The Mediation of Autocratic Regimes: How Local Officials Shaped Authoritarian Systems in Rwanda and Sudan

Abstract

Local state officials impact authoritarian systems through the mediation they perform. Desrosiers and Mahé argue that these local functionaries fulfill a number of mediating functions, including translating and representing authoritarian systems at the local level. By enacting these two roles, however, local officials do not straightforwardly reproduce the system. Instead, their interpretations and choices fundamentally influence the imprint authoritarianism has on society, from how the regime is experienced at the local level to its groundings and resilience. They demonstrate this argument by looking at pre-genocide Rwanda and Sudan under President Omar al-Bashir.

Résumé

Les fonctionnaires locaux ont un impact sur les systèmes autoritaires par la médiation qu’ils effectuent. Desrosiers et Mahé soutiennent que ces fonctionnaires locaux remplissent plusieurs fonctions de médiation, y compris la traduction et la représentation des systèmes autoritaires au niveau local. Cependant, les fonctionnaires locaux ne reproduisent pas sans ambiguité le système en jouant ces deux rôles. Au contraire, leurs interprétations et leurs choix influencent fondamentalement l’empreinte de l’autoritarisme sur la société, depuis la manière dont le régime est vécu au niveau local jusqu’à ses fondements et sa résilience. Ils démontrent cet argument en examinant le Rwanda avant le génocide et le Soudan sous la présidence d’Omar al-Bashir.

Resumo

Os funcionários públicos locais inflienciam os sistemas autoritários através da mediação que desempenham. Desrosiers e Mahé argumentam que estes funcionários locais cumprem uma série de funções de mediação, nomeadamente traduzindo e representando os sistemas autoritários ao nível local. No desempenho desses dois papéis, porém, os funcionários locais não reproduzem o sistema de forma linear. Pelo contrário, as suas interpretações e escolhas influenciam de forma decisiva a marca que o autoritarismo deixa na sociedade, desde o modo como o regime é vivido ao nível local até aos seus apoios e resiliência. Os autores demonstram este argumento através da análise do Ruanda antes do genocídio e do Sudão sob a Presidência de Omar al-Bashir.

All authoritarian systems—indeed, all political systems—rely on intermediaries to project and solidify their rule. Since the “inner sanctum” of a regime cannot be everywhere all at once, these intermediaries mediate the system across the state and society. Whether located within the formal political system, or as allies of the regime beyond its confines, intermediaries translate and reproduce the authoritarian system beyond its core. ¹ Key among these intermediaries are local state officials. These local functionaries include governors, mayors, councilors, and local administrators to which the central government delegates official powers to act as its governing “face” at lower administrative levels (Carter & Hassan 2021:40). In many systems, these local functionaries are directly appointed by the regime to further strengthen the sense of verticality and the delegated nature of their work. Given their embeddedness in the system, the central government expects these local functionaries to act as direct representatives and implementers of its political order. Intermediaries, and especially local officials, therefore matter in a very fundamental way for the destinies of authoritarian regimes. Yet, they have been the focus of little systematic attention, given the predilection of comparative authoritarian literature to concentrate on national elites (Frantz & Ezrow 2011; Meng 2020; Slater 2010).

Local functionaries and the mediation they perform for authoritarian regimes need to be studied more systematically and rethought. Scholarship on intermediaries has largely looked at their self-serving strategies for reproducing or challenging the system. For the little that has been produced, more specific work on local officials has predominantly focused on how they gather information, perform repression and cooptation, and even serve to deflect criticism for national authorities. It has, in other words, focused on the roles they perform to reproduce the system.

We offer a more holistic understanding of local functionaries’ mediation of authoritarianism. Subnational mediation creates the space for authoritarianism to be co-produced by local officials. This co-production, however, rarely corresponds to the vision and expectations of the central authorities. That is, it rarely conforms to the system as projected at the top. Following recent scholarship on the complex geographies of authoritarian orders (Hassan 2020; Letsa 2019; Tapscott 2021), work on formal and informal institutions (Carter 2021; Ochieng’ Opalo 2019; Shen-Bayh 2022), and on local authoritarian enclaves (Benton 2012; Herrmann 2010), we show that the forms of translation and representation that mediation is meant to perform inherently bifurcate regime reproduction locally.

Regimes benefit, to some extent, from this bifurcation, even if they do not generally condone it. It can give them a flexibility at the local level that contributes to their resilience. ²

Yet, how local officials interpret and implement the system—not always consciously or strategically—impacts the shape of the system locally and how citizens experience it in a manner that can also create forms of disconnection between the core and the local. As the bridge between national authorities and local publics, local functionaries often also act as the agents of local constituencies as much as the agents of the authoritarian system. Their ambiguous allegiance can therefore paradoxically impact regime reach negatively, while simultaneously enacting it. Put differently, because of its very nature, mediation changes the form authoritarianism takes locally, from its shape to the depth of its groundings. Local officials’ enactment of power therefore clearly matters to authoritarian resilience and decay, even if it has been neglected by elite-centric approaches to authoritarianism. By showing how mediation fundamentally factors into the way authoritarianism is produced and experienced beyond regime elites and their powerplay, our approach fills an important gap in comparative authoritarian studies, while building on existing work on local political realities in African political studies.

We make our argument through two cases of authoritarianism in Africa: Rwanda under the First and Second Republics (1962–1973 and 1973–1990) and the al-Inqath (Salvation) regime in Sudan (1989–2019). Following a most different comparative strategy, these constitute two very different contexts, but also different authoritarian settings (soft authoritarianism and lower capacity in Rwanda’s case, and hard authoritarianism and higher capacity in Sudan’s case). ³ This variation helps us show the inherently bifurcated nature of mediation across contexts, including across the variety of authoritarian contexts that exist, notably on the African continent. This is particularly the case of the contrast between soft and hard authoritarianism we propose. Indeed, our cases show that the imperfect reproduction of authoritarian systems is not a question of regimes’ ability to control—though it does influence the degree of shaping it can produce—but is inherent to the roles played by mediating officials, no matter the system. We should therefore expect degrees of bifurcation in softer authoritarian regimes as well as in harder cases, including in some of the softer and harder forms of authoritarianism found in Africa.

In our case studies, we specifically focus on political and administrative officials, as opposed to other types of officials such as judicial personnel or police officers, because they are more straightforward representatives of political power and the system and are therefore the ones bearing the clearest expectations with regard to reproducing the system. In both of our cases, some of these local officials were appointed, and some were elected. Indeed, in Rwanda and in Sudan, turning some key positions into appointed ones was in part meant to gain greater control over them and limit their potential sense of allegiance toward local constituencies. It never fully did, as we explore below.

We first discuss the scholarship on local officials, focusing on how their roles in authoritarian systems have been conceived thus far. In two subsequent sections, we present our theoretical take on authoritarian regimes as a mediated system, as seen from the local level, and our methodology for studying mediation in Rwanda and Sudan. The final sections turn to our two cases to show how local authorities constitute a “messy middle” for authoritarian systems.

Local Officials in Authoritarian Regimes

Autocrats have long been aware of how useful political and administrative subalterns can be in projecting their rule. Colonial powers, for example, largely relied on intermediaries—such as chiefs, soldiers, and translators—to ground their presence (Mamdani 1996). Postcolonial authoritarian polities have continued to use intermediaries, including local officials, to ensure their reach in society (Münch & Veit 2018:270). Yet, comparative authoritarian scholarship has not engaged systematically with the concept of mediation and the theoretical relations and functions it entails. It has tended to focus instead on the often-investigated pillars of authoritarian stability: repression, cooptation, and to some extent legitimation, as performed by the authoritarian core or elites (Gerschewski 2013). We turn in this section to some of the ways research has captured local officials’ roles in recent debates.

Comparative research on authoritarianism has predominantly focused on elites, who are generally taken to be the shapers of authoritarian destinies. This is especially the case in African contexts, where the focus on patrimonialism inherently feeds hierarchical understandings of political realities. Yet increasingly, research has been calling attention to the actors below the level of national elites in order to better understand political systems. This includes scholarship on authoritarian brokers and work on democratization, as well as more limited research on subnational authoritarian officials (Cheeseman 2015; Hassan 2020; Maingraud 2018; Maingraud-Martinaud 2018; Zaki 2008). While formal local authorities (local councilors, mayors, or governors, for example) are not the only ones to broker power locally, these local functionaries are usually expected to do so by virtue of their official position and embeddedness in the system. The existing literature suggests that they allow authoritarian regimes to project and protect themselves in three key manners: information-gathering, delegated repression/cooptation, and deflection.

First, local officials fulfill an important information-gathering function, which once relayed to the national level enables ruling elites to design more efficient policies (Ong 2015). Autocratic systems suffer from an information deficit, since repression breeds dissimulation among the population. To dole out material or symbolic privileges, or even threaten to repress, autocrats need to monitor when and where it matters to do so. Local officials are therefore often expected to be the regime’s eyes and ears on the ground. Second, they also are commonly the ones habilitated by the regime to execute local forms of repression and cooptation (Hassan 2020). Their local embeddedness and physical proximity to the population means that local officials can enact targeted repression, which scholars have argued is more efficient than indiscriminate repression (Kalyvas 2006). Indeed, indiscriminate repression is costly and can lead to increased opposition. Similarly, through targeted local service delivery, local authorities play a major role in building and maintaining the extraction and distribution networks of central elites at the local level (Ng 2018; Poirier 2016). Their selective distribution of goods and services serves to co-opt local elites and communities, a process that has been highlighted in many different authoritarian contexts. Third, scholarship has shown that local representatives can help insulate the regime and its national rulers from criticism by deflecting criticism to lower rungs (Cai 2008; Landry 2008). Because they are the face of the system locally and are its most obvious tool, at least for local populations, local officials can be made the target of blame for regime shortcomings and failures. In the same way, by delegating more unsavory tasks such as repression to the local level, autocratic systems are also able to channel popular resentment toward local officials (Boone 2003; Hess 2013). The fundamental role local officials play in mediating the regime at lower echelons explains why autocrats have readily adopted decentralization policies since the 1980s, even if decentralization has often been understood by international organizations and Western donors as a key tool of democratization.

This focus on the roles of local officials in authoritarian systems has, however, so far largely come with the assumption that these functionaries are committed to reproducing the system, and even actively support it in many instances. In combination with the tendency in studies of the African state to ignore political systems’ deployment beyond the capital, this has led to a rather limited understanding of the roles of local officials in authoritarian systems. The dominant “reproduction assumption” can be found, for example, in scholarship focused on understanding how regimes control local functionaries’ performance of mediation. From a principal-agent standpoint, researchers have focused on the means autocratic regimes use to ensure that local officials effectively play the roles expected of them. This can be done through surveillance. But other management strategies studied by researchers include having a centralizing hegemonic ruling party in charge of officials’ careers (Landry 2008), using deconcentrated agencies to monitor the behavior of local agents (Dickovick & Riedl 2010), and avoiding the transfer of meaningful or substantive power to lower levels (Poteete & Ribot 2011). Such strategic management can be especially salient when demands for political change become pressing at the local level, and when local officials are confronted with the choice of acting on behalf of the regime or allowing local politics to play out even if they are detrimental to the regime (Benton 2012). Thus, while this focus on controlling the mediation performed by local officials suggests they can stray, scholars of this line of thought have overwhelmingly focused on how they are brought into the fold.

Another area of investigation promoting this assumption centers on the decision-making process of local officials. Some scholars have insisted on the instrumental reasons that drive local functionaries’ decision to be effective mediators of the regime. For many, local officials do so because the system is essential to their status and privileges (Hamad 2018; King 2007). In other words, as self-interested agents, they reproduce authoritarianism because it is to their advantage to do so. From this standpoint, reproduction of the system rests as much on the overt efforts of national elites to keep local agents in the fold as on a tacit “clientelist pact” that local officials understand and act upon (Herrmann 2010). Here again, while the literature implicitly suggests that local functionaries can choose to reproduce the system or not, the tendency among scholars has been to focus on what strategically leads local officials to act in support of the system.

There is nonetheless increasing work—though far from systematic—on how local functionaries can undermine the hold of regimes on local constituencies and even challenge regime reach. Given their role in the system and to be able to perform their functions, local officials have some degree of autonomy and influence (Ochoa-Reza 2004). In some instances, this leads local representatives to take advantage of the power and autonomy they have to create their own local authoritarian fiefdom. This is notably what literature on authoritarian enclaves demonstrates (Gelman 2010; Gibson 2013; Herrmann 2010). Instead of looking at what national authorities expect of and do with local authorities, this literature speaks to local authorities’ influence in establishing or promoting their own local authoritarian orders in hybrid or semi-authoritarian political contexts. This scholarship suggests that we need to avoid treating authoritarianism as a strict hierarchical chain of command when it comes to local functionaries. Looking at what local officials do or how they choose to practice authoritarian power points to the existence of multiple expressions of authoritarianism within a single regime (Morelle & Planel 2018). Yet, this scholarship tends to insist on the competitive nature of subaltern authoritarian officials, as creators of alternate political spaces. In so doing, it understates their embeddedness in the broader political system and therefore how they also engage and impact it.

We are therefore missing a complete understanding of mediation in authoritarian systems. Neither the focus on the reproductive roles of local functionaries, nor the work on local manifestations of authoritarianism aptly captures what it means to convey power through functionaries. With few exceptions (e.g., Hassan 2020), existing scholarship tends to promote the assumption that local officials are “all in” or basically “all alternate,” but mediation in the middle is messier than these two binary positions. In contrast to assumptions on the reproduction of the system by local functionaries, and pushing further reflections on the forms of production they perform locally, we insist instead in the next section on the inherently imperfect reproduction of the system that comes with mediation. Demonstrating this bifurcated nature of mediation requires us to move, however, beyond the conventional focus on repression and cooptation and explore some of the more symbolic forms of authoritarian production that local officials engage in.

Conceptualizing Mediation

Scholarship has so far tended to study categories of local functionaries on the basis of their position in the system or specific functions (e.g., higher level regional officials, judges, or community councils). All, however, are expected by autocrats to act as faithful mediators of the system. For those who do choose to enact mediation, it also empowers them to build it as they enact it, that is in a manner that rarely conforms to how the system is envisioned at the top . ⁴ In this section, we propose our theoretical take on mediation.

This paradox of being reproducers but also inherent subverters of the system exists wherever local functionaries stand within the formal authoritarian hierarchy. Right below the national level, local officials such as governors, like the one we study in the Sudanese case, are in close proximity to the regime. They are therefore more obviously watched by its inner core. But, as higher-ups in the administration, they generally enjoy greater power or influence in the system, and hence more sway in terms of its destiny. At the very local level, village representatives or local councilors, such as the ones we look at in the Rwandan case, enjoy less power and hence less influence, but their remoteness from the center of power can, in many instances, make them less monitorable. They are also often more closely embedded in local communities, unless they were parachuted in from elsewhere. Similarly, systems where local functionaries are appointed rather than elected aim to foster a stronger, more vertical chain of command to tie local officials to the core. Across all levels, however, all are empowered by their functions to enact the system. Through the very act of mediation, higher and lower political officials benefit from an influence on the shape of the system locally.

We therefore understand mediation as the process by which local functionaries convey the authoritarian system, its structures, institutions, and norms, from the top down to the citizens. Surprisingly, despite its centrality to the establishment and entrenchment of all political systems, the concept of mediation has remained mostly conceptualized in democratic contexts. Its closest equivalent in scholarship on authoritarianism is brokerage. Yet, mediation has a hierarchical dimension that conceptually distinguishes it from brokerage (Stovel & Shaw 2012). Brokerage is about the creation of complex networks, where brokers are neither dominant nor dominated. Through their brokering, they gain power and enhance their position within systems (Lindquist 2015). By contrast, local officials performing mediation are embedded in pre-existing hierarchical power relations, which entail clear expectations on the part of autocrats regarding their reproduction of the system.

These expectations are rarely met in the strictest sense. While local functionaries enact the system, they continuously reconfigure it locally. Since mediation is the “art of being in-between” the rulers and the ruled (Yannakakis 2008), it is fundamentally ambiguous. ⁵ Because it relies on understanding, interpretation, and context and relations, the act of enacting is never perfect or complete. Mediation can therefore be measured in terms of the degree to which it mirrors the expectations of the regime—in other words as a continuous process—but it never achieves perfect reproduction. Political systems, including at the local level, are therefore always the reflection of those who embody them.

This is scarcely captured in the roles scholarship generally attributes to local functionaries. Information-gathering, delegated repression/cooptation, and deflection are straightforwardly aimed at reproducing the system—though the literature discussed above usually acknowledges that agents can choose not to follow expectations. Understood from the standpoint of these functions, reproduction seems relatively straightforward. This neglects how local forms of adaptation and bifurcation can help make regimes more palatable at the local level and contribute to their resilience. But the literature also neglects two other key mediation roles and how they implicitly contribute to local bifurcation that set regimes on imperfect local foundations. The first is translation, or acting as the transmission belt. The second is representation, or being the ongoing face of the regime, as already captured to some extent in work on the legitimation of authoritarian regimes. Neither of these auxiliary functions can be performed in as straightforward a manner as information-gathering, delegated repression/cooptation, or deflection.

By virtue of functioning in or with the system, local functionaries contribute to legitimating the regime by expressing its values, appropriating its rhetoric, or, in its most manifest forms, by publicly endorsing the rulers (Luft 2020; Maingraud-Martinaud 2018). However, this translation of language, norms, and practices at lower echelons is often also the reflection of the interpretation, choices (conscious or not), and agendas of local agents. Those choices are in part constrained by the various repertoires of legitimation available to local agents, such as those relying on tradition or on service to the community (Poirier 2016). Inherent to translation is therefore a bifurcation of the system, at least from the standpoint of how it is understood from the top. This translation is also much harder to monitor effectively on the part of the top of the system, as a result of its normative nature.

Local authorities are also expected to represent the system within their constituencies. On the one side, local politicians and bureaucrats act as direct agents of the state and its regime, as they enact and implement its ideology, structures, and policies. It makes the system a concrete reality, even if there is distance from the capital. For citizens, this regular contact with local officials is a symbolic reminder that the regime exerts its presence and influence from the top down. But, quite commonly, they also serve to represent their constituency within the system, a role that shapes how they respond to authoritarian demands from the top. Local authorities are recipients of a variety of demands from their constituents, including some that may go against the grain of the system. This is the case even in closed political systems (Rowell 2006). This makes local officials the bearers of both top and bottom expectations, and hence often the agents of the two levels, which further changes the shape and experience of authoritarianism at the local level. In other words, by virtue of the extent to which they also reflect bottom-up demands, local officials produce their own political realities.

Overall, local officials have a necessary and important imprint on the system. While some local authorities can act as straightforward one-way transmission belts, we argue that in most cases mediation entails the bifurcation of the authoritarian order at the local level. This is the case because local functionaries may not be able or willing to gather information, enact repression and cooptation, and serve as deflectors efficiently. This has partly been addressed by the existing literature. There is, however, a more innate form of bifurcation in mediation. Even if local officials are able and willing to serve the regime effectively, given that they need to translate and be representatives of complex normative components (structures, institutions, norms, and the expectations that accompany them), they are also very likely to change how authoritarian realities are expressed and lived locally. Local officials modulate the system, which may at times produce an alternate if not competing authoritarian reality. It may be softer or harsher; it may be more formalized or on the contrary more arbitrary. Paradoxically, though necessary to the reproduction of authoritarianism and even its adaptability to local realities, mediation also plays against the establishment of clear reach from the top, and as a result plays a part in the destinies of the regimes themselves.

Our take on mediation in authoritarian systems therefore covers a problematic theoretical gap left by the existing literature on comparative authoritarianism and African political studies, divided between work on formal institutions and functions, and their reproductive bias, and scholarship on the normative, informal, and personal, stressing arbitrariness. Existing scholarship on African autocracies has produced a “false dichotomy” between formal institutions and personalist power (Hassan 2022). A focus on mediation shows instead how the personal is always a vector of the formal, in ways that fundamentally shape political realities. We turn to Rwanda and Sudan to illustrate how we studied mediation in practice, and then demonstrate the inherent bifurcation it entails.

Studying Mediation in Rwanda and Sudan

To illustrate our argument, we chose two cases: pre-genocide Rwanda and the Sudanese regime under President al-Bashir. Following a most different cases comparative strategy, we opted for two contrasting cases in terms of their domestic and international contexts, as well as in terms of the type of authoritarianism they represented (see Table 1). Rwanda during the First and Second Republics was akin to a “soft authoritarian” regime, relying “more centrally on the means of persuasion than on the means of coercion” (Schatz 2009), despite insistence in Rwandanist literature on the reach of the state prior to the genocide and the compliance it supposedly fostered. The regime and state were low capacity. By contrast, Sudan under the Inqath military and Islamist dictatorship was characterized by the pervasive use of repression (Deshayes & Mahé 2020), as well as stronger state capacity. To further contrast the cases, we focused in Rwanda on very localized forms of mediation, quite removed from Kigali, the capital. In Sudan, we looked at different officials, but predominantly discuss an upper-echelon executive, namely, a governor. In both cases, specific policies were enacted over time to make some of these positions appointed ones, such as bourgmestres in Rwanda starting with the Second Republic, and governors in Sudan, as discussed below.

Table 1. Characteristics of the case studies

* We consider that the Second Republic, in its original form, ended in 1990, as a war began in Rwanda that resulted in a peace agreement, as well as political liberalization that fundamentally altered the shape of the regime.

^** Though the al-Bashir regime began in 1989 and the Cold War ended in 1991, the regime predominantly consolidated in the decades that followed the end of the Cold War.

Given their characteristics, we could expect ambiguous mediation to have been much more prominent in pre-genocide Rwanda than in Sudan, given the softer nature of the system and the symbolic distance between local officials and central authorities. In contrast, we could expect Sudan to have shown that stronger, harder regimes are better at ensuring that mediation performed by local functionaries meets regime expectations. In other words, we also built into our comparative strategy a comparison between a most likely case of bifurcated mediation (soft test: Rwanda) and a least likely case of bifurcated mediation (hard test: Sudan). Our aim, however, in maximizing the differences across these cases was to illustrate how the imperfect reproduction that mediation produces can be found across regime types, though factors such as capacity, the nature of appointments of local officials, and types of electoral systems may shape degrees of bifurcation. In other words, given the inherent nature of bifurcation in mediation, our framework should be transferrable to other authoritarian contexts.

Both these case studies draw on extensive multi-method research, from formal archival research for the Rwandan case (Rwandan National Archives, Belgian and French Diplomatic Archives) to qualitative in-country research for the two settings (over one hundred interviews in Rwanda and Sudan combined, conducted with local officials and some of their beneficiaries to assess local authoritarian forms).

Given the more historical nature of the Rwandan case, primary archival sources included reports from key Rwandan ministries (e.g., Ministry of the Interior and Presidency), as well as regular reports on the country’s political context, actors, and institutions by Rwanda’s two main partners for the period, Belgium and France. Interviews were conducted in 2015. Interviewees in Rwanda included ordinary citizens from eight urban and rural locations outside of Kigali. Participants were required to have lived in Rwanda for all or part of the First and Second Republics. Seven interviewees were women. Participants were not asked to identify along ethnic lines, but ethnic identity, surmised from answers, reflected common but rarely problematized estimates among the general population, with a slight overrepresentation of Tutsi than average numbers and no Twa participants. A number of participants occupied local administrative positions, including responsables (cellule level), a conseiller (secteur level), and a bourgmestre (commune level). Other Rwandan interviewees were ordinary citizens, most defining themselves as subsistence farmers or, in urban centers, as low-skilled workers.

In the Sudanese case, data collection included semi-structured interviews, participant observation, and surveying official documentation during in-country research in 2015 in Khartoum and El Obeid. Interviewees were individuals involved with the country’s development policy, either as participants or as implementers. Participants had to originate from North Kordofan. Socio-economic profiles were varied. While some interviewees, such as cattle herders, came from modest socio-economic backgrounds, most participants were local elites. This included ordinary citizens with varied professional occupations, such as businessmen and teachers, as well as members of the local government and local state institutions in North Kordofan, including a mayor, civil servants, and members of the Chamber of Commerce. All interviewees were men. No information about their ethnicity was provided, nor did it come up during interviews.

Mediation in Practice

In this section, we show how mediation played out in practice in pre-genocide Rwanda and Sudan. While both systems held high expectations in terms of local officials faithfully mediating the system from the top to local constituents—indeed in these two cases national authorities seemed intent on trying to strictly manage mediation—both pre-genocide Rwanda and Sudan show instead how mediation necessarily entails the imperfect reproduction of the system. In both cases, translation and representation ended up bifurcating the system and therefore not affording it the deep and stable reach it was expected that intermediaries should build.

Pre-genocide Rwanda’s Very Local Officials

Rwanda’s current stringently authoritarian government has lent itself to assumptions about control. Built around a former rebel group, it is seen as operating following a military ethos, with a hierarchical and tight grip over the country, and to be obsessed with surveillance, which it has purportedly extended deep into Rwandan society (Purdeková 2011; Purdeková, Reyntjens, & Wilén 2018; Sundberg 2016). These presumptions regarding the extensive reach of the regime and its ability to coerce Rwandans, including local officials, to fall in line are not new to Rwanda. Rwanda was authoritarian before the current regime, and in yesterday’s Rwanda like today’s, researchers saw local officials as subsumed by an overwhelming authoritarian system.

Yet, despite all these assumptions regarding the straightforward enactment of the system on the part of intermediary actors, the reality of mediation in pre-genocide Rwanda was always more complicated, even as successive regimes strove to expand authoritarian control. ⁶ Pre-genocide Rwanda shows how the local authoritarian realities state officials embodied always amounted to an imperfect mediation of the system. Keenly aware of local functionaries’ importance for their rule, especially in terms of bringing authoritarian power to the local level, national authorities strove to control mediation and even expand it. Yet, the ones who performed it regularly strayed from the expectations at the top, not only in terms of how they projected the system, but also in terms of how they navigated relations with their constituents. They often acted in the everyday as their representatives—or their overlords—thus creating very local patterns of authoritarianism distinct from the expectations of Kigali.

The pre-genocide Republics, the First under President Grégoire Kayibanda and the Second Republic under Juvénal Habyarimana, understood early on the need to foster points of contact with Rwandans and to ensure the system was translated and represented locally. To do so, the regime sought to project itself and its power within society under the guise of a supposed decentralization of power. As early as the first years of the First Republic, President Kayibanda regularly spoke of public servants’ role in contacts with what he called the “popular masses.” ⁷ The regime sought a “deep implantation” (implantation profonde) in the words of the Second Republic, for which local officials were essential. ⁸ Local political authorities and bureaucrats, generally drawn from the population at the very local level, were regularly described as those tasked with being the face of the regime beyond Kigali, and with symbolically representing national authorities locally as well as performing state services for the population.

But local officials were to be more than state representatives; they were also understood to serve the regime’s authoritarian system. These local state representatives were to act as “antennes émettrices” or local antennas, expected to transmit regime ideology and expectations to Rwandans, while also relaying information back to the regime. ⁹ This surveillance function was an essential one, in a country and at a time where the means for watching citizens were limited. The regime needed local agents to act as informants, in communities often made remote by a lack of roads and communication infrastructure. Local officials translated and embodied the regime locally, its programs as well as its ideology, but were also expected to collect information for it to allow a better hold on local realities.

The reliance of the national authorities—and awareness of their reliance—on local functionaries transpired in their efforts, especially obvious under the Second Republic, to multiply forms of mediation, which translated into a proliferation of layers of local administration. Decentralization under the Second Republic meant, for example, a focus on developing the structures at the level of the commune, an administrative level identified by national authorities as the main locale for the implementation of their policies (see Figure 1). ¹⁰ The commune was to be the landing point of development policies, a priority of national authorities under the Second Republic. In return, it was also to be where regime expectations with regard to Rwandans were made most obvious. Recalling that the commune was “the administrative echelon the closest to the population,” Habyarimana explained that it “remained the only center radiating all our political decisions and the only hub connecting the prefecture and the government to the countryside in political, economic, and socio-cultural fields,” making it, as explained on another occasion, the “most important unit” of the political system. ¹¹ To consecrate the importance of the commune for the regime, the Second Republic made the bourgmestre, the administrator in charge of the commune, an appointed position. Bourgmestres therefore became one of the most important vectors of the system, with expectations that they would be more clearly and hierarchically tied to the core and hence more effective vessels of its demands.

Figure 1. Subnational Political and Administrative Levels in Pre-genocide Rwanda

In addition to this greater control over who ran the commune, the Second Republic expanded the types of state agents at the very local level beyond what had existed during the First Republic, from creating local cells of the state party, the Mouvement révolutionnaire national pour le développement (National Revolutionary Movement for Development, MRND) to involving greater numbers of actors such as animateurs, agronomists, and state-run cooperatives in encouraging patriotism among Rwandans. All served the dual function of ensuring a grounded development in the country, while simultaneously multiplying the contacts Rwandans had with the regime and state, ensuring that constituents were constantly reminded of the state and regime’s presence in their lives. In other words, the commune also became a hub of representation. Over time, the regime tried to extend local mediation by focusing on other administrative levels, including creating a new one, the sous-préfecture in the 1980s.

The regime also seemed intent on ensuring that local agents dependably reproduced the system. Indeed, as conveyed in national level rhetoric at the time, authorities in Kigali seemed to nurture an obsession with reminding officials of their duties and the need for their loyalty to the system. This came in the form of local officials’ growing involvement in normative displays of authoritarianism, from weekly sessions dedicated to games and patriotic songs called animation, to community work known as umuganda undertaken for the sake of the country. Local officials were expected to perform these, as well as to watch who performed. They were both acting as exemplars of patriotism and surveyors of citizens’ patriotic displays. Through local functionaries’ performance of animation and umuganda, national authorities also hoped intermediaries would help regularize and legitimize their practice.

Ensuring local officials’ performance also came in the form of their routine retraining. The Second Republic regularly promoted what it called the recyclage of administrators, as transpired from speeches given by its president. Recyclage was described as a means not only to ensure the quality of local staff, understood as reining in those who performed poorly, but also to train local functionaries to the ideals of the system. It therefore served to remind local officials of how to perform their translation function with respect to regime expectations, vision, and ideology, as much as it served to ensure their effective performance of local political or administrative tasks.

Yet, as the obsession with recyclage and the regular public chastising of bureaucrats and poorly performing administrators over the course of pre-genocide Rwanda suggests, the relationship local officials maintained with the regime was never straightforward, and their role in anchoring the system never mechanically fulfilled. The intensification of efforts on the part of the regime to turn local officials into cogs may have contributed to the sense that the system subsumed completely those tasked with implementing it. But the regime and state never had the ability to completely control local representatives. As a result, Rwandan local functionaries regularly strayed from the regime’s vision, hierarchical structure, and expectations. They skirted or adapted, consciously or otherwise, their expected official functions. Some even avoided mandatory community work to the great frustration of the country’s top authorities. Indeed, throughout the decades prior to the genocide, the paradoxical reliance on local officials and their empowerment allowed them to enact authoritarianism in their own way.

As early as the eve of independence, the Belgian Résident général at the time, the highest colonial administrator, commented on the important imprint local authorities had on the territory they oversaw. He described the situation in the different Rwandan territories as “confused” and “varying locally depending on the personality of the Administrator and the prefect, on the proximity to Gitarama, home of the Parmehutu, [the dominant party at the time], on the local political context, even as a result of accidental factors.” ¹² The trend continued to prevail under the two successive independent Republics, despite the efforts of national authorities to turn local authorities into cogs in their authoritarian system (Reyntjens 1987:92–93). From local functionaries’ lack of experience to their outright abuse of influence, the issue of local administrative leeway in terms of how they enacted the system regularly featured in discussions of local political and administrative performance during the Second Republic, the regime keenest on developing a rigid authoritarian system. Even more so than its predecessor, the Second Republic strove to achieve a vertical chain of authority, which tied the local to the national. But it never succeeded, and the space for local interpretation and hence bifurcated reproduction remained. Indeed, despite the multiple efforts deployed by Second Republican authorities, Filip Reyntjens (1987:92–93) described efforts to rein in administrators as an “unfinished quest.” By the mid-1980s, after decades of attempted control of official authorities at the local level, the Belgian ambassador at the time noted that administrative practices remained very variable across the country, depending on who was in place locally. ¹³ Similarly, by the late 1980s, Danielle de Lame (1996:45, 56, 65) described the exercise of power locally—and we can surmise the embodiment of the authoritarian system—as “extremely personalized.”

Many Rwandans who lived and interacted with local authorities at the time explained this variability by pointing to the meager salaries at the local level, raising once again the issue of state capacity. ¹⁴ Some locals sought out these positions, even if they did not have the needed competencies for the position, in order to enjoy the advantages afforded to those in the administration. But in many instances, disappointed by their salary or benefits while in their function, local administrators used their influence for their own interest and therefore strayed from being compliant authoritarian representatives. In other words, they also began building their own system of influence and enacting representation for themselves.

This personalization directly impacted the shape of authoritarianism locally. Personalities and patterns of governance on the part of local officials mattered to the relations they fostered. This played out in the form of local authorities’ corruptibility; individuals could be bought to circumvent authoritarian expectations such as communal work, or quotas in terms of access to school. In other words, local officials sometimes worked to undermine regime expectations in an obvious manner, when they could be bought to avoid them. But the relations local officials maintained locally also sometimes constituted an alternative to the stringent relations the authoritarian core hoped to foster. As they built ties with local communities in the performance of their functions, local officials were often seen as part of the community—indeed, they often stemmed from the community. Some were seen as quite personable, dedicated to their constituents first and foremost, their power exercised to the benefit of both the regime and Rwandans. Rwandans even sometimes described functionaries at the most local levels as working for and representing citizens rather than the authoritarian system, or as being the “eye of the community”.

Experienced locally, and especially through these local intermediaries, authoritarianism was therefore rarely conceived as strictly emanating from the top. Local officials filtered and hence shaped the system. If anything, local authoritarianism could in many locales constitute an alternate authoritarian space, often experienced as a kinder, paternalistic form of guidance, in contrast to the starker rule at the top. This may explain why so many Rwandans blamed political problems and crises on national authorities, from ethnic dynamics to intense political competition, in contrast to the softer political context local officials appeared to foster in the minds of ordinary Rwandans. Harsh and cutthroat authoritarian politics resided among the authoritarian core. In the hands of local functionaries, many felt authoritarian realities could be made to work for the local context and citizens. Rather than acting as the vectors of national authorities, local officials were seen as the creators of their own local political realms, which citizens could engage with differently and more personally than with national authorities.

Local Officials Under the Inqath Regime in Sudan

Under the rule of Omar al-Bashir (1989–2019), Sudan was often portrayed as a failed, weak, or fragile state. ¹⁵ This characterization largely reflected the country’s struggle with multiple civil wars at its periphery, illustrating the state’s inability to impose order and control in the face of rebel movements. Yet, paradoxically, the Sudanese state has at times been described as “totalitarian” (El-Affendi 2013) and a “successfully failed state” (Prunier & Gisselquist 2003). This paradox is the very consequence of mediation: it produced a complex geography of power below the authoritarian core. It is through local officials’ work that the regime’s authoritarian attempts at transforming Sudanese society were at the same time expressed, modulated, and undermined, including through ambiguous translation and representation.

Local functionaries and mediation were of utmost importance for the regime. Born of a coup undertaken by the Islamist movement led by Hasan al-Turabi, the National Islamic Front, and a faction of the military led by Omar al-Bashir, the regime promoted a “civilization project” aimed at transforming Sudanese society and citizens according to its specific religious and political vision (Salomon 2016; Verhoeven 2015). To do so, like pre-genocide Rwanda, it insisted on the need for a deep implantation within society, a policy known as tamkin. Usually translated as “empowerment,” tamkin referred to the idea of penetrating pre-existing structures, capturing society, and deeply rooting the regime in society and the lives of individuals (Revilla 2021:420). It is no coincidence that the symbol of the National Congress Party (NCP), the hegemonic ruling party from 1999 to 2019, was a tree.

To reach deep into society, the regime increased the number and type of local officials through the development of a hegemonic political party and decentralization, diversifying those able to monitor the population and perform translation and representation in its name. In 1991, the new rulers formally adopted federalism, dividing the country into nine states (wilayat), 65 provinces (muhafazat), and approximately 328 localities (mahaliyat) (see Figure 2). In 1992, the regime created the Popular Committees (PCs) at the level of the neighborhood. Tasked with providing basic services, most importantly distributing essential subsidized goods, PCs were therefore a strong feature in people’s daily lives, and their elected members were often, if not always, affiliated with the ruling party.

Figure 2. Subnational Political and Administrative Levels under the Inqath Regime in Sudan

While decentralization has a long history in Sudan, dating back to the creation of the Native Administration in 1922 by the British colonial power, the logic of decentralization under the Islamist regime was one of inflation: in 1994, a new constitutional decree increased the number of wilayat to 26. Between 1989 and 2003, the number of provinces was raised to 98—although they were abolished in 2003—and the number of localities to 493 (El-Battahani & Gadkarim 2017:7). ¹⁶ Creating new localities was not only a way to spread surveillance and translate ideology, it was also a means to buy off certain groups, because having a locality meant gaining access to important basic services. Competition between groups over those local offices was militarized in some spaces, such as Darfur, and often articulated along ethnic lines, a process fostered by the regime as a divide-and-rule strategy.

In official discourses, the president criticized the idea of the tribalization of local offices and their perception as spaces for representation of specific groups. In 2015, al-Bashir reverted to appointing state governors (wali). Governorship had been turned into an elected office in the 2005 interim constitution. ¹⁷ Al-Bashir argued that these elections had contributed to “the use of regionalism and tribalism as means to gain access to positions of power at the expense of competence and citizenship” (Sudan Tribune 2014). Yet, the regime had encouraged tribalization in some areas such as Darfur. The reform was therefore aimed at regaining full control over the management of tribal divisions among the elites. ¹⁸ In addition, as the country slid into a deepening economic crisis from 2011 onward, those offices turned into lucrative rents for those ready to support the regime in these difficult times. ¹⁹

The strategic use of local offices to support the regime signals how crucial their allocation and role in mediating the regime was for national authorities. Yet, as in Rwanda, mediation as performed in practice did not serve the regime as straightforwardly as might have been hoped. This was evident notably in the role played by governors. Governors, as important officials, often embodied their own interpretations of authoritarian norms, in a form of bifurcated translation. They also nourished ambiguous relations with their constituents. Indeed, from the perspective of the citizens, governors were consistently conceived as a key point of contact if one wanted to transfer demands to the higher level. Despite the authoritarian and repressive system, citizens expected governors to be able to represent them and their interests to some degree.

Ahmed Haroun, who was governor of North Kordofan between 2013 and 2019, illustrates the complex position local officials occupied in the system and in people’s daily lives. Born in North Kordofan, Haroun nurtured local loyalties by launching a large development program, the Renaissance of North Kordofan. His speeches about this program emphasized the importance of the province. He stated, for instance, that “we should be the number one wilaya and nothing less.” ²⁰ Official videos promoting the development policy claimed that the province had a specific identity, characterized by its peacefulness and its diversity, illustrated by images of various cultural practices (Mahé 2020). In the context of a national regime that had repressed cultural practices deemed outside of the confines of the specific Islamic identity it promoted, this worked to distance Haroun to some extent from the regime and its ideology and legitimized him as someone who belonged in North Kordofan, valued local people and customs, and could consequently properly represent them.

Relatedly, his nomination to the governor’s office was seen as a change by many people because of his close connection to the region, and he was consequently perceived as a local actor as much as an agent of the broader authoritarian system. One religious leader, for instance, argued that relations between his congregation and authorities had improved after the governor’s nomination because “he is from here.” ²¹ The fact that “he knows everybody” locally was seen in a positive light. ²² He was closest to the people and had preexisting personal networks that made it possible to discuss local issues with him, and therefore solve them.

Yet, Haroun’s direct predecessor had also been from North Kordofan. If Haroun represented some form of change, it was because of his ties to the regime’s core rather than his identity. He had indeed been an early supporter of the Islamist movement and the regime, played a central role in the establishment of the regime’s security apparatus after the 1989 coup, and most notably coordinated the government’s counterinsurgency campaign in Darfur between 2003 and 2005. His role in some of the regime’s darker episodes made him largely untouchable despite the controversies. He was, for instance, at some point strongly criticized by NCP members for his role in the outbreak of the conflict in South Kordofan, where he was governor between 2009 and 2012 (International Crisis Group 2013). He was nonetheless appointed governor of North Kordofan and replaced Omar al-Bashir as the head of the NCP in March 2019, days before the overthrow of the regime.

This proximity to the inner core enabled Haroun to negotiate some degree of freedom from the system and gain support for his own initiatives in the province. According to a story that circulated in El Obeid in 2015, he phoned the president after citizens complained about a speech al-Bashir had made in which he criticized the population of North Kordofan for its demands regarding basic services. Haroun arranged a meeting between a delegation from North Kordofan and al-Bashir in Khartoum and, months later, the president came to El Obeid and declared that the central government would financially support the Renaissance of North Kordofan. ²³ A picture of this event was taken and later used in promotional material for the program.

Such stories played up Haroun as a representative and even defender of North Kordofan within the authoritarian system. Yet the Renaissance itself was an act of translation of the regime’s values, rhetoric, and practices at the local level. This development policy was based on the idea of popular participation and mobilization and the tenets of tamkin, which it expressed by referring to “nafir,” a Sudanese tradition of communal work and mutual aid especially common in the western parts of the country and valued in North Kordofan. ²⁴ While this reference to tradition made participation locally intelligible, it also stood in continuity with the regime’s practices. Indeed, the Inqath had reinterpreted and implemented similar nafir many times since its inception, for instance, to build universities or roads (Mahé 2018). The nafir of the regime, just like that of the Renaissance, were implemented in practice as a form of taxation that implied accrued surveillance and control, usually using the PCs as boots on the ground. Rather than relying on the PCs to organize participation, the governor oversaw the creation of nafir committees at every level of the wilaya, a process that demonstrated his willingness to build his own network of surveillance, mobilization, and co-optation in parallel to the central authorities.

This is the ambiguity of mediation at work. Because of his in-betweenness, which was discursively built and instrumentalized, Haroun was able to translate the tamkin policy in a way that fostered his own vested interests, as well as to support his own local networks. But it also benefitted the local community in very tangible ways: roads were asphalted overnight, a hospital was renovated, and a new stadium inaugurated. Mediation in this case shaped the authoritarian system in ways that made it possible for people to support their local official while criticizing the authoritarian core, and to engage with the former on the basis of good faith in favor of development. People never forgot that Haroun was a loyalist to the regime. But this ambiguous positioning and role in mediating the regime meant that many citizens inhabited an “ambivalent middle” (Wedeen 2019:3). Citizens saw both the role of the governor in a violent authoritarian system and how the wilaya nonetheless benefitted from having Haroun as its representative.

Sought After but Never Fully-Controlled Mediation

In Rwanda and Sudan, autocrats’ emphasis on the necessity to mobilize the population, to extend control down to the local level, and to root the regime in local realities, made the work of local functionaries an essential component of the authoritarian system. It is evident from both cases that the two regimes fully understood the utility of local functionaries. Local officials were meant to enact surveillance and produce conformity locally. In addition, they also conveyed the regimes’ vision, structures, and policies. For the center, local functionaries were the regime’s tool to build its ties and relations locally, which national authorities saw as essential to their groundings. The regimes’ keen awareness of the importance of mediation transpired in the growing multiplication of local officials, as well as efforts to keep them in the fold. Yet ultimately, mediation never perfectly reproduced authoritarianism locally.

In Rwanda, national authorities pushed decentralization to ensure a greater footprint of their authoritarian system at the local level. But instead of fashioning local officials into cogs diligently reproducing the system, Kigali struggled to keep them from exercising power in their own way. Despite the successive authoritarian regimes in Kigali, local politicians and administrators often continued to exercise their power idiosyncratically, and through clientship ties and personal networks. Some local functionaries proved more coercive, some amenable. Indeed, for many Rwandans, these local officials worked as much for them as for the system, existing therefore not as clear executors of Kigali’s authoritarian order, but as managers of both local and national expectations. Ultimately, even though it was conceived as one of the main tools for grounding autocratic rule, the complexity of mediation at the local level never allowed the system to achieve the “implantation profonde” it sought. Instead, the alternative local authoritarian spaces that were produced allowed Rwandans to feel a distance from Kigali, and to feel that their actual political realities were the local ones.

In the case of Sudan, the study of the Inqath regime shows how the authoritarian system used the tightly enmeshed structure of decentralization to try to control local officials as much as the population, and to make them into effective mediators of the system. Yet, looking at the kind of work local officials engaged in on the ground illustrates the limits of this control. Even some of the most embedded local officials, such as Haroun, the governor of North Kordofan, cultivated ambivalence, translating and representing the bottom to the top, as much as the top to the bottom, straddling both the world of power circles and the local. In Sudan too, mediation contributed to the projection of the system locally, but paradoxically never afforded it the deep local roots it sought.

Conclusion

No junta, ruling family, or political party leadership can extend its rule alone. All authoritarian regimes need intermediaries to reproduce themselves beyond the core. To do so, authoritarian regimes rely in large part on local functionaries, embedded in the system and therefore its most obvious purveyors. Scholars have begun to understand the role local officials play, especially in terms of some of the key functions they fulfill for autocrats: intelligence-gathering, delegated repression, and cooptation, as well as serving as the target of discontent. Yet, the focus on these has reinforced an implicit bias toward understanding local officials as inherent reproducers of the system. Even if they are known in scholarship to be able to oppose authority or to act in their own self-interest, the predominant tendency has been to focus on how local officials choose to faithfully reproduce the regime.

Through our work, we have shifted the focus to the roles of local functionaries in translating and representing authoritarianism locally. By virtue of the influence mediation gives to local officials in terms of conveying its norms, policies, and practices, as well as by enacting its relations with the population, local officials are never straightforward reproducers of the system. Both translation and representation entail forms of local bifurcations that ultimately create alternative, if not competitive, political realities. This ultimately means that, though they are the most obvious subjects of regime expectations with regard to building authoritarian resilience and reach, local functionaries always build bases for the regime that are imperfect.

Because they engage with citizens and matter to their daily lives to a much greater extent than national political elites, local officials also matter in a very real way for the destinies of authoritarian regimes. Even under systems expecting strict compliance, these local actors straddle the boundary between reproducing and undoing the system when they enact it locally. To some extent, bifurcated mediation may be to the advantage of regimes. Mediation brings flexibility and adaptability to the system; it is shaped by how local functionaries feel they need to embed themselves locally to perform their reproductive tasks. It gives the system a local edge that local populations may come to recognize better. Yet, as our two cases illustrate, mediation also comes at the cost of a straightforward, deep, local implantation. Personalization of power at the local level, even in the form of positive relations between local agents and their constituency, impacts citizens’ relations with the broader system, and ultimately its core. Mediation may not be what dramatically breaks down an authoritarian system, but as it builds the system locally it also inherently magnifies some of its frailties. At its worst, if mediation breaks down, core elites can become increasingly insulated from what goes on at the lower levels. This is why, as the cases of Rwanda and Sudan further illustrate, a significant part of authoritarian regimes’ efforts is dedicated to creating and, just as importantly, managing mediation and the local officials who perform it. Mediation may have been neglected by scholarship, but regimes have long been aware of their reliance on it and of the type of influence it gives those who perform it. Indeed, given its importance, mediation is ultimately at the heart of how we understand and study authoritarianism.

Though we argue for the ubiquity of bifurcation in mediation processes, one of the limitations of our study was not exploring further the factors that influence the degree to which intermediaries matter within a political system, as well as what influences the degree to which mediation bifurcates regime expectations. We suggest these may be the most promising directions for future research into the mediation of authoritarian regimes. Key factors potentially stand out with regard to both the importance of intermediaries within a specific authoritarian regime and the degree to which mediation can potentially stray from expectations at the top. These include state capacity, the nature of local functionaries’ positions, and especially whether they are appointed or elected, as well as some of the features of the political system. Federal systems, notably, given their propensity toward devolution, may lead to more substantive patterns of bifurcation. This may be something they encourage, in order to build greater accommodation of federated states’ differences. In some contexts, however, where tensions exist between federated states and the federal system, these bifurcations may contribute to nurturing divisions. Similarly, although we have shown that bifurcations exist across both soft and hard forms of authoritarianisms, the existing literature seems to suggest that certain types of regimes may create different patterns of mediation at the local level. Hybrid systems may lend themselves to greater variation at the local level. What this entails, however, in terms of the resilience of these hybrid systems, warrants further investigation.