The Origins and Legacies of Unpredictability in Rebel-Incumbent Rule

ABSTRACT

Many rebel groups ‘govern’, becoming increasingly institutionalised, accountable, and predictable. This is now well-accepted; however, less attention has been paid to another common observation: some rebel orders—and rebel-incumbent regimes—are more aptly characterised as unpredictable. We find that this is because they adopt vague mandates and delegate provisional authority. Our analysis shows that, in some cases, this (1) allows rebels to accommodate potentially incongruous pre-existing authorities and institutions, which can integrate unpredictability into early governance arrangements; (2) helps rebels cultivate social control from a comparatively weak position; and (3) has enduring legacies for rebel-incumbent rule. We illustrate this argument with Uganda’s National Resistance Movement (NRM) and explore broader relevance with the Afghan Taliban.

Introduction

A great innovation of the burgeoning rebel governance scholarship has been to study how rebels create novel forms of political order. Drawing on theories of state formation and consolidation, this scholarship has been significantly shaped by the idea that the most efficient organisation of violence is a predictable organisation (al-Tamimi 2015, Péclard and Mechoulan 2015, Arjona 2016, Stewart 2018).¹ Rebels build social contracts with civilians both to enhance their own legitimacy, and to undermine the state (Wickham-Crowley 1987, p. 478, Arjona 2016, p. 171, Revkin and Ahram 2020). This view links political order to political predictability (Worrall 2017, p. 711). Political unpredictability is thus interpreted as an indicator of limited capacity and weak institutions, attributable to a rebel group’s embryonic quality, short time horizons, or limited presence (Arjona 2016, pp. 10, 13, 161).

In contrast, we argue that political unpredictability is a hallmark of some rebel orders and the rebel-incumbent regimes that emerge from them. We define unpredictability as the inherent quality of being unable to reliably foresee or anticipate outcomes, making them challenging to depend on or plan for. While there is not a deep academic consensus on this term, as we elaborate in our section on ‘Bringing Unpredictability Back In’, our definition is informed by the Cambridge Dictionary (2023): ‘tending to change suddenly and without reason or warning, and therefore not able to be depended on’. As we elaborate subsequently, we diverge from scholarship that sees unpredictability in governance as primarily an issue of either personalisation or poor coordination deriving from weak or disrupted institutions. Instead, we emphasise unpredictability as a product of how governance institutions are structured and organised,² deriving from institutional multiplicity, vague and ambiguous mandates, and provisional authority.

Our contention is not that rebels necessarily set out to establish unpredictable political environments. Instead, we start from the observation that rebels often must contend with pre-existing authorities and institutions. We find that in some cases, they manage these relationships through deferral and ambiguity, rather than establishing clear lines of authority and accountability. Indeed, during war, there are particular incentives to structure governing arrangements in this way: unpredictability creates wiggle room for insurgents to continually redefine their scope of authority, producing an environment in which it is easier to reallocate responsibility and blame, redefine loyalty and betrayal, and redistribute opportunity and onus. Once set in motion, what may have begun as compromises or half-measures can become the foundations of incumbent rule.

Unpredictability is not unique to rebel regimes; rather, our analysis emphasises first, that increasingly consolidated rule does not necessarily correlate with increasing predictability, and second, shows why unpredictability may be particularly prominent in orders emerging from rebel rule. While political unpredictability can have many sources, we identify several that may be especially pertinent to rebel governance: rebels may lack the clout to alienate pre-existing authorities, and the know-how to govern without enrolling members of the former regime. In other cases, they may even find potential synergies with pre-existing governing arrangements. They thus may forge institutional partnerships of convenience, for example, leaving in place state administrative structures or forms of traditional authority that then become enmeshed in the post-rebel governing order.³ Even when rebels implement their own governance structures, they may leave the authority of these structures provisional and ambiguous to maintain rebel control and limit the likelihood that new governance structures will become autonomous spaces from which to challenge the rebel leadership. As a result, it can be structurally unclear which authority (and which set of rules) will be definitive in a given situation, and relatedly, which authority will have the power to implement said decision.

These initial conditions can have long-lasting legacies: the vast array of overlapping and competing public authorities tolerated during conflict and in its immediate aftermath do not necessarily settle into a predictable relationship (for example, where one’s ethnicity determines which set of rules applies, or where formal and informal institutions are combined to form a synthetic set of compatible rules).⁴ Instead, these originally ambiguous and nominally provisional arrangements may preclude the emergence of clear lines of accountability between public authorities (whether state or not) and citizens. These legacies provide fresh insight into how some rebel-incumbent regimes are likely to evolve, with implications for regime longevity, democratic practice, and collective claim-making among others.

We develop the argument based on Uganda’s National Resistance Movement (NRM) and set out suggestive parallels based on the case of the Afghan Taliban. We selected these insurgencies as both ‘most different’ and ‘least-likely’ cases for our inquiry, helping establish the plausibility of our argument (Eckstein 1992, p. 158). In many respects, the Taliban and NRM could not be more different in the variables thought to shape rebel governance – in terms of ideology, battle-ground tactics, and international networks. However, following Nelson Kasfir (2015), both groups nonetheless meet the bar for rebel governance, defined as contexts in which rebels first, ‘hold some territory within the state against which it is rebelling, although its control over specific territory may fluctuate temporally and spatially. Second, civilians must reside in that area. Third, the group must commit an initial act of violence to become rebels and then either continue hostilities or credibly threaten them in territory it governs’ (Kasfir 2015, p. 25).

They are also ‘least likely’ cases, in the sense that they have been comparatively successful and therefore, in theory, should operate with a baseline level of predictability. Said differently, if some of the most successful rebel groups develop and sustain (eventually sovereign) regimes characterised by political unpredictability, then we might wish to re-evaluate views that see unpredictability as an indicator of weakness or superficiality, for instance due to limited capacity or short time horizons. Our data is drawn from a careful re-reading of secondary source material and interviews; we also have research experience that helped inform our theoretical framework, argument, and analysis (Tapscott in Uganda since 2014 and Urwin in Afghanistan since 2013).⁵

To be clear, our contribution is not identifying that rebel orders are sometimes unpredictable. This observation has been widely documented (see, e.g., Mampilly 2011, Reno 2015, Hoffmann and Verweijen 2019). Instead, our contributions are (1) to conceptualise and analyse unpredictability not as a product of institutional weakness, but as an element of institutional design that in turn has its own important political effects; and (2) to point to the potential implications for rebel rule and rebel-incumbent regimes. In this sense, we see institutionalisation and predictability as distinct concepts, such that a political ‘order’ can constitute strong institutions where a single group may hold (fully or relatively) undisputed authority, but which may nonetheless produce politically unpredictable outcomes from the vantage point of its subjects.

The remainder of the article proceeds as follows: first, we discuss prevailing views of unpredictability in studies of rebel orders. We find that, generally speaking, scholars either emphasise predictability over unpredictability, or attribute unpredictability to an embryonic quality of rebel orders that is assumed to gradually resolve if and as rebel institutions consolidate. Next, we set out our definition of unpredictability as a question of institutional design rather than institutional weakness; we then draw on historical institutional scholars and the concept of layering to discuss how studies of rebel orders might productively bring unpredictability ‘back in’. The article then turns to the NRM in Uganda, showing how it developed a governance arrangement characterised by political unpredictability that has allowed this regime to establish and maintain sovereignty and social control from a comparatively weak position. Fourth, we briefly discuss suggestive parallels with the Afghan Taliban. We conclude by observing how early accommodations of different and potentially incompatible rules, vague mandates, and provisional authority can have long-term legacies, marking rebel-incumbent regimes with foundational sources of political unpredictability.

Unpredictability in Rebel Rule: The State of the Field

Scholarship on rebel governance starts from the insight that life under rebel rule can be surprisingly orderly (Arjona 2016, p. 2, Revkin 2021, p. 47). For example, Matthew Bamber-Zryd documents how the Islamic State developed its governance strategy over iterative periods of rule, each time governing greater amounts of territory, with increasingly complex institutions, for longer periods of time – and producing more efficient results than its predecessors (Bamber-Zryd 2022, pp. 1316–17). In this view, deepening rebel control allows for the institutionalisation of the use of violence, such that coercion becomes ever-more impersonal, regularised, and efficient. For some, this understanding of predictability is informed by theories of state formation and consolidation that link increased control to institutionalisation of governance and, concomitantly, to an increase in predictability (see, e.g., North et al. 2009). Indeed, recent literature reviews have also emphasised both the underlying teleological assumptions underpinning much of the literature on rebel governance, as well as the need to further explore the effects of complex and overlapping rule systems (Loyle et al. 2023; Teiner 2022).

Broadly speaking, scholars theorise a common process: First, rebels establish basic security and policing services, with the dual benefit of securing their military bases and tamping down on dissent (Wickham-Crowley 1987, pp. 482–83, Mampilly 2011, pp. 17, 63). This is typically followed by justice services, whether through the presence of brick-and-mortar courts like the Liberation Tigers of Tamil Eelam in Sri Lanka, or roaming judges on motorcycles like the Taliban in Afghanistan (Stokke 2006, Terpstra 2020). Rebels may then begin to provide a broader range of public services such as hospitals, schools, and sanitation facilities, and intervene to regulate social relations and cultural practices (Mampilly 2011, Stewart 2018).

Citizens, of course, also interact with rebel governance initiatives, at times resisting or shaping them in their own right. For instance, in Afghanistan, Jori Breslawski documents how citizens maintained Shuras to resolve disputes before rebels became involved in local affairs (Breslawski 2021). Ashley Jackson also demonstrates how Afghan civilians strategically exercised their agency to shape the Taliban’s governance initiatives (Jackson 2021). In Somalia, Michael Skjelderup traces how traditional authorities represented community concerns and were even at times able to influence Islamic rulers’ decisions (Skjelderup 2021). Others have highlighted how pre-existing networks – whether clientelist or social – shape whether and how rebels can penetrate the communities they seek to control (Rubin 2020, van Baalen 2021).

When effective, rebel governance can undermine the state’s ideal-typical monopolies (e.g., on the legitimate use of force, taxation, law enforcement), challenging state authority. For instance, rebels may gain legitimacy for courts that deliver justice – more so if the courts’ jurisdiction extends beyond civilian disputes to those that arise between insurgents and civilians (Loyle 2021). Extensive and inclusive rebel institutions can help rebels accrue both domestic and international legitimacy, while also demonstrating (and reinforcing) unique territorial control, making rebels the de facto and uncontested rulers of an area.

Literature also points to a correlation between monopolistic control of territory and greater investment in governance (Metelits 2009, Staniland 2012, Arjona 2016). Territorial control is associated with longer time horizons, and an expectation of increasing cooperation with civilians that helps justify investing in service delivery (Arjona 2016, pp. 48–50). Variation in rebel rule, from ‘exert[ing] control from a distance’ to ‘holding territory to create proto-states’ has been described ‘as a broad spectrum along which groups will move, as circumstances – opportunities and constraints – allow’ (Worrall 2017, p. 716).

A stylised depiction of this process (Figure 1) depicts two extremes: at the far left, rebels have limited territorial control, are inconsistently present, weakly institutionalised, offer few services, and make executive decisions. At the far right, rebels have established territorial control, and their institutions can be further developed, predictable, and bureaucratic, and they can effectively provide diverse services. Unpredictability is assumed to be highest at the left-hand side of the figure where rebel institutions are the weakest, and lowest at the right-hand side where rebel institutions are strong and consolidated.

Figure 1. A Stylised Depiction of Rebel Governance as a Progressive Process.⁶

To be clear, scholarship on rebel governance offers important nuances and caveats not captured in Figure 1. For instance, rebel governance can vary substantially over space and time, resembling a ‘dynamic patchwork’ rather than a synthetic governance system as found in ‘more homogenous, unitary nation-states’ (Mampilly and Stewart 2020, pp. 8, 23). Furthermore, empirical studies point to cases where rebels control significant territory but do not govern (Kasfir 2015, p. 26), or where they exercise social but not territorial control (Jentzsch and Steele 2023). It is also widely recognised that the presence of rebel services and institutions does not necessarily correspond to democratic or liberal outcomes (Revkin and Ahram 2020), and that institutions born from conflict can significantly shape the governance trajectories of post-conflict states (Loyle etal. 2023) – however, there remains an underlying notion that democratic outcomes are difficult to sustain without institutionalisation.

Scholarship on rebel governance offers several explanations for unpredictability in rebel orders, which link back to weak or superficial institutions: First, rebels may only sporadically be present, thereby introducing unpredictability through violence and impunity but never establishing a social contract in a given community (Caris and Reynolds 2014). Second, rebels may have short time horizons, and therefore lack incentive to establish a new order (Arjona 2016, pp. 50–55; see also Magaloni et al. 2020 on similar dynamics in criminal governance). Finally, unpredictability is part of a transition from the previously existing political-social order to a new, rebel-dominated one (Worrall 2017, pp. 716–17). Each explanation points to an embedded assumption that unpredictability will decrease as rebels consolidate control and institutionalise a monopoly on (or majority control of) violence. Our analysis, however, demonstrates that by implying a gravitational pull towards order and predictability, the analytic assumptions underpinning this diagram (especially those linking predictability to institutionalisation) can be unhelpful and limiting. In contrast, we argue that political unpredictability reflects a broader institutional logic of some rebel orders, embedded in the very foundations of their governance arrangements – whether by choice or necessity.

Bringing Unpredictability ‘Back In’

Scholars have variously used the concept of unpredictability in studies of politics and governance, typically attributing unpredictability to weak institutions, and the resultant exercise of personal and unchecked power. For example, Andreas Schedler studies uncertainty in authoritarian regimes, and sees institutions as the core technology to manage social (and future) uncertainty. Weak institutions beget high uncertainty; strong institutions, low uncertainty (Schedler 2013, pp. 23–24). Schedler’s work shows that authoritarian regimes are characterised by uncertainty deriving from regime threats, and their inability to calculate these threats, because the role of institutions is ambiguous and their strength is endogenous to the (contingent, contested, changing) political context (pp. 8–12). Others conceptualise institutions primarily as coordination devices rather than as restraints on personalised power; in this case, unpredictability derives from the sudden failure or disruption of institutions. For example, Anastasia Shesterinina studies uncertainty faced by ordinary citizens during the onset of civil war, when everyday patterns are disrupted, and before people can establish expectations about the use of violence (as shown in, e.g., Arjona 2016, when citizens develop expectations over time even in the midst of ongoing conflict). Shesterinina (2021, p. 2) is concerned with how people navigate this uncertainty, and finds that they turn to spaces that retain legibility – family networks and shared histories – to develop shared expectations.

Other scholars show that unpredictability can be created not through personalisation, but instead through institutional design. It is in this latter tradition that we understand unpredictability. For example, Alisha Holland (2016) examines forbearance as ‘the intentional and revocable non-enforcement of law’ (2016, p. 232).⁷ By establishing the possibility that state law could be applied post hoc, forbearance can inject a degree of unpredictability into everyday life, causing people to adapt their behaviour to mitigate that risk (for example, supporting a particular political candidate). In a different vein, Marie-Eve Desrosiers (2020) studies the ambiguities of authoritarianism in pre-genocide Rwanda. In addition to highlighting how personalisation introduces (unpredictable) variation in implementing authoritarian rule at a local level, she notes that local Rwandan officials operated simultaneously as representatives of the state and the people (also see Tapscott 2016 on a similar phenomenon found in Uganda’s community policing initiatives).

For us, unpredictability is thus primarily neither an issue of personalisation nor of poor coordination in the face of weak or disrupted institutions. Instead, in contexts of significant institutional multiplicity, where mandates are vague and authority is provisional, unpredictability about which rules will apply, to and by whom, and with what consequences, can be further heightened such that unpredictability becomes a condition of political life. Political unpredictability can then be the sustained result of a particular institutional set up that stands to be reinforced rather than undermined as institutions gain capacity and functionality.

Starting from this point, we turn to historical institutionalist scholars including Mahoney and Thelen (2010). Like many other institutional arrangements, rebel orders may change only incrementally, and in this sense, are often strongly bound by choices made early on in the insurgency, as well as by the pre-existing institutional arrangements they encounter when extending territorial control. Initial decisions to compromise authority, layer rebel command on top of existing power structures, or co-opt services provided by other authorities can have long-term implications for post-conflict rebel rule. This means that as rebel orders persist over time and even gain sovereign control, there is no reason to assume they will become increasingly coherent and predictable. Instead, increased resources and capacity may only reinforce a governing environment characterised by a plurality of actors with overlapping, vague and provisional mandates. The resulting governing orders may look fragile due to fragmentation and persistent political unpredictability but can actually be surprisingly resilient.

How can we study unpredictability as something other than a failure to achieve an institutional goal? We propose emphasising what rebels do, over what they say they want to do – and in service of this, to focus on how different institutions or components of a governing structure interact to produce real-world outcomes. The difference between focusing on rebel discourses and stated goals versus rebel actions and institutions becomes readily apparent when studying Uganda’s NRM – a movement that was originally shaped by pan-African ideology and overtly stated its commitment to democratic values. But in practice and from the outset, it established a governing arrangement that kept real power highly centralised while outsourcing responsibility for everyday governance to its citizens. If we focus on the NRM’s rhetoric and self-proclaimed ideology, the eventual outcome of an authoritarian and repressive regime is surprising; if we study the structures they implemented from the get-go, today’s governance outcomes appear much more of a continuity.

We understand rebel rule to encompass any institutions and actors that contribute to rebel governance, and in our cases, eventually incumbent governance. Some of these may not be directly under rebel control, but work with or alongside rebels. Consequently, we look beyond rebel-made institutions and actors. As we reveal empirically, when rebels encounter pre-existing institutions, they sometimes layer their control on top, seeking to co-opt and manage pre-existing authorities more than to generate fundamentally new distributions of power.

Our proposal corresponds in many ways to the concept of institutional layering articulated by Mahoney and Thelen (2010, 20): ‘With layering, institutional change grows out of the attachment of new institutions or rules onto or alongside existing ones’. While rebels may aim to reform or displace existing institutions in other ways, they are often faced with local-level resistance to entirely new institutional arrangements, or significant costs and logistical hurdles of ensuring that new institutional structures do not merely reproduce existing power structures. For rebels with their eye on the statehouse, there may be limited incentive to invest in uprooting and replacing local-level institutions if they can instead be dominated or bent to rebel needs. Additionally, the existing complexity and sheer number of institutions can create opportunities for distributing numerous power positions as war spoils, potentially leading to the preservation of incoherent or irrelevant institutions in order to offer ministries and mandates to war victors.⁸ In this sense, the rebel orders we examine opportunistically combine different pre-existing governing structures with rebel-generated structures to gain and maintain control.

Different governance structures may relate to one another in a variety of changing ways: through avoidance, collaboration, contestation, or simply co-existence. As noted by Mahoney and Thelen, ambiguity can be a permanent feature of institutional arrangements,

even where rules are formalised. Actors with divergent interests will contest the openings this ambiguity provides because matters of interpretation and implementation can have profound consequences for resource allocations and substantive outcomes … competing interpretations of one and the same rule can mobilise quite different coalitions. (Mahoney and Thelen 2010, 12)

This ambiguity, which may be particularly pronounced when different rules are layered on top of or alongside one another, can offer points not just of fragility but also of flexibility and adaptability. By allowing rebels to continually renegotiate their scope of authority both during and after war, vague mandates and structures of provisional authority can allow rebels to cultivate social control from a position of comparative weakness. This carries important implications for understanding both rebel orders and the regimes that may emerge from them.

Our argument dovetails with that of Kasfir et al. (2017) on ‘multi-layered governance’. As they note, ‘The creation of many actors creates the possibility that multiple streams of influence require examination in order to explain governance’ (p. 274). We similarly concur that it is necessary to ‘avoid any presumption that they necessarily entail a system in which multiple actors share a goal of cooperation or are expected to work together’ (p. 263). We build on this intervention by investigating the nature of the relationship among different layers and showing how provisionality and vague mandates can create governing arrangements that (1) are unpredictable for those participating in and living under them; and (2) find resilience rather than fragility in that ambiguity.

The following section examines the case of Uganda’s National Resistance Movement, which fought a guerrilla-style insurgency beginning in 1981, successfully taking control of the State House in January 1986. As we will show, early in its insurgency, the NRM delegated provisional authority along with vague and wide-ranging mandates to village level councils called ‘Resistance Councils’ (RCs); these dynamics have continued to structure governance arrangements today, allowing the regime to maintain control while tasking civilians with many of the tasks of everyday governance. The next section first sets out the history of the NRM’s RCs in relation to our argument. The three sub-sections trace: (1) the structurally ambiguous design of RCs and their provisional authority, such that they were both agents of the people and the state; and how these vague mandates played out in (2) security and (3) justice sectors. Each sub-section covers war to post-war years to show institutional layering and continuities over time.

Sources and Legacies of Unpredictability in Uganda’s National Resistance Movement

Uganda’s NRM is widely known for initiating a system of civilian governance in rebel-held territories during its Bush War (1981–1986), which it then extended to the entire country after taking power in 1986. In 1995, RCs were enshrined in the constitution as the backbone of local government and renamed ‘Local Councils’. They continue to form the basis of a five-tiered governance structure, with elected representatives at the village, parish, sub-county, municipality and district levels (see Figure 2: Resistance Council Structure).⁹

Figure 2. Resistance Council Structure (Tidemand 1994, p. 6).

Using existing analytic frameworks, RC Councils might reasonably be described as an innovative form of rebel governance, which would become increasingly predictable and efficient as the NRM consolidated control. They were ‘inclusive’ – all Ugandans over 18 were given a vote for village level representatives; ‘transformative’ – positions included representatives for often-marginalised groups like women and the disabled; and ‘effective’ – meaning ‘political institutions fulfil[ed] the role they [were] assigned’ (Mampilly and Stewart 2020, p. 38). Indeed, Weinstein described the RC structure as ‘a revolutionary reconfiguration of local power’ that ‘empowered civilians to make decisions about their own communities’, including the unprecedented move to allow constituents to recall RC members if their performance was unsatisfactory (2006, p. 177). Nonetheless, the mandate and authority of RCs were ambiguous across law, practice, and rhetoric.

The legacies of these early dynamics continue to characterise the government today, such that many have described the regime as unpredictable – a place where ‘confusion’ and ‘uncertainty’ are central to everyday life and especially encounters with the state (Khisa 2013, Goodfellow 2014, Tapscott 2021). Some have described a resulting ‘subjunctive’ subjectivity, in which citizens’ political encounters with the state are ‘conditioned by pragmatic considerations made in light of simultaneously doubtful and hopeful expectations of the future’ (Alava 2017, 198, Whyte 2002). As a result, citizens seek to mitigate or manage this uncertainty themselves. Many respond by self-policing, curating their opinions and actions in a way that they believe will reduce their risk of facing potentially costly repercussions (Tapscott 2021).

We attribute political unpredictability in part to the NRM’s wartime approach to governance, when the NRA rebels needed to mobilise the population and incorporate them into the war effort, while seeking to build and maintain political control. This is evident in the role of RCs. For example, the NRM devolved responsibility for everyday governance extensively to RCs, initiating local elections for the first time in Ugandan history and delegating wide-ranging powers of legislation, implementation and adjudication. The NRM also consulted civilians in their recruitment of ‘fighters and other persons in the movement because if unpopular elements were recruited they would have damaging effects on the image of the movement’ (Munyambibi-Tumusiime 1992, p. 43). In this ‘people’s war’, the RCs liaised between rebels and the community, providing a conduit for intelligence and supplies.

After the NRM took power, RC elections continued to mobilise the population. However, the NRM ensured that RCs would remain local experiments in democracy that would not challenge national-level politics. In early elections, party competition was banned, as was public campaigning (Burkey 1991, p. 16). While direct elections took place at the village level, the tiered electoral college structure meant that fewer and fewer people voted at each successive level, limiting the ability of the population to influence national-level politics (Burkey 1991, p. 16). Dan Ottemoeller also notes the enduring link between the RCs and the NRM as a rebel actor: ‘The initial justification and definition for the RCs since 1986 is rooted in the war-time experience of the system. The RC’s war-related legitimacy is also closely paralleled by the NRM’s own claims to legitimacy and accountability which are also rooted in a warrior legacy of courage and sacrifice’ (Ottemoeller 1996, p. 45). Summing up these tensions, Ingvild Burkey reflects in her 1991 thesis on RC committees post-1986:

The NRM needed to gain legitimacy by allowing people a measure of authority over their own local affairs, but had no intention of losing control over the political process … the NRM not surprisingly kept a tight rein on the process of democratisation. The powers it conceded to the people were hedged with undemocratic safeguards: what it gave the RCs with one hand, it took away with the other. (Burkey 1991, p. 5)

Resistance Councils were thus framed as simultaneously of the people, the state, and the party, creating ambiguity about what role they were fulfilling at any given time. The legacies of this are evident today, such that ordinary people can be categorised as citizens with access to justice through (fairly) independent courts, or equally, they can be branded as pro-state militia or anti-state traitors, subject to martial law (Burnett 2011).

In the following sections, we show how RCs were layered on top of pre-existing public authorities, and were given extensive, vague, and provisional mandates. The resulting governing structure has continued to shape state-society relations, creating a mode of governance in which authority is often fleeting, provisional, and unpredictable. In this context, pervasive political unpredictability limits political organisation and mobilisation, and obstructs lines of accountability associated with a liberal-democratic state-society contract.

The Ambiguous Role of RCs: Agents of the NRM or of the People?

The first RCs, established during the Bush War, were given wide-ranging and general responsibilities in the areas of local governance and organising civilians in relation to the war effort. In the words of one former member of the NRA resistance:

as we gradually captured more areas, we used to call them semi-liberated areas, when we chased away government. That meant that there was no administration at all … you know, people cannot stay without administration. So that is how the core idea came up. But also, they were very critical in gathering intelligence, in looking for food for fighters, shelter, cover – like if we had to relocate from one end to another, they were the ones who could find safe forests where we could hide. (Zoom interview, 8 February 2023)

From the outset, RCs therefore were envisioned as acting both as civilian administrators and as members of the NRA resistance. Even after the NRM took the State House, and RCs became part of the incumbent state political structure, this ambiguity between civilian and NRM ‘cadre’ has remained blurry, creating and sustaining uncertainty around their source of authority and lines of accountability.

The earliest elections of RCs were held during the war in ‘liberated zones’ in 1982 and offered unprecedented opportunity for local participation. In addition to the intrinsic good of local democratic participation, there were also instrumental reasons to have civilians select their own RC leadership:

The elections permitted [the NRA] to avoid responsibility for unpopular decisions of village committees. ‘Villagers complained about how they were governed’. If chairs took advantage of their positions, villagers expected the NRA, which had chosen them, to do something about it. (Kasfir 2005, p. 286, internal citations omitted)

In this sense, even while the NRA maintained authority and veto power, they adopted an institutional structure that devolved responsibility for many daily tasks to the village level. During the war, the structure of RCs therefore demonstrated the rebels’ rhetorical commitment to local-level democracy, while placing responsibility for everyday governance tasks with civilians themselves.

In the years after the NRM took power, the new government quickly established RCs across the country, with an extensive and vague mandate which included: assisting the police in maintaining law and order; maintaining security, encourage self-help projects; recommending residents to the armed forces; communicating between the government and the people; overseeing government policy in the area; electing ad hoc committees as necessary; electing members of the Tax Assessment Committee at the sub-county level; and monitoring the administration of the area and report any misconduct to the appropriate authority (Ddungu 1989, p. 12). However, as is apparent in many early narratives about RCs, they often lacked authority to implement these extensive agendas. For example, Frederick Golooba-Mutebi documents that RCs warned chiefs against ‘mistreating’ people; but democratically-elected RCs lacked the power to motivate community members to engage in public works or to overcome collective action problems (Golooba-Mutebi 2004, 296). When asked why he did not use his authority to enforce local by-laws, one sub-county chairman responded:

It is very difficult to do that sort of thing. People may know that what they are doing is wrong, but if you punish them or report them to higher authorities, they will hate you. Some may even start plotting to harm you. I don’t want to create enemies for myself. If I ask them to do something and they refuse, I just leave them. (Sub-county Chairman, cited in Golooba-Mutebi 2004, 296)

RCs thus occupied a politically ambiguous space between the people who selected them and the NRA/M, which had created the institutional justification for their existence. In this sense, they were structurally set up to have responsibility both to the people and the state, without the concomitant power to enforce decisions on behalf of either.

RCs also functioned in a legally ambiguous space, where their authority was kept provisional. Their wide-ranging and ambiguous powers led ‘to quite a number of conflicts between the RCs and state organs such as Magistrates, Chiefs and Police’ (Tidemand 1994, p. 86). In 1987, the new NRM government passed the Resistance Councils and Committees Bill, defining RCs as ‘popular elected local government institutions’. This was followed by the Judicial Amendment Bill in 1988, formalising RCs function as courts. However, this legislation left significant ambiguity about the nature of the system. As elaborated by Ottemoeller:

If focusing on the RCs policy making and legislative powers the RCs could be defined as organs of the people. But if one focuses on the roles and duties of the [District Administrator], it appears that the RCs might be defined as a branch of central government. The NRM appears to have wanted to have it both ways, as long as the RCs did not unduly challenge state policies they were granted wide leeway in political and policy terms. (Ottemoeller 1996, p. 70)

According to Ugandan political scientist Mahmood Mamdani, there remained ‘no clear agreement on what this [the role of RCs in society] should be … are the RCs to be organs of the state, of the NRM or of the people?’ (Mamdani 1988, p. 1176). Mamdani further elaborated that RCs were characterised by competing narratives that made them difficult to pin down:

The bureaucratic point of view sees the RCs as no more than appendages of the civil service, created to implement government policy more effectively – in other words, as organs of the state. The democratic point of view, on the other hand, sees the RCs more as popular organs created to counter and hold in check abuses of the civil service and all other state functionaries; as organs of the people, whereby RCs could legitimately be the site of a healthy debate between points of view that cover the whole range of ideological positions within Ugandan society. The third point of view, which may be termed sectarian, sees RCs as organs of one single political group, the NRM. (Mamdani 1988, p. 1176)

The poorly-defined relationship between RCs, the state, and the NRM/A – not to mention forms of non-state authority – meant that even as the NRM implemented local and democratically-elected forms of governance within the areas it controlled, lines of accountability remained unclear, in turn producing unpredictability about who would enforce which rules on whom.

The ambiguous role of RCs was also reflected in the NRM’s rhetoric about and symbolic treatment of the lower RCs. On one hand, the NRM issued general statements like ‘you, the RCs are the government’, ‘RC have the power’ and informally left certain legislative powers (and responsibilities) to them. On the other hand, the NRM and Museveni often referred to the ‘backwardness’ of the people, legitimating direct interference by the NRM Secretariat or the President in RC’s affairs (Tidemand 1994, p. 164). In an interview, a former NRM political commissar and self-described ‘freedom fighter’ in the NRA Bush War, recounted how, in the first five years of NRM rule, the village-level RC chairman was given significant symbolic deference:

If the president went anywhere, they would first call the chairman RC1 to allow the president to speak, and he would give a speech and say, ‘Ok, Mr President, here we are and the security is good, and I allow you to speak in my area’. And people would laugh and say you’re funny, but it was motivation – an RC1 chairperson allowing the president to speak in his area!

To further elucidate the extent and nature of the NRM’s ambiguity in delegation to RCs, as well as how it has produced a political system characterised by political unpredictability, we discuss their functions as security and justice providers.

Protecting Friends and Fighting Foes: Security Provision During and After the War

RCs were in part established to provide security to civilian populations during the war, and they were given autonomy to do so. Each council had a Defence Secretary, tasked with policing and security. Weinstein notes that the NRA protected civilians from the UNLA, and helped set up civilian watch schemes to warn villagers of the presence of UNLA soldiers (Weinstein 2006, p. 180). These village militias received some limited military training, making them more ‘amenable’ to war (Rukooko 2005, p. 215). Village militias also mobilised political opposition against Obote’s forces and politics, and supported the NRA in the war effort, for instance collecting intelligence, helping to transfer civilians to safety (including forcing local people to relocate if necessary), and acting as a rear force protecting civilians going to collect food (Rukooko 2005, p. 215). Expedit Ddungu elaborates that RCs

scrutinised and recruited people, boys and girls to join the NRA. They staged road blocks day and night together with local militias. They advised people on when and where to hide, when to reappear, where to put roadblocks, etc. They issued pass documents in case one wanted to move a reasonable distance within the war zone. They informed on enemy positions and liaised with the NRA. (Ddungu 1989, p. 15)

Beyond this, there is limited information about if and how RCs provided security to civilian populations during the war (Baker 2007, p. 370). Bruce Baker notes that ‘“popular justice” was the only method of justice and policing known during the Bush War period. Law and order at the local level was in the hands of the community alone’ (Baker 2007, 370).

After taking power, the NRM maintained the practice of recruiting village militias, now called local defence units or LDUs, as a type of pro-government militia that served as auxiliary forces to the NRA as well as to RCs. In contexts of relative peace, LDUs were mobilised to support development programmes, enforce the law, and gather intelligence for security organisations. In contexts of conflict, including the rebel insurgencies in the north and east of the country, LDUs captured deserters, recruited for the NRA, and prevented other rebel groups from recruiting fighters (Rukooko 2005, p. 217).

LDUs were often poorly paid or unpaid, and their mandate was loosely defined. Because the NRM defined security as ‘everyone’s responsibility’ (Rukooko 2005, p. 216), it remained unclear the extent to which LDUs were acting as agents of the regime versus of their own accord. The NRM has continued to informally authorise militias and vigilantes to provide security over the decades of its rule, while at times publicly denying responsibility for them and their activities (Omach 2010, pp. 431, 446, Tapscott 2016). These dynamics have kept the role of LDUs and other militias ambiguous and uncertain, further blurring the line between state and society. This has created pervasive unpredictability for ordinary citizens about who has the authority to use violence and under what conditions, causing citizens to self-police – for example, avoiding public rallies or demonstrations and exercising caution about criticising the government (Tapscott 2021).

In addition to providing security and intelligence, RCs engaged in policing, including for social behaviours. For instance, RCs often established community by-laws, some prohibiting card playing, restricting drinking hours, and requiring the cultivation of specific staple foods like cassava (Tidemand 1994, p. 98). Women’s representation on RCs gave voice to issues like domestic violence and divorce. Some RCs regulated the behaviour of unmarried women, for instance, requiring that they announce their romantic partners to the committee (Tidemand 1994, p. 101). RCs outlawed witchcraft: they evaluated evidence and expelled those determined to be witches. Policing in this context became a parallel system to the state. While many of the by-laws that RCs enforced did not contradict state law, others did, for example, those mandating corporal punishment.

Writing about local policing and security, Baker notes that the RC system replaced the traditional authority of chiefs and had the legacy of extending the state’s control of policing to the lowest administrative level by linking RC procedures to formal court processes. ‘The state also maintained its local policing presence through the militias it has formed to counter terrorists, rebels and cattle thieves. The state police force may still be small, but central control over local policing has been strengthened in Uganda as a result of the war’ (Baker 2007, 384). These competing formal and informal legal orders continue to create unpredictability in Uganda today about which laws will be enforced, and with what consequences (Goodfellow 2014, Tapscott 2017).

Sharing Accountability and Safeguarding Authority: Rebel Approaches to Justice and Dispute Resolution

From the outset, the RC1 Chairman was also granted judicial powers, for instance, over land disputes, theft, household disputes and divorces. RCs typically handled cases including thefts, moving without a letter of introduction from an RC, or adultery. More serious issues were brought to the military. While RCs generally focused on reconciliation, punishments could also include caning or expulsion from the community (Burkey 1991, pp. 46–52). RCs were further given some jurisdictional authority over NRA soldiers. For instance, when a soldier came to an area, he was obliged to report his presence to the RC Chairman, as well as whether he had a gun and how long he planned to stay (Tidemand 1994, p. 139). RCs could question orders from NRA soldiers, report misbehaviour of NRA soldiers to higher authorities, or even disarm soldiers (Ottemoeller 1996, p. 303, Tidemand 1994, p. 141). These punishments could be substantial. Tidemand writes:

Punishment of the NRA soldiers most often included a stay in a ‘ditch’ for a couple of days or weeks, kandoya (being tied with the arms behind the back in a way that may cause severe damage), or in rare instances death. (Tidemand 1994, p. 81)

The RCs’ capacity to discipline soldiers contributed to societal acceptance of RCs, and in this sense represents a form of civilian oversight of the military.

After the war, tensions related to accountability and authority manifested in diverse ways. For instance, though the NRM granted RCs the right and responsibility to monitor the state’s bureaucracy, ‘it requires in the same clause that in case of malpractices RCs report to higher organs of the same bureaucracy’ (Ddungu 1989, cited in Tidemand 1994, p. 32). RCs did not replace state institutions, such as the administrative parish chief, but instead worked in parallel to them. For instance, as articulated in speeches by Museveni and in the NRM’s Ten Point programme, RCs were tasked with acting as watchdogs against corruption in the civil service, calling out state agents who abused their powers. However,

This role and the generally vaguely defined roles of civil servants in relation to RCs led to conflicts at various levels. At the district level between the RCs, the District Administrator (DA) and various civil servants, at sub-county level between RCs, chiefs, magistrates and police, and at village level mainly between the chiefs and RCs. (Tidemand 1994, p. 161)

Further exacerbating these tensions, Ugandan civil servants and centrally appointed political actors were reluctant to accept the authority of elected politicians. NRM leadership at times exercised its powers directly though the District Authority. This contradictory attitude towards governance simultaneously gave power to the people through democratic institutions, and also clawed it back in arbitrary moments when the regime determined that this elected leadership would not or could not hold district executives accountable.

The structural legacies of a system designed around these competing goals continue to shape governance in Uganda today. For instance, LC5s (elected representatives at the district level) and RDCs (presidentially appointed administrators at a district level) may adjudicate the same issue and come to conflicting decisions, causing low-level insecurity to prevail amongst local communities (Tapscott 2021). At a local level, communities under the LC continue to establish and update by-laws that regulate activities ranging from the sale of alcohol and opening hours of discos, to prohibiting theft, witchcraft and prostitution. When these local laws come into conflict with the state, they are negotiated through a process in which state authorities continually redraw their claims to jurisdictional authority, often leaving LCs on the back foot (Tapscott 2017).

Assessing Plausibility: Vague Mandates and Provisional Authority Under the Afghan Taliban

To probe the plausibility of our argument, we consider the Afghan Taliban – a rebel group that like the NRA/M has successfully taken state power (twice) and that engaged in civilian governance both as insurgents and incumbents. While scholars note that the Taliban seemed overwhelmed by governance demands in the initial phase of the Taliban’s first uprising in Kandahar in 1994, it subsequently went through various iterations – taking state power, losing it again, and reorganising as an insurgent movement that, over time, became increasingly coherent (Terpstra 2020). During its post-2001 insurgency, the Taliban published a code of conduct, established courts, and taxed citizens and industry (Giustozzi 2019).¹⁰ In the 2010s, a shadow government began to form, with commissions replicating the forms of government ministries in sectors including finance, health, and education (Jackson 2018, Jackson and Weigand 2020). The Taliban maintained a considerable communications apparatus, including a spokesperson and an official website detailing Taliban policy. However, even while developing governance structures and policies, the Taliban has maintained institutional multiplicity, vague mandates, and provisional authority in a way that has allowed them the flexibility to adapt to different audiences and meet short-term goals without making significant governance concessions.

Institutional multiplicity is woven into the fabric of governance in Afghanistan. Multiple parallel systems have always operated simultaneously (at the very least a monarchy or central government and tribal or customary rule). In more recent years, massive statebuilding projects led by the Soviets in the ’80s and the Americans in the ’00s and ’10s have multiplied governance structures alongside those developed by insurgents. For much of the past decade, the government, Taliban, and customary authorities have frequently delivered services side-by-side, especially in crucial areas like justice and security provision (Baczko 2013, Giustozzi 2014, Jackson and Weigand 2020).

The Taliban has consistently kept its mandates vague, in particular, by governing with a set of rules for fighters (layeha) rather than implementing a constitution. Although the Taliban formed a constitutional committee in 1998, the constitution was not made public until 2005, a year after the American-backed Republican government of Afghanistan ratified its new constitution (Lombardi and March 2022). Instead of implementing this document, the Taliban released its inaugural layeha in 2006, outlining the regulations and guidelines for their soldiers and supporters (Clark 2011). While the constitution provided a framework for governance and state functioning, the layeha primarily focused on the conduct of Taliban members during armed conflict and maintaining discipline within their ranks, rather than administering a country. As the insurgency evolved into a ‘government in waiting’ (Jackson 2018), the Taliban continued to revise and update their layeha, rather than reintroducing the constitution they had produced (Johnson et al. 2018). When the Taliban took power in 2021, several interviewers sought clarification from Taliban leadership regarding which constitution would frame its governance project. Taliban officials responded: ‘the 1964 one’, (referring to the last monarchy of Afghanistan). No written legal code or framing document outlining the division of power or any other governance aspects has been made public (Gul 2021).

Provisionality has also been central to the Taliban’s governance strategy. When they first came to power in 1996, they formed an interim government that was never officially established (Strick van Linschoten 2016). As insurgents and then again when they returned to power in 2021, the Taliban followed a similar pattern, forming an interim cabinet that, even three years later, has yet to be replaced or made official. The Taliban describes their rule as a temporary, using the term ‘caretaker government’ (Watkins 2022). This self-narration helps the Taliban sidestep criticism for poor governance, while mitigating against challenges to their interpretation of Islamic and Afghan values.¹¹

These dynamics have fostered an ambiguous relationship between the Emir in Kandahar (who possesses absolute and undefined power) and his appointed Taliban cabinet in Kabul. Without a constitution, a significant question arises: Who makes policy – Kandahar or Kabul? In the first six months of Taliban rule, it seemed that the cabinet, led by the Prime Minister’s office, would be responsible for shaping governance policies. However, ministers increasingly found their decisions overridden by the emir (Watkins 2022). For example, in March 2022, the Emir overruled the decision to reopen girls’ high school at the last moment (Barr 2022). Many observers of the Taliban struggle to understand the dynamics of power within the Taliban’s Islamic Emirate, leaving questions about how power is exercised and decisions are made (Watkins 2022).

The Taliban has repeatedly framed divisive issues as temporary. Consider bans on girls’ education: this was first justified as a short-term measure due to the ongoing war with the National Resistance Front of Afghanistan (NRF). Later, the Taliban stated that restrictions would remain in place until schools comply with the ‘principles of Islamic law and Afghan culture’ (see Barr 2022). By making temporary claims or promises, the Taliban has helped mute the impact of divisive policies, which could lead to internal divisions or external backlash. These tactics, reminiscent of their insurgent rule, have persisted in shaping the Taliban’s second Emirate, underscoring the critical role of ambiguity and provisionality in their governance approach – both in their insurgent past and their present incumbency.

Conclusion: Bringing Unpredictability ‘Back In’ to Studies of Rebel Governance

Both the NRM and Taliban could be considered relatively successful rebel orders in that each transitioned from rebels to ruling parties. However, the cases are otherwise very different across numerous important factors, suggesting that our argument may provide insights for diverse cases. For instance, Afghanistan has been the recipient of enormous quantities of foreign aid, and the subject of a large-scale international state building intervention. Uganda received comparatively less foreign support during the years in which the NRM was both fighting and its earliest years in the State House. These cases thus present an opportunity to explore rebel institutional forms and their characteristics for groups that sought to govern a sovereign territory and the people in it, but otherwise exhibit significant variation.

Our analysis shows that each of these different rebel orders is importantly characterised by unpredictability arising from different and sometimes incompatible rules, and a context where it is often unclear which rules or institutions will apply. As a result, these rebel orders produced unpredictable outcomes for both public authorities and citizens living under them, allowing these regimes to claim responsibility for successes, shift accountability for failures, and keep civilians in a state of uncertainty as to whether the regime will protect or prosecute them. This helped rebels manage populations that could constitute both friend and foe, and established the adaptability and resilience necessary to govern in the midst of war. But these strategies were not easily reversible in the aftermath of conflict, instead having implications for post-conflict rule.

In Uganda, a system of parallel governing orders, with poorly defined mandates and a complex web of accountability, initiated a system in which it is continually ambiguous where real power lies – and in fact, it is not obscured so much as constantly shifting and relocating in contingent and unpredictable ways. The design of these local governance systems made them both independent from and accountable to the state. In Afghanistan, the relationship between law, custom, and codes of conduct were ambiguous by design. The Taliban maintains competing centres of power, such that governing decisions are always provisional and subject to revision, allowing them to claim progress in Islamic state-building without alienating key factions and elites.

Our study of two rebel orders that succeeded in gaining sovereignty reveals that vague mandates and provisional authority were important aspects of rebel rule in each; and that this in some ways became baked into the foundations of governance with post-conflict legacies. In addition to studying the quantity and quality of rebel services and institutions, we can deepen our understanding of the way these rebels rule both during and after war by bringing unpredictability ‘back in’. This means studying how these rebel orders sustain different and sometimes incompatible rules, such that they can appear in some ways to be both democratic and autocratic, integrated with society and distant from it, pursuing both transformative and traditional agendas, and including civilians even as they exclude them.

Political orders that sustain unpredictability can offer governance benefits, particularly in the context of waging an armed conflict, in which effective institutions must be autonomous enough to survive when rebels must fight or flee, but also sufficiently subservient to the rebel order such that they do not offer a platform for collective civilian opposition to rebels when rebel control returns. They are also helpful to rebels who must be able to treat civilians as both allies in their war against the state, and potential state sympathisers and government informants that could pose a threat to the rebel operation. Such conditions favour ambiguity and overlapping mandates, which in turn allow rebels to reassign responsibility and authority to gain advantage from a relatively weak position.

In addition to thickening our understanding of rebel governance, a focus on unpredictability can help improve policy response. For instance, attributing unpredictability to the embryonic quality of rebel regimes undergirds an assumption that contradictions will resolve on their own as rebel orders consolidate. Development assistance has often been tied to this assumption. Since the NRM took power, Uganda has been an outsized recipient of foreign aid, including substantial statebuilding and development programmes. Underpinning this programming has been a belief that, over time and with appropriate incentives and investments, the regime would become more formalised, more predictable, and – perhaps even – more democratic. However, if the very foundations of these regimes are structured to accommodate different and potentially incompatible rules, such a belief is likely misguided. Instead, further investment may strengthen a structure that is built to contain and even foster ambiguity and unpredictability, making these regimes more resilient. These foundational structures and their legacies may help explain why, despite massive foreign investment, these regimes have not developed into the predictable, formalised states that donors envisioned.

Acknowledgments

We would like to extend our great thanks to the many scholars who have engaged our work, helping to nuance and sharpen our thinking. We had the opportunity to present this work and receive comments at the International Studies Association and the Graduate Institute’s Centre for Conflict, Development, and Peacebuilding. A great thanks go to members of the Non-State Armed Governance Group, especially Rebecca Bell-Martin, José Antonio Gutiérrez Danton, and Zachariah Mampilly. We also would like to thank Matthew Bamber-Zryd, Jasmine Bhatia, Deval Desai, Nelson Kasfir, Romain Malejacq, William Reno, Megan Stewart, and Henning Tamm, each of whom engaged thoughtfully with our ideas as this project evolved. We also would like to thank the convenors of this special issue, the editors of Civil Wars, and our anonymous reviewers. The research was supported by the Gerda Henkel Foundation’s Special Programme for Security, Society, and the State and the LSE’s Centre for Public Authority and International Development.

Disclosure Statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was funded by the Gerda Henkel Foundation [AZ 17/KF/18] and the LSE’s Centre for Public Authority and International Development [ES/P008038/1] and [ES/W00786X/1].

Notes on contributors

Rebecca Tapscott

Rebecca Tapscott is a Lecturer in the Department of Politics at the University of York, UK, and a visiting fellow at the Firoz Lalji Centre on Africa at the LSE.

Eliza Urwin

Eliza Urwin is the Head of Research at the Geneva Graduate Institute’s Centre on Conflict, Development and Peacebuilding.

Notes

1. Some have also criticised comparisons between rebel and state rule. As Mampilly writes, ‘The problem for analyses of rebel governance is that by analogizing the existence of political order outside the state to the process by which states are formed, analysts are often forced to see a state where none exists’ (Mampilly 2011, pp. 35–36).

2. We use the term ‘institution’ in a general sense to refer to constraints that structure social interactions (North 1990). Recognising that rebels’ initial attempts to govern might not meet the standard of ‘institutions’ we also use the term ‘governance structure’.

3. For example, José Antionio Gutiérrez documents an interesting case in Colombia where the FARC-EP guerrillas were able to advance their political agenda, reinforcing their organisational work in rural communities, by repurposing institutions originally designed to contain the insurgent movement (Gutiérrez 2021).

4. Indeed, we take the opportunity to note that both formal and informal institutions can be sources of order-making, whether by restraining personalised power or serving as a coordination device. We do not equate unpredictability with informality, but rather with structures that produce unpredictable outcomes in terms of what rules will be applied, by who and on whom, and with what consequences.

5. Research drawn on in this article received ethical review and approval from George Mason University and the Geneva Graduate Institute as well as Gulu University’s Research Ethics Committee and Uganda’s National Council for Science and Technology (UNCST).

6. Figure 1 is original and compiled by authors.

7. Holland focuses on cases where elites use forbearance to create and strategically distribute resources either as corruption or informal welfare provision (p. 236). In this sense, she focuses on how forbearance essentially moves populations of people from one institutional logic (that of the formal law, police, judges, and bureaucrats) to another institutional logic (that of informal rules of exchange and reciprocity) when they do not fall into line. In contrast, we describe a thicker and more pervasive unpredictability, in which people are uncertain which authority will apply which set of rules in any given situation.

8. There can be other reasons to maintain existing institutions as well – for example, a senior Afghan analyst working for an international organisation, explained that Taliban may have been concerned that dissolving certain governing bodies implied that the Taliban lacked the technocratic expertise to run them (February 2023, virtual interview). Rebels may also prefer to leave in place defunct organisations rather than ruffling feathers to formally dismantle them.

9. Each RC council from village to district level elect an executive RC committee with nine members: the Chairman; Vice-Chairman; Secretary; Secretary for Youth; Secretary for Women; Secretary for Information; Secretary for Mass mobilisation and Education; Secretary for Security; and Secretary for Finance. In 1989, RCs were amended to require that the secretary for women be a woman, and the secretary for youth be a youth (Tidemand 1994, p. 7).

10. For a detailed description of the Taliban’s first emirate see Terpstra (Terpstra This issue).

11. This may reflect some lessons learned over decades of insurgency. For instance, Terpstra has noted how in the late 1990s, the Taliban’s ‘social and religious policies were widely unpopular among the general public’ particularly in cities, resulting in backlash when North Alliance militias killed and captured thousands of Taliban soldiers (Terpstra 2020, p. 1155).