OXFORD STUDIES IN AFRICAN POLITICS AND INTERNATIONAL RELATIONS General Editors NIC CHEESEMAN, PEACE MEDIE, AND RICARDO SOARES DE OLIVEIRA OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi Oxford Studies in African Politics and International Relations is a series for scholars and students working on African politics and International Relations and related disciplines. Volumes concentrate on contemporary developments in African political science, political economy, and International Relations, such as electoral politics, democratization, decentralization, gender and political representation, the political impact of natural resources, the dynamics and consequences of conflict, comparative political thought, and the nature of the continent’s engagement with the East and West. Comparative and mixed methods work is particularly encouraged. Case studies are welcomed but should demonstrate the broader theoretical and empirical implications of the study and its wider relevance to contemporary debates. The focus of the series is on sub- Saharan Africa, although proposals that explain how the region engages with North Africa and other parts of the world are of interest. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi Arbitrary States Social Control and Modern Authoritarianism in Museveni’s Uganda REBECCA TAPSCOTT 1 OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 3 Great Clarendon Street, Oxford, OX2 6DP, United Kingdom Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide. Oxford is a registered trade mark of Oxford University Press in the UK and in certain other countries © Rebecca Tapscott 2021 The moral rights of the author have been asserted First Edition published in 2021 Impression: 1 Some rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, for commercial purposes, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, by licence or under terms agreed with the appropriate reprographics rights organization. This is an open access publication, available online and distributed under the terms of a Creative Commons Attribution – Non Commercial – No Derivatives 4.0 International licence (CC BY-NC-ND 4.0), a copy of which is available at http://creativecommons.org/licenses/by-nc-nd/4.0/. Enquiries concerning reproduction outside the scope of this licence should be sent to the Rights Department, Oxford University Press, at the address above Published in the United States of America by Oxford University Press 198 Madison Avenue, New York, NY 10016, United States of America British Library Cataloguing in Publication Data Data available Library of Congress Control Number: 2021933370 ISBN 978–0–19–885647–4 DOI: 10.1093/oso/9780198856474.001.0001 Printed and bound by CPI Group (UK) Ltd, Croydon, CR0 4YY Links to third party websites are provided by Oxford in good faith and for information only. Oxford disclaims any responsibility for the materials contained in any third party website referenced in this work. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi http://creativecommons.org/licenses/by-nc-nd/4.0 To my parents OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi Acknowledgements This book is a product of countless conversations, observations, exchanges of ideas, and encouragement that occurred over the better part of eight years. The book began as a PhD thesis, and the final manuscript owes a great deal to my supervisors. Alex de Waal, Dyan Mazurana, Jenny Aker, and Will Reno—each contributed substantially and uniquely to this project. Alex encouraged me to pursue interesting questions even when answers appeared elusive, and has con- tinued to patiently and generously engage with and support this project through its many iterations. Dyan’s attention to detail and thoughtful readings have critically developed and nuanced my work, in these pages and elsewhere. Jenny’s support and enthusiasm have been indispensable; her dual commitment to scientific inquiry and compassion continue to inspire me. I owe a special thanks to Will, who offered detailed comments on multiple versions of this manuscript, often articulating key aspects of my argument more clearly and concisely than I had been able to do myself. Will encouraged me to be ambitious and adventur- ous in my thinking, and continually pointed me toward the broader theoretical implications of my findings. The research would not have been possible without support of numerous donors. Fieldwork was supported by two research consortia—the Justice and Security Research Programme (JSRP, grant contract PO5729) and the Centre for Public Authority and International Development (CPAID, reference number ES/P008038/1)—both funded by the United Kingdom’s Department for International Development. I received additional funding from the Gerda Henkel Foundation’s Special Programme for Security, Society and the State. JSRP and CPAID also offered me access to a bright and accomplished community of researchers studying authority, justice, and security in eastern Africa and beyond. Particular gratitude is owed to Tim Allen, who showed great faith in my potential as a young scholar, encouraging me to push forward on this project. Tim also helped me secure research funding that allowed me to return to Uganda over several years, and introduced me to the vibrant research community at the Firoz Lalji Centre for Africa at the London School of Economics. I also received funds from the Morris Abrahams Foundation and the Fletcher School that allowed me to attend conferences and supported additional fieldwork costs. Revisions on this manuscript took place at the Graduate Institute of International and Development Studies in Geneva, where I was hosted at the Albert Hirschman Centre on Democracy. The support of the Centre’s team was indispensable. Thank you to Shalini Randeria for welcoming me to the Centre, OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi and to Christine Lutringer for fostering a friendly and collaborative working environment. In Uganda, I was very lucky to meet and work with a diverse group of highly skilled researchers and scholars, each of whom brought their own perspectives and opinions to my project, and provided excellent company, valuable advice, moral and intellectual support, and encouragement in the field. I have opted not to include their names here out of what I hope is an abundance of caution for the possibility of a changing political environment. Most especially, I thank the many Ugandans who participated in this project. I hope they see their views reflected in these pages. Many other scholars have engaged with my ideas over the past five years, encouraging me to look at my questions and data in new ways. I cannot possibly name them all—a few people who were particularly generous with their time and thoughts include Rita Abrahamsen, Henni Alava, Cathy Boone, Adam Branch, Pierre Englebert, Julian Hopwood, Mette Lind Kusk, Gabrielle Lynch, Anna Macdonald, Dipali Mukhopadhyay, Dennis Rodgers, Nancy Rydberg, Philipp Schulz, Nora Stel, and Koen Vlassenroot. Others have offered incisive comments on parts of the text, including Jean-François Bayart, Jonathan Fisher, Jonathan Goodhand, Samuel Hickey, Ben Jones, Matthew Kandel, Keith Krause, Milli Lake, Didier Péclard, Holly Porter, Sungmin Rho, and Aidan Russell. I had opportunities to present this work and receive comments at the Conflict Research Group’s ‘Governance at the Edge of the State’ summer school in 2014, Northwestern University’s Program of African Studies Lecture Series in 2016, Uganda Elections Workshop hosted by Richard Vokes and Sam Wilkins in 2016, the Harvard-MIT-Tufts-Yale Political Violence Workshop in 2017, the Graduate Institute’s Political Science Faculty Seminar in 2018, as well as numerous inter- national conferences. The project further received valuable comments at two workshops that I co-organized on the politics of disorder—one at World Peace Foundation in 2017 and the other at the Democracy Centre in 2019. These workshops were generously funded by the World Peace Foundation, the Swiss National Science Foundation, and the Gerda Henkel Foundation. I owe a huge thanks to Nic Cheeseman, who has continually engaged with my ideas and encouraged me to be ambitious and bold in my endeavours. With Ricardo Soares de Oliveira and Dominic Byatt, Nic shepherded this project through the publishing process, and for that I am extremely thankful to all of them. The book also benefited from the insightful and constructively critical comments of three anonymous reviewers. Michelle Neiman and editors at Oxford University Press improved the prose with excellent editorial work. Parts of this book have appeared in journal articles: an early version of the argument appeared in an article in Development and Change; a version of the ethnographic vignette in Chapter 5 appears in an article OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi viii  in African Affairs; and parts of Chapter 6 on crime preventers appear in an article in the Journal of Eastern African Studies. Most of all, my gratitude goes to friends and family whose belief that I would complete this book allowed me to do so: my husband and colleague, Deval Desai, joined me at the early stages of this project. His unflagging engagement has added depth and nuance to every page of this book. My parents, Susan Weinstein and Stephen Tapscott, and my sister, Leah, have unwaveringly supported and encour- aged my intellectual pursuits and aspirations. Finally, I thank the family that I have gained along the way—my son and my family-in-law—who have made my world bigger and my life fuller, and for whom I am immensely grateful. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi  ix OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi Contents List of Figures xiii List of Tables xv List of Abbreviations xvii 1. Violence, Governance, and Uncertainty: An Introduction to Citizens’ Lived Experiences of the Ugandan State 1 2. Arbitrary Governance and Modern Authoritarianism 17 3. Institutionalized Arbitrariness in Uganda (1986–2016) 47 4. Violence, Sovereignty, and the Uganda Police Force 73 5. Claiming Jurisdiction: Local Vigilantes and the Struggle to Consolidate Power 98 6. Whither the State? Surveillance, Crime Preventers, and Potential State Presence 126 7. Varieties of Arbitrary Governance 151 8. Arbitrary Governance in Africa and Beyond 188 References 207 Index 225 OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi List of Figures 0.1. Gulu, Mbarara, Moroto, and Soroti, Uganda xviii 2.1. Dynamic oppositions in institutionalized arbitrariness 31 3.1. District creation in Uganda (1959–2019) 55 4.1. Residents carry the deceased from the health post toward the police station 90 5.1. Two batons used during night patrols 109 6.1. Crime preventers at pass-out ceremony with President Museveni presiding 133 6.2. Protesters walking toward Karuma Bridge 139 7.1. Total fatalities per year by district 154 7.2. Vote share for President Museveni by district 155 OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi List of Tables 1.1. The oppositions and factors of institutionalized arbitrariness 12 1.2. Case selection to probe alternative explanations for the perception of arbitrary state governance 13 1.3. Varieties of institutionalized arbitrariness 14 7.1. Case selection to probe alternative explanations for the perception of arbitrary state governance 156 7.2. The four oppositions describing state capacity and their contribution to arbitrary governance 157 7.3. Hypothesized values for four factors of institutionalized arbitrariness 158 7.4. Types of arbitrary governance 180 OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi List of Abbreviations ADF Allied Democratic Forces DPC District Police Commissioner DRC Democratic Republic of the Congo EPRDF Ethiopian People’s Revolutionary Democratic Front FDC Forum for Democratic Change IGP Inspector General of Police ISO Internal Security Organisation LC Local Council LDU local defence unit LRA Lord’s Resistance Army MP member of parliament NAADS National Agricultural Advisory Services NCPF National Crime Preventers’ Forum NGO non-governmental organization NRA National Resistance Army NRM National Resistance Movement POMA Public Order Management Act RDC Resident District Commissioner RPF Rwanda Patriotic Front UPDF Uganda People’s Defence Force US United States ZANU(PF) Zimbabwe African National Union (Patriotic Front) OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi S O U T H S U D A N DEM. REP. OF THE CONGO KENYA TANZANIA KAMPAL A Mbarara 0 0 0 50 50 100 mi 33 33 3330 0 100 km RWANDA Gulu Soroti Lake Kyoga Victoria Nile Lake Kwania Lake Albert Marghenta Peak Equator Lake George Lake Victoria Lake Edward Moroto Victoria Nile Al be rt Ni le Fig. 0.1 Gulu, Mbarara, Moroto, and Soroti, Uganda Source: Wikimedia Commons, modified by author OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 1 Violence, Governance, and Uncertainty An Introduction to Citizens’ Lived Experiences of the Ugandan State Around midnight in June 2014, in a village in northern Uganda, an unarmed 37-year-old man was shot in the back as he fled from a police officer. He died hours before he was found, lying face down in a compound a few hundred metres from his home. That night, he had been guarding his community as part of a local vigilante group. A year previously, the village had experienced a surge in violent crime. The police and district-level authorities did not have the capacity to secure the village; instead, they had called on the community to form a vigilante group for their own protection. The vigilantes patrolled with rudimentary weapons, including machetes, sticks, and ropes to restrain suspects. They began to organize themselves, implementing code words and rules of conduct, enforcing a curfew, and collecting taxes from community members. But their work was controversial: they failed to catch criminals and were accused of drinking on the job as well as beating and extorting community members. The morning after the shooting, community members learned that a resident of the village had falsely reported to the police that the vigilantes were armed with guns and robbing passers-by. Community members speculated that the resident had lied to prompt the police to arrest the vigilantes, which would lift the curfew and its penalties. Instead, the police arrived and shot into the dark, killing one vigilante and maiming another. William Odera, the brother of the deceased vigilante, sought justice.¹ He wanted retribution for his brother’s death and compensation to support the slain man’s children. He decided to open a police investigation into the murder. But when Odera went to the police, they replied that the shooting was caused by a miscommunication within the village and so he should resolve his complaint ‘at home’. A government official sided with the police, arguing that the vigilantes were in fact armed robbers, and thus their demise was warranted. Odera expressed frustration—it is well known in Uganda that to solve such a problem ‘at home’ ¹ All names have been changed to protect the identity of respondents. Names of specific villages have been changed or generalized to reflect the municipality level or higher. Where necessary to protect the identity of the respondent, I have omitted interview locations or dates. Unless otherwise noted, respondents are male; with the caveat that I have also obscured the gender of respondents if it might reveal their identity (for example, in the cases of former government officials). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi Arbitrary States: Social Control and Modern Authoritarianism in Museveni’s Uganda. Rebecca Tapscott, Oxford University Press (2021). © Rebecca Tapscott. DOI: 10.1093/oso/9780198856474.003.0001 would require Odera’s family to attack the family of the youth who had made the false report. Instead, Odera gathered witness statements and took them to police headquar- ters, first in Gulu Town and later in Kampala. While facing the aftermath of arbitrary state violence and impunity, Odera did not disengage from the state. Instead, he turned to it for justice—even though the state’s agents were responsible for the death of his brother. In response to his complaint, the state’s institutions ground into action. Over the better part of a decade, Odera followed a lengthy bureaucratic procedure for internal investigations of police behaviour, holding out hope that it would deliver justice even while his efforts were met repeatedly with obstacles. A mutual friend reflected on Odera’s actions: Odera would not survive either choice [solving the matter at home or going to the criminal justice system]. If he had retaliated that very night, he would do it, fine—kill people, kill livestock, destroy property—to finish his own interests. But then the government would come in and say, ‘We have the court of law, you should have brought it to us.’Odera, on the other hand, said, ‘I cannot retaliate— let me take the right path of law . . . It is their job to protect and not destroy the community members.’ Would the government really accept that? Never. The only possibility is to calm down Odera and at the end of the day they frustrate him: ‘Your case file is lost, we don’t know where it is.’ They’re just trying to confuse the case . . . So the man [who reported the vigilantes to the police] is a free man now, he doesn’t have anything to answer, and Odera does not have any right to retaliate. It has been three years. The government is frustrating Odera time and again. (32-year-old male, phone interview, 5 March 2017) Odera faced a predicament. He was caught between the exhortation to take the matter into his own hands, and the continued relevance of the ‘right path of law’. He responded to this contradiction by submitting to government induced ‘confusion’. He allowed himself to be ‘calmed down’ in response to the numerous obstacles that ‘frustrated’ his interests. The police held off Odera’s claims not through intimidation or direct refusal to help him, but rather by engaging him in a lengthy bureaucratic process that kept him involved because it could plausibly deliver justice—especially compared to the available alternatives. This case took place under the modern authoritarian regime in Museveni’s Uganda. This authoritarian regime can be categorized as ‘modern’ in the sense that it governs in large part by law, rather than unrestrained violence and executive decrees. For instance, to allow Museveni to remain president for over 35 years, the constitutional provisions for presidential term limits, and later age limit, had to be removed. These changes were made not by fiat, but by acts of OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 2   parliament that were challenged and upheld in court.² On one hand, such an approach has benefits. In Museveni’s Uganda, it has allowed the ruling regime to frame itself as a fledgling democracy, deserving of foreign aid and investment. On the other hand, it carries risks. Ordinary citizens can use the resultant civic spaces and democratic institutions—however limited—to organize and make claims on the regime. For example, Ugandans seeking to challenge Museveni have done so through civic protest, electoral campaigns, the press, the courts, and Parliament. Live opposition means that the regime is always engaged in a balancing act. As a result, these modern authoritarian regimes have been widely described as inher- ently unstable and prone to conflict. Scholars continue to debate how these regimes govern, given that they weaken their own institutions in pursuit of unchecked executive control (c.f. Levitsky and Way 2002; Schedler 2006; Scheppele 2018). To contribute to this ongoing debate, this book identifies a new form of modern authoritarianism—arbitrary governance—that focuses more on weakening com- petition than on maximizing control. Such arbitrary states allow for pockets of civic organization and pathways for citizens to make claims on the state; however, they make these spaces fragile by intervening in them violently and unpredictably. Such interventions—in which state authorities assert their right to control a situation, activity, people, or territory—supersede or displace other authorities operating there. The effect is to make the role of local public authorities³ contin- gent and reconfigurable in relation to the regime. Arbitrary states can therefore paradoxically encourage non-state and local actors to use violence to control people, resources, and territory, while limiting the capacity of such actors to consolidate power. The state’s ability to reinforce its interventions with over- whelming and unaccountable violence makes it difficult for citizens and local authorities to calculate and assess the risks of possible intervention, which causes them to self-police. I call this mode of governance institutionalized arbitrariness. It helps to explain how today’s authoritarian rulers can project power even as their pursuit of an unchecked executive threatens their own institutional capacity. I use the terms institutionalized arbitrariness and arbitrary governance throughout this ² The extent to which these amendments were procedurally proper was challenged in Uganda’s Constitutional Court; in the end, the amendment was upheld by a 4–3 vote. Opposition Members of Parliament (MPs) charged that ‘the amendment is invalid because it was passed during parliamentary sessions marred by brawls and plain-clothes security operatives assaulting and dragging MPs out of the chamber. They also argue that Mr Museveni wants to be president for life and the amendment violates a basic democratic principle, that power belongs to the people. His allies say the amendment removes age discrimination, and was passed after widespread public consultation’ (BBC News 2018). The process might be characterized as plausibly compliant with a procedural vision of the rule of law. See Chapter 3 for more details. ³ The term ‘public authority’ refers to actors who hold ‘legitimate authority’ and provide public goods, such as security and justice. The term focuses attention on practices of authority, viewing it as an ‘emergent property, always in production’; as well as on the dynamics of competition, conflict, and contestation that contribute to its production (Hoffmann and Kirk 2013, 2, 6). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi , ,   3 book, with institutionalized arbitrariness emphasizing the regime’s governing strategy, and arbitrary governance emphasizing citizens’ perceptions of that strategy. Based on a study of the micro-dynamics of violence and governance in Uganda,⁴ such as the activities of vigilante groups like the one described in the first pages, this book makes three main contributions to studies of modern authoritarian rule. First, it identifies a new type of authoritarian regime that is more concerned with undermining threats to its authority than with monopoliz- ing violence. In contrast to scholarship that sees governance as a quest for ever increasing control, institutionalized arbitrariness relies on unpredictable asser- tions and denials of authority that fragment and weaken civil society and local public authorities. Second, this book describes the mechanisms by which such regimes work, presenting them in a four-part framework. I use this framework to analyse how arbitrary governance shapes state–society relations by fracturing civil society, limiting the space for political claim making, and causing citizens to self- police. Third, the book uses sub-national variation in Uganda to probe key alternative explanations for citizens’ perceptions of arbitrary governance, includ- ing conflict-affectedness and political relationship to the ruling regime. It finds that these alternative explanations do not account for arbitrary governance, and yields a typology of arbitrary governance that illustrates how the same factors in different combinations similarly fragment civil society and cause citizens to self- police. The remainder of this chapter draws on the case of Odera and his deceased brother to elaborate institutionalized arbitrariness in the context of new global trends in authoritarian rule. It then describes the research question and method- ology, before previewing the book’s major contributions and outlining its organization. 1. Arbitrary Governance and Social Control in Authoritarian States In recent years, scholars of authoritarianism have noted a trend in which institu- tions designed to implement the rule of law and democratic governance have been hollowed out to facilitate the ruler’s ability to exercise arbitrary power. They point to the rise of authoritarian and populist tendencies in countries like Hungary and Poland, Venezuela and Brazil, India and Myanmar, and the United States (US). Even as those who study authoritarianism grapple with how to analyse these ⁴ This study traces the local-level workings of state power to investigate who can use violence, how, and with what effect. Many scholars examine the micro-dynamics of violence to locate the sources of coercive power that underpin state authority and the state–society contract (e.g., Barkey 1994; Blok 1975; Kalyvas 2006; Volkov 2002). This book seeks to build on that tradition. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 4   observations, scholars of sub-Saharan African politics have studied the disjunct between state institutions and executive power for decades. Both sets of scholars study regimes characterized by the ruler’s use of arbitrary power, democratic- looking institutions, and resultant political unpredictability. Since the wave of democratization on the African sub-continent in the 1990s, scholars of the post-colonial African state have increasingly noted that today’s African rulers diverge in character from those of the early post-independence period. Rather than espousing pan-African and revolutionary ideologies, today’s African statesmen adopt the rhetoric of democratic institutionalists, calling for universal suffrage, the rule of law, checks and balances, and transparency and accountability (Cheeseman 2015). At the same time scholarly depictions of the post-colonial African state as a shadow state (Reno 1995) governed by politics ‘of disorder’ (Chabal and Daloz 1999) or ‘of the belly’ (Bayart 1993) seem to retain relevance. They also bear an uncanny resemblance to some forms and operations of arbitrary power increasingly identified in the twenty-first century and framed as ‘right wing’ and ‘populist’ in the global North, from the US to Poland (Norris and Inglehart 2019; Scheppele 2018). These ‘hybrid’ regimes—neither democratic nor authoritarian—somehow balance democratic institutions and repressive tenden- cies to project power over their populations, without buckling under the pressure of it all. To help answer this puzzle, this book combines hundreds of interviews conducted in Museveni’s Uganda with the rich literature on the post-colonial African state to understand how such authoritarian rulers project arbitrary power and how it plays out in the daily lives of ordinary citizens. It is grounded in and focused on scholarship theorizing the African state, while contributing to a broader debate about contemporary forms of authoritarianism. This book identifies unpredictability as a key tool that helps today’s authori- tarians maintain a balance between democratic institutions and arbitrary power. These arbitrary states cultivate an institutional environment structured to accom- modate numerous authorities with overlapping and contestable jurisdictional remits. The result is a governance landscape of fragmented and competing sources of power. To maintain its authority, the regime intervenes with overwhelming and unaccountable violence at unpredictable moments. In an environment so charac- terized by unpredictable state intervention, ordinary citizens cannot manage or ignore the possibility of state interference—but neither can they rely on it. Instead, state authority is unpredictably present and absent. Rather than eliminate civic organization, such an approach makes civic spaces fragile, undermines collective action, and destabilizes those who would seek to challenge the state, allowing the regime to remain the most powerful governing actor among many. This modern form of authoritarianism produces chaos and competition outside the ruling cadre. Nevertheless, arbitrary governance is able to accommodate—and even relies upon—stable bureaucratic institutions at a local level. These bureaucratic institutions are neither mere façades nor are they overrun by society. They often OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi , ,   5 work—sometimes even in an impersonal way—to deliver services and project state power. Functional bureaucracies distinguish this type of governance from both neopatrimonialism, in which weak state institutions are colonized by per- sonalist rule, and also from traditional understandings of authoritarianism, which often assume that rulers will seek to eliminate institutions with autonomous function in order to maximize their control (Fukuyama 2013). Institutionalized arbitrariness is a mode of governance through which the state produces a self-policing population that can be alternately demobilized and remobilized. Though citizens may seek to subvert, challenge, or engage the state on a daily basis, pervasive uncertainty dilutes their efforts. For instance, the vigilante group that Odera’s brother joined sought to substitute for state police in their village, but their authority evaporated when the police intervened and shot at them. At the same time, the persistent possibility that state institutions will function as they ought means that even when citizens become disenchanted and suspicious, many continue to engage with the state’s formal governance institutions. For example, for Odera, the government’s lengthy and complex process kept him engaged, if also apprehensive, so that he remained ready to jump into action if the case showed signs of movement. The possibility that he would be arrested himself if he took the law into his own hands only made him more invested in the state’s process. Institutionalized arbitrariness allows such regimes to project power, cultivating a citizenry that generally abides by the law, votes, and—more often than not—submits to state authority even when it lacks regular presence, consistent coercive control, or the regular ability to provide services. By illustrating the micro-dynamics of violent contestation between local actors and state authorities, the case of Odera and his brother draws our attention to the fundamental importance of disorganization and unpredictabil- ity as key tools of authoritarian governance—the core thesis advanced in this book. The term ‘institutionalized arbitrariness’ highlights how the seemingly contra- dictory notions of ‘institutionalization’ and ‘arbitrariness’ can be combined to create productive tensions that enable the state to project authoritarian power over a population and a territory. ‘Arbitrary’ refers to a ruler’s unchecked and unaccountable power, exercised in such a way that those affected cannot predict or understand how power is wielded and have no means of questioning or challenging it. It is thus necessary to look not just at how rulers exercise discretion in an arbitrary way, but also at the institutional apparatus through which power is exercised, and potentially restrained or tempered, as well as how these institutions shape citizens’ expectations about how power might be used in the future (Krygier 2016, 203–4). Arbitrariness is institutionalized in that it has become a regular part of how such authoritarian regimes function. The resultant predictable unpredict- ability shapes the behaviour of ordinary citizens and offers important insights into modern authoritarianism. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 6   When writing analytically about violence, it can be difficult to acknowledge its personal costs. The death of Odera’s brother and its aftermath offer a small reminder of the long-term effects for friends and family, communities and clans, of what might otherwise seem to be merely an impersonal instantiation of state violence. This book does not aim to make a normative claim about what constitutes ‘good’ governance. However, the cases described herein illustrate questions of justice, morality, and dignity confronted by people living in modern authoritarian states. Though this book focuses primarily on physical and material aspects of security, it does so with the conviction that this has far-reaching and deep-seated social, cultural, and interpersonal implications. 2. Research Questions and Motivation This study began as a comparative inquiry into vigilante governance in contexts of limited statehood.⁵ I originally planned to ask where vigilante groups exist and why, what form they take, and what determines group cohesion and longevity. This inquiry was based on the hypothesis that vigilante groups arise to fill a security and governance gap. My questions were intended to isolate the micro- dynamics of violence and governance and document grass-roots processes of the consolidation of power in the absence of the state. I believed that the functioning of vigilante groups would reveal the process by which violence is incrementally institutionalized at a local level and offer insights into the micro-processes of state formation. I expected that, unlike the aspiring rulers in Charles Tilly’s model of European state formation, Uganda’s twenty-first-century vigilantes would be savvy about state authority and how to use it to their advantage. As a result, I hypothesized that they would strategically adopt the symbols, rhetoric, and forms of a modern state to strengthen their position without making the pre- sumed linear and sequential transition to a formal and bureaucratized institution (Hagmann and Péclard 2010; Lund 2006b). I anticipated that vigilantes would thus produce and enact bureaucratic institutional forms, while in practice relying on personalized relationships to conduct business. What I found was much more interesting. I began my field research in northern Uganda in 2014. In this recently post- conflict environment, political order was fragmented and contested, offering opportunities for new forms of authority to emerge and consolidate control. Gulu Town—the main urban centre in northern Uganda, located approximately 100 kilometres from South Sudan—had been the epicentre of a brutal and violent ⁵ The concept of ‘limited statehood’ refers to countries with ineffective state institutions, that lack the capacity to implement central decisions and monopolize force, but that still are able to govern, for example relying on partnerships with private or non-state actors (Risse 2011). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi , ,   7 conflict from 1986 to 2006, during which the Lord’s Resistance Army (LRA) rebels fought Uganda’s government, victimizing the civilian population. The conflict is often framed in ethnic terms, with the Acholi LRA rebels emerging to challenge the new regime of Yoweri Museveni, who had violently seized power in 1986, thereby displacing several decades of northern rule. Nearly ten years after the informal end of the LRA war, the conflict-affected northern region still faced the highest poverty rate in the country⁶ and had the lowest levels of infrastructure development, with limited access to markets and services (World Bank 2016). Northerners—and residents of Gulu in particular—opposed the ruling National Resistance Movement (NRM) regime on regional and ethnic grounds (Branch 2011). As a study location, Gulu offered a rich environment to observe entrepre- neurial politics in the struggle to establish new public orders in the aftermath of war. In particular, local vigilante groups were common. Such groups have been shown to use violence to make governance claims at a local level. They would thus be a valuable lens to examine the dynamics I wished to study (see Abrahams 1987; Buur 2006; Pratten and Sen 2008). I reasoned that tensions between the centre and the periphery would actively be unfolding and thus ripe for observation and documentation. During my first month in Uganda in February 2014, I conducted 18 interviews with 23 respondents using a combination of convenience and snowball sampling. I also discussed my research agenda with several scholars working in the area.⁷My exploratory study showed that my original questions would be difficult to answer—and that those answers might reveal very little about governance and state authority. The vigilante groups active in the area were hard to categorize in terms of their history, form, function, use of violence, and even participants. For instance, asking when a group was founded elicited different answers from different respondents. Upon further investigation, I discovered that most groups had been formed and dissolved repeatedly; their membership and mandate shifted over time. Low-level crime and insecurity paired with a long history of civilian militias meant that, over many years, either a ‘new’ vigilante group was formed or an ‘old’ one was resurrected. The form of these groups, too, was fluid: such groups were ‘non-state’ in that their members received none of the benefits afforded to formal state employees, yet they reported to and received orders from local councillors.⁸ At times they received payment from informal systems of local ⁶ In 2013, the northern region had a poverty rate of 43.7 per cent, compared with 24.5 per cent in the eastern region, 8.7 per cent in the western region, and 4.7 per cent in the central region. The annual percentage reduction in the north was also the lowest of the four regions at 3.1 per cent (World Bank 2016, 5). ⁷ I owe many thanks to scholars including Tim Allen, Teddy Atim, Ronald Atkinson, Adam Branch, Julian Hopwood, Holly Porter, and Alex deWaal for early insights and encouragement in the field. ⁸ Local councillors, or LC1s, are most local administrative position in Uganda. The Local Council (LC) structure was established as a part of the NRM’s early strategy to build grass-roots support across the country. Today, they have a tiered structure representing the village, parish, sub-county, OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 8   taxation. Further study revealed that numerous different actors, from locally elected authorities to the police and even sometimes the military, could call on these vigilantes for manpower. As volunteers, vigilantes often left for paid work; others were arrested or fell out of favour with the community and thus were banned from security activities. It became apparent that any issue related to violence and the state—such as who can use violence to enforce laws, what level of violence is acceptable, and what activities merit punishment—encompassed multiple contradictory dimensions. Such contradictions frequently figure in ethnographies of vigilantes, gangs, self- help groups, militias, and the like. Scholars emphasize the twilight or boundary nature of such groups (Lund 2006b) and the fluidity of individual roles in the group and group roles in the community (Bøås and Dunn 2017; Göpfert 2012; Shah 2008). These observations pushed me to ask new questions oriented around the ‘how’ and ‘why’ of governance. Instead of trying to explain away the uncer- tainty and ambiguity that characterized security and governance in my field site, I opted to make it the focus of my research. Why is the relationship between violence and governance ambiguous in northern Uganda? What does that tell us about how the Ugandan state governs ordinary citizens?My endeavour became an inductive theory-building project on governance in authoritarian regimes. Questions of ‘how’ and ‘why’ inevitably raise the question of intent—are these processes purposeful or merely incidental? Whether or not arbitrary governance is intentional is extremely difficult to ascertain. To do so decisively would require intimate knowledge of the thinking of authoritarian rulers and their associates, not just at the present moment, but throughout their rule. Such an inquiry is outside the scope of this book.⁹ Still, the preponderance of the evidence marshalled here suggests two conclusions. First, arbitrary governance shapes the behaviour of ordinary citizens, and is thus critical to the functioning of these regimes regardless of the ruler’s intentions. Second, as elaborated in Chapter 2 and Chapter 8, arbitrary governance is a shared characteristic of several authoritarian regimes municipality, and district. They comprise a council of ten representatives, one of whom is the chairman. Today, the LC chairman functions as a kind of village-level leader. A full explanation of the local council system and its historic ties to the NRM state is provided in Chapter 3. ⁹ Answering the question of intent via a ‘smoking gun’might be achievable through other methods, such as elite ethnography. In the case of Uganda, such a study would be challenging in the current political environment. However, even if possible, such approaches face methodological challenges, as what respondents recount can be shaped by current interests and concerns, and in politically charged environments can sometimes only be accessed by making allegiances with particular interlocutors, which may bias findings (Malejacq and Mukhopadhyay 2016). Instead, this study collected hundreds of in-depth interviews, not as records of the truth, but to identify experiential patterns in the views of respondents (Fernández-Kelly 2015, 14). It is thus a view from below, augmented with approximately a dozen elite interviews with former members of Museveni’s government to triangulate data and help interpret the validity of findings. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi , ,   9 that manipulate the relationship between arbitrary power and the rule of law, despite their different political and historical trajectories. The findings suggest institutional intent. It is not that individual authorities design their own actions to be unpredictable, but rather that the structure of the system produces uncertainty among different authorities. Unpredictability derives from uncertainty about which authority will be relevant and what rules she (or more likely he) will apply. One might then ask if arbitrary governance is the result of historically contingent factors that the state accepts and leverages, versus a mode of governance that it actively produces. I chart a middle course between these two interpretations. Arbitrary governance is indeed tied to historical factors, such as the post-colonial nature of the state, limited resources, and the global political aid economy since the Cold War. At the same time, the regime has regularly made choices that reinforce rather than counter unpredictability and arbitrariness—for example, pursuing institutional fragmentation, raising citizen militias, and relying on the security services to police the domestic population with excessive and unaccountable violence. Intent therefore is not a grand strat- egy, but rather is reflected in the structural design and maintenance of the system. These details are treated in Chapter 3 and Chapter 8, while questions of intent are taken up at the end of Chapter 2 and in Chapter 7. 3. Researching the Arbitrary State Uganda—once characterized as an emerging democracy and lauded for its seem- ingly democratic reforms addressing fiscal policy, decentralization, and gender quotas—is now widely recognized as a ‘hybrid’ state, in which democratic insti- tutions and practices are manipulated to further centralize authoritarian control (Tripp 2010). The hybrid state combines aspects of democracy and authoritar- ianism, with (almost) regular elections, (formal) separation of powers, and (mostly) free speech. Institutionalized arbitrariness emerges from a paradox that characterizes such states and that Uganda illustrates well: governing institutions are highly fragmented and relate to each other stochastically—even while citizens widely perceive the state as a coherent entity with significant regulatory control over violence. Conceptualizing ‘the state’ based on empirical evidence in Uganda only high- lights this paradox. The state is not monolithic; it consists of diverse actors, institutions, and practices, each with varying interests, that can act alternately in a public or private capacity. To respond to the definitional challenge associated with ‘the state’, some scholars focus on what the state does and how, rather than on a deductive notion of what it is (Bierschenk and de Sardan 2014; Mitchell 1991). My inquiry is similarly concerned with practices of governance, and specifically on the institutionalization of violence. At the same time, my research OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 10   revealed that rejecting the notion of state institutions in favour of practices of statehood would create its own challenges. In Uganda, citizens’ conceptualization of the state shapes how they understand themselves as subjects. Their imaginary of the state encompasses the opaque workings of central power, citizens’ belief in the existence and importance of these levers of power, and citizens’ expectations that something akin to governing is happening in the Statehouse. Moreover, Uganda’s ruling regime has worked hard to make government, state, and party synonymous. The image of ‘the state’ is fundamental to this story—the individuals, institutions, and ideas that comprise the state are typically closely tied to power. I therefore use the terms ‘ruling regime’ and ‘state’ in a limited sense to refer to those individuals who control Uganda’s formal and public state institutions as citizens imagine them. The majority of my research took place in and around Gulu Town in northern Uganda between 2014 and 2016. Gulu constituted a ‘most-likely’ case in which to observe unpredictable and arbitrary state intervention.¹⁰ Because residents of Gulu have been staunchly opposed to the NRM regime and the area is highly conflict-affected, it is likely that the state appears unreliable and unpredictably present, making it an ideal location to disentangle the mechanisms of institution- alized arbitrariness. Over 10 months of field research with Gulu as the primary site, I conducted 303 unstructured and semi-structured interviews with approxi- mately as many respondents. From these, I developed a framework to analyse arbitrary governance, and in 2018, I returned to Uganda to apply that framework over six weeks and 76 interviews in three additional locations in Uganda. Interviews were on average an hour in length, although they varied based on the topic of conversation and practical considerations, ranging from the respondent’s comfort level to inclement weather. For many interviews, I employed Ugandan researchers to help locate respondents and translate when necessary. I supplemented these interviews with informal conversations as well as observa- tions of security trainings, political events, daily life, and dispute resolution in municipal or village-level courts and by local mediation. Overall, my interview respondents represent a broad cross-section of society, weighted towards those involved in informal and local security provision—that is, poor young men. 3.1 A Framework and Typology for Studying Arbitrary Governance From my study of local security and the micro-dynamics of violence in Gulu, I developed a framework to analyse arbitrary governance. The framework, ¹⁰ ‘Most-likely’ or ‘least-likely’ cases are ‘cases that ought, or ought not, to invalidate or confirm theories, if any cases can be expected to do so’ (Eckstein 1992, 158). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi , ,   11 elaborated in Chapter 2, is comprised of four oppositions that characterize state capacity: (1) the use of lawful versus exceptional violence; (2) the state’s defined jurisdictional claim versus lack thereof; (3) state presence versus absence; and (4) state fragmentation versus consolidation. Existing scholarship often assumes that each of these oppositions has an internally stable or stabilizing relationship. Take the third opposition of state presence versus absence as an example: a given state might be present in some places and absent in others; or it might be partially present; or effectively absent (O’Donnell 1999). However, regimes that employ strategies of institutionalized arbitrariness manipulate these oppositions, unpre- dictably and repeatedly collapsing and reinstating them, such that the oppositions become unstable. In the same places and time periods, the state might alternately appear present or absent; partially present; and—less common but importantly— create doubt about the very meaning of any difference between state absence and presence.¹¹ Destabilizing these oppositions produces four ‘factors’ that together constitute institutionalized arbitrariness. Table 1.1 lists each opposition and the factor that is produced when the opposition is destabilized. Combined, the destabilization of these oppositions produces an environment of seemingly arbitrary intervention that makes the government ever present in citizens’ imaginations, despite its general material absence in terms of daily security provision or law enforcement. The resultant system promotes societal fragmentation and political demobilization. The destabilization of these four oppositions is reflected in the experience of Odera and his brother. The case shows how the regime at times allows the exercise of overwhelming violence outside the law (as when the police shot the vigilantes in the dead of night), and at other times reasserts the claim of lawful violence (as when the government official defined the police shooting as a lawful response to criminal activity). The capacity to arbitrarily deploy extralegal violence limits citizens’ abilities to reasonably ignore or manage interactions with the state. By Table 1.1 The oppositions and factors of institutionalized arbitrariness Destabilized opposition Factor in institutionalized arbitrariness 1 The use of lawful versus exceptional violence Citizens’ perceptions of the state’s capacity for overwhelming and unaccountable violence 2 The state’s defined jurisdictional claim versus lack thereof Citizens’ perceptions of a fluid state jurisdiction 3 State presence versus state absence Citizens’ perceptions of potential state presence 4 State fragmentation versus state consolidation Citizens’ experience of a coherent and consolidated state ¹¹ This may be difficult to imagine in the abstract, and is elaborated with numerous examples in Chapter 2, as well as Chapters 4–6. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 12   unpredictably claiming and denying jurisdictional authority, state authorities destabilize citizen claim making and undermine other non-state authorities. In this case, the vigilante group was unable to consolidate its position in the com- munity because the police’s violent intervention undercut the vigilantes’mandate. In the aftermath of his brother’s death, Odera waited for years for the police to respond to his case, confident that the higher authorities knew about his situation and would eventually respond. The perception that the state could be present at any time and is informed about local problems makes government non- intervention appear to be as much a choice as intervention. Much of the literature on state formation assumes that all states fundamentally seek to consolidate control over territory—and that claiming jurisdictional authority, increasing state presence, and regulating the use of violence are key to this process. In contrast, my analysis shows that it is actually by rendering jurisdictions, state presence, and uses of violence unstable and fluid that some modern authoritarian rulers are able to project power. 3.2 Alternative Explanations To probe the validity of the four-factor framework that emerged inductively from my research in Gulu, I returned to Uganda in 2018 to examine alternative explanations for arbitrary governance. I selected three additional locations that varied on key alternative explanations for citizens’ perceptions of an arbitrary state: conflict-affectedness and relationship to the ruling regime (see Table 1.2).¹² Conflict-affectedness could produce a perception of state arbitrariness through citizens’ memories of war and its violent unpredictability, as well as through the Table 1.2 Case selection to probe alternative explanations for the perception of arbitrary state governance Highly conflict-affected Less conflict-affected Pro-opposition Most-likely case: Gulu Mixed case: Soroti Pro-regime Mixed case: Moroto Least-likely case: Mbarara ¹² All four study locations are medium-sized towns with relatively developed infrastructure, includ- ing potable water, schools, hospitals, and security services such as the police. Each is also fairly homogenous in terms of ethnic identity, with only one case—Mbarara—where many respondents share the same ethnic identity as the president. In each location, I interviewed several respondents with non-majority ethnic identities and I also inquired specifically into the role that ethnicity plays in politics. Though it is clearly a foundational cleavage in Ugandan society and politics, ethnicity did not emerge as a salient factor in civilian perceptions of arbitrary governance. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi , ,   13 social, economic, and political devastation that violent conflict often causes. Similarly, citizens’ reported experiences of an arbitrary government could reflect anti-government sentiment and the challenges of embedding central state institu- tions in politically hostile locales. This would mean that arbitrariness is evidence of uneven implementation, not a governing strategy. Sub-national variation in conflict-affectedness and political support for the regime allowed me to probe whether arbitrary governance characterizes Uganda’s ruling regime or is simply a locally held perception induced by a history of violent conflict or anti-government sentiment. I found that though the implementation and effects of arbitrary governance varied depending on contextual factors, citizens experienced the state as arbitrary in all four places. Moreover, the same four factors were present, though they manifested in different ways. Exploring these variations produced a typology of institutionalized arbitrariness, summarized in Table 1.3 and detailed in Chapter 7. The typology helps explore how different legacies of conflict and varying levels of political support for the regime shape citizens’ perceptions of arbitrary state power. Each type of arbitrary governance causes citizens to self-police by destabilizing citizens’ expectations of local authorities and the central state. Together, the findings suggest that governance through unpredictability is not merely a result of happenstance, the state’s low capacity, or outside researchers’ difficulty in understanding a society that is unfamiliar or illegible to them. Rather, institution- alized arbitrariness constitutes an approach to authoritarianism today—one that is based more on fragmenting alternatives to state power than on exercising iron- fisted control. 4. Organization of the Book This book proceeds as follows: Chapter 2 elaborates institutionalized arbitrariness, drawing on a wide range of scholarly work to develop the framework and establish its external applicability. It highlights the contributions that institutionalized arbitrariness makes to the study of modern authoritarianism, as well as of state formation and consolidation. Chapter 3 traces the emergence of a system of arbitrary governance in Museveni’s Uganda and describes the historical and Table 1.3 Varieties of institutionalized arbitrariness Highly conflict-affected Less conflict-affected Pro-opposition Atomization Contested assemblage Pro-regime Violence at scale Discipline OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 14   institutional factors that make it a modern authoritarian regime. The chapter illustrates how Museveni’s regime is both personalist and institutionalized, and highlights key moments that have shaped citizens’ perceptions of Museveni and his regime. Chapters 4 to 6 individually explore three of the four factors of institutionalized arbitrariness and how each is produced: (1) the state’s capacity for overwhelming and unaccountable violence, produced by a fluid relationship between exceptional and lawful violence; (2) a fluid state jurisdiction, produced by state authorities unpredictably claiming and denying their authority; and (3) potential state pres- ence, produced by a fluid divide between state presence and absence. Together, the chapters underscore the fourth factor in institutionalized arbitrariness: a frag- mented institutional environment co-existing with the perception of a coherent and consolidated state. Each chapter highlights the presence of multiple and competing authorities as well as the seemingly contradictory, yet commonly held perception that the ruling regime maintains at least marginal control over state and society across the country. Together, these four factors produce institu- tionalized arbitrariness, which focuses more on fragmenting alternatives to state power than on fully consolidating authority and violence. Chapter 4 focuses on the Uganda Police Force as a lens to examine how the Ugandan state produces and sustains the perception among citizens that it has access to overwhelming violence. The chapter shows how the Ugandan state strategically links and delinks its governing institutions to overwhelming violence, making the deployment of violence unpredictable in both its intensity and accountability. The result is an ambiguous relationship between lawful and excep- tional violence that keeps citizens fearful of the state, on one hand, and marginally engaged, on the other. Chapter 5 examines local vigilantes and the micro- dynamics of violence, focusing on the nature of the ambiguous space between state and society. It shows that this space is hostile to the putative authority of non-state entities, in large part due to the state’s unpredictable jurisdictional claims. Chapter 6 addresses the question of state presence, tackling the puzzle of how a largely absent state can appear to citizens to be present. The chapter examines how Uganda’s ruling regime manipulates the relationship between its presence and absence to keep citizens in an ambiguous position as potential agents and subjects of the state. Chapter 7 uses sub-national variation to probe alternative explanations for arbitrary governance. Evidence from the cases shows that violent conflict and political leanings shape the ways in which institutionalized arbitrariness mani- fests, exaggerating certain components and attenuating others. Such differences result in ‘varieties’ of institutionalized arbitrariness that, taken together, bolster the book’s argument that it is a distinct type of authoritarian rule. The resultant typology of four varieties of institutionalized arbitrariness, each corresponding to a different study location, illustrates some of the different outcomes produced by OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi , ,   15 changing combinations of state violence, fluid state jurisdiction, unpredictable state presence, and institutional fragmentation. Lastly, the chapter uses these variations to examine some limitations of the theory, including questions about the regime’s intent and citizens’ agency. Chapter 8 returns to questions of external validity and the theoretical contri- butions of arbitrary governance. It probes the broader applicability of institution- alized arbitrariness by surveying three regimes that display similar characteristics to Uganda’s NRM: the Ethiopian People’s Revolutionary Democratic Front; the Rwandan Patriotic Front; and the Zimbabwe African National Union (Patriotic Front). Drawing on scholarship on these regimes, I find aspects of arbitrary governance in all three cases, though its precise manifestation is shaped by contextual factors, notably the strength and independence of state institutions. Next, I discuss the theoretical and empirical implications of arbitrary governance in today’s world of increasingly transnational movements and fragmented sub- national power. The book concludes with a reflection on avenues for future research, including how international aid and improvements in surveillance technologies shape arbitrary governance, and the resultant dynamics between state and society in this type of modern authoritarian regime. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 16   2 Arbitrary Governance and Modern Authoritarianism How do modern authoritarian rulers project power? Scholars have noted that today’s authoritarian rulers have added new tactics to their playbooks. In addition to coercion, patronage, and delegation, these rulers often adopt democratic insti- tutions and use rule of law-compliant reforms to maintain control (c.f. Lührmann and Lindberg 2019; Schedler 2006; Scheppele 2018). For example, rulers might make legal reforms to censor the media, as Recep Tayyip Erdoğan did in Turkey; to selectively allocate parliamentary seats in order to marginalize political oppos- ition, as Rafael Correa did in Ecuador; and to modify the electoral calendar in order to hamper political rivals, as Abdoulaye Wade did in Senegal (Bermeo 2016). In this book, I use the term ‘modern authoritarian’ to refer to regimes that use rule of law-compliant reforms to undermine checks on executive power.¹ I describe in more detail how this term relates to existing scholarship later in this chapter. Modern authoritarian regimes are characterized by a tension between authoritarian rule and democratic institutions. Though authoritarian rulers weaken democratic institutions, the continued presence of these institutions offers repeated opportunities for challenges from the political opposition. As a result, these regimes have been described as structurally unstable (Levitsky and Way 2002). Over the past three decades, political science scholarship on modern authori- tarian regimes has mushroomed (Ezrow 2018). However, it has rarely engaged with scholarship on post-colonial neopatrimonial states which grapple with many of the same phenomena. I engage these literatures with original field research to help explain how such seemingly unstable regimes are able to control their populace. Until now, scholars have generally attributed the success of modern authoritarian regimes to a careful balancing act between coercion and patronage on one hand and democratic institutions on the other (Geddes et al. 2018). In such systems, democratic institutions can allow rulers to assess and respond to public opinion (Gandhi 2008), to gain domestic or international legitimacy (Levitsky and ¹ Uganda’s ruling regime has made every effort to elide government, state, and party, and thus when referring to the Ugandan case, I use the terms ‘state’ and ‘ruling regime’ interchangeably to describe an apparatus that encompasses the government, the ruling party, and the state’s administrative institu- tions. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi Arbitrary States: Social Control and Modern Authoritarianism in Museveni’s Uganda. Rebecca Tapscott, Oxford University Press (2021). © Rebecca Tapscott. DOI: 10.1093/oso/9780198856474.003.0002 Way 2002), or to diffuse political opposition (Brownlee 2009). Scholarship on post-colonial and neopatrimonial states draws attention to the importance of indirect rule, where control relies on accommodating elites who bring with them the support of their constituents (Bayart 1993; Mamdani 1996; Van de Walle 2007). I offer a third explanation. Instead of delegating authority, modern authoritar- ian regimes can stabilize control and project power directly into the lives of ordinary citizens through unpredictable assertions of authority that undermine the autonomy of those who would otherwise challenge it. As described in Chapter 1, this type of governance, which I call institutionalized arbitrariness, rests on an institutional arrangement that fosters competing low-level security and governance actors who, together, create a governing environment characterized by unpredictability for ordinary citizens and local authorities. This unpredictabil- ity pervades the public space, fragmenting civic organization and weakening alternatives to state authority. Unpredictability is a motif of Museveni’s Uganda. In my research, Ugandans broadly described the state as unpredictable and volatile, with effects they char- acterized as disorienting and depoliticizing. A common refrain among respond- ents across the country was that when it comes to the government, ‘things have to be confusing’. Some argued that this was because the government sought to create a confused population that could not identify shared goals to act on politically. Other scholars of Uganda have noted similar phenomena, describing this state as using ‘arbitrariness and unpredictability’ to restrict media workers (Tripp 2004, 12); as creating ‘seemingly deliberate confusion’ around civil militias (Janmyr 2014, 212); as ‘produc[ing] “security” and “insecurity” simultaneously in a con- stant aporetic relationship’, and as fostering ‘ambiguity or double-faced meaning of things’ (Verma 2012, 57).² The Ugandan state has even been compared to ‘the dry season rains—something occasional and potentially destructive’ (B. Jones 2009, 3). Others have noted the president’s ‘tendency to keep things around him as disorganized as possible to avoid the formation of any ordered arrangement that might possibly be turned against his own, personal raw power’ (Carbone 2008, 29). Despite the emphasis these scholars place on fragmentation, arbitrary and unpredictable state power, and resultant uncertainty, they stop short of examining these as components of a broader system of governance. ² Cecilie Lanken Verma approaches uncertainty in northern Uganda through the emic notion of lakite, or ‘somehow’. She writes, ‘Lakite was a notion I only gradually came to notice, due to its modest translation into the adverb “somehow” in English, but, as it turned out, when used in Acholi in the form of an adjective, it carried a much profounder meaning as the expression of uncertainty, even to the extent of the extreme. Lakite indicated everything considered “tricky”, often dangerous, and which would cease understanding. It was related to secrecy in the way that it was seen to contain something hidden, something not to be grasped, an ambiguity or double-faced meaning of things, expressed as the way in which the government or certain people would “show different faces all the time” or hide their true intentions “behind the face” ’ (2012, 10–11). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 18   Analogous phenomena can be seen in authoritarian settings globally. The Chinese government, for example, has blocked online political organization not through censorship but by flooding social media platforms with random and seemingly innocuous messaging to create so much noise that no single message can gain traction (King et al. 2017). In Lebanon, Nora Stel has studied how the state manages refugee populations by keeping their status informal and limiting or ignoring authoritative knowledge about them—their numbers, encampment situ- ation, and their representative structures—such that they cannot claim political rights (Stel 2020). And in American prisons, officials have been shown to inten- tionally disorient inmates to make them more manageable, for example by enforcing changing and often contradictory rules to make prisoners feel like they are in the twilight zone (Doolittle 2017). Each of these strategies of unpredictable governance are contextually specific, relying on particular institutional and elite power arrangements. However, they share the principle of governing not by crushing opposition outright, but instead by destabilizing, fragmenting, and diluting it. This makes civic spaces fragile and splinters collective action such that, to ordinary citizens, the ruling regime appears to be the most coherent and strongest governing entity. Political unpredictability thus allows such regimes to project power directly to the grass roots, causing local authorities and ordinary citizens to self-police. Institutionalized arbitrariness is distinct from other forms of authoritarianism in two key ways. First, it uses uncertainty to produce a plausible-enough percep- tion that the regime has a stable hold on power, especially compared to other actors. Institutionalized arbitrariness makes other actors look weak or irrelevant. As a result, citizens see the ruling regime as the only viable option and view collective action as impractical or even futile. The regime can thereby outsource many governance and security responsibilities while limiting principal–agent problems associated with indirect rule. Second, like modern modes of authoritar- ianism that hollow out state institutions, or strategies like ‘coup proofing’ that seek to prevent coups by fragmenting potential loci of power, arbitrary governance explains how regimes sustain an incongruence between state institutions and the organization of violence. However, arbitrary states are distinct because they do not seek, a priori, to weaken state institutions. Instead, they can tolerate functional— and even relatively strong—state institutions by multiplying them and creating confusion among them. This chapter develops the theoretical foundations of institutionalized arbitrari- ness as an explanation for how modern authoritarian regimes project power. It first offers a brief summary of the expectations set out in research on state formation and consolidation in order to identify what remains to be explained in modern authoritarian regimes. It then turns to scholarship on modern auth- oritarianism and neopatrimonialism. These fields of study, rarely put in conver- sation with one another, offer complementary views about how and why state OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi      19 institutions are decoupled from enforcement power. Together, they provide new insight into the projection of arbitrary power in authoritarian states. The chapter then details how institutionalized arbitrariness contributes to these literatures. It elaborates a four-part framework to identify and analyse the functioning of arbitrary governance and gives examples of the four factors in a variety of contexts, in Africa and beyond. The chapter concludes with an examination and rebuttal of three alternative explanations for observed unpredictability in the relationship between citizens and the state: corruption (arbitrary governance is really just the product of the cumulative acts of self-interested individuals); illegibility (arbitrary governance is really just a hidden order); and happenstance (arbitrary governance is not intentional, and therefore not a mode of governance). 1. Arbitrary Power and State Formation Today’s dictators and authoritarians are far more sophisticated, savvy, and nimble than they once were . . .Modern authoritarians have suc- cessfully honed new techniques, methods, and formulas for preserv- ing power, refashioning dictatorship for the modern age. (Dobson 2012, 4–5) Today’s authoritarian rulers have adapted to survive democracy’s advance. The resultant modern authoritarian regimes raise questions about what we thought we knew about state formation. These regimes do not follow a teleological process in which the struggle for control and resource extraction produces state institutions that regulate and restrain arbitrary power in pursuit of efficiency, as detailed by scholars like Charles Tilly and later adopted by new institutionalists, like Douglass North, John Wallis, and Barry Weingast. Instead, they simultaneously exhibit the institutional forms of modern democracies and an executive that wields arbitrary power with seemingly few de facto constraints. As noted by Nic Cheeseman and Jonathan Fisher: Legally, authoritarian states usually subscribe to the separation of powers: the preeminence of the rule of law, and respect for freedom of expression and organization. In practice, though, they are characterized by over-mighty presi- dents who maintain excessive control over all branches of government, enforce the arbitrary suspension or uneven application of laws, and implement unpre- dictable crackdowns on perceived opponents—sometimes within the confines of the law, sometimes outside it. (Cheeseman and Fisher 2019, xxv) This section elaborates how modern forms of authoritarianism contradict expect- ations described in widely accepted theories of state formation and consolidation. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 20   It then addresses two stylized strands of research, each of which could be under- stood as critiquing the prevailing views on state formation: first, recent political science scholarship on modern authoritarian regimes, and second, a long- standing literature on the post-colonial neopatrimonial state, derived mainly from studies of African states.³ 1.1 State Building and Consolidation: The Bias toward Identifying Political Order State-building theories are broadly oriented around identifying political order. FromMaxWeber’s treatises on the rational bureaucratic state to the seminal work by Charles Tilly and James Scott, much political science scholarship on the modern state focuses on how rulers strive to organize and institutionalize violence to maximize efficient resource extraction. These theories focus on formal govern- ing institutions as technologies through which arbitrary power is contained, managed, and deployed in the modern state. Such theories therefore understand the rational bureaucratic state, in which governing institutions have a monopoly or near monopoly on the use of force, as a by-product of a synergistic relationship between ruler and subjects. Analytically, this literature assumes the existence of institutional teleologies: from fragile to stable, personal to impersonal, and unpredictable to predictable (Tilly 1992). To the extent that it has a normative strain, this research sees institutionalization as a good that should be pursued in policy interventions (North et al. 2009). Such state formation theories have also informed studies of politically fragile and non-democratic states, which are presumed to have encoun- tered obstacles to this ordering process despite the apparent efforts of political elites. Scholarship in this area examines barriers to state formation posed by the international order and juridical sovereignty (Jackson and Rosberg 1982); by international legal sovereignty and a derivative domestic authority (Englebert 2009); by post-colonial border drawing (Herbst 2014); by the enduring political and economic legacies of colonialism (Acemoglu et al. 2001; Acemoglu and Robinson 2012; Mamdani 1996; Nunn 2008; Young 1998); and by the historic and contemporary international economic order (Amsden 2003; Wade 2004), among other factors. Studies of informal or non-state governing entities similarly emphasize political ordering processes—like those described in key theories of state formation—but they focus on processes that take place outside the state governing apparatus. ³ Much of the scholarship on neopatrimonialism focuses on African states; however, scholars have also applied neopatrimonialism beyond Africa to countries worldwide (for example, see Bach and Gazibo 2013). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi      21 These studies conceive of such non-state authorities and institutions in much the same way as do those that focus on state institutions as formal ‘rules of the game’ that regulate behaviour. Both presume that iterative processes of domination and submission by rational actors will eventually produce a de facto contract between authorities and their constituents (Bratton 2007; North 1990; Raeymaekers et al. 2008). A recent literature on rebel governance has extended these same lessons— applying the logic of political order making to the supposed disorder of rebel group operations—and in this way likening their approaches to those of statesmen (Arjona 2016; Arjona et al. 2015; Mampilly 2011; Mukhopadhyay 2014; Péclard and Mechoulan 2015).⁴ These and other studies fundamentally focus on political ordering and barriers to it, understanding unpredictability, contingency, and political disorder as remainders or noise that can be minimized by identifying the correct explanatory model.⁵ Instead, this book focuses on unpredictability and disorder as distinct approaches to governance that require further explanation. 1.2 Modern Authoritarianism: Contending with State Institutions Scholars have critiqued both the analytic and normative stances found in litera- ture on state formation and consolidation. In addition to elaborating cases in which highly arbitrary, violent, or repressive systems have been institutionalized (Arjona et al. 2015), research has identified cases in which institutions have been decoupled from the deployment of arbitrary power—whether to mask such power or to facilitate it. As Larry Diamond notes, modern dictators have innovated in order to operate in the ambiguous space between authoritarianism and democracy by manipulating the division of powers, extolling democracy and human rights, and fostering civil liberties, even while restricting political organization and centralizing power such that their states ‘have the form of electoral democracy but fail to meet the substantive test, or do so only ambiguously’ (Diamond 2015, 166). Today’s authoritarian rulers find power in cultivating an unsettled and dynamic relationship among institutional ‘types’ which allows for sometimes grafting rules from one institutional repertoire to another and, at other times, for preserving purely bureaucratic or patrimonial institutional environments. For example, ⁴ Earlier versions of such an argument examined how criminal organizations, such as the mafia or gangs, use similar strategies of ordering and organizing the populace under their control to maximize efficient and sustainable resource extraction (Bardhan 1997; Blok 1975; Venkatesh 2008). ⁵ In his study of the econometric modelling of conflict, Christopher Cramer notes that ‘uncertainty’ functions as a kind of ‘adhesive’ holding these models together. ‘Imprecision and inconsistency in the application of this variable . . . suggests that it is something of a residual used to patch up the holes in a model and stop it from collapsing’ (Cramer 2002, 1848). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 22   leaders from Hungary and Russia to Ecuador and Venezuela have centralized executive control by making legal reforms that appear legitimate because they were copied from liberal democracies (Scheppele 2018). However, by selectively adopting laws from diverse sources, such leaders cobble together an illiberal bricolage that effectively undermines checks and balances on arbitrary executive action (Krygier 2016). Such moves appear to be liberalizing, while having the effect of further concentrating power in the executive. For example, Viktor Orbán in Hungary and Erdoğan in Turkey adopted measures to expand the jurisdiction of their courts, effectively flooding the courts with minor cases and rendering them unable to function (Scheppele 2018, 551). Another widespread strategy, justified in democratic terms as bringing government closer to the people, imple- ments policies of decentralization in order to fragment the opposition and recen- tralize power under the ruling regime, as has been documented across Africa, in countries including Ethiopia, Kenya, Nigeria, Zimbabwe, Malawi, Senegal, and South Africa (Green 2008; Fessha and Kirkby 2008; Lewis 2014). Other scholars have studied how authoritarian rulers shape the electoral landscape—for example by using elections as a ‘safety valve for regulating societal discontent and confining the opposition’ (Brownlee 2009, 519) or by fostering the formation of tame political parties to channel political opposition and stabilize the regime (see March 2009 on Putin’s Russia). These, and similar phenomena, have been diversely described as hybrid regimes (Diamond 2015), competitive authoritarian regimes (Levitsky and Way 2010), and electoral authoritarian regimes (Schedler 2013).⁶ The broad concern is to understand how states can simultaneously exhibit the characteristics of developed democracies (such as multiparty elections) and authoritarian regimes (such as restrictions on public expression), and to disentangle how these seemingly con- flicting features interact to produce ‘in-between’ states (Van de Walle 2002). In such regimes, ‘elections and other “democratic” institutions are largely facades, yet they may provide some space for political opposition, independent media, and social organizations that do not seriously criticize or challenge the regime’ (Diamond 2015, 169). In such cases, democratic institutions are used to mask, or even legitimate, the reality of authoritarian domination. ‘In-between’ regimes make the political playing field itself uneven, so that the same democratic rules create unexpected and unfair results (Levitsky and Way 2002). As Aili Mari Tripp notes, these regimes ‘embody divergent impulses: they promote civil rights and political liberties, and yet they unpredictably curtail those same rights and liberties’ so that such states are neither truly democratic nor authoritarian (Tripp 2010, 1). Arbitrary power and political unpredictability are characteristic of such non-democracies. ⁶ See also review essays by Matthijs Bogaards (2009) and Yonatan Morse (2012). OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi      23 Scholarship on hybrid governance and modern authoritarian regimes suggests that institutionalization is not normatively good in its own right. Instead, scholars turn our attention from formal state institutions to power and the real politics of its institutionalization. Much of this scholarship can be understood as describing situations in which arbitrary power and state institutions are paired and decoupled in new and unexpected ways that allow such regimes to appear to have the liberal institutions of a democratic state, while enabling the weakly fettered exercise of executive power. At the same time, scholars often focus on how unique combinations of state institutions and arbitrary power allow rulers to claim ever more control—a phenomenon I call the productive work of arbitrary power. Institutionalized arbitrariness offers two contributions to scholarship on mod- ern authoritarianism. First, it contributes to the literature on hybrid regimes by offering a dynamic model that sees the repeated pairing and decoupling of violence to state institutions as unpredictable, stochastic, and ongoing, and exam- ines the governance effects of this process. It explores how repeated processes of alternately channelling arbitrary power through state institutions (such that it both enforces the rule of law and is subject to it) and decoupling it from them (such that it does neither) enables regimes to project power over local authorities and ordinary citizens. Such an approach understands political disordering as a discrete political phenomenon on a par with ordering—one that shapes the actions of authorities and the experience of citizens in such regimes. Second, along with assertions of arbitrary power, it shows that it is necessary to examine instances in which such rulers use new institutional arrangements to deny or withdraw their authority—what I call the negative work of modern authoritarian- ism. Simultaneously asserting and denying authority by combining productive and negative work dissolves background conditions for producing political meaning and creates political subjects governed by uncertainty about their position vis-à-vis the state. 1.3 Post-Colonial and Neopatrimonial States: A Hidden Political Order While recent scholarship on authoritarianism—much of it based on cases from the global North—grapples with how to understand arbitrary power without assuming its necessary institutionalization, a second strand of literature on post- colonial and African states has been doing so for decades. This second strand of literature seeks to explain how and why formal state institutions often appear decoupled from arbitrary power. This literature emphasizes an empirical and historically conditioned disjuncture between state institutions and the organiza- tion of violence. Within the resulting gap, it argues that rulers exercise arbitrary OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 24   power that is constrained by the rules, norms, and expectations that arise from interpersonal social relations, which are identified as the real site of politics. Because of their personal nature, social relations can be difficult to uncover and may remain hidden from the view of outsiders. As a result, scholars have noted that these systems can be mistaken as chaotic or ungoverned, when in reality they rely on unique and context-specific governance arrangements among formal and informal actors and institutions. This literature therefore asks scholars to focus on potentially hidden background relations, institutions, and organizations to find political order (Chabal and Daloz 1999). In these studies, institutionalization itself cannot be judged as normatively good or bad—rather it is the content of these background social relations that must be evaluated on a case-by-case basis. In this second strand of literature, scholars use terms such as ‘hybridity’ (Tripp 2010) and ‘multiplicity’ (Goodfellow and Lindemann 2013) to describe ‘boundary’ (Lund 2006) or ‘shadow’ (Reno 1995) institutions that characterize non-western arrangements among formal and informal governance institutions. Such scholar- ship captures a series of complex and messy relationships between unstable and unsustainable conceptual binaries, such as public and private organizations; formal institutions and informal associations; and state and non-state actors. It variously describes ways in which a state’s organization of violence does not match its governing institutions. These theories see patrimonial (informal) institutions as holding the real levers of power (Chabal and Daloz 1999), or see patrimonial and bureaucratic (formal) institutions as sharing them (Bach and Gazibo 2013; Cheeseman 2018a; Tripp 2010). Violence might be fragmented and dispersed, for example in ‘security assemblages’ (Abrahamsen and Williams 2010), or per- sonalized, for instance behind powerful warlords (Reno 1998; 2011). In these explanations, state institutions are just one of many places where politics play out—and often one of relative insignificance. Understanding state form as largely determined by colonial institutions, much of the literature on neopatrimonial African states sees decolonization as a moment of contradiction. This literature explains that colonialism destabilized and in some instances obliterated indigenous authorities, while decolonialization drove a wedge between governing institutions and governing authority. Newly independ- ent African nations gained control over colonial governance institutions but lost the corresponding organization of violence needed to translate them into effective rule. Pierre Englebert summarizes the observation: ‘Leaders of historically non- legitimate, incongruent, or mismatched states are indeed faced with a peculiar challenge and are limited in the options available to them to address it . . . From colonialism, they inherited the instruments of statehood but not the power that came with it in colonial days’ (Englebert 2000, 11). As a result, some have proposed that the state’s institutional form required an underlying organization of violence (usually a Weberian monopoly on the legitimate use of force) that simply was not present in post-colonial African states. According to some, these OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi      25 states became ‘vacuous and ineffectual . . . an empty shell’ (Chabal and Daloz 1999, 14). Weak state institutions fostered strongman rule or hybrid governance arrangements in which individuals coupled formal state institutions with private armies, militias, or gangs to extend control and maximize access to resources and rents. Some have categorized this as indirect rule by the metropole, where organized violence remains in the hands of an assemblage of powerful corporate interests, international institutions, and former colonial powers (Ferguson 2005; Mamdani 1996; Mbembe 2001). Others argue that state forms are more or less surface shell games alongside evolving forms of local violence which are the real sites of governance (Lombard 2013; Mukhopadhyay 2014; Reno 1998). For both sets of scholars, the ‘mismatch’ between state-like forms of governing institutions and privatized forms of violence is the fundamental pathology of post-colonial African states, making them vulnerable to fragility and failure. This mismatch between governing forms and governing violence has also been used to explain fragmented, hybrid, multiple, or complex institutional environ- ments. Such an explanation suggests that in the immediate post-colonial moment, elites continued to draw on and manipulate pre-existing ideas about authority and jurisdiction because they represented agreed-upon modes of exchange, in which private exchange was based on reciprocity and public exchange was based on bureaucracy. Scholars have argued that a combination of path-dependency, the advantage of incumbency, and resultant widespread opportunities for elite pre- dation have rendered the post-colonial state impotent (Bayart 1993; Jackson and Rosberg 1982). This tradition has been taken by some to suggest that such states are characterized by weak or even dysfunctional formal political institutions that are entangled with informal and personalized institutions (critiqued by Cheeseman 2018a). Even scholarship that nominally focuses on political disorder tends to find and reify some hidden order that is then used to explain the operations of power. In their work on the African state and the ‘political instrumentalization of disorder’, Patrick Chabal and Jean-Pascal Daloz describe a background political order based on socio-cultural relationships that—once revealed—makes sense of the seeming disorder of formal state institutions. They write: ‘[T]he disorder of which we speak is in fact a different “order”, the outcome of different rationalities and causalities. It appears as disorder only because most paradigms are based on a notion of a form of social, economic, and, therefore, political development which reflects the experience of Western societies’ (Chabal and Daloz 1999, 155). Chabal and Daloz note the possibility of the very argument developed in this book—that some elites benefit from the production and exploitation of what they call ‘political disorder’. However, they abandon the notion of disorder tout court in favour of understand- ing African governance in terms of a hidden order explained by the particularities of African culture. OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi 26   In contrast, this book places an African case in conversation with existing literature on modern authoritarianism in order to develop a comparative frame for contemporary analyses of arbitrary power that avoids the othering of African politics. Unlike the ‘hidden order’ described by Chabal and Daloz, institutional- ized arbitrariness recognizes that a state’s approach to governance can be unpre- dictable, and therefore indecipherable, for its citizens as well as for outsiders. Arbitrary governance is not limited to African states and does not rely on uniquely African political and social traditions. Internal illegibility is exactly what makes institutionalized arbitrariness an effective approach to projecting arbitrary power for any regime. Attributing political unpredictability to a hidden order assumes a stable and discernible relationship between patrimonial and bureau- cratic institutions. This assumption overlooks evidence that rulers alternately and unpredictably reify and destabilize the imagined divisions among institu- tional orders. Ignoring such phenomena leads us to neglect a key tool of modern authoritarianism. 2. A New Category of Modern Authoritarianism: Institutionalized Arbitrariness Rather than imagining governance as complex, networked, and navigable only by insiders, institutionalized arbitrariness describes a state that is illegible, unpredict- able, and capricious for those inside as well as outside the system. Citizens— subject to the state’s constantly changing jurisdictional claims, the incalculable risk of overwhelming state violence, and the possibility but never certainty of state presence—come to view the state as coherent and authoritative, even while experiencing it as woefully fragmented and low capacity. The result is that state authority is ever present in citizens’ imaginations despite the state’s limited provision of traditionally recognized governance and security services, such as access to courts, presence of police, and collection of taxes. Even as low-level non-state governing authorities carry out certain functions locally—for example, resolving disputes, organizing roadworks, andmanaging com- munity vigilante activities—they lack sufficient political space to consolidate control over their own jurisdictions and constituencies. They thus continue to function as fragile and weakly institutionalized authorities, able neither to become autonomous from the state, nor to make claims on state authorities or the ruling regime. The unpredictable and dangerous Leviathan found in this type of modern authoritarian regime creates pervasive uncertainty, precluding relationships of accountability between citizens and public authorities—whether state or non-state. When citizens cannot discern any pattern or regularity in how and when state actors claim or deny authority, and cannot understand what rubric they will use to OUP CORRECTED AUTOPAGE PROOFS – FINAL, 17/5/2021, SPi      27 do so, they cannot establish shared expectations for state behaviour. Instead, citizens are left to navigate a system of authorities whose roles and responsibilities are constantly shifting and are thus difficult to predict. These dynamics have been described in diverse contexts. For example, in Guinea Bissau, Henrik Vigh describes how prolonged political turmoil forces citizens to replace their habits with flexible routines that can be constantly re-evaluated in light of the changing socio-political landscape, wherein players reconfigure their affiliations all the time (Vigh 2006). In Pakistan, Laurent Gayer describes how residents of Karachi see the city as ‘undecipherable’ and a ‘foreign land’ because of growing uncertainty around the threat of violence. He recounts how this unpredictability produces a general sense of malaise. ‘As every group in the city denies involvement in violent or illegal activities, the identity of the perpetrators and their rationales are any- body’s guess. This opacity sustains a form of “epistemological uncertainty”—a doubt about what one knows about one’s social relations and environment—that obfuscates norms and relationships’ (Gayer 2014, 248). The absence of shared expectations means that when a state authority uses his or her power to redefine responsibility, there is no meaningful ground upon which citizens can demand accountability. Thus, the state’s unpredictable and harsh interventions constitute an approach to governance that allows non-state entities to function at a low level, both in terms of authority and capacity, while the state retains its position as the most powerful actor. Institutionalized arbitrariness diverges from indirect rule managed by local warlords. As described by William Reno in western Africa (1998) or Dipali Mukhopadhyay (2014) in Afghanistan, such indirect rule entails sub-national authorities consolidating resources and violence to govern a discrete geographic space. This is not possible when mandates are continually redefined or reconfig- ured. Institutionalized arbitrariness also differs from theories of neopatrimonial- ism, according to which citizens have personal relations with individual big men and unpredictability comes from the temperament and interests of this patron (Chabal and Daloz 1999; Erdmann and Engel 2007). Instead, in the cases exam- ined here, unpredictability and arbitrary power have been institutionalized into the state, dissolving the efficacy of background social relations and undermining even hidden logics of governance. In this way, institutionalized arbitrariness destabilizes local authorities in relation to the central state. Such an environment fosters the projection of state power even while the state regularly outsources key governance tasks, including security and policing. While institutionalized arbi- trariness helps explain a phenomenon similar to that of Scheppele’s ‘legalistic autocrats’, who use rule of law to centralize authoritarian control, it goes a step further, explaining the mechanisms by which arbitrary power is projected. This book counters a view of political ordering as the predominant governing strategy and stabilization as the prevailing logic of state form. It thus illustrates how political unpredictability and instability work together to produce a new state form—one that favours modern authorita