The study of rebel governance has made substantial progress in recent years. Scholars working in this area have brought together disparate strands from across the social sciences to create a coherent and fast-moving research program (Mampilly 2011; Staniland 2014; Thomas and Bond 2015; Jo 2015; Arjona et al. 2015; Huang 2016a; 2016b; Henshaw 2016; Arjona 2016; Wood and Thomas 2017; Baczko et al. 2018; Jo and Niehaus 2018; Loyle and Bestvater 2019; Henshaw et al. 2019; Brenner 2019; Asal and Jadoon 2020; Florea 2020; Griffiths 2021; Breslawski 2021; Mampilly and Stewart 2021; Stewart 2021; Revkin 2021; Cunningham et al. 2021; Loyle 2021; Loyle et al. 2021). The goal is to understand the causes and consequences of rebel governance with respect to a range of issues. In this article, I wish to explore a particular topic regarding rebel governing organizations (hereafter referred to as rebel groups) and the socio-political order they create over civilian populations. Framed as a general question, what factors lead rebels to choose between forms of civilian control that are fleeting and disorderly, and other more ordered relationships that may or may not permit a level of local civilian autonomy? Put differently, how do rebels choose the form of rule?

I contend that a richer understanding of this topic may benefit from a comparative analysis with the literature on empire (Lenin 1939; Gallagher and Robinson 1953; Fieldhouse 1961; Cohen 1973; Louis 1976; Doyle 1986; Muldoon 1999; Maier 2006; Nexon and Thomas 2007; Barkey 2008; Pitts 2010; Burbank and Cooper 2010; Go 2011; Gill 2016; Luttwack 2016). This may seem like an odd comparison. After all, in relation to rebel groups, empires are usually much bigger, more elaborate, and more enduring socio-political structures. Yet, there is a striking similarity that, to my knowledge, has not been explored systematically.Footnote 1 To survive and sustain their conflict with the state and other opponents, Rebel groups have to determine their socio-political relationship with surrounding civilian populations. Recent scholarship has explored these choices and begun to specify the connection between key factors and the resulting relationship (Arjona 2016; Breslawski 2021; Mampilly and Stewart 2021). However, there is a much older literature on the form of rule that empires adopt with respect to peripheral groups and political units. As I will show, both literatures have developed quite similar decision models, and have come to similar conclusions, with little apparent knowledge of the other. My wager is that the research on rebel governance will benefit from a comparative analysis and potential cross-fertilization; indeed, the empires literature, although older, would also benefit from the comparison with the work on rebel governance. To that end, this is not an article that will present and then test a core argument. Rather, the purpose of this article is to synthesize two very different and largely unconnected literatures, and then highlight paths for research.

The article proceeds as follows. I first define my terms, and pause to consider the comparability of the phenomena. I next turn to an analysis of the rebel governance literature on the form of rule, followed by a similar analysis with empires. I then conduct a more comparative analysis and outline a set of key explanatory categories involving push factors, pull factors, intergroup competition, and time. The payoff is that I develop hypotheses on rebel governance for future research. In all, this article contributes to this special issue by emphasizing the theme of control with respect to rebels and empires and the populations that surround them.

Rebel groups and empires

I begin by defining the two units for comparison. For rebel groups, I adopt Kasfir’s fairly broad definition of “consciously coordinated groups whose members engage in protracted violence with the intention of gaining undisputed political control over all or a portion of a pre-existing state’s territory.”Footnote 2 To count, Kasfir says that such groups must hold territory in which civilians reside, and that they committed an initial act of violence and continue or threaten to do so. Scholars working in this area have defined the phenomenon in various ways, often using other terms like insurgents or non-state combatants interchangeably. There are various questions of scope such as: where is the line between rebel groups and criminal organizations? In the end, this is a contested term. But for my purposes, Kasfir’s widely used definition is appropriate. It identifies non-state, territory-possessing, violence-wielding organizations that possess and seek to possess political control over local populations. The definition encompasses a fairly broad set from Hezbollah to Tamil Eelam to the Wa State (Mampilly 2011; Brown and Hermann 2020). Many of these groups will be discussed in the accompanying articles in this special issue.

Empires are also a contested concept, and may even be more difficult to pin point than rebel groups. This could be surprising given its common usage, but that commonality is partly a reflection of the great variety of ways in which empires are conceived. Some conceptions of empire stress ideology, while others emphasize process and/or structure (Maier 2006; Pitts 2010). In a colloquial sense, the terms “empire” and “imperialism” are used to connote power—take for example, two Economist articles on the might of Silicon Valley and Facebook.Footnote 3 As with rebel groups, the choice of the definition of empire depends on the project. As such, I adopt Doyle’s definition of empire as “effective control, whether formal or informal, of a subordinated society by an imperial society.”(Doyle 1986, 30) Doyle sees empires as a type of state that engages in practices that create imperial structures. Already there are concepts here that require some unpacking—and I return to them below—but suffice it to say that Doyle’s definition is fairly common in the literature. Like Kasfir’s definition of rebel groups, Doyle provides a general and portable concept.

This general conception of empire has existed in a variety of contexts. It would include the classic cases such as the Ottomans, the British, the Romans, along with the major Chinese dynasties. It would encompass lesser-known cases like Majapahit, Bornu, and Durrani. But it is not a set that shares a common name. History records Augustus as the first Roman emperor, but he called himself “princeps” and during its first century Rome was referred to as the Principate (Maier 2006, 37; Elden 2013). The word “emperor” implied different things to different people during the European Middle Ages (Muldoon 1999, 17; Go 2011, 6–7). Moreover, the term does not translate perfectly across languages and cultures and in many ways it is an historical convenience to call large states of antiquity empires when they did not call themselves that. Therefore, it is not the name that matters, but how the political unit ruled. According to Doyle’s definition, there is an imperial unit or society (a core) that attempts to subordinate and rule over adjacent, peripheral, and sometimes distant populations. The definition centers on that relationship.

There exists a rough symmetry between the two phenomena: rebel organization and empire. In each case there is a politically motivated organization that controls territory and wields violence and/or the threat of violence. In each case, there is a population that sits outside the organization, but proximate to it, over which the organization may choose to impose its will. It sounds like I am also describing the state, and, indeed, scholars working on rebel governance have explored the relationship between rebels and states (Mampilly 2011).Footnote 4 According to Tilly, rebels, states, and empires can all be placed on a continuum that also includes bandits, pirates, and gangs, among others (Tilly 1985, 170; Spruyt 2017, 77). However, as Mampilly and Stewart noted, the relationship between rebel organizations and empires is particularly fitting (Mampilly and Stewart 2021, 37). Both phenomena have an expansionist character; rebels typically aim to increase territorial control as a way to oppose the state; empires expand their imperial reach because they are, well, empires. There is a sense in which both are less static than states. Moreover, a consequence of that expansionist character is a need to determine how to rule over subject populations, especially as new populations come under their control.

Scholars working in both literatures have identified the different forms that rule can take, and the factors that shape those choices. For precision, I define the form of rule as the choice that the expanding organization (rebel group or empire) makes in terms of how to govern (control) neighboring populations. That choice can take a set of values ranging from no relationship at all to one of complete governance provision under the direct administration of the rebel group or empire. My aim is to compare these choices across the two literatures.

As some readers will point out, these two organizational types / phenomena are not isomorphic. It is important to weigh their differences and similarities, and consider to what extent the lessons from one domain can be applied to the other. A good place to start is with purpose and/or ideology. Rebel groups and empires typically have different end goals. Whereas rebels are trying to retain and win territorial control from the state, and often look to overthrow the state or secede from it, empires are already state-possessing. There are some goals that may be more exclusive to rebels, such as overthrowing the landed classes, just as there are goals that can drive empires, like regional or world domination, that are ill-fit to rebels. Nevertheless, there is an intersection. Empires often begin as rebel groups—indeed, they are successful rebellions. There is a set of potential purposes that can drive rebels and empires alike, such as the spread of a given faith or ideology. But zooming out, it is important to differentiate between the purpose of rebels/empires—i.e. the push factors that drive them—and the process by which they expand and structure their rule—i.e. the form of rule they adopt. It is the latter issue of structure that I am comparing in this article; it is my dependent variable. The purposes of rebels and empires are not fundamentally different; they overlap in some ways and are different in others. This is a space where a comparison can be useful.

What about the competitive environment in which the two organizations operate? It is not clear that it should vary in a qualitatively important way. Rebel groups are nearly always in competition with the state, and are quite often competing with other rebel groups. For example, ISIS was competing with both the Free Syrian Army, the Kurds, and the Syrian Government at various points (Baczko et el. 2018). Likewise, empires quite often faced stiff competition over territorial control. As Doyle argued, competitive pressures between empires in the nineteenth century had global consequences. Although the scale of these competitive dynamics may be greater for empires, it is not clear that they are qualitatively different. Indeed, this is another domain where a review of the empire literature may shed light on rebel governance.

Another potential point of difference has to do with the distinction between the unit and surrounding population—that is, the distinction between rebels and civilians and imperial cores and peripheries. One interesting critique of the rebel governance literature is that the distinction between rebels and civilians may be overly stylized. After all, rebels have to come from somewhere. In some cases, they may be choosing how to control civilian populations that are the same base population from which the rebels emerged. In other cases, this merging dissipates as the rebel group expands its territorial control and encounters new populations. Is this different from empires? In broad terms, the answer is no, because empires also vary in this regard. Whereas the Russians tended to expand into neighboring territory and gradually blur the distinction between core and periphery, the British and French created complex administrative structures to rationalize their maritime realms (Gill 2016). Given their greater size and reach, empires were more likely to rule over dissimilar populations, but this is a matter of degree. Like purpose and the competitive environment, it is not clear that we should expect systematic differences in the form of rule on account of dissimilar populations. Indeed, a shared characteristic between rebel groups and empires is the tendency to gradually assimilate peripheral populations, thus transforming them into rebels or imperial subjects. Doyle referred to this transformation as crossing the Caracallan threshold, a liminal moment in the life cycle of empires. By analogy, when Amilcar Cabral and the African Party for the Independence of Guinea and Cape Verde (PAIGC) brought peripheral populations into the rebel cause, they were crossing a similar threshold.Footnote 5

A more substantive difference pertains to size and longevity. Empires can be enormous and they can last for centuries. Although rebel groups can be quite large—e.g. the Confederate States of America—they can also be very small, and most readers will accept that they are, on average, smaller than empires. Given their greater size and longevity, empires are typically more highly institutionalized. For the most part, it is not clear that these differences should undermine the comparison between how they rule over peripheral populations. The one notable exception may be the time horizon. Rebel groups do not have relatively high longevity given that they either die out, dissolve, or win their rebellion and become a state. They are transitory by nature. This gives them a shorter time horizon, on average, a factor that can affect their choice over the form of rule, and one that I will revisit.

Finally, one factor that is different is the existence of the sovereign state that the rebels oppose. Rebels seek territorial control against a sovereign state that nearly always has legal jurisdiction over that territory, at least in the modern international system (Griffiths 2021). There is no analog for empires, unless the empire is also a rebel group that seeks to overthrow or separate from a larger state. Although this represents a clear qualitative difference, it is uncertain whether it would yield systematic differences with respect to the form of rule.

In sum, although empires are generally different from rebel groups, these differences do not undermine a useful comparison of the form of rule. On the contrary, the differences actually help guide the comparison. The distinction between purpose and process (a common one in the Empire literature), the core/periphery population distinction, the competitive environment, the time horizon in particular, as well as the potential significance of the sovereign state, are useful considerations to bear in mind. I now turn to an analysis of the form of rule within each literature.

Rebel groups and the form of rule

How do rebel groups choose the form of rule over civilian populations? To attempt an answer to this question, it is useful to begin with Arjona’s foundational work on social order in the context of rebel governance (Arjona 2016). She theorized three different types of social contract that rebel groups make with civilian populations: disorder, aliocracy, and rebelocracy. Whereas disorder implies no social contract, the difference between aliocracy and rebelocracy is the intensity of the form of rule. When rebelocracy is established, rebel groups intervene broadly to manage not only security and taxation, but also justice provision. Where aliocracy holds, rebel groups intervene more narrowly and permit local civilian control over justice. To foreshadow one of the discussions below and a core theme in this article, her conception of rebelocracy and aliocracy can be roughly mapped on to the distinction between direct and indirect imperial rule.

Arjona posits several factors/assumptions that shape the choice over the form of rule. The first is the time horizon of the rebel group. When rebel groups have a longer time horizon, they are more likely to form a social contract because of a set of benefits (obedience, support, resources, access, etc.) that follow (Arjona 2016, 55–62). But when their time horizon is short on account of indiscipline, armed competition with the state and other rebel groups, or other macro factors, rebel groups will forego the establishment of a social contract because of the cost it entails, and disorder will follow. Here, Arjona assumes that once rebel groups have a longer time horizon, they prefer rebelocracy, the most direct and intense form of rule, because it carries greater benefits than aliocracy. Yet the driver of that outcome (rebelocracy vs aliocracy) comes down to the civilians and their ability to overcome the collective action problem and resist the rebels. When civilian resistance is sufficient, rebel groups settle for aliocracy as a middle position. Where it is lacking, the rebels choose rebelocracy as the form of rule. Figure 1 provides an illustration of this decision tree and the key separating factors.

Fig. 1
figure 1

Rebel group decision tree (Arjona)

Full size image

In sum, Arjona has mapped out the architecture of rebel governance. Each endpoint represents a type of social contract between a specific rebel group and a specific civilian population. Although a rebel group will face multiple civilian populations, typically organized around a village or town, they have to develop a social contract with each population in a dyadic manner. Arjona’s study focused primarily on Colombian rebel groups like the Revolutionary Armed Forces of Columbia (FARC) and the National Liberation Army (ELN). Much like empires, these rebel groups possessed a territorial core in a landscape of dissipating control and sometimes ambiguous civilian support. The parsimonious decision tree in Fig. 1 shows the key separating factors that, like shifting levers, push in the direction of a specific form of rule.

A related theory on rebel groups and the form of rule is given by Breslawski (2021). As Fig. 2 shows, she develops a decision tree with the same structure. Although her three outcomes are given different names (no institutions, inclusive institutions, exclusive institutions), they are roughly the same as Arjona’s and can be read from top to bottom as intensification in the form of rule. A relationship without institutions implies no social contract (disorder). And whereas inclusive institutions mean that rebel groups allow civilians to control local affairs, rule based on exclusive institutions entails direct rebel rule and the exclusion of civilian administration. Breslawski used this decision tree to understand rebel decisions on the form of rule in both a large-N study and in a close analysis of rebel governance in Aceh.

Fig. 2
figure 2

Rebel group decision tree (Breslawski)

Full size image

Despite the similar structure, there are important differences in the theoretical mechanisms. Breslawski’s first decision node pertains to rebel constituency, asking “whether or not rebels view civilians as members of the group they are fighting for.” (Breslawski 2021, 458) Unlike Arjona, who emphasizes the rebel time horizon, Breslawski stresses affinity. But both factors shape the resulting form of rule in the same way. Her second decision node focuses on community cohesion, “the extent to which political, economic, or social divisions exist within the community.” In contrast to Arjona, who assumes that rebel groups will maximize the form of rule given a longer time horizon, Breslawski assumes that rebel groups will evaluate the costs and benefits of inclusive and exclusive institutions. When the civilian community is cohesive, rebels can afford to rule indirectly; but where the community is not cohesive, rebels will have to create more direct forms of rule that exclude civilian participation. These are subtle but important differences. My purpose here is not cross-evaluate them, but rather to catalog them and see how they fit (or not fit) with similar decision nodes and mechanisms that are identified in the literature on empire.

A third model, more comprehensive but less parsimonious, is provided by Mampilly and Stewart (2021). They include many of the same decision factors, such as the tradeoff between coercion and co-optation and the importance of local support, which shapes the ability of rebel groups to utilize civilian institutions and integrate civilians. To this they add an additional feature regarding revolutionary purpose, and, as a result, whether rebel groups demand changes in local civilian institutions. In general, they theorize a greater design element in the choice of rebel rule, taking seriously the revolutionary intentions that rebels sometimes possess.

Overall, although these frameworks are meant to explain rebel governance and the form of rule, there are key differences. While Arjona begins by noting the importance of time horizons, Brewslawki stresses the importance of affinity. Whereas Arjona highlights the institutional and organization strength of the civilian population, and its resulting ability to resist, Breslawski models civilians as a more passive actor. However,Breslawski does not assume that rebel’s preferred outcome is to maximize their rule (rebelocracy); rather, they take a more neutral position on the relative merits of inclusive versus exclusive institutions and condition their choice on the characteristics of the civilian population. Meanwhile, Mampilly and Stewart cast this more clearly as a rational process of decision-making. They also bring in the importance of revolutionary intent.

In various ways, these differences touch on some of the earlier considerations regarding the comparability of rebel groups and empires. Of the three models above, Mampilly and Stewart is the only one that stresses purpose and revolutionary ideology as a push factor that can shape the resulting form of rule. Meanwhile, Arjona highlights the importance of both intergroup competition and the time horizon. Although, interestingly, she connects them causally by arguing that increased competition reduces the time horizon. For her part, Breslawski underscores the role that ethnic affinity plays. Finally, both Arjona and Breslawski draw attention to the manner in which features of the civilian population (institutional strength, cohesiveness) are pull factors in the form of rule.

How rebels choose the form of rule is a complex process with many moving parts. It is difficult to know not only what factors should be included and prioritized, but also how they should stand in relation to one another causally. To gain perspective on these questions, I now turn to the literature on empires which, as it turns out, has modeled these decisions in strikingly similar ways.

Empires and the form of rule

Doyle wrote one of the most comprehensive works on empire and the choice over the form of rule (Doyle 1986). He specified several different relationships between empires and their periphery, including informal empire, and formal empire consisting of either indirect or direct rule (see Fig. 3). A relationship of informal rule holds where the peripheral political unit is formally independent, and yet engaged in a subordinate relationship with the empire. Scholars of this topic have argued that the USA has commanded informal empire over some Latin American states, just as the Soviet Union had an informal empire over the Warsaw Pact states. Doyle argued that the British Empire practiced informal empire in various regions of the world prior to the 1870s. Note that for completeness, I have included an upper branch on Fig. 3 that denotes a non-relationship between empire and periphery, or at least not a hierarchical one. Although Doyle does not discuss this relationship in direct terms, he does treat it as a residual category of potential imperial relations.

Fig. 3
figure 3

Empire decision tree (Doyle)

Full size image

Formal control holds when the empire legally rules the peripheral society—i.e. it is not formally independent (Doyle 1986, 38). The next node on the tree indicates whether the peripheral unit is ruled indirectly or directly. Whereas direct rule exists when personnel from the empire govern the district, indirect rule occurs when governance of the district is allocated to native elites under imperial supervision. For example, crown colonies in the British Empire were typically ruled directly; in contrast, protectorates experienced indirect rule (Griffiths 2016, 84). This distinction is similar to that made by both Arjona and Breslawski; it captures the intensity of the form of rule. Will the peripheral unit (or civilian population) maintain a level of autonomy over local affairs, or will personnel from the controlling unit (empires, rebel groups) rule directly?

What factors shape imperial decisions over the form of rule? To answer this question, Doyle synthesized a set of explanations from the literature on empires, grouped into three categories. The first, which he calls metrocentric theories, locate the cause of expansion in the dominant society. This view was particularly popular in the 19th and early twentieth centuries, and exponents like Lenin and Schumpeter sought to explain expansionism as either the result of class struggle or social-militaristic atavisms (Lenin 1939; Schumpeter 1950). The second type of explanation, or pericentric theories, locates the cause of expansion in the periphery. Here, it is socio-political conditions in the periphery that incites the would-be empire to expand. These theories became popular in the mid-twentieth century, and an early example was given by Gallagher and Robinson, who conceived of imperial expansion as a policy option for core states to pursue when alternative methods of exploitation were less cost effective (Gallagher and Robinson 1953). Finally, the third category locates the source of expansion in the international system itself and contends that it is interstate competition that compels core states to acquire territory. Prominent theorists here include both Cohen (1973) and Fieldhouse (1961).

Doyle argued that all three explanations play a role and were active in driving imperialism in the 19th and early twentieth century. Moreover, they help explain imperial decisions at the first node in Fig. 3, between no control, informal control, and formal control. Empires are driven to expand by metrocentric and pericentric factors, but the choice between formal and informal control is shaped by interstate competition. It was the increased competition of the latter nineteenth century and scramble for territory that led to a shift in the preferred form of rule from informal to formal empire. Informal rule in conditions of increased imperial competition is problematic because it permits peripheral units to play empires off one another more easily (Doyle 1986, 343–344). Thus, each empire has greater incentive to take formal control over other territories.

However, once formal control is established, imperial rulers need to determine the intensity of rule. Doyle, following others in the literature, holds that this decision was based in part on the institutional strength of the peripheral unit. Where there was an existing state-like structure, as the British encountered among many of the Indian Princely States, indirect rule was preferred because it was cost effective. But where state institutions were relatively absent, more direct forms of rule were necessary. As Griffiths (2016) shows, it was the strength of local institutions that often determined the administrative relationship across different empires.

To sum up, there are several key mechanisms that shape imperial rule. There are metrocentric (push) factors related to the sociological and ideological content of the imperial society. There are pericentric (pull) factors that draw empires out to control peripheral populations and/or acquire resources. There are competitive factors that, contrary to Arjona’s assumption on rebel rule, push empires to intensify the form of rule. Finally, the strength of the local institutions shapes the cost–benefit analysis of how directly empires will rule. To this we can add two more factors that come up in Doyle and elsewhere (Gill 2016). First, different empires displayed different institutional habits/tendencies with respect to the form of rule. Relative to the British, the French tended to opt for more centralized and direct rule, and the Russians rarely practiced indirect rule. Second, over the long run, the general tendency with empires was toward a gradual drift in intensification. The longer the relationship, the more likely the empire was to be drawn in institutionally to take more direct control.

Push, pull, competition, and time

I am now in a position to speculate on the lessons that may be drawn from the empires literature to shed light on rebel governance and the form of rule. To do so, I organize the factors that have been discussed thus far into four categories: push, pull, competition, and time. I explore their relationship, and I establish working hypotheses for further research.

The empires literature draws a useful distinction between metrocentric and pericentric explanations. To reiterate, metrocentric theories locate the cause of expansion in the metropole—i.e. its class divisions, its desire to spread civilization, its religion, etc. These are push explanations that can be rendered without any contextual knowledge of the peripheral unit. In contrast, pericentric explanations bring in the content of the peripheral unit to explain why the empire chooses a specific form of rule. Is the peripheral unit socially cohesive, is there an existing state-like structure, and how does the answer to these questions shape rebel choices in terms of governance? Crucially, whereas metrocentric theories should expect greater uniformity in the form of rule across diverse peripheral groups, pericentric theories would yield greater heterogeneity in the form of rule depending on local context. This distinction can be applied to rebel governance; we can shed the imperial language of metropoles and settle for the simpler language of push and pull.

There are two push explanations where rebel governance is concerned. The first pertains to the point made by Mampilly and Stewart (2021) on revolutionary purpose. Like empires with transformative ideologies, rebel groups possessing revolutionary ideologies are more likely, all else equal, to push for more direct forms of rule as a means to transform the civilian society. Consider two rebel groups, one that espouses a conservative ideology and aims to overturn a more liberal government, and one that is Leninist and hopes to restructure societal institutions. The Leninist group would be more likely to engage in direct forms of rule because it seeks deeper changes that require implementation by rebel personnel (Stewart 2021). Of course, revolutionary zeal can take different forms, but, on balance, it should act as an intensifier where rule is concerned. This can be stated as a general hypothesis.

H1

Revolutionary rebel groups are more likely to engage in direct forms of rule.

A different push factor has to do with the past practices and habits of rebel groups. One of the lessons of imperial rule is that different empires displayed different practices and habits where the form of rule is concerned. British imperial governors often opted for indirect rule because that choice was common to them, and amounted to a form of institutional memory. I conjecture that similar patterns ought to hold where rebel groups are concerned. That is, they are likely to choose forms of rule that resemble past choices. This is particularly so for rebel groups that have dealt with multiple civilian populations. This expectation can be stated as follows.

H2

For a given rebel group, the choice of rule will correlate with past choices on the form of rule for other civilian populations.

Note that both of these hypotheses are metrocentric, or rebel-centric, in nature. They ought to hold as intensifiers of rule in general, and, as opposed to pericentric (or pull) explanations, do not depend on the characteristics of the civilian population. It is when we bring in pull explanations that we can expect to see greater variety in the form of rebel rule because rebel leaders are calibrating those choices according to the group with which they are interacting. Pericentric explanations in the empire literature tend to see empires as engaging in a different, more periphery-focused cost–benefit analysis where the form of rule is concerned. Given their overall purpose to control resources and peripheral societies, empires determine the degree of intervention based on peripheral conditions such as local institutional strength, social cohesion, and trust. When these factors are lacking, empires are pulled in further.

In the rebel governance literature, Arjona and Breslawski varied in their treatment of pericentric (pull) factors. Whereas Arjona argued that rebels prefer direct rule (rebelocracy) and only settle for less (aliocracy) when organized peripheral groups can mount a successful resistance, Breslawski posited that rebels were basically neutral on the form of rule but conditioned their decision on whether peripheral populations were socially cohesive. Here, Breslawski’s approach hews much closer to Doyle’s argument on imperial rule, particularly if it were expanded to include not just local cohesion, but also affinity and the quality of civilian institutions. To paraphrase Doyle, if rebel groups can rule indirectly and still extract key resources and effectively control the external orientation of a civilian population, then, all else equal, they will. A researcher could explore this distinction to gain a richer understanding of rebel preferences. If Arjona is correct, then indirect rule (aliocracy) should only occur when the civilian population is strong enough to resist. However, if Breslawski is right, then we should find instances of indirect rule even when the civilians cannot resist or perhaps do not even want to. Both scholars see a cost–benefit calculation, but they begin with different underlying rebel preferences. Boiled down, this variation is captured with the following hypothesis:

H3

Rebel groups will rule indirectly only when civilian resistance is sufficient.

There are numerous factors that influence hypothesis 3. For instance, the preferences of revolutionary rebel groups will shift in the direction of more direct rule. Institutional habit can also act as an intensifier. Moreover, as I discuss below, there are other factors that shape these decisions. But, once they are held constant, peripheral features of the civilian groups shape the costs and benefits of direct versus indirect rule. Hypothesis 3 is designed to detect the preferences of rebel governors.

As a research strategy, the distinction between push and pull factors (metrocentric and pericentric) can be examined and teased out in situations where rebel governors face multiple but diverse civilian populations. If push (metrocentric) factors are dominant, then we should expect to see greater homogeneity with respect to the form of rule. But if pull (pericentric) factors are the driving force, then we should expect more heterogeneity in their relationships with civilian populations. As an illustration, imagine a setting in which a rebel group is proximate to three different civilian populations (A, B, and C). Whereas A and B consist of socially cohesive, institutionally developed populations, C is more internally riven and weaker institutionally. If the rebel group is driven primarily by push factors, then it is more likely to govern C in the same manner as A and B. However, if the rebel group is driven primarily by pull factors, then it ought to govern them differently, perhaps choosing direct rule for population C and direct rule for A and B. These hypotheses can be stated as follows.

H4a

When push factors are dominant, there should be greater homogeneity in the form of rule.

H4b

When pull factors are dominant, there should be greater heterogeneity in the form of rule.

These hypotheses could, of course, be further developed. For example, Arjona’s and Breslawski’s respective models point to different motivating factors and underlying assumptions. A detailed study could zoom in on these differences. But overall, the relative importance of push and pull can be analyzed in settings with multiple civilian populations.

This brings us to the issue of intergroup competition. According to Doyle, the late nineteenth century competition between empires produced an intensification in the form of rule because each empire wanted to stake its claim against competitors. Interestingly, Arjona drew what appears to be the opposite conclusion when she theorized that high levels of intergroup competition motivate rebel groups to choose disorder. For Arjona, competition is a dampener, not an intensifier. In part, this is because she models competition as a variable that influences the time horizon, one of her master variables. While, as I discuss below, she is correct on the importance of time, competition is a vital factor that should be dealt with separately. In fact, it is not clear that increased competition would necessarily decrease the expected time horizon; it could motivate rebel groups to take a longer view of their struggle. On its own, competition with other rebel groups and/or the state should affect the form of rule. If the dynamics of rebel governance are similar to empires, then, all else equal, higher competition should correlate with more direct forms of rule.

H5

Higher levels of competition with the state or other rebel groups correlates with more direct forms of rule.

But perhaps competition among rebel groups has a fundamentally different character than competition among empires. In Doyle’s description of nineteenth century imperial competition, formal control over a peripheral region carried weight. For example, German control over eastern New Guinea was accepted by other empires as part of a negotiation process, one that was typically quid pro quo. That kind of agreed upon formal recognition requires a level of order between actors that may often be absent between rebel groups. Where practices and norms of recognized control exit, then competition may drive rebels to intensify the form of rule. But where it is lacking, competition may push in the other direction if rebels are unable to control and govern the population in an empirical sense. An exploration of hypothesis 5 opens a way to explore these intriguing questions.

Time plays an important role in the form of rule, and here I identify two hypotheses. The first is essentially Arjona’s, but stripped of the sub-factor related to competition. If a rebel group has a short time horizon, then it ought to prefer less direct forms of rule. Taking over the governance apparatus of a civilian population is an investment of sorts, one that is meant to pay off for reasons stated above. But, if a rebel group has an expectation of imminent defeat or victory, then why would it make that investment? There are various factors that can shape these expectations, but, in the aggregate, I hypothesize that the length of the expected time horizon will correlate with the form of rule.

H6

Rebel groups with longer timer horizons are more likely to engage in direct forms of rule.

A second time-related hypothesis relates to the longevity of the rebel group. Doyle and others in the empires literature (Gill 2016; Griffiths 2016), pointed out that imperial rule tended over time to intensify on its own. In a kind of ratchet-like manner, imperial governorship could be gradually drawn in to cover more and more competencies, and, once gained, were less likely to decentralize. Time itself tends to increase the form of rule for each dyad. I conjecture that the same dynamic will hold with rebel governors. All else equal, protracted governance over civilian populations will gradually become more direct.

H7

The longer the given rebel/civilian relationship, the more direct the rule becomes.

Collectively, these hypotheses are wagers on key dynamics with rebel groups and the form of rule. They are parts of the machinery, and ought to hold when the other factors are held constant. Although they are organized around the themes of push, pull, competition, and time, each could be further specified and potentially sub-divided. Each represents a direction for further research.

Conclusion

Recent work has explored how rebel governors choose the form of rule over civilian populations. This is a non-trivial research area, one with far-reaching consequences for the populations involved. My approach to this conversation was to bring in the much older literature on empire and the form of rule to see if lessons and concepts drawn from that body of work could shed light on rebel governance. To be sure, rebel governors and empires are different phenomena. Yet there is a rough similarity in that both entities need to determine their ruling relationship with peripheral populations. Indeed, scholars working in both literatures have theorized the architecture of those decisions in remarkably similar ways, and without an apparent knowledge of one another. To use an analogy, they are two distinct populations speaking different languages that are based on the same underlying syntax. My contribution was to distill the various mechanisms that shape those decisions and sort them into four categories: push, pull, competition, and time. Using those categories, I developed a set of hypotheses for rebels and the form of rule that may be useful for further research.

Some readers will want more, perhaps pointing out that each literature is much bigger than what I discussed. The literature on empire is particularly large and quite nuanced, and there are many other fruitful avenues for comparison that I elided. But drawing together two large literatures is no easy task and can too easily become a laundry list of terms and categories. I contend that the shared architectures on the form of rule provide a fascinating and useful focal point for comparison. Further work can be built upon that intersection. For example, the preferences of civilian/peripheral populations could be further problematized; there may be scenarios in which they prefer rebel rule over self-rule. Moreover, these actors—rebels and empires—are not monolithic. As Brenner showed in his study of the Kachin and Karen,Footnote 6 and Cunningham showed in her analysis of self-determination efforts,Footnote 7 rebel groups are typically divided. Likewise, empires can be quite compartmentalized, far-flung, and subject to “man on the spot” decisions. These are additional directions to explore, and my hope is that this analysis serves as a springboard for further comparison of the two research areas.