States of mind: The role of governance schemas in foreign-imposed regime change

Hypotheses

Four hypotheses flow from the preceding discussion. First, when interveners undertake statebuilding after invasions, they should be predisposed to promote political structures that fit their own schemas. Second, and conversely, they should either reject or have difficulty recognizing governance structures that do not conform to their schemas. Third, these decisions should be more reflexive than reflective – that is, based more on the ‘logic of habit’ than on cost-benefit calculus. Finally, interveners’ governance schemas should resist change. Only in response to a major shock, or another unusually urgent situation, should we expect interveners to question their own schemas in earnest.

These hypotheses are subject to several caveats and clarifications. First, I am not suggesting that interveners will always seek to reproduce their own ‘regime type’ – democracy, autocracy, etc. – in the target state. The boundaries established by governance schemas are broader than that. They are defined by the interveners’ conception of political authority, including answers to the questions listed above: Is political authority centralized or dispersed, personalized or depersonalized, formalized or informal, and what is the relationship between political authority and territory? Second, schemas do not determine outcomes; rather, they ‘orient’ action by making certain options appear natural and others undesirable or unthinkable. ⁴² In this sense, schemas give rise to what Barnett and Duvall call ‘internalized constraints’. ⁴³ Third, as noted above, maladaptive governance schemas are unlikely to be abandoned entirely, even if new schemas are adopted.

Recognizing the power of governance schemas does not invalidate the findings of the scholarship on foreign-imposed regime change. Specifically, I do not deny that cost–benefit calculations may play a part – even an important part – in interveners’ strategizing. I posit, rather, that taken-for-granted governance assumptions delimit the array of institutional options that outside powers readily conceive of promoting. Within these limits, decision-making processes may follow rationalist cost–benefit calculations, or other decision-making logics. Nevertheless, problematizing these boundaries is critical if we wish to understand why interveners do what they do, and why they promote certain kinds of governance structures, and not others, in countries undergoing regime change. This analysis may also shed light on the difficulty that interveners sometimes have in recognizing – and correcting – the shortcomings of their own statebuilding strategies.

Case selection and methodology

This article focuses on two recent cases of foreign-imposed regime change: the US-led invasions of Afghanistan and Iraq, which were followed by efforts to create functioning systems of governance in both countries. Using the method of ‘structured, focused comparison’, ⁴⁴ I pose six questions within each case study: First, what were the principal features of the interveners’ governance schemas? Second, what were the main elements of the statebuilding strategy that the outsiders pursued? Third, did this strategy mirror the interveners’ schemas of governance? Fourth, did the interveners give serious consideration to alternative strategies that did not match their schemas? Fifth, how did the interveners respond to the target society’s governance practices: did they embrace, reject, or ignore them? Finally, did they come to question their own governance schemas – and, if so, under what circumstances? Based on the answers to these questions, if the interveners (a) promoted political structures that mirrored their own schemas, (b) failed to give serious consideration to strategies falling outside their schemas, (c) rejected or ignored ‘non-conforming’ governance practices in the target society, and (d) avoided questioning or revising their assumptions until they faced a major shock or unusually urgent situation, then the hypotheses set out in this article will be supported.

However, the Afghanistan and Iraq case studies (which could be viewed as variants of a single case, because they both involved the same primary intervener and took place at roughly the same time) do not, in themselves, constitute a comprehensive test of these hypotheses. Rather, they offer what Arend Lijphart calls ‘theory-confirming’ case studies, ⁴⁵ or what Harry Eckstein terms ‘plausibility probes’, ⁴⁶ which allow us to determine whether the hypotheses hold true in some cases of post-invasion statebuilding. I selected these cases because they are arguably the most prominent instances of foreign-imposed regime change since the end of the Cold War, but it is important to note that, in both cases, there was a wide gap between the interveners’ governance schemas and political practices in the target societies. This gap probably biases the results in favor of a positive finding because, all other things being equal, the effects of schemas on statebuilding strategies are more likely to be visible when the difference between the interveners’ schemas and local governance practices is widest. However, the same conditions allow us to evaluate the hypotheses’ plausibility in circumstances where we should expect to see the effects of governance schema, and a positive result would provide a basis to pursue additional research, to refine hypotheses, and to conduct additional empirical tests. This article, in other words, should be seen as the first step in a multi-stage endeavor. Like all plausibility probes, it focuses more on theory-building than theory-testing; and, accordingly, it ends with a roadmap for future research.

In addition to the Afghanistan and Iraq cases, I conduct one further plausibility probe: the Mongols’ invasion and occupation of northern China in the thirteenth century. I selected this case for three reasons. First, if governance assumptions have the effects that I have posited, they should also be visible in interventions where modern conceptions of statehood were not the dominant schema of the interveners. As we shall see, the archetypal form of domestic governance for the US-led coalition was the Westphalian–Weberian state, a centralized, hierarchical, bureaucratic and geographically delimited model of political authority. The Mongols, by contrast, tended to think of governance as personalized, informal, clientelistic, clan-based and non-territorial (or not linked to a fixed location) – virtually the mirror image of the Westphalian–Weberian schema. Second, thanks to contemporaneous Chinese scholars, there are documentary records and accounts of the Mongols’ invasion and occupation practices in existence, which unfortunately is not true for many other pre-modern interventions. Third, both the Americans and the Mongols faced a similar challenge – that of establishing functioning systems of governance in territories they had invaded and whose regimes they had deposed – and both enjoyed an unusual degree of military supremacy relative to the conquered societies. These similarities provide a basis for asking the fundamental question of this article: What role did the interveners’ governance schemas play in these historically distant episodes?

The Mongol case also highlights the value of ‘historicizing’ the study of intervention and statebuilding, and of the international relations discipline more generally. ⁴⁷ This article is certainly not the first to examine pre-modern or non-Western cases in order to shed light on contemporary international theories and practices. Other scholars have, for example, examined the thought, institutions and actions of ancient China in relation to modern international relations; ⁴⁸ the ‘international relations’ of the ancient Near East; ⁴⁹ balance of power dynamics in ancient Assyria, India, East Asia and elsewhere; ⁵⁰ and various ‘international systems’ in the pre-modern era, including empires, city-states and nomadic tribes, ⁵¹ including patterns of governance, cooperation and rivalry among indigenous polities in the pre-Columbian Americas. ⁵² Such studies, however, have not been typical of the field. Research on statebuilding has, in the words of David Lake, been ‘surprisingly ahistorical’, ⁵³ the study of intervention has similarly focused overwhelmingly on Western and modern cases (mostly since the end of World War II), ⁵⁴ while the international relations discipline, more generally, has tended to relegate non-Western and pre-modern historical episodes to the ‘remote and incomprehensible past, casting them as idiosyncratic and transient preludes to the “real” international politics that emerged after 1945’. ⁵⁵ Yet, there appears to be a growing interest in historical sociology within the field of international relations, ⁵⁶ including among students of intervention. ⁵⁷ Given that, as Thomas Biersteker points out, ‘forms of state, meanings of sovereignty and conceptions of territoriality are neither fixed nor constant across time and space’, this is a welcome trend – one that this article seeks to advance. ⁵⁸

American schema of governance: the Westphalian–Weberian state

The US entered Afghanistan and Iraq with taken-for-granted conceptions of political authority and governance. Although American officials talked publicly about transforming both countries into electoral democracies, other elements of their schema were so fundamental that they went largely unspoken, including the presumption that restoring governance meant constructing Westphalian–Weberian states. This model of governance is ‘ingrained … in the modern imagination’ as the preeminent form of political organization in much of the world. ⁶⁰ In its archetypal form, it is ‘Westphalian’ in that it includes a central authority capable of exercising exclusive control over a bounded territory – a concept of sovereignty emerging from the seventeenth-century Treaties of Westphalia – and ‘Weberian’ in that this authority enjoys a monopoly on the legitimate use of violence and displays ‘rational–legal’ institutional characteristics, such as a centralized, hierarchical, governmental apparatus, in which political authority is attached to formal offices and administrative tasks are carried out by a specialized bureaucracy.

There is no reason, in principle, for statehood to be equated with the Westphalian–Weberian model. ⁶¹ Other kinds of states – including feudal states, ancient states, ‘composite’ states (involving a mix of heterogeneous administrative arrangements), ⁶² and the ‘moving state’ of the nomadic Mongols, which we shall examine below – have been organized around different principles. Nevertheless, the sovereign territorial state ideal has largely crowded out its competitors. ⁶³ This is not to say that all contemporary states fulfill the Westphalian–Weberian ideal in practice. The point, rather, is that this model has emerged as the predominant standard against which states are typically judged. As sociologists in the world polity school have pointed out, this is fundamentally a cultural phenomenon: ⁶⁴ Ideas about what governance is, and what it looks like, have become embedded in ‘the institutional and cultural structure of the wider international and world system’, ⁶⁵ and the Westphalian–Weberian model has come to serve as a sort of ‘cognitive script’ for the organization of political space. ⁶⁶ International organizations seeking to rebuild governing institutions in countries after civil wars, for example, have followed this script. ⁶⁷ So, too, did the US-led coalition in Afghanistan and Iraq, as we shall see.

Afghanistan: the Bonn process

After the fall of the Taliban in Afghanistan, the task of planning for a post-Taliban regime fell largely to the US, the United Nations (UN) and their principal Afghan partners. Consultations by UN special envoy Lakhdar Brahimi led to an international conference in Bonn, Germany, in December 2001, which produced an agreement to appoint Hamid Karzai as Afghanistan’s interim president and to launch a phased process of reconstituting a functioning central government. ⁶⁸ This process began with a Loya Jirga, or Afghan tribal assembly, to appoint a transitional government, followed by the drafting of a new constitution and national elections in 2004, which confirmed Hamid Karzai as Afghan president.

By the end of the process, however, Afghanistan had acquired ‘one of the most centralized’ systems of government in the world. ⁶⁹ Formal powers were concentrated not only in the national government, but especially in the office of president. As it happened, this model of governance was a decidedly poor fit for the country’s circumstances. The principal loyalty of most Afghans (80% of whom live in the countryside) was historically not to the central state, but to their own kin, village, tribal, ethnic, or professional group. These collective attachments and identities – loosely called qawms – have historically served as primary governance structures in Afghanistan, offering a measure of security, livelihood and community in a society that has been subject to recurring shocks through its history, including drought, war and the collapse of central government institutions. ⁷⁰ Indeed, it was not until the late nineteenth century that Afghanistan experienced the centralization of power in a Western-style state, under the leadership of Abdul Rahman Khan. His efforts to concentrate power in the capital were not ultimately sustainable, however, in part because they ‘went against the grain of Afghan tradition’ by attempting to eliminate the independent authority of Afghan qawms. ⁷¹ By contrast, most periods of successful Afghan governance (i.e. those not resulting in economic collapse, internal war, or both) saw a delicate political balance maintained between center and periphery, and between formal and informal actors. During these periods, national-level authorities ruled largely through informal negotiation with local power-brokers, patronage arrangements, tribal maneuvering and the occasional use of coercive force to assert central control over strategic locations, such as major population centers, while leaving more remote areas largely to themselves. ⁷²

In the end, the Bonn process yielded a governmental system that was paradoxically too strong and too weak. ⁷³ It was too strong, as noted, because of the extraordinary concentration of authority in the formal institutions of the central government and presidency. Yet, it was also too weak: the new Karzai regime lacked the basic fiscal and coercive instruments to perform the state’s minimal and traditional roles of patron, broker, negotiator and enforcer within the Afghan political context. This weakness stemmed, in part, from the initial unwillingness of the US and its international partners to provide the Karzai government with the money and military support it needed to exercise influence beyond the capital. ⁷⁴ The result was arguably the worst of both worlds: a governance system that was simultaneously over-centralized and under-resourced.

What did this outcome reveal about the governance strategies and assumptions of the interveners? On one hand, the Westphalian–Weberian state model was not foreign to Afghans, who had lived through a number of attempts to consolidate power in the capital, from Abdul Rahman Khan through to the period of Soviet occupation in the 1980s. ⁷⁵ Hamid Karzai, himself, reportedly favored the establishment of a centralized state in which most formal powers would be concentrated in the presidency. ⁷⁶ From this perspective, the ‘completely centralized state’ ⁷⁷ that emerged from the Bonn process could be seen as an expression of Afghan governance preferences, rather than foreign ones.

On the other hand, this assessment overlooks two countervailing considerations: First, most Afghans did not conceive of the central state as holding a monopoly of political authority, or even as providing the principal locus of political authority in the society. In practice, rural Afghans have ‘rarely depend on formal government institutions’ to regulate their affairs ⁷⁸ and have tended to view the state not as a ‘trustee of their common interest, but another particular interest like a tribe or clan’. ⁷⁹ Put differently, much of the society has historically eschewed the monolithic Westphalian–Weberian state model and has not thought of governance ‘in state terms’. ⁸⁰ Hence, while the modern territorial state has figured in the institutional repertoire of Afghan governance, it has been only one of several actors – most of which are local and informal – performing governance functions. For this reason, it would be misleading to suggest that the outcome of the Bonn process represented the ‘Afghan view’ of governance, even if Karzai (and those immediately around him) had a direct interest in a highly centralized state. Second, Karzai would not have been able to achieve this ambition were it not for the strong support of outsiders, most notably the US and the UN, both of which ‘adamantly opposed devolving power to the regional or provincial level’ and offered ‘uncritical support’ for concentrating power in the office of the president, ⁸¹ apparently in order to retain the benefits of working with a ‘clearly identifiable’ Afghan partner: namely, Karzai. ⁸²

There is no evidence that American or UN officials gave serious consideration either to other models of governance or to ways of accommodating informal solidarity structures at the regional and local levels. Apart from providing for Loya Jirgas to approve the interim administration and a new constitution, the Bonn document read like a ‘script’ in which outside interveners and the UN reiterated the institutional features of a Westphalian–Weberian state. ⁸³ The default assumption, apparently unquestioned and unacknowledged, was that post-Taliban Afghanistan would be governed by means of a formal state apparatus legitimized by democratic elections. James Dobbins, who was the American representative Bonn conference, later explained that his instructions were:

to get an agreement and almost any agreement would do, so long as it resulted in an Afghan government that would replace the Taliban’s, unite the opposition, secure international support, cooperate in hunting down al Qaeda’s remnants, and relieve the United States of the need to occupy and run the country. ⁸⁴

In a subsequent interview, he observed that US officials gave no consideration to alternative models of governance for Afghanistan. Asked if any policy planning or deliberation had taken place within the US government on this question, Dobbins answered: ‘No, nothing at all’. ⁸⁵ Barnett Rubin, an American expert on Afghanistan who served as an advisor to Brahimi at the Bonn Conference, offered a similar assessment, reporting that there was ‘no discussion whatsoever’ in the US government on different Afghan governance models. ⁸⁶ The basis for the new Afghan constitution, he explained, became the country’s 1964 constitution because ‘nobody had any other ideas’. ⁸⁷

The question of how the central state would interact with Afghanistan’s informal structures of governance was not answered, nor was it apparently even posed. Informal qawm-based networks, it seems, did not correspond to the interveners’ schemas and were not treated as ‘governance’ in the design of the post-Taliban political order. ⁸⁸ Noah Coburn, an astute observer of Afghan village life and of the international effort, characterized this approach as a kind of cognitive blinkering: For outside actors in Afghanistan, including international military and political actors, ‘all of whom were accustomed to operating in a political paradigm that conceptualized the state using the Westphalian model, thinking about the Afghan state in different terms was simply impossible’. ⁸⁹ ‘Impossible’ may be an over-statement, given that the interveners eventually revisited their assumptions, but their initial promotion of a centralized, formalized, bureaucratized governance model did seem reflexive and unquestioned.

The interveners’ disinclination, or inability, to think beyond their own governance schemas was visible not only in the overall design of the post-Taliban polity, but also in various specific institutional reforms. For example, efforts by outside experts to draft new laws and legal procedures in areas such as money laundering, financing of terrorism and bribery were ‘skewed from the start by a lack of sociocultural awareness of Afghanistan’s legal history, principles of Islamic law and Afghan customary law systems’. ⁹⁰ For a country in which social norms called for the resolution of disputes as much as possible outside official institutions, ⁹¹ and where most people placed greater trust in traditional justice bodies than the formal legal mechanisms of the state, ⁹² writing new state laws based on ‘transplanted foreign concepts’ was a recipe for these laws to be ignored – or, worse, actively resisted. ⁹³ The failure of outsiders to understand and accommodate the informal structures of Afghan justice sometimes verged on the tragicomic, such as when a senior Italian magistrate, tasked with drafting an interim criminal procedure code, ‘put the Afghan law aside and produced a brief version of the Italian’ legal code. ⁹⁴ These and other initiatives can be traced back to the Bonn agreement, which helped to ‘lock in a model of state building in which the new Afghan state was substantially to mirror the structure and functioning of developed Western states, irrespective of whether this would prove sustainable in the Afghan context’. ⁹⁵

Iraq: invasion and aftermath

Meanwhile, American policy-makers had already turned their attention to Iraq. After deposing the Saddam Hussein regime, the US established a Coalition Provisional Authority (CPA) with instructions to ‘restore conditions of security and stability … including by advancing efforts to restore and establish national and local institutions’. ⁹⁶ Although different parts of the US government – including the Department of State, Department of Defense, and US Agency for International Development – had begun examining postwar challenges in the weeks and months before the invasion, little of this work was incorporated into the formal planning process for the operation itself. ⁹⁷ Instead, the invasion strategy appeared to assume that existing institutions would ‘survive the invasion and remain sufficiently intact to continue to administer the offices of government and provide the Iraqi people with essential services’. ⁹⁸ As it turned out, however, Iraqi civil servants did not return to their jobs after the invasion – their ministries were looted and largely abandoned.

Once again, the US was confronted with the question of what type of governance arrangements it wished to promote. Coalition officials summarized their objective as the creation of a ‘sovereign, democratic, constitutional’ Iraq. ⁹⁹ As in Afghanistan, the Westphalian–Weberian assumptions of this project did not appear to be questioned. The CPA’s primary goal was to restore the central state as an effective apparatus of rule, albeit one that would embrace the principles of electoral democracy and free-market capitalism. The memoires of the CPA’s head, Paul Bremer, are revealing in this respect. ¹⁰⁰ When Bremer described the political reforms that he believed Iraq required, he framed this analysis entirely in terms of Iraq’s variance from an idealized democratic Westphalian–Weberian state: ‘[T]hirty years of tyranny had gravely distorted civil administration, jurisprudence, and any semblance of representative governance’. ¹⁰¹ Hence:

Iraq needed a new constitution, written by Iraqis … A modern constitution, I believe, was essential to define and set the boundaries for political activity … It would provide essential checks and balances in the new political system. ¹⁰²

When Bremer met with President Bush in the White House, he also explained how important it was to strengthen Iraqi ‘civil society’, but he defined this concept, too, in decidedly Western terms: ‘a free press, trade unions, political parties, [and] professional organizations’. ¹⁰³ According to Bremer, Bush replied: ‘I’m fully committed to bringing representative government to the Iraqi people’. ¹⁰⁴ Both men seemed to believe that a ‘liberal order was Iraq’s destiny’ and to assume that Western conceptions of statehood and civil society were readily transferrable to Iraq. ¹⁰⁵

However, these assumptions were deeply problematic. They overlooked important features of the Iraqi political system, including the role of religious leaders (Sunni and Shia), ethnic groups (Arabs and Kurds) and tribes and tribal confederations (mainly in Sunni Arab rural areas) in the informal governance of the country. ¹⁰⁶ The image of Saddam Hussein’s Baathist state as all-powerful had been a deliberately cultivated illusion. When the Baath Party came to power in the 1960s, it had attempted to use the state’s power to eradicate Iraqi tribes as social and political actors, not least because they were seen as a threat to party rule. ¹⁰⁷ However, these efforts had failed, leading the Baathists to pursue a different strategy: Rather than destroying the tribal structures, they would manipulate them for political advantage. ¹⁰⁸ Iraqi state officials maintained regular contact with tribal chiefs through ‘low-profile personal visits’ and ‘modest presents’. ¹⁰⁹ In exchange for cooperation, tribal chiefs could obtain government jobs and other benefits for their supporters, as well as ‘payments, access to weaponry, and a blind eye from Baghdad to smuggling and other illegal activities’. ¹¹⁰ (Those who did not cooperate might find themselves in mortal peril or facing a challenge from rival chiefs orchestrated by Baathist officials. ¹¹¹ ) Furthermore, Saddam’s reliance on the tribes increased over time. ¹¹² In the latter years of his rule, when the Iraqi state was suffering under international sanctions, his government ‘devolved functions such as judicial powers, tax collection and law enforcement to the resilient local tribal or kin networks it detected’. ¹¹³ Although the Iraqi state prior to the 2003 invasion still retained considerable coercive power, tribal networks represented a ‘parallel hierarchy that overlap[ped] with formal government structures and political allegiances’. ¹¹⁴

American officials appeared to have little knowledge of these dynamics prior to the 2003 invasion, or even afterward. ¹¹⁵ David Kay, head of the Iraq Survey Group, a group of more than 1000 linguists, interpreters and aides combing Iraq, later described the time immediately following the invasion as the period of ‘our maximum ignorance’, adding: ‘It is hard to believe how little we understood’. ¹¹⁶ Some American observers implored the coalition to pursue more ‘tribal-based’ governance policies in post-invasion Iraq, but their ideas were largely ignored. ¹¹⁷ Instead, the CPA focused on building a strong Weberian-Westphalian state apparatus, eschewing tribes and other non-state orders in Iraqi society, and blocking occasional attempts by US personnel to set up meetings between tribal leaders and senior officials. ¹¹⁸ This was not simply a strategic preference; it appeared to reflect deeper assumptions about what counted as genuine governance – assumptions that drove the plan to build a ‘modern’ state in Iraq, defined as a centralized, bureaucratized, legal–rational, nontribal state. ¹¹⁹ In the words of Peter Mansoor, who commanded a US army brigade at the time (and who later wrote two books about the American invasion and occupation of Iraq), the ‘CPA’s policy toward the Iraqi tribes showed its lack of cultural awareness and understanding of the complex relationships that existed in Iraqi society’. ¹²⁰ The CPA, he contends, viewed tribes as ‘an outdated part of Iraq’s past’ ¹²¹ and as ‘incompatible with a modern democratic state’. ¹²² Put differently, they did not fit the governance schema.

This blinkered approach proved costly, contributing to the alienation of the Sunni tribes, which were concentrated in an area to the west and north of the capital – the so-called Sunni Triangle. ¹²³ As resistance to the American occupation grew, fueled partly by the political sidelining of the minority Sunni population, security began to collapse, creating an opening for Al Qaeda to gain a foothold. ¹²⁴ Participants in the insurgency ‘came from a mixture of groups and included former senior Baathists, tribesmen and foreign fighters’. ¹²⁵ Anbar province, in the southwestern corner of the Triangle, was one of the worst areas for insurgent attacks and American casualties. It also illustrated, perhaps better than any other part of the country, the limitations of the US approach to – and understanding of – governance in post-invasion Iraq. Anbar’s tribal confederations had long performed critical informal governance roles, such as ‘levying taxes, providing social goods and generally acting as a political entity with religious and ethnic overtones’. ¹²⁶ These were some of the groups that the Baathists quietly relied upon to perform governance functions alongside the formal state.

Yet, American officials largely ignored Anbar’s tribal sheiks – or treated them as enemies. Engaging these leaders in a sustained political dialogue was not part of the plan. The tribes were at best ancillary, and at worst inimical, to the Iraqi statebuilding effort in Baghdad. The CPA devoted little attention to governance arrangements at the regional and local levels – a ‘lost opportunity’, according to several observers. ¹²⁷ Meanwhile, as US policy downplayed the political role of Sunni tribes and local governance, focusing instead on killing insurgents and building up the institutions of the central government, violence in the country worsened, leading to the deaths of 30,000 Iraqis in 2006 alone. ¹²⁸

Reimagining governance in Iraq

Faced with a deepening crisis, some US military units began experimenting with different approaches to achieving stability. Much attention has since focused on the innovative efforts of US divisional commanders, including Gen. David Petraeus and Gen. James Mattis, who worked directly with local leaders and to ‘replac[e] military operations with civic action’ within their respective areas of operation. ¹²⁹ These efforts sought to win the support of the local populations, or at least to blunt their opposition to the coalition presence and the US-led statebuilding process. But similar experimentation was also taking place at much lower levels of command and in individual units from 2005 onward, including in Anbar province, ¹³⁰ where conditions by mid-2006 were so dire that the intelligence chief for the US Marine Corps in Iraq reportedly told his superiors that there was almost nothing the US military could do to improve the political and social situation. ¹³¹ With little left to lose, some American military officers began negotiating deals with tribal sheiks, including agreements allowing armed tribal militias to guard their own neighborhoods, an approach previously ‘expressly forbidden by US policy’. ¹³² Some of these tests seemed to work. In a number of instances, Sunni tribal chiefs split with Al Qaeda and received US payments and permission to police their own communities.

These initiatives contributed to a sea change in US strategy in Iraq, which ultimately shifted toward ‘bottom-up political accommodation at the local level’ with informal actors, including tribal groups. ¹³³ Experimentation and adaptation by US military units – initially ‘with little direction from higher military and civilian authorities’ – was one driver of this change. ¹³⁴ Another was the recognition at the highest levels of the US military and government that the existing strategy was failing badly. This situation of great urgency created an opening for policy entrepreneurs inside and outside of the US government to press for a rethinking of the strategy. ¹³⁵ Petraeus, whose personal experiences as a divisional commander in Iraq had produced encouraging results, ¹³⁶ became the unofficial spokesperson for this group. In 2005, he returned to the United States as the commander of an army think tank, where he assembled military and civilian experts to write a new field manual for counterinsurgency operations, released in December 2006. ¹³⁷ The manual identified the core objective of counterinsurgency as ‘effective governance by a legitimate government’, and it presented ‘legitimacy’ as a matter to be defined by the local populace. ¹³⁸ Indeed, the document described ‘Western liberal’ conception of legitimacy as one of several examples, which also included medieval monarchies, theological societies, and imperial China. The implication of this historical comparison was clear: American (or Western) ideas about the state were neither timeless nor universal, and it was vital to understand and build upon local conceptions of legitimate governance. The document (and earlier field-level experiments in tribal accommodation) had the effect of calling attention to hitherto taken-for-granted assumptions about governance.

The preparation and release of the counterinsurgency field manual also coincided with a major review of Iraq policy in the White House. After years of mounting casualties and frustratingly little progress on building effective Iraqi governmental institutions, many American political and military leaders (including the commanders of the Iraq operation) were arguing that the US military should disengage from the country. ¹³⁹ Others, joined by several people who had been involved in preparing the manual, pressed instead for a recommitment of US forces and a change in strategy along the lines of the counterinsurgency doctrine. ¹⁴⁰

President Bush ultimately decided to increase the number of US troops in Iraq – the ‘surge’ – and to appoint Petraeus as the new overall commander of the mission. The US military henceforth ‘radically changed’ its mode of engagement in Iraq. ¹⁴¹ In addition to dispersing troops into local communities in order to provide better security to the population, the new approach witnessed an ‘expansion of outreach to the tribes’ and more extensive ‘partnering with tribal militias that formed localized security forces’. ¹⁴² While US officials continued their efforts to strengthen the administrative capacity of the central state and to encourage national-level politicians to reach consensus in Baghdad, they now also embraced the importance of ‘delving into intricacies of Iraq’s religious, ethnic, and personalistic factions … to see what kind of levers of persuasion and influence might be effective in bringing them together’. ¹⁴³ Although there were reasons to be skeptical about the sophistication of this approach – during the Cold War, US counterinsurgency approaches may have obscured, rather than revealed, local social and political realities in some cases ¹⁴⁴ – the shift toward counterinsurgency in 2006 seemed to represent a ‘cultural revolution’ in America’s Iraq strategy: Understanding Iraqi society and engaging its informal political actors were suddenly high priorities for the coalition. ¹⁴⁵

This change could be seen as a pragmatic, if belated, response to a deteriorating security situation and the failure of the Baghdad-centric governance strategy, ¹⁴⁶ but it also entailed the ‘reimagining’ of governance assumptions that had gone largely unquestioned and unacknowledged until then. As Toby Dodge notes, the new approach placed an ‘almost anthropological focus on the culture of the populations among whom US forces were now operating’ – a body of knowledge that US officials believed was ‘crucial to understanding how societies were organized, how they gained security, and how insurgents sought to mobilize them’. ¹⁴⁷ The occupiers, in other words, had shifted their conception of governance, which now included a role for kinship networks and tribal leaders. Indeed, this counterinsurgency model has even been described as an implicit form of ‘constitutional design’, in that it recognized and empowered individuals and groups as de facto political authorities. ¹⁴⁸ ‘Constitutional’ or not the interveners’ conception of governance had undergone a transformation and now extended well beyond the Westphalian–Weberian ideal.

In the short term, the strategy appeared to work: violence declined markedly. American officials were less successful, however, at convincing the Shia-dominated Iraqi government to treat informal Sunni groups as legitimate political actors in their own right. Although the Baghdad authorities agreed to create a new organization to determine which Sunni groups they were prepared to work with, ¹⁴⁹ once the US presence in the country started to wane, the Iraqi government arrested and imprisoned many of the local Sunni leaders who had come forward. ¹⁵⁰ Efforts to engage informal Sunni political structures, therefore, provided only temporary respite from the unrest unleashed by the US-led coalition when it toppled the Saddam Hussein regime.

A new strategy in Afghanistan

As the US surge unfolded in Iraq, conditions in Afghanistan were worsening. American policy-makers and other international officials had assumed that democratic elections would legitimize the new central government and provide a basis for political stability in the country. However, ‘the relationship between elections and political legitimacy was less clear for Afghans’, who looked to their leaders primarily to deliver ‘security, economic improvement, and a functioning government’. ¹⁵¹ In practice, the Afghan state that emerged from the Bonn process – which, as noted above, was simultaneously too strong and too weak – failed this test in many parts of the country. The Taliban, now reconstituted in neighboring Pakistan, skillfully exploited this situation by mocking the Karzai regime as an ineffectual stooge of foreigners, while reinfiltrating operatives into the countryside where there were still few Afghan or foreign troops. In 2006, international forces were deployed to the predominantly Pashtun southern and eastern provinces of Afghanistan near the border with Pakistan, where the Taliban had enjoyed its greatest support. Violence mounted steadily in the ensuing years, ¹⁵² but Washington remained distracted by the dire situation in Iraq. Only near the end of the Bush presidency, in 2008, did the US Administration show signs of acknowledging the mounting crisis in Afghanistan. ¹⁵³

Barack Obama succeeded Bush in January 2009 and sent a new commander, Gen. Stanley McChrystal, to Afghanistan. McChrystal analyzed the situation and reached a grim, if by then obvious, conclusion: the United States and its allies were losing. ¹⁵⁴ He called for a profound change in the coalition strategy in Afghanistan, clearly informed by the new counterinsurgency doctrine and the apparent lessons of Iraq. ¹⁵⁵ Among other things, he noted that a ‘top-down approach to developing government capacity [had] failed to provide services that reach local communities’ and that the Afghan government had ‘not integrated or supported traditional community governance structures’, which he characterized as ‘historically an important component of Afghan civil society’. ¹⁵⁶ McChrystal’s willingness to embrace customary political structures was explicit throughout the text, including in his call for greater support of ‘local informal justice systems’. ¹⁵⁷ Many of the same themes also appeared in the ‘integrated campaign plan’ issued to US military commanders and civilian officials, which called for a ‘bottom-up’ approach to security and governance-building, including pilot ‘community security arrangements’ and ‘traditional dispute resolution mechanisms’. ¹⁵⁸

This reorientation of US policy continued when Petraeus replaced McChrystal as Afghanistan commander in mid-2010. ¹⁵⁹ Upon arriving, Petraeus required senior officers in the country to read a paper, ‘One Tribe at a Time’, that a special forces major had written based on his own experiences in Afghanistan. ¹⁶⁰ The paper advocated a ‘complete paradigm shift’ in US policy toward ‘strategy of tribal engagement’ on the grounds that ‘the tribal system is the single, unchanging political, social and cultural reality in Afghan society’. ¹⁶¹ Such formulations may have conveyed a reductionist understanding of Afghan society and a monolithic conception of ‘tribe’, but the fact that this type of analysis was now embraced by the US command underscored the shift in American strategic thinking and in its conceptions of governance since the days of the Bonn agreement. The interveners had come to recognize the need to accommodate, not ignore, the informal qawm-based networks that were so important to Afghan social cohesion and governance.

The coalition did not abandon the idea of building up the central state. Far from it, the US and its partners continued to espouse the principle of sovereign statehood in both Iraq and Afghanistan, including the need for effective formal governments. Now, however, they explicitly recognized the value of ‘non-state’ modes of governance, from informal local militias to traditional justice mechanisms. In this sense, they had also come to doubt the Westphalian–Weberian schema that had previously gone largely unquestioned and unacknowledged, and they had adopted a new schema that conceived of governance in considerably broader terms. The result was a hybrid policy – one that still aimed to strengthen the capacities of the central state, but also underscored the importance of acknowledging and engaging with qawm-based entities as genuine forms of governance in their own right.

In sum, both the Iraq and Afghanistan cases appear to support the hypotheses set out above. Specifically, the interveners promoted political structures that fit their own taken-for-granted governance schemas, they either rejected or failed to recognize local governance practices that did not conform to their schemas, they appeared to make these decisions reflexively rather than than reflectively, and they avoided questioning their schemas until they faced a situation of unusual urgency – namely, imminent defeat.

Mongol schema of governance: the ‘moving state’

Mongol social and political forms were well adapted to their nomadic lifestyle. Extended family groups – or ‘tribes’ ¹⁶² – moved with their livestock, sometimes covering great distances as they followed seasonal migration routes between pasturelands. Although individual tribes occasionally combined into loose confederations based on kinship, they had few reasons to unite permanently or to pledge loyalty to higher leaders. However, there was one thing they could not do on their own: conquer and extract wealth from neighbouring societies. Accomplishing these goals required coordination and cooperation among the tribes, a role that came to be performed by supra-tribal leaders, or ‘khans’. Loyalty to the khan rested only on his ability to manage fractious intertribal politics and to provide a regular supply of war booty and land fiefs, or ‘appanages’, to his followers. This was how Chinggis Khan acquired power and united the Mongol tribes between 1186 and 1209. ¹⁶³

Political authority in this society resided primarily in interpersonal and inter-group relationships, both within and between tribal units. The Mongols, in other words, had ‘no abstract concept of loyalty to the office of tribal chieftain’. ¹⁶⁴ In contrast to more depersonalized structures of authority (in which offices exist independently of the people who occupy them, such as in the Westphalian–Weberian ideal), continuation of the great khan’s office depended on the person who claimed that title and his ability to maintain patronage networks. ¹⁶⁵ Partly as a result, Mongol governance remained largely decentralized, even as the great khan gained in power. By granting appanages to lower-ranking leaders and allowing them to acquire spoils on their own initiative, he recognized and reinforced the decentralized power of ‘local’ chieftains, while simultaneously increasing his own prestige. ¹⁶⁶

The resulting polity consisted of ‘a combination of tribal and feudal’ structures, or extended family networks criss-crossed by patronage arrangements. ¹⁶⁷ The decision-making practices of the supra-tribal polity also reflected this dispersed structure of political authority. Occasional consultative gatherings, which brought together tribal leaders from great distances, served as ‘the only legitimate forum for political and military issues’, including the acclamation of the great khan himself. ¹⁶⁸ Nor did the Mongols have a formalized legal code, or even a written language, until Chinggis Khan took steps to introduce both. As Hidehiro Okada points out: ‘In Mongol society all organizations were built on the personal, rather than the bureaucratic, principle’. ¹⁶⁹

Moreover, because Mongol tribes were continually in motion, territory held a different political significance than in the modern conception of sovereignty. Although access to rich pasturelands was valued, as was control of conquered territories that provided opportunities for plunder, political authority was largely dissociated from physical space. ¹⁷⁰ Some contemporary Chinese writers called this form of governance a ‘moving state’, an apt phrase capturing the mobility, fluidity and informality of Mongol rule. ¹⁷¹ Even after storming across Eurasia, Chinggis Khan continued to rule his empire from horseback, both figuratively and literally. He roamed the steppe followed by a bodyguard that also served as the empire’s informal administrative apparatus. To the extent that the empire had a central administration, it was located wherever the great khan ‘chose to alight’ from his horse. ¹⁷²

This was the Mongol way of governance, which they saw as rightful and sacred, a belief rooted in a sense of cultural supremacy and a conviction that the sky-god Tengri had fated the Mongols to rule the world. ¹⁷³ Put differently, they regarded their own governance traditions and institutions as universal and a reflection of the divine. ¹⁷⁴ The ‘moving state’ was not just a superior form of governance. In Mongol eyes, it was governance.

Early attempts at governing China

When Chinggis Khan died in 1227, he bequeathed to his successors an empire that consisted mostly of nomadic peoples on the Eurasian steppe. ¹⁷⁵ The major Asian centers of wealth and power – China and Persia – were still mostly beyond the Mongols’ control. Both were sedentary, populous, agricultural societies, whose governing institutions were profoundly different from those of the Mongol empire. China, in particular, had long been ruled by a centralized, formalized bureaucracy, quite unlike the Mongol method of decentralized, personalized, informal governance. Even though Chinggis Khan had campaigned in nearby regions of China, conquering settled states was not his priority at all. He preferred to keep most of his people on the grazing lands of the steppe, where they could maintain their traditional nomadic lifestyle. ¹⁷⁶ In the five decades following his death, however, Chinggis Khan’s successors undertook to conquer China in earnest, leading to the defeat of the Jin dynasty in 1234.

The Mongols now faced a challenge: What type of governance arrangement would they establish in the settled Chinese territories they occupied? During and immediately after the Jin conquest, they simply looted China, employing ‘terror as a tool to discourage resistance and destroy[ing] whole districts, regardless of economic consequences’. ¹⁷⁷ Thereafter, however, they began to introduce their own governing models and methods, less by design than by default. The practice of granting appanages to military commanders and prominent followers, for example, created de facto Mongol fiefdoms in northern China, with political control devolving to the rulers of these fiefdoms, according to Mongol tradition. ¹⁷⁸ Tax collection was similarly decentralized and irregular, amounting to ‘an endless series of extraordinary requisitions and extractions’. ¹⁷⁹ This, too, was a distinctive feature of the Mongol economic and political system, wherein senior leaders repaid their followers’ loyalty by giving them license to plunder. The occupiers also began to apply Mongol customary law to the territory, ¹⁸⁰ and they employed non-Chinese foreigners to perform local services, such as tax farming. ¹⁸¹

However, these methods would soon prove ineffective and terribly costly to both the Chinese and their Mongol overloards. The decentralized approach to administration may have functioned well for a nomadic society on the unsettled steppe, but it was a ‘disaster’ for the complex Chinese agricultural economy, which went into a steep decline. ¹⁸² The Mongols neither understood nor evinced much interest in the practices and economics of cultivating the soil for food. ¹⁸³ On the contrary, they regarded farmers and sedentary peoples as little better than their own grazing animals (and used the word ‘herd’ to describe both). ¹⁸⁴ In localities under their control, Mongol overlords undermined agricultural production by driving Chinese peasants from their farms, enslaving them, by imposing arbitrary and impoverishing taxes, and by returning tilled fields to grazing land. ¹⁸⁵ In the resulting economic crisis, the Mongols faced mounting resistance and disorder in northern China, to which they responded with even more brutality. The results were devastating, as Frederick Mote explains: ‘People fled to escape their enslaved status; human resources were wasted, cultural norms were destroyed, and the society was rendered well-nigh ungovernable’. ¹⁸⁶

Mongol methods of administration, in other words, were very poorly adapted to the requirements of ruling this sedentary society. ¹⁸⁷ Recognizing this, some Chinese paradoxically implored the occupiers to strengthen their central management and control of their society. In fact, the invaders’ informal and decentralized legal system was so fluid and inconsistent, many Chinese, who were accustomed to more formal codes, viewed it as ‘no law’ at all. ¹⁸⁸ One Chinese scholar, who served as an advisor both to Chinggis Khan and his successor, is said to have warned the Mongols: ‘Although you inherited the Chinese Empire on horseback, you cannot rule it from that position’. ¹⁸⁹ But this is precisely what the Mongols did for many years, despite growing evidence that their methods of governance were ill-suited to ruling China and were destroying the very wealth that the Mongols wished to extract.

The puzzle is why the Mongols adopted and persisted in this approach. There are three possible answers. One is that the Mongols did not really have a governing system of their own; they were raiders and pillagers who conquered for spoils. According to this view, it should not surprise us that the Mongols lacked the capacity to govern China, a society with ‘complex administrative traditions of long standing’. ¹⁹⁰ This perspective, however, misconstrues the Mongol system of governance, which was not only real, but also complex and sophisticated, reflecting the pastoral–nomadic conditions in which it had emerged. ¹⁹¹ Navigating the intricacies of Mongol tribal and feudal relations would have required considerable skill. Although Chinggis Khan is remembered for his military talents and his brutality toward his enemies, he was, first and foremost, a brilliant politician.

A second possible explanation is that the Mongols did not care if their governance system failed in China because they considered conquered peoples to be chattel. This is true, but only to a point. Subject populations were seen as possessions to be exploited, ¹⁹² but the problem with this explanation is that Mongol governance of China failed even as a mechanism of exploitation: it undermined the agricultural economy and provoked local resistance. Yet, in spite of this failure, the Mongols stubbornly persisted with their own governance methods for many years, and paid little heed to the administrative models and methods that had served the Jin dynasty.

The best explanation for this behavior is not that the Mongols lacked a governance system, or that they simply sought to exploit their conquests, but rather, that the Chinese administrative system did not register in the Mongol schema of governance as a legitimate form of rule that deserved to be appropriated, replicated, or even seriously considered by the Mongols. The Jin dynasty’s governing institutions – geographically fixed, centralized, hierarchical, and largely depersonalized, including specialized agencies, formal tax systems and codified law (not unlike the Westphalian–Weberian state model that developed in Europe centuries later) ¹⁹³ – were ‘utterly alien’ to the Mongols, whose political universe was defined by the moving state of the nomadic steppe. ¹⁹⁴ Although they gradually came to rely on foreigners to perform administrative roles in parts of their empire (including as scribes and translators) and borrowed some institutions from conquered societies, the Mongols had little understanding of Chinese governance, nor did they seem interested in learning about it. Sedentary peoples and their governing institutions were ‘no more part of [their] political universe than were the domestic animals of the steppe’. ¹⁹⁵ Even though the wealth of their growing empire eventually stimulated ‘a rudimentary form of steppe urbanization’, Mongol tribal leaders and khans typically choose not to live in these proto-towns, preferring not to ‘associate themselves with sedentary urban culture’. ¹⁹⁶

In this sense, the Mongols did not choose to replicate their own governance forms in China, any more than they chose to eschew Chinese institutions. Rather, it ‘never occurred’ to Chinggis Khan or his immediate successors ‘to link Mongol government into China’s historical tradition’. ¹⁹⁷ For the same reason, most Chinese appeals for ‘judicial and institutional rationality and centralism’ fell on deaf Mongol ears. ¹⁹⁸ Indeed, in the years after the Jin conquest, the invaders made only minor adjustments to their ruling methods, ¹⁹⁹ despite the confusion and chaos that resulted from their failed attempts ‘to apply the principles of nomadic federation to the Chinese’. ²⁰⁰

Khubilai Khan and the reimagining of governance

The Mongols were aware of the mounting crisis in China, but seemed unable to conceive of governing in a different manner. It would take someone in a position of influence to lead the process of questioning and rethinking the governance schema. That person was Khubilai Khan. In 1236, just two years after the Mongols deposed the Jin dynasty, 21-year-old Khubilai received an appanage in northern China with a population of 10,000 households. ²⁰¹ Soon afterward, in his corner of the Mongol empire, Khubilai began experimenting with new approaches to governance. He recognized that the decentralized and informal Mongol system of taxation – which relied largely on subcontracted tax farmers and extraordinary requisitions, combined with the impressing of Chinese peasants for forced service at the behest of Mongol overlords – was driving Chinese farmers from their fields, heightening tensions with the local population, and destroying the local economy. ²⁰² These problems reflected the larger mismatch between Mongol governance and the socioeconomic circumstances of sedentary China.

He consulted widely with Chinese experts. Although it was not uncommon for Mongol rulers to retain indigenous advisors, Khubilai displayed an unusual level of comfort and willingness to seek local counsel, even at a young age. ²⁰³ He then undertook a series of reforms. Among other things, he transferred the power to manage Chinese farmlands from individual Mongol overlords to Chinese technical advisors, thus creating a proto-civil service in the traditional Chinese model. He also rationalized and centralized the tax system within his domains. By 1251, when he gained responsibility for most of northern China, he had set up ‘what amounted to Chinese-style central government agencies to engage in all aspects of civil governing’. ²⁰⁴

These methods departed significantly from prevailing Mongol practices. Khubilai’s grandfather, Chinggis, had been unwilling to recognize or accommodate Chinese ways of governance. ²⁰⁵ He and his immediate successors had been ‘deeply imbricated’ in the traditions of nomadic rule. ²⁰⁶ Although some of Khubilai’s predecessors borrowed ideas and technologies from other peoples, his experiments represented more than minor adjustments to Mongol governance methods. He was, in effect, stepping outside the normative and cognitive limits of the orthodox Mongol schema and ‘seeing’ governance in a different way.

For precisely this reason, his experiments put him on a collision course with the Mongol leaders elsewhere, who not only criticized his behavior as imprudent, but also castigated his use of Chinese institutional forms as a form of apostasy – a corruption of ‘universal’ Mongol norms. As Khubilai’s power grew, Mongol traditionalists back on the steppe saw him as a ‘threat to their way of life’. ²⁰⁷ By the early 1250s, Khubilai’s elder brother, Möngke, had become the great khan. Some of Möngke’s allies attempted to turn him against Khubilai, arguing that the younger lord had flouted traditional Mongol precepts, but Möngke refused to take action against his brother. Möngke died soon afterward, setting a stage for a power struggle between Khubilai and his rivals. Those who sought to preserve Mongol ways ‘uncontaminated’ rallied around another one of Khubilai’s brothers, Arigh Böke. ²⁰⁸ In 1260, both brothers were named as great khan in competing ceremonies, giving rise to an ideological civil war over the ‘proper ordering of the Mongol Empire’. ²⁰⁹ After a series of battles, Khubilai prevailed.

Khubilai Khan now ruled the whole empire. The bulk of his attention, however, remained focused on China, where he extended his earlier governance reforms, now on a much larger scale. Among other things, he dispensed with informal tax farming and systematized tax collection – payments that peasants had previously made to local lords would now be remitted directly to the central government ²¹⁰ – and he waived or reduced taxes in areas experiencing economic difficulties. ²¹¹ He also created a bureaucracy to promote agriculture and more productive uses of farmlands, gave official standing to local groupings or councils, and granted some village-level entities with the power to make decisions on local matters. ²¹² Most dramatically, he declared a new Chinese dynasty under Mongol rule, and announced that the capital of the Mongol empire would be located on Chinese territory. Khubilai Khan understood that he needed to adopt Chinese institutional structures and symbols in order to achieve political quiescence and to promote wealth-creation in occupied China. All of these measures were aimed at establishing ‘stable administration’. ²¹³

He did not, however, entirely abandon the Mongol governance schema. He continued to perform Mongol rites and rituals at his court, forbade intermarriage between Mongols and Chinese, and banned Chinese from learning the Mongol language. ²¹⁴ He also prohibited Chinese advisors from occupying the highest administrative posts, refused to reinstate traditional Chinese civil service examinations (which would have reduced his control over the bureaucracy) and preserved a ‘conciliar, deliberative style of decision-making’ in the distinctively Mongol manner of a tribal assembly. ²¹⁵ This resulting system was neither Mongol nor Chinese, but ‘consisted of a mixture of different political and cultural elements’. ²¹⁶