent expansion and to highlight some of the broad patterns the phenomenon displays over space and time; and third, to analyze the data using statistical modeling techniques to show support for some of the claims made in the theory of inadvertent expansion laid out earlier. What the quantitative analysis mainly shows is what is distinctive about cases of inadvertent expansion within the broader universe of great power territorial expansion observations.

The qualitative component, presented in chapters 3–7, consists of a series of paired theory-testing case studies of successful and failed inadvertent expansion by great powers between 1818 and 1932.63 Each of these chapters pairs a “positive” case of successful inadvertent expansion (where unauthorized peripheral expansion leads to subsequent central authorization by the capital) with a “negative” case of failed inadvertent expansion (where peripheral expansion ultimately leads to rejection of the fait accompli by the capital). Including cases of both successful and failed inadvertent expansion helps mitigate the risk of selection bias that might result from including only one or the other.64

These historical cases rely on the comparative method and process tracing. The inclusion of paired cases allows me to engage in a controlled comparison across cases.65 Through comparing cases of successful inadvertent expansion with cases of failed inadvertent expansion, I show that, in each case, the outcomes covary with the key explanatory variables highlighted by the theory. In each case, I also engage in process tracing.66 Each case study presents ground level evidence of the argument and its causal mechanisms at work to show that the outcomes are produced for the reasons specified by the theory.67 While this qualitative component thus provides evidence for all parts of the theory, its primary value is in addressing the question of why leaders in the capital agree to accept the territorial fait accompli of their peripheral agents in some cases, but not others.

Finally, a few words on scope conditions and case selection. This book focuses on acts of territorial expansion carried out by great powers from 1816 to the present. This is not to suggest that minor powers have never engaged or do not engage in inadvertent expansion, as they most certainly have and do.68 But, as I detail in chapter 2, I limit the empirical scope to great powers because they have carried out most of the territorial expansion, their expansion is most consequential, and the historical record for these actors is most complete. I also limit my temporal scope to the post-Napoleonic (1816–) international system. This is similarly not intended to suggest that inadvertent expansion is a strictly modern phenomenon. In fact, given the importance of weak central control over the periphery to inadvertent expansion, it is likely to be more common the farther back we go in time. However, for reasons of empirical tractability and to build off of existing data collection efforts, I focus only on the post-1815 period.

The book includes ten cases, with at least one from each of the great powers that experienced successful or failed inadvertent expansion: France, Germany, Italy, Japan, Russia, the United States, and the brief introductory example of the United Kingdom.69 The inclusion of this many cases both increases the representativeness of the sample selected and helps illustrate the generalizability of the theory.70 In the book’s online appendix (https://doi.org/10.7910/DVN/JBGNNH), I include the results of a balance test between the cases selected and the broader population of inadvertent expansion observations, which shows that the sample of cases selected is relatively well balanced on most important variables.71

The chapters’ cases span over a century of history, from the United States in Florida in 1818 to Japan in Manchuria in 1932. They occur in five of six major world regions, including the Western Hemisphere, South and Central Asia, the Asia-Pacific, Europe, and sub-Saharan Africa. The cases also show considerable variation in terms of the great power’s domestic political regime type, including the reasonably democratic regime of late nineteenth-century France, the mixed-regime of the early nineteenth-century United States, and the harshly autocratic regime of mid-nineteenth-century Russia. Moreover, they include great powers with continental orientations, such as the Russian Empire, as well as maritime orientation, such as the Japanese Empire. The cases also span great powers that differ in terms of their relative power, such as late nineteenth-century Germany, which was then the most powerful continental European state, as well as “barely great” interwar Italy and Japan. Finally, there is variation across the cases as to whether the expansion itself is an extension of state borders, as in the case of Russia in Central Asia, or the growth of overseas empires, as in the cases of France in Tonkin or Germany in East Africa. All of this variation should increase our confidence that the theory, and the phenomenon of inadvertent expansion more broadly, is not simply the product of a particular actor or a particular context.72

Chapters 3–7 each consists of a pair of highly comparable cases that were selected and paired to facilitate controlled comparison and to highlight important aspects of the theory.73 For instance, chapter 3 includes a pair of US cases in which it is the same great power, operating in the same region, in roughly the same time period, but with varying levels of geopolitical risk leading to divergent outcomes. What is notable about these US cases is that it is also the very same individual—Andrew Jackson—who is the peripheral agent in one case (Florida) but the leader in the capital in the other (Texas). Having Jackson figure so prominently in both cases also well illustrates the notion of the “view from the frontier” and the “view from the capital,” showing how it manifests in the very same individual in both key positions.

Chapter 4 consists of a pair of Russian cases in which it is the same great power, operating in the same region, and with the cases separated by only five years, yet with divergent outcomes. Both Russian cases feature nearly all of the very same leaders in the capital across the two cases, who, due to varying geopolitical risk, found reason to make different authorization decisions across the two cases. Similarly, chapter 5 includes a pair of French cases that also involve the same great power, operating in the same region, and separated by only a decade, yet with varying outcomes. These cases, however, deal with the very same territorial entity (Tonkin) and involve an almost-identical process of peripheral expansion yet, also for reasons of geopolitical risk, have differing outcomes. Chapter 5 is presented as a pair of comparative cases but, in other ways, can also be seen as a single longitudinal case.74

Chapter 6 involves the only cross–great power comparison, presenting one Japan case and one Italy case. These two cases are very well matched on many important factors such as relative power, time period, domestic political regime type, and the proximity of the expansion to their respective capitals. The confluence of domestic political pressure and geopolitical risk in the Japan and Italy cases also powerfully illustrates the dilemma of inadvertent expansion.

Chapter 7 includes a pair of German cases involving the same great power, operating in the same region, separated by only five years, but with varied outcomes for reasons of geopolitical risk. These two cases involve the very same peripheral agent, whose fait accompli is accepted in one but rejected in the other. The German cases also present a unique inferential opportunity in the form of a timely change of leadership in Berlin: the fall of Bismarck in March 1890. This transition shows, first, how reluctant leaders can be pressured to accept territorial claims for domestic political reasons, whereas enthusiastic leaders can be discouraged by geopolitical conditions. It also shows how different leaders with very different orientations toward empire can be similarly influenced by their expectations of geopolitical risk. Table 1.1 sums up the variations across these cases.

Overall, the careful pairing of the qualitative cases provides inferential leverage by controlling, to some extent, for any idiosyncrasies associated with individual great powers, time periods, regions, territorial entities, leaders in the capital, and/or peripheral agents across cases. This helps control for a number of potential confounders by effectively holding many factors constant. It is also worth noting that, while the quantitative data presented in chapter 2 includes both armed conquest and political annexation, the qualitative component is primarily focused on conquest, with only a single chapter—Germany in East Africa (chap. 7)—including cases of annexation. This is partly because these cases tend to be the most consequential but primarily in order to facilitate cross-case comparison.