Decentralization in Africa: The Paradox of State Strength 9781626373266

Citation preview

Decentralization in AFRICA

Decentralization in AFRICA The Paradox of State Strength edited by

J. Tyler Dickovick James S. Wunsch

b o u l d e r l o n d o n

Published in the United States of America in 2014 by Lynne Rienner Publishers, Inc. 1800 30th Street, Boulder, Colorado 80301 www.rienner.com and in the United Kingdom by Lynne Rienner Publishers, Inc. 3 Henrietta Street, Covent Garden, London WC2E 8LU © 2014 by Lynne Rienner Publishers, Inc. All rights reserved

Library of Congress Cataloging-in-Publication Data A Cataloging-in-Publication record for this book is available from the Library of Congress. ISBN: 978-1-62637-053-1 (hc. : alk. paper) British Cataloguing in Publication Data A Cataloguing in Publication record for this book is available from the British Library.

Printed and bound in the United States of America The paper used in this publication meets the requirements of the American National Standard for Permanence of Paper for Printed Library Materials Z39.48-1992. 5

4

3

2

1

Contents

vii

Acknowledgments

1 Decentralization: Theoretical, Conceptual, and Analytical Issues James S. Wunsch

1

2 Botswana: Political and Economic Obstacles to Decentralization Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

23

3 Burkina Faso: Limited Decentralization Under Tight Oversight Pierre Englebert and Nestorine Sangaré

45

4 Ethiopia: Ethnic Federalism and Centripetal Forces J. Tyler Dickovick and Tegegne Gebre-Egziabher

69

5 Ghana: Decentralization in a Two-Party Democracy Joseph Ayee and J. Tyler Dickovick

91

6 Mali: Incentives and Challenges for Decentralization Susanna D. Wing and Brehima Kassibo

113

7 Mozambique: Decentralization in a Centralist Setting Beatrice Reaud and Bernhard Weimer

137

v

vi

Contents

8 Nigeria: Issues of Capacity and Accountability in Decentralization Dele Olowu and James S. Wunsch

159

9 South Africa: Decentralization and the Apartheid Legacy Louis A. Picard and Thomas Mogale

183

10 Tanzania: Devolution Under Centralized Governance Per Tidemand and Nazar Sola

205

11 Uganda: Decentralization Reforms, Reversals, and an Uncertain Future Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

229

12 African Decentralization in Comparative Perspective J. Tyler Dickovick and Rachel Beatty Riedl

249

List of Acronyms References The Contributors Index About the Book

277 281 301 303 317

Acknowledgments

The first acknowledgment we wish to make is an expression of gratitude to one of our authors. Along with Susanna Wing, we express our appreciation and thanks for the life and work of Brehima Kassibo, who passed away before publication of this book was complete. Brehima was a wonderful scholar and colleague who contributed much to our understanding of Mali, and he will be missed. In a multiauthor volume, there are many people whose support is indispensable, regardless of whether they are listed as authors. We are grateful for those who contributed their energy, time, and ideas to this book. The standard disclaimer applies, that any errors of fact or interpretation are our responsibility and that of the authors, not of those acknowledged below. The initial research for the project was supported by the Africa Bureau and the Office of Democratic Governance (now Democracy, Rights, and Governance) at the United States Agency for International Development (USAID). We are thankful that Ryan McCannell, at the Africa Bureau, originally envisioned a project to ask comparative questions about what policymakers need to know regarding decentralization experiences in Africa. He followed up on his vision with guidance and support impressive in both quantity and quality. Ryan and Jeremy Meadows, his colleague at the Africa Bureau, offered insightful suggestions, critiques, and corrections throughout the project. We convey our gratitude to Ed Connerley, who, during his time at USAID and after, engaged with the authors and us in long-running conversations about the significance and meaning of decentralization. With his extensive knowledge, his wealth of experience and insights, and his analytical mind, Ed contributed substantially to this project as well as to many others over the years. We express gratitude to everyone who assisted us at

vii

viii

Acknowledgments

USAID, including Kellie Burk, and to Kathy Stermer and Chelsea Keyser at ARD for their support and organization of much of the research behind this project. We thank Jesse Ribot and Jonathan Rodden for extensive and thoughtful commentary on a draft of the final comparative chapter. We also thank Jesse Ribot, Larry Schroeder, and Brian Wampler, as well as contributors Louis Picard, Paul Smoke, and Per Tidemand, for their input on the initial research questions, their critiques of drafts, or both. We would also like to thank those who have helped the book through the publication process. We are grateful to Lynne Rienner for her willingness to consider the manuscript and to work with us to improve it, to Sandy Thatcher for his support of the project at the early stages, and to Karen Williams for skillfully shepherding the book to fruition. We thank as well the anonymous reviewers, whose insights and critiques helped improve the work. All of the contributors thank their respective families and their home institutions for support throughout this process. Tyler Dickovick acknowledges the support that the Lenfest Sabbatical Award and Lenfest Grants at Washington and Lee University provided in facilitating the transformation of the project from its initial stages to the finished book. —J. Tyler Dickovick and James S. Wunsch

Decentralization in AFRICA

1 Decentralization: Theoretical, Conceptual, and Analytical Issues James S. Wunsch

Decentralization is one of the leading political reforms that developing countries have undertaken in the two decades since the end of the Cold War, and it has arguably been undertaken (at least in name) in more countries in Africa than anywhere else in the world. In the past twenty years, the majority of African central governments have initiated or deepened the transfer of power and resources to a range of subnational government (SNG) bodies, including regions, states, provinces, districts, and municipalities. This book uses a structured comparative approach to bring together and examine a wide range of decentralization processes across Africa; we focus on ten countries, examining each of them according to a common conceptual framework and preestablished set of criteria. Our structure enables us to shape a coherent debate on the impact of decentralization while exploring the topic’s empirical breadth on a continent where the term itself has taken many forms and had many meanings. We conclude that the legal, fiscal, and administrative authority of SNGs has increased in all ten country cases, though country experiences have varied. While the authority of SNGs has increased, there are fewer achievements for them in other dimensions of governance, including autonomy, accountability, and capacity. Through comparative examination of these characteristics of governance that matter to both scholars and practitioners, we believe this volume makes a unique contribution to academic and public policy discussions on modern-day decentralization. We ask two key questions throughout the book. First, what have been the characteristics of decentralization in Africa? Specifically, how has it been done and how has it varied from place to place? This raises a question of how we conceptualize and classify decentralization in Africa as countries have undertaken rather different processes under this rubric. To make sense 1

2

Decentralization: Theoretical, Conceptual, and Analytical Issues

of the empirical variation and to render comparisons useful in an area of such conceptual complexity, all of our chapter authors adopted a common framework that highlights how decentralization in each country fared by measuring subnational governance on four dimensions: authority, autonomy, accountability, and capacity (see USAID 2009 for an original elaboration of this framework). We believe this conceptual and empirical treatment is inclusive yet parsimonious, and can serve as an underpinning for the future of debates on African decentralization by integrating varied country experiences. Our second question goes beyond empirical detail to implications for politics and public policy: what have been the consequences of decentralization processes in Africa? Looking at individual cases and comparatively across countries, we find that results have been mixed, but that findings can be illuminated according to our four criteria. There have been considerable achievements in the decentralization of legal authority to SNGs and subnational administrative units, but less positive impacts in other areas. Subnational autonomy remains quite restricted by several top-down forces, especially the control exercised by state authorities and by the dominant political parties in many African countries. Similarly, there is some downward accountability of SNGs to local citizens through increasingly institutionalized subnational elections, but accountability in many cases continues to flow mostly upward to actors at the center. Finally, most of the country studies note that capacity remains a challenge, but this is true at all levels of government and not just at the subnational level. Taking these findings together, we argue that decentralization in many countries in Africa exhibits a paradox: it can actually strengthen or entrench national-level actors, even as it changes little in governance at the local level. There are a number of reasons that central governments might reconcile themselves to decentralization, or even embrace it. Decentralization appears to have helped stabilize and maintain several African regimes where the hold on power may have been precarious in the absence of decentralization. In other cases, it may well be helping to entrench certain governing parties in power, especially in dominant party systems. Still other cases have witnessed central states and governing parties that have found ways to delimit decentralization when it risked an actual dispersion of power. In short, decentralization has strengthened central authority as much as it has weakened it. The chapters in this book show how such dynamics have played out (with a number of variations) in ten countries. The structure of the comparison means that the country chapters “speak to one another,” a feature reinforced by the early establishment of common themes that enable both empirical breadth and theoretical contributions. Since the groundbreaking work by Diana Conyers (1983) and Dennis A. Rondinelli, John R. Nellis, and G. Shabbir Cheema (1984), decentraliza-

James S. Wunsch

3

tion has been understood to include three forms or strategies: deconcentration, delegation, and devolution. Each one moves governmental prerogatives and initiative from the center to the periphery, with the last of them including local accountability, usually through elections (Conyers 1983). This volume considers all three strategies in its coverage of decentralization in Africa although, as it shows, there seems to be a good deal more deconcentration and delegation than devolution, both in law and in fact. While many studies have emphasized only devolution—sometimes called “democratic decentralization”—in their analyses, we have chosen this broader understanding (Olowu and Wunsch 2004). We believe a narrower definition limited to devolution would be an error for a truly comparative study of Africa—conceptually, theoretically, empirically, and operationally. Conceptually all three forms fit the standard definition of decentralization, which involves the transfer of authority, powers, resources, and responsibilities from central to lower-level actors. Thus, while devolution may be a prerequisite for democratic local governance, it is not the only form of decentralization. Regarding the outcomes of decentralization in theoretical terms, it is possible for delegation and deconcentration to provide many of the beneficial outcomes that devolution is intended to provide, such as more local responsiveness and improved service delivery, particularly if issues of providing effective accountability can be resolved. These issues remain challenging under devolution as well, leading it at times to failure in achieving these very goals. For example, in many cases devolution plus local elections has led to elite capture, high levels of corruption, continued poor administrative performance, exclusion of citizens, and so on (Crook and Manor 1998). In these circumstances, it is unclear how democratic decentralization advances democracy any better than delegation or deconcentration, notwithstanding the existence of subnational elections. Also, when one considers the goals of economic development and stability, the theoretical case for devolution is again unclear, given that deconcentration and delegation may offer greater personnel and institutional capacity, opportunities for coordination, economies of scale, and consistent enforcement of the rule of law (Connerley, Eaton, and Smoke 2010). Of course, as research by scholars such as Goran Hyden (2006) and Pierre Englebert (2000, 2009) suggests, including Chapter 3 by Pierre Englebert and Nestorine Sangaré in this volume, much depends on the specific nature and efficacy of the national regime relative to SNGs. To be sure, however, neither does this book presume that devolution— when it occurs—results in necessarily worse outcomes. This is an empirical question that the cases are intended to explore. These questions help illustrate how complex and uncertain theories of decentralization are at this point. In part to help disaggregate this theoretical complexity, the volume places considerable emphasis on determining the extent to which the four key dimen-

4

Decentralization: Theoretical, Conceptual, and Analytical Issues

sions of governance—authority, autonomy, accountability, and capacity— have actually moved downward to SNGs and with what consequences. Our reasons for including all three forms of decentralization in this study are also empirical. In virtually all of the ten cases in this volume, African governments have undertaken mixes of devolution, delegation, and deconcentration. In making policy and implementing decisions concerning decentralization, for example, even those African decisionmakers placing a high priority on democratic outcomes are commonly constrained by consideration of the likely effects of decentralization for national integration and stability (Eaton and Connerley 2010). Thus, their efforts generally also include deconcentration and delegation. While possible exceptions might be seen among the federal states that have more extensive forms of devolution, even in these cases the central state (or party) continues to weigh heavily on subnational governance, as seen in all three of our federal state cases, Nigeria, Ethiopia, and South Africa. The outcomes of decentralization are strongly influenced by the mix of these strategies. In looking at the ten countries, we can see this mix of strategies. All have subnational elections, some degree of formula-based transfers, and legal frameworks that transfer responsibilities in major social service areas to SNGs, which are elements of devolution. However, the ten countries also have provisions that allow substantial ongoing central governmental administrative and political oversight, which are elements of deconcentration and delegation. That is, they all reflect significant political, fiscal, and administrative aspects of decentralization and these are expressed in operational mixes of devolution, deconcentration, and delegation. Thus, in including all three forms of decentralization, this book matches the empirical facts of decentralization in Africa. Operationally, all regimes seek and need varying forms of legitimacy in the eyes of their citizens and external actors. This requires continued institutional innovation to try to improve performance, and it may generate incentives for the pursuit of improved economic and social outcomes through decentralized governance. Because this is an additive process, the creation or empowerment in recent years of elected SNGs—devolutionary democracy—usually occurs in the context of previously or contemporaneously deconcentrated or delegated units of government. Thus, regimes utilize a variety of mechanisms to delegate powers or responsibilities of the central government to SNGs and nonstate actors. These patterns can be seen in all of our cases. The assumption of the equal importance of devolution, delegation, and deconcentration as forms of decentralization thus facilitates empirical description and analysis of each form as it is encountered in African decentralization. The case studies in this book describe the multistructural realities of African decentralized governance, and comment on the interactive

James S. Wunsch

5

effects of and tensions among deconcentration, delegation, and devolution. The objective of the book is to present a balanced discussion of decentralization’s three dominant forms and an accurate portrayal of their mixed results among its cases.

Theorizing Decentralization: Issues of Principal-Agency and Collective Action One cannot understand the strength and duration of Africa’s decentralization efforts apart from the economic and governance crises that the continent has gone through over the past thirty to forty years. The economic decline of the 1970s that ended with the crash of several economies and the wrenching structural adjustment era in the 1980s, along with the crises of governance where single-party and highly centralized regimes lost legitimacy and at times also collapsed, led African civil society and leaders alike, as well as donors, to search for new governance strategies (Ndegwa 2002). Even in recent years as international investment flows returned to Africa, particularly from China, decentralization continues to be seen, and at times pushed, by donors and developing country governments as a way to pursue a variety of goals. Prominent among these goals are democratization, social and economic development, and stabilization of new or existing governing regimes (Connerley, Eaton, and Smoke 2010; USAID 2009). We recognize that such goals are ideals, and are at times more rhetorical than they are faithfully pursued in practice by policymakers or donors. As noted in the concluding chapter, the third goal—regime stability—appears to be the most consistent, notable impact of decentralization in Africa. Much academic theory links decentralization to these goals (Wunsch and Olowu 1990; Grindle 2007; Ndegwa 2002). It has been linked to such pressing issues as reducing the costs of central government, mobilizing more resources from the grassroots, and, in particular, resolving some of Africa’s chronic governance problems. The governance problems revolve around chronic corruption, rent seeking, inability to deliver services to the grassroots, poor maintenance of public infrastructure, poor implementation of public programs, and lack of initiative at local levels. One key reason that decentralization has been seen as means to resolve some of these problems is its potential to reduce problems of principal-agency. Principal-agent problems are caused primarily by ineffective accountability mechanisms and structures (see Hiskey 2010). They short-circuit the relationship in democracies between the citizens as principals and the government officials as their agents, which is necessary to keep the latter working in the interests of the former. These kinds of problems are a primary

6

Decentralization: Theoretical, Conceptual, and Analytical Issues

cause of moral hazard, which occurs when government officials have opportunities to secure rents or otherwise use their office for private gain. When uncontrolled, this leads to poor and irresponsible policymaking and implementation as well as vast opportunities for corruption. Administrative hierarchies which are top down, overly centralized, with multiple levels, and/or which lack balancing downward and lateral accountability, leave principals unable to control their agents. This is because these structural factors frequently lead to information leakage and effectively to weak administrative sanctions. Nominal agents then can use their power and authority simply to pursue their own interests (Tullock 1965). This is accentuated when relevant information and managerial and supervisory capacity are limited by organizational weakness and when the public— either as individuals or as members of civil society organizations (CSOs)— has little or no control over its officials. This is typical of politics in Africa, even when there are active party systems, and it leads to abuses at all levels of government, particularly at the grassroots as noted by Englebert (2009). The logic of decentralization as a remedy for this is that enhancing lateral, local, and at times grassroots control over local officials, along with strengthening key elements of upward accountability, will reduce these problems (Grindle 2007). The dilemma here is in ensuring that local political and administrative leaders will not also fall victim to moral hazard and become irresponsible agents (Englebert 2009), which would simply reestablish the same governance problem at the local level. One way of addressing this problem is to mobilize the grassroots and subnational civil society, assuming that they can assert control over local administrators and that the new heads of local administrative hierarchies will respond to these accountability mechanisms. This also assumes that local officials will have sufficient information and rewards/sanctions to be effective managers and supervisors (Hiskey 2010). The challenge for accomplishing this is in assuring that competent and honest supervisors can be put in place, that effective management and information mechanisms can be developed and instituted, and that local public constituencies can be mobilized and empowered. Further, the latter must not become fragmented by their heterogeneity or divided by side payments, but instead must hold local officials accountable. This is in part a collective action problem and the rarity with which it has been surmounted suggests that it is a significant one, even at the local level where smaller numbers of people to organize and fewer issues to deal with would seem to make it easier (Olson 1965; Wunsch 2008). Collective action problems are thus a serious impediment for effective subnational governance in Africa. While some issues and areas have seen conspicuous success in overcoming these problems such as public schools in the Ugandan city of Jinja, villages in rural Uganda and Mali, the Sungusungu

James S. Wunsch

7

movement in rural Tanzania, community associations in southern and eastern Nigeria, education initiatives in rural Chad, development initiatives in rural Senegal, community regulation of scarce water in northern Nigeria, and emerging governance systems in Somaliland, most African states are generally characterized by passive local governments and fragmented communities (Wunsch 2013; Dauda 2004; Wunsch and Ottemoeller 2004; Coulibaly, Dickovick, and Thomson 2010; Heald 2005; Barkan, McNulty, and Ayeni 1991; Smock 1971; Fass and Desloovere 2004; Galvan 2007; Fiki and Lee 2004; Shivakumar 2003). A parallel result of this, and the ineffectiveness of many centrally established formal government organizations, is that collective goods are generally underprovided at local levels (Wunsch 2008, 2009). This speaks directly to the potential of decentralization to enhance development and stability. Roads, public education, public health measures, trash pickup, security, and the like are woefully short in the rural areas and even much of urban Africa. Public choice theory suggests that these problems are partly explained by the costs of getting diverse people who are dispersed across wide areas to agree on the issues and agree to work together in addition to the issue of resolving problems of free riding on public goods (Olson 1965). Decentralization, as a means of enhancing the institutional capacity and resources of local units of government as well as reducing their area and population, is in theory a mechanism to overcome these problems and encourage collective action. It would accomplish this by its abilities to reduce decisional and organizational costs through smaller and less heterogeneous populations; to utilize social capital, social pressure, and existing social infrastructure to discourage free riding; and to reduce the operational costs of large organizations by enhancing the capacity of smaller, subnational governance institutions (Wunsch 2008, 2013). To understand the theoretical potential of—and obstacles to—decentralization, it should be recalled that effective political institutions can be understood to be public goods themselves, entities whose actions benefit all in an area whether those residents help pay the initial costs in establishing them or the sustained costs involved in supporting them over time (Ostrom 1990). Given an understanding of political institutions as public goods and the problems of mobilizing the resources to pay for them, the relative rarity of effective local governments in Africa can be more easily understood. Elinor Ostrom (1990) shows how new local institutions frequently grow from existing local institutions and shared understandings among members of communities. However, the large and heterogeneous areas of most local governments and the decades of top-down and authoritarian rule—both colonial and postcolonial—may have suppressed the ability of the existing social and political institutional infrastructure to develop into new institutional arrangements able to resolve contemporary public goods and com-

8

Decentralization: Theoretical, Conceptual, and Analytical Issues

mon property problems, including the development of new institutions. As a result, formal local institutions able to facilitate collective action, as well as manage principal-agency, are weak. Reports of effective governance in Somaliland emerging out of the historical collective choice institutions of the Somali demonstrate the potential of this process, at least in some circumstances such as the absence of juridical sovereignty and other opportunities for rent collection by political elites (Jackson and Rosberg 1986; Englebert 2009; Shivakumar 2003). However, these enabling circumstances often do not exist and long-since compromised historic governance institutions frequently work poorly (Lund 2003, 2007). Researchers who have found historic institutions important in effective local governance emphasize their growth through a syncretic process where principles and understandings embedded in the institutions are used by contemporary actors to solve current governance problems, rather than by specific, historic actors or organizations that have reemerged (Wunsch 2000, 2013; Shivakumar 2003; Galvan 2007; Heald 2005; Fass and Desloovere 2004). Insofar as decentralization policies and programs help reduce the costs of establishing and sustaining effective local governance institutions, one might reasonably expect them to strengthen and eventually lead to greater provision of local public goods. However, unless communities can be persuaded to believe that these institutions will be accountable to them, it is unlikely they will be willing to pay the costs—in time, resources, cooperation, and trust—necessary to sustain decentralized political institutions. This can easily become a self-fulfilling prophecy and, as a result, local political institutions are likely to be captured by smaller (privileged) groups of persons who can easily organize to reap private goods (rents) from decentralized governments (Bueno de Mesquita et al. 2001). Chapters 6 and 8 on Mali and Nigeria, respectively, are examples in this volume of the reappearance of rent-seeking behavior at formal decentralized units. When this happens, there is no reason to expect that such institutions will work for the benefit of anyone but such small groups, and the public goods provision problem will remain as well as issues of principal-agency (Bardhan 2002; Bierschenk and Olivier de Sardan 2003; Lund 2007). The difficulty that SNGs have in raising local taxes, as seen in all the cases in this volume, may reflect this problem. Much of the same logic applies to the management of common property, be it natural resources and ecosystems or local government budgets, vehicles, buildings, and so forth. Whether or not Africa’s decentralization reforms can resolve these problems is one of the most important questions relating to the likelihood of these countries to make progress on such broader governance issues—a core topic of this volume. The successes of organic institutions, such as those in Mali’s villages described in Chapter 6 and as Dan Ottemoeller and James Wunsch (2004) found in villages in Uganda, suggest that decentralization has this potential

James S. Wunsch

9

but will require continued innovative design to accomplish it. When organizations such as these begin to solve these problems, decentralization’s potential to occupy empty political space will begin to contribute most effectively to development and democracy. The rarity with which this is seen among the cases presented in this volume may have to do with the overly large size of many decentralized units, the continued limits on local autonomy, and the upwardly focused accountability found in most of these cases. If this is true, placing more focus on enhancing the capacity for communities to engage in collective choice and less on their technical and administrative capacity might be a fruitful strategy to pursue. Establishing local collective choice institutions—here, decentralized governance arrangements that have grassroots confidence in them—is anything but simple. This cannot be done with the wave of a hand, or the passage of a law. Political elites must first find it in their interest to establish them. Then, common understandings among key local actors about rules and their meanings must be reached, credible commitments to follow them must be made, credible sanctions to buttress local decisions must be agreed on, role occupants must be selected and supervised, workable boundaries as to who is included in the decisionmaking system must be agreed on, and a track record of success must be achieved before any institutions can be expected to develop broad local support and trust, produce collective goods, manage common property, or resolve issues of principal-agency (Ostrom 1990, 2005; Treisman 2007). These challenges apply alike to devolution, deconcentration, and delegation.

Disaggregating Decentralization In reality each form of decentralization, while conceptually distinct and an important way of defining and setting a strategy, is a mix of several critical dimensions of governance. In part to pry apart, assess, and analyze the diverse and complex mixes of decentralization found in real-world cases, the authors of these studies have not tried to force them into categories which, while useful at a general level, might better be seen as Weberian heuristic constructs (deconcentration, delegation, and devolution) than as summary labels that can describe a state’s strategy. Standard usage of the three forms of decentralization reflects an expectation, in the above order, of generally increasing levels of authority, autonomy, capacity, and downward accountability by SNGs. However, these three general terms mask the subtleties of the decentralization process as found in the empirical world. As noted, none of the ten cases can be accurately described by any one of these terms alone. Thus, the book assesses the extent to which four key dimensions of governance—authority, autonomy, accountability, and

10

Decentralization: Theoretical, Conceptual, and Analytical Issues

capacity—are found in each case among diverse elected SNGs and subnational administrative units. This framework was first developed by the United States Agency for International Development (USAID) in its evaluations of decentralization in the developing world, and it serves as a useful organizing principle for this volume where it is reflected in the structure of each of the country chapters (see USAID 2009). According to standard usage, in this book we mean the following by these four dimensions of governance: Authority is the legal right to undertake actions in the name of government. It can be understood as prohibiting, authorizing or facilitating actions by officials or citizens. In rational-legal orders, it must be allocated to officials or citizens through such mechanisms as constitutional or statutory provisions; court cases and administrative actions; or, as authorized by these measures, by elections. Important dimensions of authority for decentralization include the number and range of activities authorized as well as their scope of independence from oversight. Autonomy refers to the extent to which one level of government—here, SNGs—may undertake actions without the permission of other levels of government. There are of course national, legal provisions that must be observed by subnational levels of government, but significant levels of autonomy exist when subnational levels of government are allowed to initiate actions subject to post hoc oversight by administrative and judicial institutions rather than being required to obtain permission before action. Autonomy also is determined by SNGs’ fiscal ability to undertake actions on their own. The latter can be compromised by a lack of authority to raise revenue on their own, a lack of local resources to tax, and detailed prior control by superior governments over how SNGs are allowed to spend their fiscal resources. Accountability is the ability of actors to reward or sanction in a meaningful way officials or institutions that initiate actions. It can be understood as flowing in three directions: upward to administrative, judicial, or political institutions and officials; downward to the public or civil society; and laterally (or horizontally) to other units of government and, in some cases, to professional bodies with standard-setting and regulative functions. In rational-legal systems it is also regulated by law, which sets forth the actions that can be sanctioned, the circumstance when sanctions can be applied, the procedures that must be followed, the sanctions that may be applied and who may apply them (Ostrom 2005). Capacity refers to the presence of resources and skills adequate for governmental units and personnel to be able effectively to undertake the functions that have been assigned to them. For decentralized governments, capacities would also be important in regard to those of superior levels of government; professional bodies; other decentralized entities; civil society;

James S. Wunsch

11

individual citizens where they share authority; and, for the latter two, in their roles as mechanisms of accountability. Authority, autonomy, accountability, and capacity are thus dimensions of governance important to understanding the extent and nature of decentralization. In real-world decisionmaking about decentralization, achieving a theoretical maximum level of any one—or all four—of these characteristics is not possible. This is partly because decentralization is crafted over time as a complex series of incremental decisions made by diverse actors, dealing with these various dimensions, and is usually made within severe resource constraints. It involves numerous trade-offs among valued options as well as high degrees of uncertainty about future outcomes. The relationship among the four dimensions is a product of an evolutionary process. According to the research reported in this book, African decentralization efforts are universally mixtures of varying intentions, strategies, and outcomes.

The Dynamics of Decentralization Our reasoning with respect to the importance of these dimensions in describing and analyzing the development of effective subnational levels of governance is as follows: (1) Without some degree of authority, subnational units of governance of any sort, whether the goal be devolution, delegation, or deconcentration, cannot lawfully act. (2) When devolution or delegation are sought, without autonomy subnational units of government become mere extensions of central government bureaucracies and represent no change from existing systems. (3) Without accountability they become powers unto themselves and vulnerable to serious problems of moral hazard in which SNGs might act in ways where offices are used for private gain and disregard the governance roles of other levels of governance and nonstate actors. (4) Without capacity to act, subnational units of government are mere placeholders and, therefore, meaningless and irrelevant. Only when appropriate mixes of these qualities are present can we expect units of government to be effective and responsible vehicles for action. Various forms of decentralization initiatives emphasize different levels of these qualities. For example, deconcentration initiatives might emphasize subnational capacity and upward-oriented accountability, but limited subnational authority and autonomy. Similarly, delegation might emphasize greater levels of subnational authority, carefully defined autonomy, and enhanced subnational capacity, but only upwardly oriented accountability. These dimensions can be used as valuable analytical tools in moving toward a deeper understanding of the complexity and dynamics of decentralization for each of the ten case studies, both because they assist in prying apart the various tactical and strategic options state leaders have taken and

12

Decentralization: Theoretical, Conceptual, and Analytical Issues

because in reality these dimensions exist in tension. By breaking these dimensions out, the operational tensions are highlighted. For example, in pursuing devolutionary decentralization—a strategy arguably important at least by law in nearly all of the cases—it would appear valuable for SNG units to have significant levels of authority and autonomy. But in pursuing improved governance as an overall outcome of decentralization, these very qualities must be balanced against the need to prevent SNGs from causing negative externalities (e.g., drawing down or ruining another jurisdiction’s water shed or dumping waste in another jurisdiction) and to maximize positive ones (e.g., linking road networks, efficiently locating schools, reinforcing public health programs, and synchronizing market towns). They also must be balanced against the danger that jurisdictions thereby empowered might discriminate against political and ethnic minorities, women, and the poor; fail to maintain quality control and professional standards; act to erode national initiatives in key areas such as road networks, flood control, education, health, and agriculture programs; lead to waste and corruption; and so forth. In part, the remedy for these problems lies in enhancing accountability. However, accountability is also problematic in operation. For example, to whom should SNG officials be accountable, and regarding which functions? As noted above, accountability in decentralization can always be understood as potentially going in three directions: downward to citizens and civil society; upward to superior levels of government and courts of law; and laterally to other branches of government and to professional organizations regarding professional standards, qualifications, and practices. In some situations, lateral accountability also reaches to neighboring units of government regarding negative and positive spillovers and regional initiatives. While this seems complex enough, the reality is that any office—and many individual officials—are likely simultaneously accountable in two or all three directions, and to multiple actors in each. For example, executives in devolved governments are accountable to their publics via elections and regarding performance in office. They are also accountable to superior levels of government and courts for their technical performance, conformity to national policies, respect for and support of national prerogatives, and in general regarding the faithful execution of laws that apply to them such as those regarding hiring and firing personnel, procuring supplies, and awarding contracts. At times, of course, officials receive contradictory directions from the several actors at this level. Also, some wishes of the local population should not be met, such as ones that entail waste and corruption, or disregard the rights or interests of minorities and the politically weak and marginalized. Downward and lateral accountability also raise issues of the need for effective CSOs and local councils or legislative bodies, institutions that often have been weak at subnational as well as at national levels and are critical checks on malfeasance and misfea-

James S. Wunsch

13

sance by each level of governance. Similarly, professional and administrative personnel face equally complex patterns of multidirectional accountability, ones that may pull them in contradictory directions regarding any policy, program, or personnel issue. Finally, as Stephen Ndegwa (2002) and others note, effective upward accountability has generally operated poorly in Africa (Olowu and Wunsch 2004). Capacity would seem without controversy for all three general forms of decentralization, but it too is problematic, if only because resources are limited and needs are vast. Where should governments deploy those limited resources, and what strategies best enhance subnational capacity? Traditional training programs frequently have been ineffective. Also, the tendency of those with training and skills to leave outlying areas for the capital cities or other large urban areas means the sustained impact of training rural personal on local institutions has been limited. Furthermore, expanding subnational responsibilities generally requires stronger if different roles for national and regional governments, which requires their capacity to grow as well. Thus, the demand for resources is great, but which level should be emphasized first is unclear. Finally, what priorities should be set among competing local needs for enhanced capacity? How should choices be made amid enhancing skills among local legislators and elected executives in devolutionary decentralization as well as managers in all forms of decentralization; developing civil society to take on accountability roles; and strengthening offices of planning, audit, contracting, community development, agricultural outreach, public health, medical services, and so forth? A cross-cutting issue is that, even where there are trained personnel, SNGs often lack the resources to pay enough to attract them to poorer and more isolated areas, and national resources are too limited adequately to fill these local gaps. States that seek to establish multiple levels of SNGs to engage the grassroots, such as Ethiopia, Nigeria, and Uganda, particularly, face this problem. It is important for several reasons to emphasize the mixed outcomes of and trade-offs among the four dimensions of governance as they relate to decentralization (authority, autonomy, accountability, and capacity). First, the different mixes of devolution, delegation, and deconcentration found among our cases imply emphasizing these four aspects to different degrees. Second, there are clear trade-offs among them regarding the governance values that they maximize. In other words, none of these is without the risk of some adverse outcomes; neither can all four dimensions be maximized without introducing contradictory tensions into subnational governance. Third, for any single dimension, there are ambiguities, trade-offs, and negative consequences that policymakers must understand, and among which they must choose. Policymakers are forced to make these decisions even when the trade-offs are ambiguous and the outcomes unclear. This is part of why decentralization and its outcomes are so idiosyncratic to each state. In

14

Decentralization: Theoretical, Conceptual, and Analytical Issues

brief, an increase in each of these dimensions is not necessarily possible or, for some of them, even always good. African governments have made choices regarding their approaches to the tensions and challenges inherent in these four dimensions of decentralization as well as the sequence in which they were pursued. What were these choices? What were their consequences for development, stability, and, when relevant, local democracy? For example, Nigeria chose to enhance local capacity through extensive training and large grants of monies derived from its oil exports before developing effective structures of accountability. Other states such as Mali, Ethiopia, and Uganda pursued a sort of big bang strategy where they pushed all aspects rapidly and simultaneously. What were the consequences of these choices? For Nigeria, Chapter 8 suggests that this led to extensive corruption. For Ethiopia, Mali, and Uganda, Chapters 4, 6, and 11, respectively, suggest that it led to severe problems of SNG capacity. The consequences of choices regarding accountability, for example, have also been important. As noted in Chapter 7, Mozambique’s municipalities, which had a greater emphasis on downward accountability than districts, seemed to perform much better than the districts in delivering services. Describing, analyzing, and evaluating the choices made by African governments regarding these four dimensions of decentralization are major goals of this book. A number of the case studies in this book describe the constraints placed, often deliberately, by central governments on devolved levels of government by deconcentrated units at the same or higher levels. As shown in Chapter 9, South Africa reflects these tensions. This is also seen in francophone Africa such as the description in Chapter 3 of Burkina Faso’s conflicted struggle to decentralize. Burkina Faso is a remarkable example of stark differences between policy as reflected in sweeping national legislation and the functional realities of rural SNGs. Though national decentralization legislation was enacted in 1998 (under an acknowledged policy of gradualism), implementing decrees were enacted only in 2009. Nonetheless, almost all rural communes (key units of devolved governance) are still unable to deliver the services formally devolved in 1998 because they have not yet signed required operation protocols with the responsible governor, a deconcentrated official of the central government. Remarkably, communes in Burkina Faso are responsible to three such deconcentrated authorities: a regional governor, a provincial high commissioner, and a departmental prefect. To a lesser extent this is also true in Mali, as described in Chapter 6. Mozambique (see Chapter 7) also uses a rather complex mix of strategies emphasizing deconcentration among other forms of decentralization, which seems to have hampered development of effective SNGs. This use of deconcentration to constrain devolution reflects an important running theme

James S. Wunsch

15

in the book: African decentralization has resulted in robust changes to legal frameworks, but has been circumscribed by strong centripetal forces.

The Political Economy of Decentralization in Africa At the theoretical level, one can understand why choices among autonomy, authority, capacity, and accountability are challenging and frequently problematic. If sorting through these challenges and trade-offs to make rational choices among them were all that affected these dimensions, the situation would be difficult enough. However, in the real world of applied politics, theory gives way to the options and choices dictated by political considerations. Thus, another goal of this book is to understand why African leaders have made their choices in implementing decentralization. The importance of this challenge can be seen, for example, in the reality that every state reviewed in this book—some more than others—has made constitutional or legislative and policy commitments to devolutionary decentralization, as well as substantial efforts to achieve it, yet few have made much progress in reaching it. Why? While the answers at times hinge on balancing the sorts of tension referred to above, they are also intrinsically tied to Africa’s unique political economy. Overwhelming evidence regarding the African political economy points toward the state and access to it as the primary determinant of wealth for virtually all Africans (Englebert 2000; Hyden 2006; Wunsch and Olowu 1990; Reno 1999; Villalón and VonDoepp 2005; Bayart 1993). It occurs at the top levels of the state where access to rents associated with sovereignty in the international arena, such as export taxes, natural resource revenues, international aid, and access for investors and exporters to the domestic economy, provides great wealth. It occurs within state bureaucracies where control of contracts, jobs, promotions, and location of field offices and intermediate and local government headquarters are decided. It also occurs at field levels where local personnel eagerly seek pathways to urban postings, and even poorly paid personnel are able to extract rents from the grassroots through their control of access to licenses, official papers, permits, or local services. One way or another, the state, control over elements of it, and the opportunities it provides largely determine who prospers, who just gets by, and who lives in poverty. Even in post-2000 conditions, when investment flows returned to Africa, there is no evidence that the state-centric political economy has substantially changed, as Englebert (2009) shows in his study of state sovereignty and rent-seeking behavior at all levels of government. Decisions regarding decentralization are intrinsically involved in this political economy. With relatively weak private economies and civil societies—ones

16

Decentralization: Theoretical, Conceptual, and Analytical Issues

in any case usually dependent on access to or tolerance by state actors—the pathways that decentralization reforms take and the goals that they are able to support are largely driven by the interests and relative influence of diverse state-based and state-focused actors. Thus, as seen in this volume, the impetus for decentralization will grow from the interests of those at the center, rather than how effective it may be made as a tool for development or democracy. As our cases illustrate, if state leaders in Mali, postrevolutionary Ethiopia, or Nigeria find decentralization useful to deal with regionalism or to legitimize military rule, those at the center will support it. If leaders in postcoup Ghana, postrevolutionary Uganda, or transitional South Africa find it useful to consolidate and legitimize their rule, they too will support it. However, if leaders in postapartheid South Africa, postsettlement Mozambique, or contemporary Ethiopia and Botswana find it a likely base for opposing parties or regional and ethnic factions, they will limit it in practice if not in law. Similarly, subordinate actors in each state will support it insofar as it serves their interests. A major goal of this volume is to describe and explore those choices, the reasons for them, and their consequences. Beyond the impetus of the political center, the political-economic interests of various ministries that stand to gain or lose control over resources such as appointments, budgets, personnel, and substantive areas will lead them to block, slow, or try to direct it in their favor. Local-level officials will make similar decisions, as will members of the public, insofar as they can affect it. Of course, these are broad generalizations, and interests at any level in any state may express themselves in varying reactions to certain general forms of decentralization (delegation, deconcentration, devolution), and by support for specific institutional arrangements that affect authority, accountability, autonomy, and capacity. For example, local executives will usually support strong-executive models over those with strong local legislative bodies. Thus, they will likely encourage steps that enhance general local authority, autonomy, and capacity if they enhance executive authority, but not ones that encourage local legislative capacity or any forms of local popular accountability such as to civil society (Crook and Manor 1998; Wunsch and Ottemoeller 2004). This can be seen in nearly all of our cases. Ministries of planning may well support local authority and capacity to do decentralized planning, but only if they retain effective control over final capital investments. This is particularly the case if by doing so they can capture budgetary control from a ministry of finance or from sector ministries such as agriculture, education, health, and transportation. Thus, they will likely also oppose local autonomy in choosing projects and accountability by the local populace. We can see this in Uganda. Finally, as Englebert (2009) suggests, local officials who collect rents from the public will often be enthusiastic about greater authority and autonomy decentralized to them. However, they will not be

James S. Wunsch

17

enthusiastic about measures that enhance their accountability to anyone, particularly about those that enhance local government’s capacity to supervise and manage them. The circumstances and situations of political parties have also strongly affected decentralization in the ten cases. The highly competitive parties in many Nigerian states along with massive oil rents have caused decentralization to be intrinsically connected with intense local-level rent seeking as well as patronage. Nigerian local governments in turn were used to build bases to capture control over state governments and their oil accounts. Ghana also has highly competitive political parties, but its economic system historically has not been distorted by lucrative oil rents. There, each party has supported decentralization as a general policy because it provides the party with regional bases to build on after national electoral defeats. Alternatively in states where there are single dominant parties, such as in Burkina Faso, Ethiopia, Mozambique, South Africa, and Tanzania, dominant party influence can negate nominal legal local autonomy and accountability. National party structures have preempted local concerns in the nomination of candidates and have considered SNGs primarily as mechanisms for clientelism to reward local party cadres and to keep competitors weak. In different ways a similar dynamic plays out in Botswana and Uganda, which have clear leading parties in the case of the Botswana Democratic Party (BDP) and the National Resistance Movement (NRM), respectively. Mali, prior to the 2012 coup, had a more fragmented party system. Each state’s pathway toward decentralization is a result of these kinds of diverse political patterns and interests, and the effectiveness of actions taken in support of them. As a result, the progress of decentralization along the four dimensions varies substantially among our cases. Similarly, their success in achieving democracy, development, and stabilization also varies substantially. In substantial measure a major purpose of this book is to describe and explore the circumstances that nonetheless encourage effective progress along these dimensions toward broader goals, and to develop generalizations regarding when, in these state-focused political economies, they are most likely to occur.

Issues of Finance, Administration, and Political Structure Several recurring and tangible choices regarding specific policies and institutional arrangements are likely to deeply affect authority, autonomy, accountability, and capacity. These include fiscal issues such as financing SNGs, administrative ones such as the nature (or indeed existence) of a unified SNG civil service–personnel system, and structural choices among fed-

18

Decentralization: Theoretical, Conceptual, and Analytical Issues

eral and unitary governmental systems as well as the size, design, and number of tiers of local government. Existing political-economic factors certainly influence these choices but, once made, they have their own evolving and profound influence on the functioning of SNGs. This is because these choices allocate power and influence; generate new and evolving interests; and, thus, modify the incentive structures that local officials and citizens face, thereby influencing what their subsequent actions are likely to be (Ostrom 2005). The way that decentralized governments are funded deeply implicates the meaningfulness of any authority granted to them as well as their capacity, accountability, and autonomy. In particular when one considers delegation and devolutionary decentralization, large and unconditional central grants may enhance SNG capacity (though not necessarily). However, because these grants are cost-free flows of resources from outside, they also reduce the incentive for local officials to raise local resources, seek efficient operations through good management, develop effective policies and programs, and build effective institutions to perform those tasks. They erode the likelihood of developing longer-term and enduring capacity at subnational levels, and tend to encourage rent-seeking behavior by local officials and local interests that want a share of free money, which itself also reduces local capacity. Any sort of grant is also likely to erode downward accountability by reducing the incentives for local citizens to invest energy in holding local governments accountable over taxes that might otherwise have been raised locally and for local officials to listen to them. Conditional grants may reduce rent-seeking opportunities, but also reduce local autonomy, downward accountability, and local leaders’ abilities to respond flexibly to particular local needs. This will also likely reduce the incentives for local dwellers to pay much attention to local governments. These dynamics mean that the probability of decentralization leading to local development and possibly enhancing security is reduced. Nonetheless, the local poverty typical in most of rural Africa means that local governments will have few resources and little capacity without such grants. A key topic in this book is exploring how African states have dealt with these financing dilemmas and what the consequences of their choices have been. It is arguable, for example, that South Africa, Uganda, and Ghana’s systems of substantial grants have enhanced immediate local capacity while undermining aspects of the other three qualities: autonomy, effective local authority, and accountability to local citizens and CSOs. Nigeria’s massive grants appear to have set off intense rent seeking at local and state levels. The grants themselves and these rent-seeking behaviors have seriously undermined autonomy and accountability as well as damaged capacity because of how the rent seeking has affected Nigerian subnational personnel in general. Mozambique’s grants have stimulated local corruption and rent seeking as well as partisan patronage at local levels. Ethiopia’s grants have led to clientelism.

James S. Wunsch

19

One of the recurring administrative issues among the ten cases is human resources management (HRM). Tanzania, Botswana, Nigeria, Uganda, and South Africa, among others, have wrestled with these issues. These systems have been difficult to design without mixed and often problematic consequences for all forms of decentralization since they have important implications for capacity, accountability, and autonomy. Theoretically, Africa’s choices include a single national personnel system where employees of all levels of government work under the same terms of service and are assigned to subnational units by a central ministry. This can be seen in Burkina Faso. Mozambique also has a personnel system dominated by the national government, though several agencies are involved in its hiring process. A variation of this pattern is utilized in Uganda where subnational personnel are members of a national personnel system under a single set of terms of service, but are engaged, paid, evaluated, and discharged by SNGs at their discretion. HRM systems also include dual civil service–personnel systems, where national and subnational personnel are divided. One variation of the latter can be seen in South Africa, Botswana, and Ghana where most subnational personnel, though part of a separate SNG civil service organization, are engaged, discharged, paid, and assigned by a central agency. In another variation used by Nigeria, national authorities set terms of service, but local authorities engage and discharge their personnel. Finally, as in a system that was proposed for Ethiopia, there might be no national personnel system for subnational personnel who would be engaged, discharged, and paid at the discretion of subnational units of government. In reality, most systems are a bit of a hodgepodge of these models such as Mali where most local government civil servants are hired by the center or by centrally appointed local civil servants, but some are nonetheless paid by local governments. The latter frequently are paid after long delays, which means that engaging and retaining such personnel can be difficult. The advantages of a single national-government personnel system that serves localities are believed to be greater security and therefore satisfaction for personnel, more protection for personnel from politicized and arbitrary local actions, and more homogeneity in training and qualifications among personnel, and a more even distribution of personnel among poorer and wealthier areas (Wunsch 2001). Such a system might work to enhance local-level capacity. Alternatively, nationally established and revised terms of service, because of their vulnerability to pressure from organized civilservice personnel interests at the center, can become overly fiscally generous and a burden for subnational units. This in turn sustains and intensifies SNG’s dependence on central grants. This can lead to reduced local autonomy and eroded long-term capacity when fiscal burdens weigh heavily on SNGs, as can be seen in Nigeria. Similarly, national control of personnel assignment, transfer, and promotion can disrupt local authority over personnel and, thus, local plans and programs. Botswana and Mozambique have

20

Decentralization: Theoretical, Conceptual, and Analytical Issues

experienced such outcomes. Alternatively, greater local control over personnel, as seen in Uganda, seems to have led to more cohesive local governments, but also to eroding morale and greater discontent among local employees, possible loss of professionalism, and fewer protections for such personnel from arbitrary local decisions. Overall, then, personnel systems that have a stronger national role seem to enhance the fundamental personnel capacity of SNGs while reducing SNGs’ authority over employees and autonomy from national control. However, they can commit subnational units to greater personnel and fiscal burdens. This in turn can reduce funds available for supplies and capital investment to meet local priorities, thereby eroding the programmatic capacity of SNGs. Alternatively, personnel systems where subnational units have a greater role in engaging, discharging, and setting terms of service can enhance local authority over personnel who play key roles in defining and implementing their programs, thus strengthening local autonomy and, one might hope, the ability to respond to unique local needs and problems. However, more locally focused systems may also erode the capacity of these personnel and reduce the upward accountability of local political leaders and administrators with regard to their fair and professional treatment. Local or downwardly focused accountability systems and structures (e.g., elections and an involved and active local civil society) might be expected to incentivize local leaders to do a good job managing local personnel. However, the general underdevelopment of these structures and systems in Africa and the dynamics of the African political economy, which place greater power and resources in the hands of government officials, mean they are unlikely to do this well. Single national civil service systems, which are generally found in Africa with respect to senior professional staff such as medical, financial, and agricultural personnel, have many of the same trade-offs as national-local government systems, but with even less authority granted to subnational units, consequently eroding their autonomy and potential for downward accountability. In most cases in Africa, HRM has been characterized by instability and mixed arrangements, which themselves have been problematic for decentralization. A key structural issue has been the number of levels of subnational government. This has had important implications for each of the four dimensions of decentralization, and depends on decisions regarding federalism and the number of tiers for local government systems. Related to these are issues and decisions regarding the size, overall number, and territorial criteria for local government. South Africa, Nigeria, and Ethiopia are Africa’s only federal systems. Several countries, including Uganda, Mali, Ethiopia, and Ghana, have multitier systems. Mozambique is characterized by parallel devolved and deconcentrated systems of subnational governance. The number of local governments ranges from a high of nearly 800

James S. Wunsch

21

units in Nigeria and some 550 in Ethiopia to a low of 14 in Botswana. The number of subnational federal units similarly varies from thirty-seven in Nigeria to nine in South Africa, with eleven in Ethiopia. Finally, design can be based on relatively organic principles where communities with socially and economically integrated centers of balance are designed, as is generally found in Ghana or Uganda, to arbitrary ones such as in Nigeria where areas are frequently designed in ways that divide interdependent communities into separate local governments. The establishment and physical definition of SNGs has frequently become a political football, which has led to the loss of much of their demographic, economic, and social coherence; to units that are too small to be fiscally viable; and at times to civil violence, as in Nigeria. Issues surrounding the number and size of levels of subnational units of government include their cost; their functions; the distance between the grassroots and governmental units; and the extent to which multiplying levels of government leads to problems of ambiguity in responsibility, management and quality control, coordination of activity, and redundancy. Nigeria’s states, for example, were supposed to play major roles in supporting local governments in management and quality control of local services, but they have not done much of this (Olowu and Wunsch 2004). Instead, they have become arenas of rent-seeking behavior and serious drains on national revenue. Alternatively, Ethiopia’s regions can and do play a large role in supporting the country’s local units of government. At the lowest levels, the functionality of Uganda’s subdistrict levels has been unclear, and Mozambique’s complex parallel deconcentrated and devolved systems seem to have led to ambiguity regarding which system has the authority to do what as well as serious coordination problems. On the other hand, Mali’s bottom-up strategy for organizing villages into communes appears to have encouraged initiative and accountability at the lowest levels of government. The chapters of this book explore these patterns and their consequences.

Conclusion The complex African political economy must be navigated in ways that enable viable units of subnational governance to be established. The complex puzzles surrounding this task suggest that, even if there are valid reasons why Africans should pursue decentralization, it may take many years before these steps begin to make a difference. Nonetheless, there is evidence from studies of decentralization in Africa, as well as informal de facto initiatives that have grown alongside decentralization reforms, that decentralization can make a major difference in the lives of Africa’s peoples. In each of the ten states discussed in this book, decentralized gover-

22

Decentralization: Theoretical, Conceptual, and Analytical Issues

nance entities—some formal, some not—have made collective choices and effectively gathered the resources necessary to resolve issues of principalagency, provide collective or public goods, and manage common property. Some have done so through formal decentralized governance institutions and others have established their own decentralized governance entities without help or authority from central governments (Wunsch 2008, 2009, 2013). The most effective ones have utilized both of these and worked through institutionalized mixes of delegation, devolution, and deconcentration (Dauda 2004; Heald 2005). The conspicuous failures in local governance have not utilized such an institutional mix (Harnishfeger 2003; Baker 2002). These ten cases and related scholarly material illustrate the potential of decentralization to resolve chronic governance problems and how it has thereby facilitated local development, security, and democracy. The challenge for scholars and practitioners of decentralization and improved governance in Africa is to sort through the contemporary evidence to see what arrangements of authority, autonomy, accountability, and capacity work best to facilitate progress for Africa’s peoples in this process. The intent of this book is to help continue this work.

2 Botswana: Political and Economic Obstacles to Decentralization Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

Relative to other countries, Botswana is highly centralized fiscally and administratively (Schneider 2003). Typically, decentralized political systems have two or more tiers of elected local government with considerable discretionary authority over decisionmaking and revenue generation (Marks, Hooghe, and Schakel 2008; Schneider 2003). By contrast, Botswana’s single tier of elected local government has little formal authority. The political dominance of the Botswana Democratic Party (BDP) and the central government’s dominance of the economy have reinforced the formal powers of the central government. In this chapter, we document sources of centralization and evaluate how current political and economic developments are likely to affect the prospects for the balance among delegation, deconcentration, and democratic decentralization. We argue that the fallout of the global financial crisis encourages deconcentration to increase efficiency in service delivery while an increasingly uncertain political environment increases the political risks of greater devolution and makes it less likely. Our analysis is based on a review of documents plus research in Botswana between mid-July and early August 2010. Field research included interviews with officials in central government ministries and agencies, local-level politicians and staff, and representatives of nongovernmental and civil society organizations. We conducted interviews in five of ten districts (Central, Chobe, Kgatleng, Kweneng, and North West) and two of six urban centers (Gaborone and Francistown).

23

24

Botswana

The Institutional, Political, and Economic Foundations of Centralized Authority Botswana gained independence from Great Britain in 1966 as a unitary state with a parliamentary system. This constitutional form marked a break from the relatively low level of centralization before and during colonial rule. The British ruled the Bechuanaland Protectorate indirectly through a system of tribal reserves based on the Tswana merafe (singular: morafe, polities). The dikgosi (singular: kgosi, hereditary traditional leaders) had considerable discretion in the administration of their territories as long as they did not challenge the interests of the British (Tlou 1985). Thus, the system of colonial rule was fairly decentralized. Some dikgosi advocated a federal system based on the merafe, but they were outmaneuvered during constitutional negotiations in the 1960s (Mokopakgosi 2008). At independence, districts replaced the tribal reserves but, unlike the precolonial merafe or the colonial-era tribal reserves, the districts were granted little autonomy. Indeed, the postcolonial period has been characterized by a dramatic centralization of power (Molomo 2000). The BDP achieved electoral and legislative dominance in the self-government elections of 1965 and has won every subsequent national election; it has always controlled more than two-thirds of the legislature. Control over the economy reinforces the formal authority of the central government. Although Botswana numbered among the poorest countries in the world at independence, mineral developments in the 1970s and 1980s fueled rapid economic growth. Until the recent global financial crisis, diamonds represented a relatively reliable source of government revenues. The availability of mineral revenues reduced fiscal pressures to generate alternative revenue sources.1 Mineral revenues also allowed the public sector to expand; it rapidly became the most important employer. Despite significant private sector development, government remains the single most important economic actor. As of 2009, more than 30 percent of all formal sector jobs were in central government, with another 13.4 percent in either local government or parastatals (Matambo 2010: 6). In addition, government tenders represent an important source of business for many private companies. The Media Institute of Southern Africa (MISA) estimates that 80 percent of contracts and tenders are linked directly or indirectly to government (MISA 2010: 199). The dominant economic position of the government creates a powerful advantage for the ruling BDP while the weakness of the opposition limits resistance to centralization. Steven Levitsky and Lucan Way (2010) conclude that the conditions for political competition are sufficiently unfair as to justify characterization of Botswana as a competitive authoritarian regime. Neither political nor economic conditions, however, are static. Electoral support for the BDP has eroded since the late 1980s. In three of the

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

25

past four parliamentary elections, the BDP received less than 55 percent of the vote. In the 2009 parliamentary elections, 40.3 percent of constituencies either changed hands in the elections or could be considered competitive (Poteete 2012).2 Even with forty-five of the fifty-seven elected members of parliament (MPs), deep factional divisions within the BDP meant that the government could not count on a solid legislative majority. Less than a year after the elections, the BDP split. The breakaway faction formally registered as the Botswana Movement for Democracy (BMD) in May 2010 and became the Official Opposition two months later.3 Defections to the BMD overturned or threatened the BDP’s majority in some localities. The number of competitive constituencies and the fluidity in the party system makes it difficult to assess the relative strengths of the various parties. While there is no guarantee that these developments will translate into electoral gains for the opposition, the BDP can no longer take its dominance for granted. These cracks in the BDP’s political dominance appeared during a period of unusual economic instability. The global financial crisis that began in late 2008 led to a sharp contraction in luxury consumption, including diamonds. The crisis hit Botswana’s economy immediately. The economy shrank by 6.3 percent in the fourth quarter of 2008 and by an additional 20.5 percent in the first quarter of 2009 (Matambo 2010: 5). The government raised the value-added tax (VAT) from 10 percent to 12 percent in 2010 and has introduced a variety of other measures to increase revenues and lower costs. By reducing the resources available for government salaries, development projects, and political patronage, the financial crisis presents serious political challenges. Heightened financial pressure and political uncertainty affect interest in decentralization, the substance of proposed reforms, and their prospects for adoption and implementation. Advocates of deconcentration hope for cost savings and more efficient service delivery that might also generate political dividends. At the same time, political uncertainty raises the risks of devolving authority to local institutions by increasing the likelihood that they would challenge rather than build support for the ruling party (cf. O’Neill 2003).

Local Councils Currently, four organizations exercise authority at the district level in Botswana: local councils, land boards, district administration, and tribal administration. In this chapter, we focus on the councils, as established in the Local Government (District Councils) Act of 1965 and the Township Act of 1965. As elected multipurpose subnational governments, the councils have greater formal authority and democratic legitimacy than the other local

26

Botswana

authorities.4 There were nine district councils from independence until 2006, when the Chobe District Council was established. The government has designated four townships (Lobatse, Jwaneng, Sowa Town, Selebi Phikwe) and two cities (Gaborone and Francistown). The town and city councils do not represent a separate tier of local government. Instead, towns and cities have been excised from the jurisdiction of the district councils so that the three types of councils operate independently of each other at the same tier. The councils vary tremendously in territory and population, as shown in Table 2.1. At one extreme, the Central District Council is responsible for approximately a quarter of Botswana’s land area (142,093 square kilometers) and population (587,120 people in 2011). Among the district councils, the South East District Council has the smallest territory (1,780 square kilometers) and the Chobe District Council has the smallest population (23,449 residents in 2011). Natural resources and socioeconomic conditions also differ markedly across councils, giving rise to important disparities in the potential for local resource generation. The composition of the councils also varies considerably. The councils, like the National Assembly, consist of both directly elected and “specially elected” (i.e., nominated) members. The only exception is Sowa Town, where

Table 2.1 Key Characteristics of Local Councils in Botswana Council Rural councils Central Chobe Ghanzi Kgalagadi Kgatleng Kweneng North East North West South East Southern Urban councils Francistown Gaborone Jwaneng Lobatse Selebi Phikwe Sowa Total

Territory (km2)

Population (2011)

Councillors (2009)

142,093 20,800 117,910 105,200 7,960 31,100 5,120 109,130 1,780 26,470

587,120 23,449 43,370 50,500 92,247 304,674 59,829 158,104 92,843 198,262

174 9 24 26 27 82 23 47 24 64

79 169 100 42 50 159 570,162

100,079 227,333 18,063 29,032 49,724 3,599 2,038,228

22 35 9 14 16 7 603

Sources: Territory and population: Central Statistics Office (2011); number of councillors: Independent Electoral Commission (2009) and Ministry of Local Government (2009).

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

27

all members are specially elected. In 2009, the total number of councillors per council ranged from 7 (all nominated) in Sowa Town to 174 (140 elected + 34 nominated) in the Central District. The number of appointed councillors can and does change even between the decennial delimitation exercises.

Autonomy, Capacity, Authority, and Accountability: Half Full or Half Empty? The councils are responsible for setting development priorities, overseeing local infrastructure development, and delivering a variety of services. Specific responsibilities include development planning; road construction and maintenance; primary education facilities; waste management; parks, recreation, and cemeteries; enforcement of bylaws; regulation and licensing of some commercial activities; management of labor-intensive public works; and administration of markets and abattoirs. Over the years, the councils have gained responsibility for social and community development programs, self-housing projects, remote area development, fire extinction and prevention, and physical planning (Government of Botswana 2001). In what can be viewed as recentralization and disempowerment of councils, the government decided to transfer health services and rural water supplies from councils to the Ministry of Health and the Ministry of Minerals, Energy and Water Affairs, respectively, in 2009. Even if their scope of responsibility is relatively broad, the councils have little discretionary authority. The Ministry of Local Government and Rural Development provides policy direction on issues pertaining to governance, community development, social services, primary infrastructure provision, district development planning, coordination, policy formulation, and legislative development. Most council decisions are subject to central government approval. The ministry also mobilizes resources (e.g., human, finance, and technology) and transfers them to the local authorities (Ministry of Local Government 2009). In essence, the councils are elected bodies charged with the delivery of services as mandated by the central government. Thus, councils are better characterized as delegated, rather than devolved bodies. Advocates of the status quo argue that councils lack the capacity, in terms of both human and financial resources, required for devolution (field research, 2010). However, the limited capacity of the councils is at least partly a product of central government policies and practices. Financial dependency and centralization of personnel management restricts the autonomy of the councils. The dynamics of party discipline and asymmetries in information and expertise further advantage the center.

28

Botswana

Which Comes First? Autonomy and Administrative Capacity

The councils depend heavily on the central government for human and financial resources. Botswana’s councils are regularly characterized as having limited capacity, but these claims often are not backed up with evidence, or are based on characterizations that ignore changes over time and differences across districts. Furthermore, central government policies and practices often reduce the capacity of the councils. Their lack of capacity is at least partially a product of their lack of autonomy. A description of the legal status of the councils leads to an examination of their autonomy and administrative capacity in terms of human and financial resources. Legal status. The autonomy of the councils is limited by their status as statu-

tory bodies that exist at the discretion of the central government (Hope 2000; Dipholo and Mothusi 2005). Council representatives and members of the Botswana Association of Local Authorities (BALA) lament the fact that councils “can be abolished at any time without recourse to the public” (Government of Botswana 2001: 12, emphasis in original; interviews 2010). Even if the abolition of councils seems unlikely, there is precedent for the creation of new councils. The creation of city and town councils deprives the affected rural councils of authority over areas rich in natural or human resources.5 Likewise, the creation of the Chobe District Council reduced the territorial jurisdiction of the North West District Council; this loss has significant resource implications given Chobe’s status as a tourist center. Human resources. In our field research, we found that councils have

gained considerable administrative capacity. Most councillors and council secretaries are now content with the caliber of their staff, but chafe at the centralization of personnel management. While the early postindependence period was characterized by acute shortages of trained and skilled personnel, particularly at the local level, most councils are now headed by highly qualified and experienced officers at the level of chief executive officer (CEO) and head of department. The majority of middle management officers are degree holders; some have graduate degrees. Most of the staff in professional positions have an associate’s or bachelor’s degree. There has also been tremendous increase in staffing in specialized areas like accounting, civil and mechanical engineering, physicians and specialized nurses, and economic planning. Problems of administrative capacity arise because of vacancies in some districts and subdistricts, especially in rural and remote areas. For example, in Letlhakane (Boteti Subdistrict in western part of the Central District), most departments are managed by divisional heads and field officers instead of

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

29

more qualified mid-level managers. In the absence of qualified department heads, the senior assistant council secretary must carry out some basic administrative duties that should be done at the departmental level; the subdistrict also lacked a deputy assistant council secretary. Similar situations occurred in the newly established Maun Administrative Authority in the North West District and in the Chobe District when the new council was established in 2006. District-level officials in the North West District were instructed to reassign all of the junior staff to the Maun Administrative Authority while retaining the senior staff at the district headquarters. In this instance, the two sets of offices are separated by only a short stretch of road. Proximity may facilitate coordination, but there are morale problems because junior staff members feel that they have been demoted. The arrangement in Chobe posed practical problems. The council was launched before new staff positions were fully established and filled. The council filled some positions more than four years after it had become operational while other positions remained vacant as of 2010. As in Letlhakane, lack of staff, especially senior managers, limits the administrative capacity of the Chobe District Council. A personnel officer in Chobe who had been transferred from the Francistown City Council observed that, at 4.9 percent, vacancy rates in Chobe District were more than twice as high as in Francistown (estimated at 2 percent) (personal communication, Kasane, July 26, 2010). The problem of vacancies is not limited to new structures and rural sub-districts. Officials in the Chobe and North West Districts argue that they have more difficulty attracting and retaining staff because there is no salary adjustment to compensate for less attractive conditions in these more remote districts.6 The personnel officer for the North West District Council underlined a severe and persistent shortage of technical officers (personal communication, Maun, July 22, 2010). For example, only one of four established posts for engineers had been filled. The shortage of engineers and other technical staff hampers project planning and implementation, contributing to delays and cost overruns (personnel officer, personal communication, Maun, July 22, 2010).The Public Service Act of May 2010 centralized personnel management under the Directorate of Public Service Management (DPSM). The DPSM has delegated responsibility for most aspects of human resource management (e.g., recruitment, promotions) except transfers and training. In 2010, councils were given the authority to recruit people for middle-management positions. This resulted in dramatic reductions in vacancy rates for these positions. On the other hand, the government centralized recruitment for entry-level professional posts (usually new university graduates); the councils simply receive the officers that they are given. Despite the delegation of some human resource functions, the autonomy of the councils is limited by their high fiscal dependence on cen-

30

Botswana

tral government. Every cost with respect to recruitment, appointments, salaries, and staff welfare is paid for from central government grants. This means that councils do not control decisions on these matters. In addition, all transfer-related issues are the prerogative of the DPSM; it has the power to withdraw posts from one council and give them to other councils or other government departments without the consent of the affected agencies. In one particularly extreme example, an official was transferred from Chobe District after only a month, despite objections from the district commissioner (personal communication, Kasane, July 26, 2010). Respondents complained that frequent and ill-timed transfers impede the accumulation of institutional memory and disrupt operations (interviews 2010). Training is also centralized. Although generally satisfied with the content of the training programs, officials complained about their inability to set their own priorities. One personnel officer complained that DPSM gives priority to training people with seniority: “They take . . . even those with two years from retirement, not somebody who can grow in the organization over time” (personal communication, Maun, July 22, 2010). The councils have been required to eliminate some positions and reassign long-serving officers to new positions, without being provided with the means to retrain these officers. In one case, the councils ended up with typists in secretarial positions who had no training in customer care and relations, which resulted in members of the public complaining about their attitudes toward clients (personal communication, Francistown, July 27, 2010). More than four decades after independence, human resources remain an important constraint on the administrative capacity of the councils, but the nature of that constraint has changed. Staff qualifications are no longer the primary problem. Councils have highly qualified individuals with strong managerial and technical skills. More significant ongoing challenges arise from vacancies, especially in new or rural structures; discontinuities arising from transfers; and mismatches between qualifications and local needs. At least some of these challenges are exacerbated by centralization of personnel management, including policies for compensation that do not adequately respond to variations in the difficulty of attracting staff. Financial resources. The councils’ heavy dependence on the central government for financing limits their autonomy. But the councils are legally empowered to generate revenues from a limited array of sources. The most important are interest on money deposited in bank accounts, rentals, trade licenses, clinic fees, sewerage connection and service fees, slaughterhouse fees, water charges, and service levies (e.g., for collection of garbage) (Government of Botswana 2001). Property taxes (i.e., rates) represent an important revenue source for urban councils, but they cannot be levied by district councils. The central government provides the balance of local gov-

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

31

ernment budgets through the Revenue Support Grant. As shown in Table 2.2, grants from the central government account for about 90 percent of the budget for rural councils and around 80 percent of the budget for urban councils. Efforts to improve billing and collection are under way for various services, such as sewerage, and should bolster council revenues. However, improved collection of existing rates and fees alone will not result in financial autonomy for the councils. New sources of self-generated revenues are needed. Unfortunately, the possibilities are limited. While most existing revenue sources have a low base, especially in rural areas, central government restrictions on revenue generation are at least as important. The deter-

Table 2.2 Council Generated Revenue as a Percentage of Total Revenue, 2005–2006 to 2009–2010 Council Rural councils Central Chobe (MLG) Chobe (DDP7) Ghanzi Kgalagadi Kgatleng Kweneng North East North West (MLG) North West (DDP7) South East Southern Urban councils Francistown Gaborone Jwaneng Lobatse Selebi Phikwe Sowa Town Totals (LG data) Rural councils Urban councils All councils

2005–2006

2006–2007

2007–2008

2008–2009

2009–2010

11.7 — — 11.0 13.3 11.8 12.4 9.4 7.6 14.8 13.1 11.9

11.5 6.6 10.0 11.3 13.0 10.4 12.0 13.0 10.5 14.3 13.6 11.4

10.9 10.7 16.2 12.0 13.6 10.9 10.5 6.9 10.3 16.1 13.5 9.0

7.5 7.4 11.7 7.2 9.5 11.3 6.1 6.2 7.5 7.5 8.2 8.2

6.3 9.0 — 8.9 6.7 8.7 6.8 5.3 4.4 — 9.3 7.2

27.6 36.1 27.7 16.7 19.4 14.1

25.3 39.9 24.4 16.3 19.8 13.7

26.4 37.2 18.4 10.6 21.0 14.2

18.7 25.6 13.7 9.3 16.2 6.6

18.5 26.8 11.3 8.2 18.5 14.5

11.3 28.1 15.7

11.5 28.9 16.0

10.7 26.4 14.6

7.7 18.1 10.2

6.9 19.1 9.8

Sources: Chobe (DDP7): Chobe District (2009: 204). Includes resource royalties from tourism for 2006–2007 to 2007–2008. North West (DDP7): North West District Council et al. (2009: 202). Includes resource royalties from tourism for 2005–2006 to 2007–2008. Chobe (MLG), North West (MLG), and all other districts: calculated from data provided by the Department of Local Government and Procurement, Ministry of Local Government. Note: MLG is Ministry of Local Government. DDP is District Development Plan. LG is local government.

32

Botswana

mination of rates, taxes, and user fees is subject to ministerial approval, as are program and project ceilings for all budgetary items and the transfer of funds across programs and projects. Political expediency and the pursuit of equitable distribution of resources have resulted in centrally set fees that do not come close to covering the costs of service delivery. For example, all councils are compelled to charge a clinic fee of P5.00 (5.00 pulas, approx. US$0.75), even though the cost of operating such facilities is much higher. Councils are instructed that no citizen should be denied access to medical assistance if they cannot pay the required amount. Some important potential or actual sources of revenues have been centralized or eliminated. The local government tax, for example, was abolished in 1988. Urban councillors complained that, whenever people damage council streetlights through careless or reckless driving, they pay fines to the central government, leaving the councils with no money to repair the damaged streetlights (interview with council economic planner, Gaborone city council, Gaborone, July 4, 2010). Councils are losing another source of revenue because responsibility for water utilities is being transferred to the Water Utilities Corporation. It is taking responsibility for the most profitable areas first, with the councils retaining responsibility for areas where billing systems are underdeveloped or water is provided as a social service. The 2009 decision to centralize primary health care implies a further loss of revenue sources, although these services have been heavily subsidized in practice (see above). The transfer of resource royalties paid by companies involved in the tourism industry from the councils to the Ministry of the Environment, Wildlife, and Tourism in 2008–2009 has been a more serious blow. Resource royalties from tourism were the single most important revenue source in the North West District Council and also were important in the Chobe District (Chobe District Council n.d.; North West District Council et al. 2009).7 Unpublished data on council revenues provided to the authors by the Ministry of Local Government do not include resource royalties as a source of self-generated revenue even for years when, according to respondents and data reported in the District Development Plans (DDPs), the councils received it. This discrepancy in reporting makes it possible to compare the contribution of self-generated revenues to the total budget with and without tourism-related resource royalties for the Chobe and North West districts for several years. Table 2.2 shows that, without resource royalties, the rural and underdeveloped North West District Council performs particularly poorly in local revenue generation. With resource royalties, however, the contribution of self-generated revenues to total revenues in the North West District Council rivals that of the Lobatse Town Council and Sowa Town Council. It is not obvious what alternative sources of revenue the Chobe and North West Districts can cultivate to compensate for the loss

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

33

of resource revenues. Both councils still receive fees from domestic water users, and improvements in water billing are generating increased revenues. When centralization of water utilities extends to these rural districts, however, the councils will lose this source of revenue as well. The councils thus rely heavily on the centrally provided Revenue Support Grants. Centralized and time-consuming approval procedures impede access to these funds, causing delays that give rise to cash flow problems and cost overruns. For each project to be implemented in a given financial year, councils must submit project memoranda that provide the justification for the project; the cost, scope, and expected benefits for the citizens; and an explanation of how the project is related to national goals, among other things. The project memoranda are reviewed by both the Ministry of Local Government and Rural Development and the Ministry of Finance and Development Planning. The approval process takes one to two months. Councils can invite contractors to bid on a project only after funds have been approved. The tendering and selection process takes approximately three months and contractors take one to two months to move on site. Many projects cannot be completed within one financial year because of the delays arising from this centralized and multistep process. The argument often voiced by the Ministry of Finance and Development Planning, that failure to spend the funds within the budget year means that the councils do not have the capacity to spend all funds allocated to them, ignores these procedural delays. In fact, because of the delays associated with the project approval process, councils may have unspent funds at the end of the financial year but still be underresourced. Any unspent funds must be deposited into the General Fund, which is expected to cover any future deficits. A council also has discretionary use of up to 10 percent of its balance. The Chobe District Council, for example, drew on its General Fund to pursue infrastructure projects that would have stalled otherwise as a result of central government budget cuts in the wake of the global financial crisis. Expenditure of more than 10 percent of these funds, however, requires permission from the Ministry of Local Government and Rural Development. Because the central government provides security for the councils and has to pay their loans in the event of a default, it insists that all financial transactions between local authorities and financial institutions must be sanctioned by the Minister of Local Government and Rural Development. Financial accountability for local authorities is enhanced and controlled through two organs: the Office of Auditor General and the Local Authorities Public Accounts Committee (LAPAC). The Office of the Auditor General has overall responsibility for government’s financial accounting procedures, systems, and enforcement while LAPAC has the same responsibilities only visà-vis local authorities. LAPAC acts to find remedial ways of enhancing local

34

Botswana

authorities’ financial management capacity as well as exploring new revenue sources. There is a general consensus within both LAPAC and the Office of the Auditor General that councils must become more independent, must generate more revenues, and must have fully fledged treasuries with qualified staff that can perform reporting and auditing duties. Key to achieving these goals is expanding the sources of self-generated income such as traffic levies and parking fees. The new fees could in turn influence the configuration of other revenue sources such as income tax in a way that balances both revenue collection by local councils and citizens’ ability to pay. Instead of expanding opportunities for revenue generation, however, the center has deprived local councils of valuable revenue sources and constrained the councils’ ability to increase the productivity of many remaining revenue sources. While local economies in most of Botswana could not support much taxation, it is difficult to accept at face value central government claims about the limited capacity of the councils to generate their own revenues. The Multifaceted Role of Councillors: Issues of Authority, Accountability, and Capacity

As elected representatives, councillors represent the interests, demands, and aspirations of the citizens who vote them into office. Yet as documented above, the councils have little discretionary authority. In practice, the councils are primarily responsible for implementing centrally mandated responsibilities, for which they depend heavily on centrally provided funds. Thus, the councils face pressures for both downward accountability to their constituents and upward accountability to various central government bodies. Pressures for upward and downward accountability also arise from the important role that councillors play in linking local citizens and activists to the area MP and national party structures. In this subsection, we draw on electoral data, interviews, and documents to develop three points. First, we describe the types of individuals who serve as councillors and evaluate their capacity to fill the roles expected of them. Second, we analyze changes in the competitiveness of local elections and the implications for the effectiveness of elections as mechanisms of downward accountability as well as comment on how party discipline encourages accountability to party structures—and, sometimes, factions. Third, we highlight the role councillors play in linking citizens with the center. Diverse backgrounds and uneven preparation. Although the caliber of councillors has improved with education levels in the general population, the statutes do not define minimum qualifications and there is much variation in their educational background, age, experience, expertise, and skills. Councillors whom we interviewed included former drivers, brick layers, farmers,

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

35

teachers, accountants, sales managers, and business executives. There were primary school dropouts as well as secondary school, diploma, and degree holders. Some respondents argued that the councils should not be granted greater financial or administrative authority as long as they include people who have limited formal qualifications or lack managerial experience. Some advocated the introduction of minimum qualifications such as literacy in English. On the other hand, several respondents observed that less educated councillors often understand and deal with local concerns better than their more formally qualified peers. Improving the conditions of service for councillors would attract better qualified people. Councillors do not have a salary. Each month, they receive a base allowance and a ward allowance that total P6,700 ($972). Councillors also receive a sitting allowance of P179 ($26) per day. These allowances are supposed to cover the expenses of the councillors in carrying out their responsibilities, including expenses for transportation and telecommunications. Citizens also call on their councillors for all kinds of assistance. A councillor from the North West District explained, “What I was earning as a sales executive was higher, and that was my income. All of our allowances, we share them with other people. Destitutes come and ask for assistance. People ask to use your car” (personal communication, Maun, July 22, 2010). A councillor from Gaborone with a business background observed that “if you look at the financial incentives, you don’t want to be a councillor” (personal communication, July 14, 2010). Even people who had been drivers complained that being a councillor “is not a well-paying job” (councillor, personal communication, Gaborone, July 13, 2010). In principle, council meetings and constituency service leave time enough for other income-earning activities. In practice, most councillors have to give up their jobs and many have no other source of income. The Public Service Act bars public servants from holding full-time political positions; this affects government drivers, employees of parastatals, and teachers as well as officials working for central or local government departments. Private sector employees may also face a choice between their job and political office. The councils attract many highly competent people despite the poor remuneration, but councillors sometimes resign in the middle of their mandate to pursue their careers. Political accountability through elections. In theory, elections enhance

downward accountability to citizens. In practice, elections represent a relatively blunt mechanism of accountability, given their periodic occurrence and the imperfect link between particular decisions and election outcomes. In Botswana, downward accountability via elections is limited further by the dominant position of the ruling BDP, uneven electoral competition, the presence of nominated councillors, and upward accountability arising from party discipline.

36

Botswana

Council elections are often noncompetitive.8 The BDP consistently wins comfortable majorities on almost all rural councils. The least competitive districts are Central, Chobe, and Kweneng, where the BDP has always controlled more than 70 percent of the seats. Important exceptions include the Kgatleng District and, at least in some years, the North East and North West districts. Over the past decade, the BDP’s grip has weakened somewhat in the South East, Southern, and Kgalagadi districts. Several cities and towns became opposition strongholds in the 1980s, but the BDP regained some ground after the opposition split in 1998 and urban elections are now highly competitive. The dominance of the ruling party has limited, but not eliminated, the threat of electoral turnover; the opposition has won majorities on various councils. The formation of the BMD increased the number of opposition MPs and councillors, and increased the number of councils controlled by the opposition. As more councils become competitive, the value of elections as a mechanism for holding councillors accountable to the electorate should increase. However, provisions for specially elected councillors reduce the force of elections. The government justifies the nomination of councillors and MPs as a way to increase the representation of historically disadvantaged groups and to bring in people with important skills. In practice, the vast majority of the nominees are BDP members. The government regularly nominates people who have been rejected at the polls, further reducing the effectiveness of elections as accountability mechanisms. Nominated members sometimes determine partisan control of the council.9 Also, because councillors pay monthly subscriptions to their political party, the BDP benefits financially from this arrangement. In areas where the opposition is absent or weak, whoever wins the BDP nomination is likely to be elected. Thus, limited electoral competition increases the importance of party discipline. Party disciplinary procedures, especially the threat of being vetted out of primary contests, give the party caucus considerable force. While the party caucus can be valuable as a means of advancing a policy agenda, it can also be used to stifle debate and promote factional interests. It encourages accountability to party structures (or factional leaders) rather than local constituents. Because competitive elections decrease the importance of nomination by any particular party, they also undermine the threat of party discipline as a mechanism for upward accountability. Bridges between the center and citizens. Elections are not the only mechanisms for downward accountability. Councillors are expected to apprise their constituents of issues discussed at the council or parliament; that is, progress made in terms of development projects, newly proposed government policies or programs, and other matters of public interest. Most councillors consult their constituents by holding regular nonpartisan public meetings at the kgotla

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

37

(traditional meeting space) and with a variety of local organizations. The effectiveness of kgotla meetings as a mechanism of accountability is limited by low attendance.10 As one councillor from Gaborone laments, “If you call a kgotla meeting nowadays, 15 to 20 people—if there is good attendance—will show up. That’s not good attendance!” (personal communication, July 13, 2010). Some elected officials are developing alternative strategies for consulting with their constituents such as visiting hospitals, schools, and other public institutions. Councillors also meet with the Village Development Committees (VDCs), Village Extension Teams (VETs), and other community organizations to coordinate activities and ensure the implementation of development projects. The government created VDCs and VETs as nonstatutory, voluntary bodies to act as local partners with the district council and district administration in project development and implementation. They are an important link between citizens and councillors and provide a forum where the councillors can be asked to account for their actions or those of the government. The effectiveness of these voluntary organizations varies tremendously. Especially in rural areas, basic literacy levels can be a problem. Significant problems arise if either committee is not operational. In the Kasane Central ward of Kasane, in Chobe District, for example, the VDC chair complained that the VET was not functional and, as a result, the committee sometimes was not informed about government plans and programs (personal communication, Kasane, July 26, 2010). Citizens expect their councillors to consult with them regularly, whether at the kgotla, through meetings with the VDC and other community organizations, or by attending important social events (e.g., funerals, weddings). Citizens take councillors to task if the delivery of services breaks down or the councillor is not sufficiently present in the community. When a councillor is deemed incompetent, citizens voice their complaints during kgotla meetings addressed by national-level politicians. Sometimes citizens contact their MP or other officials directly to seek assistance when their councillor has failed to meet expectations. Councillors depend on MPs to facilitate the flow of resources into the ward while MPs depend on councillors to maintain the delivery of services and to mobilize electoral support. This mutual dependency between national and local politicians accounts for the high political salience of any reforms that alter the balance of power between local and central government.

Strengthening Local Government, or Central Control? Local authority in Botswana is an arena of ongoing transformation. Past reforms typically involved deconcentration and delegation. Devolution has

38

Botswana

been rare, limited, and—to the extent that it occurs at all—subject to reversal (Poteete 2009). Although some proposed changes involve the devolution of authority, the central government continues to emphasize reforms that extend central government control over local institutions. New political and economic uncertainties reinforce these preferences. Most of the current local institutions were established between 1965 and 1970. By the end of the 1970s, the government already felt the need for institutional reforms to improve service delivery; this led to the creation of subdistricts in the largest districts, with field offices of district administration in the subdistrict headquarters. There was no real change, however, in funding for local authorities or coordination across the various local institutions. The 2001 Report of the Second Presidential Commission on the Local Government Structure in Botswana (Government of Botswana 2001), known as the Venson-Moitoi Report after the committee chair, set the agenda for more substantial reform. The Venson-Moitoi Commission reviewed all local institutions—not just the councils—and developed proposals for procedural as well as structural reforms to improve service delivery. The commission called for strengthening councils by entrenching them in the constitution and providing greater financial resources. A number of recommendations involved greater devolution of authority. For instance, the report called for the introduction of executive mayors with the authority to implement council resolutions, elimination of the ministerial powers to appoint nominated councillors and to veto council bylaws, and more expansive responsibility for councils in specific areas (including primary education). The government’s formal response to the Venson-Moitoi Report was to reject almost all of the recommendations that would have enhanced the authority and autonomy of the councils (Government of Botswana 2003). It did approve changes to rationalize existing legislation and develop new policies, reforms related to revenue generation and collection, a few measures to enhance the capacity of local councils, creation of one new district council, creation of new subdistricts and several new forms of deconcentration parallel with and below the district level, and measures to enhance coordination across both central government ministries and local institutions. Overall, the accepted changes emphasize financial efficiency and deconcentration as a means of service delivery. Implementation of the accepted recommendations has been uneven. Some changes, such as those related to the election of council chairs, enhance democratic accountability within the councils. Most involve deconcentration and delegation through the creation of new structures, including subcouncils and administrative centers. The subcouncils are essentially committees of the full councils, constituted by the councillors who represent wards within a particular subdistrict. They make decisions about the delivery of services and implementation of projects in the villages

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

39

or settlements that they serve, but they cannot make decisions on policy issues. Administrative authorities have been established in the headquarters of the various districts and they operate in the same manner as subdistricts and subcouncils in terms of personnel, facilities, and responsibilities.11 Other recent changes involve recentralization. Centralizing reforms include the phased takeover by central ministries of responsibility for water delivery, primary health clinics, tourism royalties, and education. The proposed reforms strengthen mechanisms of upward accountability—to the national party structures as well as the central government. Amendments to the Local Government (District Councils) Act and the Township Act in 2008 increased the term of office for mayors and deputy mayors to two and one-half years and barred specially elected councillors from voting on the selection of the council leadership. By 2010, the Ministry of Local Government had drafted a local government bill that would consolidate the Local Government (District Councils) and Townships Acts, extend the term of office for council chairs and mayors to five years, allow the introduction of property rates in rural areas, and define a formal role for the BALA. Councillors interviewed on July 12, 2010, in Gaborone criticized this draft for curtailing the ability of the council to change their leadership between elections, empowering the minister of local government to remove councillors for nonattendance at meetings, and failing to incorporate or devolve authority to the councils in line with recommendations of the Venson-Moitoi Report. At its July 2010 congress, the Botswana Congress Party resolved to reject the local government bill and develop alternative legislation that would restructure and empower the councils. Representatives of the Botswana National Front and the BMD voiced similar views. The government largely ignored these complaints but, before presenting the bill to parliament in 2011, it added a provision to discourage floorcrossing. In both the Ntlo ya Dikgosi (House of Chiefs) and the National Assembly, debate focused on this provision, which would render a seat vacant in the event that a sitting councillor should change party affiliation, thus triggering a by-election. Criticism of this provision among traditional leaders and within the BDP as well as from the opposition prompted withdrawal of the bill from further consultation. The anti-floor-crossing measure appeared as a direct response to the BDP split. Although the BDP won a majority in most local councils in 2009, defections to the BMD enabled the opposition to take control of the Francistown and Gaborone city councils and the Jwaneng Town Council.12 Opposition councillors also hold leadership positions on some subcouncils. Even if the councils have little formal power, councillors play a critical role in mobilizing support for parliamentary elections, and the monthly subscriptions that they pay represent an important source of financing for polit-

40

Botswana

ical parties, especially the BDP. The newly uncertain political situation reinforces the government’s long-standing aversion to devolution of authority while encouraging efforts to reinforce upward accountability. Forcing councillors to face a by-election would make it riskier to change party affiliation, whether in response to factional disagreements or over matters of principle.13 By closing off the exit option, the measure would strengthen party discipline and upward accountability to party structures. The anti-floor-crossing proposals would also complicate efforts by the main opposition parties to federate as a new umbrella party. Thus, antifloor-crossing measures raise the risks associated with the umbrella model; if the umbrella were to collapse, or if one of the parties decided to leave the umbrella, all of its representatives would face by-elections because they had been elected under the umbrella party. The Local Government Act was eventually adopted in 2012, but without the provisions on floor-crossing.

Political Economic Changes and Decentralization The economic and political pressures confronting the government influence the pace and substance of proposals for decentralization reforms and their prospects. In the wake of the global financial crisis, the government of Botswana faces unfamiliar financial constraints. The proposals to rationalize systems for billing and revenue collection and extend property rates to rural areas respond to the need to develop new sources of revenue. As the recent centralization of resource royalties related to tourism and the phased centralization of responsibility for water utilities make clear, the government is concerned with improving revenue generation but not the financial autonomy of the councils. There is a tendency to centralize control over the most lucrative sources of revenue while leaving the councils with responsibility for the delivery of subsidized or free services. The government may have an interest in keeping the councils dependent. It has no interest in devolution. Opposition parties in Botswana have long supported the devolution of power to the councils. They see the Local Government Act as empowering the minister of local government and rural development and are worried by the recent centralization of water, health, and other services. The political context makes it more difficult to improve revenue generation or service delivery without addressing the autonomy and authority of the councils. The BDP is still able to force its legislative agenda through parliament when its caucus votes in unison. Since the 2009 elections, however, controversial proposals have prompted massive abstentions and defections on at least a couple of occasions. In sum, conditions are not favorable for legislative reforms. There is greater scope to pursue administrative reforms, insofar as these may improve revenue collection and service delivery.

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

41

The subnational arena in Botswana is highly politicized for several reasons. First, a large number of local- and national-level actors are involved in local government and are wary of changes that might alter the balance of power at the local level. Second, because the opposition controls some of the local councils, many in the BDP see decentralization—especially in the form of devolution—as favoring the opposition. Third, the central government regularly uses its authority over local institutions to gain partisan and even factional advantage. And, fourth, councillors play an important role in national politics, both by mediating the relationship between MPs and the electorate and by their influence within the national political parties. As political competition at the national and local levels intensifies, political parties have taken an increased interest in even VDC elections. The politicization of local government structures hampers the performance of local institutions and makes reform more difficult. It is not unusual for local authorities—or central government ministries—to be politically opposed to each other. When this happens, rivals can block each other by taking advantage of areas of overlapping jurisdiction and refusing to participate in the many local committees established to coordinate cross-agency implementation. Politicization also makes reform difficult to achieve since each stakeholder attempts to block changes that empower an actual or potential rival. The land boards, tribal administration, district administration, and most (if not all) central government ministries oppose strengthening the authority of the councils or even the district commissioner. Similar coalitions block other possible changes in the distribution of authority across local institutions. Botswana’s experience underlines the importance of interactions between the party system, electoral politics, and the degree of decentralization. As Kathleen O’Neill (2003) argues for presidential systems, expectations about relative electoral success at the national and subnational levels influence the prospects for decentralization. The leadership of the BDP has little interest in decentralization in part because, at least in the medium term, it has more confidence in the party’s prospects at the national than at the subnational level. 14 If an expectation that electoral prospects are greater at the subnational level is one motivation for formal decentralization reforms, this condition seems more likely in majoritarian presidential systems than in Westminster-style parliamentary systems or presidential systems that use runoffs. In this chapter, we have emphasized informal as well as formal sources of centralization and decentralization (compare Hutchcroft 2001). The concentration of economic resources in the state has reinforced the political power of the BDP while centralizing state institutions have contributed to centralization within the BDP. Centralization of authority within the BDP in turn has reinforced centralization of the state. Botswana’s experience also

42

Botswana

reveals nuances in the relationship between centralization and party system development. In general, more centralized states encourage the development of a national (i.e., more centralized) party system (Chhibber and Kollman 1998; Hicken 2009) whereas decentralization fosters the emergence of regional parties (Brancati 2006). Political contestation in Botswana has definitely focused on control over the highly centralized state. And yet the opposition has not consolidated. In fact, increased fragmentation of the opposition over the past two decades has coincided with increased centralization. Party systems are not simple or static responses to institutional incentives arising from formal (de)centralization, but dynamic manifestations of shifting electoral prospects and strategic considerations (compare Ziegfeld 2012). Shifts in electoral dynamics, and in the party system, feed back into debates about decentralization. Current changes in Botswana’s party system may create new opportunities for democratic decentralization, although the fluidity of the situation makes it impossible to evaluate the likelihood of such changes. In the end, Botswana’s experience affirms the role of political and economic background conditions in shaping not only the prospects for formal decentralization reforms, but also the realization of the potential for decentralization within existing institutional arrangements. Taking a broader perspective, Botswana’s experience underlines the double-edged character of decentralization. Decentralized bodies cannot be assumed to enhance local democracy, even when their representatives are elected. Decentralization may be designed to reinforce upward rather than downward accountability, thus reinforcing the power of the central state and dominant parties rather than empowering local communities.

Notes 1. Botswana’s value-added tax, introduced in 2000, is considerably lower than the averages for sub-Saharan Africa (16.01 percent) and the world (15.77 percent) (USAID 2010: 3). Rates for personal income taxes “are among the lowest in SubSaharan Africa and also well below world averages” (USAID 2010: 1). 2. Poteete (2012) categorizes constituencies as competitive if (1) the margin between the winner and next closest candidate is less than 5 percent of the vote, or (2) the winner receives less than 50 percent of the vote in a race involving three or more candidates. 3. The BMD lost recognition as the Official Opposition in July 2012 as a result of subsequent changes in party affiliation. From November 2012 to February 2014, Botswana’s National Assembly had no Official Opposition because the two main opposition groupings were exactly balanced and the parliamentary standing orders provide no procedures for choosing between them. The BCP became the Official Opposition following a by-election in January 2014. 4. Officials and observers in Botswana regularly refer to all four entities as “local authorities” even if most scholars would not count district administration as

Amy R. Poteete, Bashi Mothusi, and Daniel Molaodi

43

a local authority and, according to Botswana’s Interpretation Act of 1984, only the local councils count as “local authorities.” 5. The following districts have lost territory and population to city or town councils: Central (Selebi Phikwe and Sowa), North East (Francistown), Southern (Jwaneng), and South East (Gaborone and Lobatse). 6. Respondents from other councils did not mention vacancies as a problem. 7. The authors were provided with Excel spreadsheets of the budgets for 2009/2010 and 2010/2011. The 2011/2012 budget was a printout of the draft budget. 8. Although uncontested wards were once common, at least two candidates have contested nearly all council wards since the mid-1990s. 9. Where the BDP has a majority, nominations may determine the factional balance of power on the council. 10. The value of kgotla meetings as a mechanism for consultation with and accountability to the broader constituency is also limited because they attract a biased cross-section of society, with older men overrepresented while women, youth, and non-Tswana groups are underrepresented. 11. Respondents had difficulty explaining the difference between a subdistrict or subcouncil and an administrative authority. 12. The opposition also holds the balance of power on the Kgatleng District Council and the Selebi Phikwe Town Council, but is weakened by factional or interparty conflicts or both. 13. In 2011 and 2012, the government also attempted—and failed—to amend the constitution to declare parliamentary seats vacant and subject to a by-election in the event that an MP should change party affiliation. 14. Under the first-past-the-post electoral system, depending on the spatial distribution of its supports, a party may win a legislative majority with as little as 35– 40 percent of the vote.

3 Burkina Faso: Limited Decentralization Under Tight Oversight Pierre Englebert and Nestorine Sangaré

Although it is still a relatively new process, decentralization in Burkina Faso has become an essential element of the country’s political system. A large body of legislation underpins it and provides for several layers of new local institutions, at the core of which lies the commune. Every five years Burkina Faso’s 351 communes elect their municipal councils, which in turn choose their mayors. The communes are legally responsible for a wide swath of public policy and administration, including primary education, health care, water and sanitation, and local economic development, and they have the authority to raise taxes. Despite these achievements, Burkina Faso’s decentralization suffers from several significant weaknesses that undermine its transformative and democratic potential. First, its achievements are much less sweeping in reality than they are on paper. Because of a top-down gradual approach, it was not until 2006 that communes were finally installed throughout the territory and not until 2009 that they were transferred effective authority. Second, decentralization in Burkina Faso has an urban bias. Urban communes were created more than ten years before the rural ones, and they monopolize between 70 percent and 90 percent of available resources. As such, decentralization benefits mainly the more centralized urban areas. Third, the lack of financial means and autonomy dramatically reduces the effective scope of decentralization. Outside of the large urban areas, communal taxes have low yields. Because municipal councillors are unpaid, their duties compete with their need for income and suffer as a result. Most communes do not have the means to launch their development plans. Fourth, decentralized structures have little effective authority. They remain under the tutelage of numerous layers of deconcentrated authority, including regional governors, provincial high-commissioners, and departmental prefects. Despite recent transfers of 45

46

Burkina Faso

authority, the communes continue to rely on centralized administrative services for the daily performance of their duties. Finally, capacity is low, many councillors are illiterate, there are few professionally trained staff at local levels, basic equipment is lacking, and understanding of decentralization is limited. All of these features conspire to undermine the degree to which Burkina Faso’s decentralization is capable of promoting democratic deepening or the state’s developmental capacity. Some of the above limitations derive from the nature of the country’s regime. Although it has a capable administration apparently dedicated to decentralization, Burkina Faso remains an authoritarian regime that is based on personal rule, patron-client relationships, and the domination of the ruling party—the Congrès pour la Démocratie et le Progrès (Congress for Democracy and Progress, CDP). It is expecting too much of decentralization to believe that it can provide good governance and democracy in such a context. More often, it serves as the vector through which the ruling coalition spreads its influence throughout the country. Yet the work of institutional creation that accompanies decentralization no doubt is laying the foundation for a more participatory future, contingent on an eventual transformation of the Burkinabe national political system.

History and Overview of the Political System Decentralization in Burkina Faso is an ongoing and incremental process marked by the principle of gradualism.1 Aside from a brief experiment with urban communes from 1960 to 1963, Burkina Faso, like most francophone African countries (except Senegal), had no experience with decentralization until the 1990s. Instead, its political history was characterized by recurrent attempts by its central governments to further centralize and concentrate powers. In this respect, territorial reforms in 1970, 1974, 1984, and 1989 all involved deconcentration rather than devolution and promoted the territorial expansion of the state rather than autonomy for local populations. This was particularly true of the revolutionary years of the mid-1980s, during which the number of administrative circumscriptions and territorial collectivities multiplied, but remained under tight control of the central government (Sawadogo and Sebahara 2004: 68). The original impetus to decentralize arose in the early 1990s, largely as a result of pressure from donors who pushed for a new paradigm of decentralized governance, and in the wake of some retrenchment of state centralization demanded by the World Bank and the International Monetary Fund as part of Burkina Faso’s first structural adjustment program in 1991. Burkina Faso adopted a new constitution (it had not had one since 1980) on June 2, 1991, which not only ushered in an era of multiparty competitive elec-

Pierre Englebert and Nestorine Sangaré

47

tions and guaranteed individual civil liberties and political rights, but also organized the country into territorial collectivities (Article 143) and provided for the democratic management of these collectivities by local populations (Article 145). Yet it was not until 1993 that the first decentralization laws were adopted for implementation, and they fell well short of the ambitions of the constitution (and the donor community), being largely limited to a mere administrative reorganization of the territory. Lacking a clear idea of what decentralized governance should look like, the government established the National Commission for Decentralization (CND) in 1994, attached to the prime minister’s office, and charged it with “considering what Burkinabe decentralization might be” (Ouédraogo 2003: 99). The commission’s work led to the adoption of a series of decentralization guidelines in 1998, and a succession of four laws that allocated responsibilities and powers between central and local governments. The laws established two levels of decentralization. The most local was the commune (also known as the municipality), divided into urban and rural communes. Between the communes and the central state lay the provinces. Municipal authorities at the commune level and the councils at the provincial level were to be elected by universal suffrage. Donors joined efforts with the government and among themselves to facilitate the process and established a common basket of funds to accompany this first phase of decentralization. While some donors were in favor of a rapid transition to municipal rule at the local level, the government resisted and maintained its gradual approach, based in part on experimenting first in urban areas and then expanding to rural communes. The government organized municipal elections in thirty-three urban municipalities in 1995 for some 1,760 municipal council seats. These councillors elected thirty-three mayors. The ruling Organisation pour la Démocratie et le Progrès–Mouvement du Travail (Organization for Popular Democracy–Labour Movement, ODP-MT) won 65 percent of the votes. In 1996, the number of provinces was increased from thirty to forty-five, which led to new towns being designated as provincial capitals and recognized as municipalities. Several others were also granted municipal status because of their size and, by the time of the municipal elections in September 2000, the number of urban communes had increased to forty-nine (forty-two of which were won by the CDP, the successor party to the ODP-MT). The apparent implication in 1988 of President Blaise Compaoré’s entourage in the assassination of Norbert Zongo, an independent journalist, prevented the adoption of significant reform (as detailed below), including a deepening of decentralization. It was not until 2003 that the regime, having made some concessions to the opposition by adopting some of the political reforms recommended by a Council of Wise Men, was able to resume progress on decentralization. A series of new laws adopted on August 6,

48

Burkina Faso

2003 extended decentralization to rural areas for the first time. In the April 2006 municipal elections (in which the ODP-MT won 12,854 seats out of 17,786), rural voters elected councillors in 302 new rural communes, in addition to the 49 existing urban ones. In the meantime, a 2004 decree created thirteen regions and regional councils, which replaced the forty-five provinces in the 1998 legislation as the intermediate decentralized institutions. The provinces continue to exist as units of deconcentration of government services, but not as organs of democratic decentralization. Representatives to regional councils are elected by the urban and rural municipal councillors of their region. Although the elections of 2006 suggested a certain degree of political decentralization, the process suffered from a lack of effective transfer of administrative and professional competence to the communes. Thus, urban and rural communities had both the formal tools of self-rule and the elected officials to carry it out, but the decrees transferring effective authority to them in the domains legally under their control were still largely missing. Instead, the government set up a new centralized structure of interministerial coordination to facilitate the decentralization. In March 2007, it adopted the Cadre Stratégique pour la Mise en Oeuvre de la Décentralisation (Strategic Compact for the Implementation of Decentralization, CSMOD), with the goal to increase coordination among all the branches of government and donors, and to establish specific goals for continued progress. This process led to the creation of yet another centralized institution in July 2007, the National Conference on Decentralization (CONAD), which was placed under the prime minister with the minister of territorial administration and decentralization (MATD) and the minister of economy and finance (MEF) as vice-presidents. CONAD adopted its first three-year plan for the implementation of the decentralization compact in 2008. Under CONAD’s purview, the government adopted five decrees in March 2009 that transferred competences and resources to rural and urban communes in the fields of health, primary education, water, youth, culture, and sports (see the subsection Authority). However, there remain several fields covered by the 1998 decentralization laws for which communes do not have effective authority. As of June 2012 Burkina Faso had a total of 351 communes, with 359 mayors because the main cities of Ouagadougou and Bobo-Dioulasso together have eight arrondissements (subsectors), each with its own municipal council and mayor. Of these, 57 are urban communes, which correspond to cities and neighboring villages, and 302 are rural communes that include all the villages belonging to a department (a unit of territorial administration previously administered by a prefect delegated from the central government). Since 2006 all of these communes have elected municipal councillors who in turn elect among themselves the commune’s mayor, his or her two deputies, and the chairs of two permanent commissions (Econ-

Pierre Englebert and Nestorine Sangaré

49

omy and Finance, and General Affairs). They join with other municipal councillors in their region (of which there are thirteen) to elect regional councillors. After the 2006 elections, there were 17,976 municipal and regional councillors, 6,435 of whom were women. The municipal councils are elected for five years. However, the 2011 elections were rescheduled for December 2012, to be held jointly with legislative ones, allegedly to better prepare voter rolls and adopt biometric voter cards. The government passed a decree extending the term of the municipal councils accordingly.

Dimensions of Decentralization The following is our assessment of Burkina Faso’s decentralization along four dimensions used throughout this volume: authority, autonomy, accountability, and capacity. We found that, although the legal framework has advanced, there are significant limitations to the process in all four areas. Authority

De jure, Burkina Faso’s decentralization is impressive. The responsibilities of municipalities fall into several broad categories that include multiple competences.2 They are responsible for land management, territorial management, and urbanism, which include dividing land into parcels and attributing them for residency; delivering land titles; authorizing construction; creating, maintaining, and managing roads; and enforcing public safety and traffic regulations. Municipalities are also responsible for the environment and management of natural resources, such as public parks; basic sanitation and trash removal; and management of animal and forestry resources. In economic development and planning, municipalities are in charge of developing communal development plans. Additional responsibilities include: construction and maintenance of primary health care centers and adoption of public health and public hygiene measures; provision of preschools and the construction, acquisition, and management of primary and secondary schools; promotion of tourism and culture; protection of human rights; organization of assistance to vulnerable populations; provision of firefighting, funeral and mortuary services, energy infrastructure, wells and public fountains, and public lighting; and participation in the design of the regional electricity network. Finally, municipalities build, maintain, and manage markets and slaughterhouses. The regions have considerably less power; they are conceived of more as subsidiaries or support to the communes. Their responsibilities are the same except for funeral services and cemeteries, but in each case they tend to be consultative (in support of the central state or the communes) or subsidiary (in activities such as clearing bush paths, delivering wood-cutting permits in regional forests, and designing regional development plans).

50

Burkina Faso

Although the law provides for substantial devolution of authority to the municipalities, in reality this process has been slowed considerably for two reasons. First, it has taken much time to translate the decentralization laws into practical decrees and to transfer effective authority to communes and regions. Although there has been progress recently, the latter obstacle is structural and might represent a permanent hurdle to full-fledged decentralization. It is the simultaneous process of deconcentration of the state which, while supposed to assist devolution, represents in practice a parallel and centralized process creating authority structures that may supersede the communes in cases of conflict. Until 2009, the powers of the communes were severely limited by the absence of transfers of authority from the central government, whose agencies remained by default in charge of formally devolved responsibilities. As a result, communes have limited themselves to activities they can manage without legal transfer. In practice, allocation of land parcels, management of markets, and delivery of permits have probably become the communes’ most important and visible activities because these are also ones where they can generate local revenue. These actions can be controversial since they involve allocating scarce resources, collecting fees, and imposing regulations. The communes have begun testing the practical limits of these powers. For instance, in 2006 the Ouagadougou Municipal Council was forced to backtrack on regulations requiring riders of motorcycles to wear safety helmets because of widespread popular resistance (Harsch 2009: 282). Beyond this, the communes have had little authority, with an occasional bump from the participation of donors in specific projects such as school construction. With the decrees of 2009, communes are now legally and practically in charge of four of the eleven fields legally devolved to them: health care; preschools, primary schools, and literacy; water and sanitation; and culture, youth, sports, and leisure. In order to facilitate the transfer, the government delegated civil servants from the respective ministries to the communes. But so far only a few communes have signed the necessary operation protocols with the governors, and the accompanying financial transfers have been limited to urban communes (Republic of Burkina Faso 2009: 33). In other words, rural communes are not yet actually in charge of these questions. Thus, much remains to be done. A second limitation to the authority of decentralized structures lies in the parallel deconcentrated institutions of governance. While the elected municipal and regional organs of devolution are the territorial collectivities, the organs of deconcentration are known as administrative circumscriptions: the regions and their appointed governors, the provinces administered by high-commissioners, the departments run by prefects (whose territories are identical to those of the communes), and the villages (where authority is diffuse). Table 3.1 summarizes these overlapping layers of decentralized and deconcentrated authority.

Pierre Englebert and Nestorine Sangaré

51

Table 3.1 Overlapping Devolution and Deconcentration in Burkina Faso Unit

Devolved

Deconcentrated

Who Has Executive Authority?

Village Commune Province Region

No Yes No Yes

Yes Yes (department) Yes Yes

Village Development Committee Mayor, appointed prefect Appointed high-commissioner Appointed governora

Note: a. Regions also have elected regional councils, but these are deliberative, not executive, organs.

This overlap is not necessarily problematic since many of the deconcentrated organs were set up recently with the goal to facilitate and support devolution. Yet the parallel existence of these two sets of institutions produces unwelcome complexity. In addition, a different side of the law provides for the oversight and tutelage of state officials (governors, high-commissioners, and prefects) over the communes. A 2004 decree, for example, stipulates that administrative circumscriptions “represent the state” and “ensure the unity of the representation of the state.” As such they enforce the law and exercise the tutelage of the state over the local collectivities. This situation is compounded by the incomplete nature of devolution. Moreover, because municipal councillors are not paid, their mandate does not amount to employment and they must therefore pursue separate careers (farming for the majority of rural ones). Thus, communes often have no permanent administration, which puts them at a considerable disadvantage compared to administrative circumscriptions. The 2009 transfer decrees provided for the delegation of secretaries general from the MATD to 169 of the 302 rural communes (Republic of Burkina Faso 2009: 34), but the actual transfer of these civil servants was only beginning in 2010 and even this delegation may represent an extension of state authority. From the village to the region, the superimposition of these different institutions creates a significant potential for conflict and a considerable budgetary burden. Autonomy

Autonomy issues parallel the problems of authority. There is first a lack of administrative autonomy because of the oversight of deconcentrated institutions. More fundamentally, however, communes (particularly rural ones) suffer from a significant lack of financial autonomy for which there seem to be few credible fixes. Although the law defines a broad authority for communes, it also limits their political and administrative autonomy. Indeed, la tutelle de l’Etat (state tutelage) is an essential principle of Burkina Faso’s decentralization

52

Burkina Faso

inherited from the colonial period. This tutelage manifests itself first and foremost through administrative control: many communal decisions must be submitted for the approval or authorization of the state representative for the local collectivity (usually the prefect at the department level). This tutelage holds especially for local decisions that have a financial component, such that the central state keeps control over the overall budgetary implications of the actions of the communes. Moreover, the MATD has oversight over almost all delegated and deconcentrated local personnel who it recruits, assigns, and pays. Nontransferred competences also continue to be managed by their respective national ministries at the local level with the consequence that the lines between devolved and deconcentrated authority are blurred for many citizens. In Ziniaré in the country’s central region, for example, it was the Provincial Service of Domain and Taxes that managed the operations of lotissement (parcelization) of land for private residence (which raised the essential part of the commune’s budget) in the late 1990s. The collection of taxes from informal market traders is also managed by this service. The national police and gendarmerie provide public safety as the commune does not yet have its own municipal police. As Pamphile Sebahara puts it, “deconcentrated services intervene in the management of most municipal policies, either through technical assistance or through the exercise of tutelage and control . . . they play a non-negligible role in the production of rules and norms for the communal space, the implementation, and control” (2004: 334–335). In addition, the autonomy of deconcentrated services is limited since they usually take their orders from headquarters in Ouagadougou. Decentralization has brought about no noticeable change in this respect. Even when there has been transfer (in primary education, health, water, and sports and culture), the only full-time personnel in communes are the secretaries general who are administrators delegated by the central government and appointed by the MATD to assist with the establishment and functioning of local administrations. The communes’ lack of financial autonomy has been even more fundamental. Although communes adopt their budgets on their own, these budgets still must be approved jointly by the MATD and MEF. In practice, the two ministries can delegate this authority to the prefects. More importantly, the local revenue base of communes remains exceedingly limited and they are therefore dependent on subsidies from the central government. In general, they are significantly underfunded and understaffed relative to their vast legal responsibilities. Table 3.2 makes clear the dearth of local fiscal resources for the majority of communes. Although the amount of local revenues, at CFA15 billion seems to approximate the amount of government transfers (at some CFA21 billion), this comparison is highly misleading because it includes the local

Pierre Englebert and Nestorine Sangaré

53

revenues generated in Ouagadougou and Bobo-Dioulasso, which account for 64 percent of the total for all communes. The local revenues of the other urban communes combined is only 11.81 percent. In contrast, all 302 rural communes are able to generate only 13.11 percent of the total, on average CFA6,930,475 (or about US$14,143) per rural commune. Little can be achieved with such paltry resources. In fact, for the majority of communes, locally generated resources are insufficient to cover functioning expenses and minimal investment. Of 351 communes in 2008, 86 had no investment based on their own resources, and 121 of those remaining had less than CFA5 million ($10,204) for public investments. As a result, 207 of 351 communes were unable to launch their communal development plan (Republic of Burkina Faso 2010b). Clearly, the prevailing level of poverty and the lack of sufficient local economic fabric outside the main urban areas undermine the autonomy and functioning of the communes. The communes have tried to increase their rate of local revenue, but there is little to extract and there have been political consequences. Communes may raise taxes, charge service fees, and obtain receipts from administration of their domains (e.g., use of forests). Yet the taxes that they are allowed to raise generate little income: a tax on the informal sector (which is, almost by definition, hard to administer), a residency tax (which applies only to people living in parcelized zones and thus benefits mostly urban areas), a parcel tax for all persons being allocated a parcel (which thus is a one-time tax), a corporation tax (which effectively is limited to urban communes), a market tax, and a tax on weapons. Ironically, although the communes derive little revenue from these taxes, citizens feel their burden and tend to perceive decentralization as an increase in taxation (Sawadogo 2001). Increases in market taxes have

Table 3.2 Revenues of Territorial Collectivities, 2008 (CFA490 = US$1) Local revenues Ouagadougou Bobo-Dioulasso Other urban communes (47) Rural communes (302) Transfers from national budget External resourcesb Of which, portion to urban communes

15,969,215,465a 8,417,445,500 1,831,792,586 3,626,973,841 2,093,003,535 20,660,748,407 10,721,841,315 10,292,968,000

Source: Republic of Burkina Faso (2010b). Notes: a. Of this amount, about CFA1.2 billion was transferred from the communes to the regional councils. b. Loans, grants, and resources from decentralized cooperation from international nongovernmental organizations (INGOs).

54

Burkina Faso

been a frequent complaint in demonstrations and riots since 1995, according to Ernest Harsch (2009). The national government tried to alleviate this problem by suggesting in 2010 a new tax for communal development to be levied at the national level, with an annual expected revenue of CFA3 billion. The tentative announcement of such a tax met with a barrage of popular opposition, however. Nationwide demonstrations were held against its perceived injustice, which called attention to how little is known about how communes manage their current tax revenues. While the government relented, the commune of Ouagadougou did adopt such a tax in 2010, despite considerable opposition, limiting its application to vehicle owners. As a result, with the exceptions of the two largest cities of Ouagadougou and Bobo-Dioulasso, communes remain dependent on the national budget and donors for their finance (see Table 3.3), which hampers their autonomy. Allocations from the national budget have grown steadily over the years, however, and with some CFA25 billion in 2010, amounted to about 2.3 percent of the country’s entire budget (CFA1.083 trillion). These figures are, however, somewhat misleading because they include the financing of deconcentrated authorities and hide an overwhelming bias in favor of urban communes. For example, more than 90 percent of the resources for transferred competences in the 2009 and 2010 budgets (some CFA10 billion in 2010) were directed at urban communes, a bias the government declares is part of decentralization gradualism. The only rural expenses in this category were for the rehabilitation of wells. In addition, CFA5 billion was allocated both years for specific antipoverty programs in six urban communes. These two entries alone accounted for 15 billion of the 25 billion in the 2010 budget. Another 1.5 billion went to deconcentrated units, about 1 billion to the current expenses of urban communes, and about 0.5 billion to several intermediary institutions. Only 7 billion was left to be shared by all communes for other purposes, most of which are the purchase of equipment. This amounted to $40,700 per commune. The urban bias is even more pronounced in external financing, most of which is accounted for by donor grants and decentralized cooperation or direct aid arrangements with communes, mostly by foreign nongovernmental organizations (NGOs). Of the CFA10.7 billion of external financing in 2008, 96 percent went to the urban communes (Republic of Burkina Faso 2010b). Thus, underfinancing probably constitutes the most significant problem in Burkina Faso’s decentralization. Accountability

The main mechanism of political accountability at the local level is the municipal election. Elections were held in 1995 and 2000 for urban communes, and in 2006 for all communes. Communal elections scheduled for

Pierre Englebert and Nestorine Sangaré

55

Table 3.3 National Budget: Allocations to Decentralization, 2007–2010 (in thousands of CFAs)

General subsidies Current expenses Equipment Deconcentrated structures Transfers to institutions CONADb AMBFc FPDCTd Support for construction Governor’s office Governor’s residence Secretaries general’s residence Commune headquarters Resources for transferred competences Primary education Water supplye Health Subsidies from petroleum tax Budgets of urban communesf Commune headquarters Salaries of former provincial agents Priority investments in six communesg Total

2007

2008

2009

6,849,873 1,240,229 5,116,979 492,664 77,500

6,850,000 1,529,471 5,120,528 200,000

6,819,991 1,529,471 5,120,528 170,000 224,665 31,000 50,000 85,000 2,200,000 1,000,000 200,000

30,000

15,465 50,000 90,700 5,988,732 1,114,332 1,576,417 125,000 3,144,982

2,341,841 777,567 1,000,000

2,597,351 925,615 1,000,000

9,269,214

5,000,000 20,660,748

1,000,000 3,951,382 3,751,382 200,000 2,590,723 925,615 1,000,000 310,000 5,000,000 21,160,596

2010a 6,819,991 1,529,471 5,120,528 170,000 288,500 50,000 50,000 270,000 2,370,456 669,996 770,000 140,000 669,996 10,775,395 9,100,315 804,000 871,079 2,590,723 925,615 1,000,000 305,219 5,000,000 25,749,561

Source: Republic of Burkina Faso (2010a). Notes: a. Forecast. b. CONAD is the National Conference on Decentralization. c. AMBF is the Association of Municipalities of Burkina Faso, the communes’ lobbying arm. d. FPDCT is the Permanent Fund for the Development of Territorial Collectivities. e. Rehabilitation of wells. f. Support to current expenses. g. Limited to Banfora, Bobo-Dioulasso, Kaya, Koudougou, Ouagadougou, and Ouahigouya.

2011 were postponed to December 2012. For elections, communes are divided into sectors, each of which elects three councillors. All of the elected councillors form the municipal council. Elections are held in a single proportional round with closed lists, without preferential voting or the option to vote across lists. In other words voters vote for a party, not for specific candidates, and parties determine the rank order of their candidates on the list.3 From an electoral perspective, decentralization has favored the spread of democratic values. But there are two important features of the political system that considerably undermine its accountability to local populations.

56

Burkina Faso

The first is the patronage nature of Burkina Faso’s political system, which has been exported to the communes. The second is the domination of national political parties, particularly the ruling CDP, which is reproduced at the local level and reduces the degree to which municipal elections address local issues and concerns. Because of these factors, “popular participation” is more of a slogan than reality, and tends to refer to the payment of local taxes or to the reliance on NGOs for primary schools or health clinics. As Pamphile Sebahara and Flore Sindogo suggest regarding Ziniaré (the president’s hometown), “participation in decision making or in selecting priorities does not seem to be recognized as a possible form of participation for the population” (2004: 386). Moreover, because the law does not allow independent candidacies in municipal elections, candidates must belong to political parties. As a result, national politics plays a significant role in local elections. In fact, ruling party control over communes is overwhelming. In the 2006 elections, the CDP won in 320 communes out of 359, 49 out of 57 urban ones, and 271 out of 302 rural ones. The CDP tightly controls the nomination of candidates, to the extent that most of its nominees win. Hence, nominations have become one of the main tools in the distribution of state patronage. Mathieu Hilgers and Jean-Pierre Jacob (2008) note, however, that these nominations are careful not to upset local equilibria. As such, rather than opening the political space, decentralization reproduces the established order at the local level. In its selection of candidates, the CDP is first cognizant of the ethnicity of local candidates, which must correspond to the dominant local group (most communes, especially in the West, are multiethnic). Then it considers “local historical and political equilibria” (Hilgers and Jacob 2008: 174). Thus, leading candidates typically are heads of local dominant lineages. In practice, longtime members of the ruling party are prioritized in nominations. Many of them, however, although they originally come from the commune, actually live in Ouagadougou and typically work for the government or state enterprises (Hilgers and Jacob 2008: 177). The second stage of central state intervention is in the election of mayors and their deputies by the municipal councils. Considerable jockeying takes place, with delegates from party headquarters coming to adjudicate disputes among contenders. The election of the mayor takes place with secret balloting and under the supervision of the department’s prefect, also usually a member of the CDP. It is the CDP’s explicit goal to exercise a monopoly on the positions of mayors, first and second deputies, presidents of important commissions, and delegates to regional councils. Hence, the modal mayor would be a member of the CDP, head of an important local lineage, and a resident of Ouagadougou. As summarized by Hilgers and Jacob, “All the efforts

Pierre Englebert and Nestorine Sangaré

57

of the party in power tend to neutralize the potentially devastating effects [of decentralization] on the established social order” (2008: 180). Because of these features, local accountability of local governance structures is severely limited. The democratic nature of Burkina Faso’s decentralization is thus weakened. The extent to which decentralization has failed to empower local citizens outside the established local structures of power, especially in rural areas, might be reflected in the decline in participation rates in municipal elections over time. Turnout declined from 69.61 percent participation in 1995 (when only urban communes voted) to 49.12 percent in 2006 (after rural communes were included in the electoral process) (Hilgers and Jacob 2008: 188, fn.37). There is also an interesting phenomenon by which decentralization may undermine or displace village governance and accountability. A 2004 decree created the Comités Villageois de Développement (Village Development Committee, CVD). Each village held a constitutive assembly under the authority of the municipal council of all voting-age inhabitants to elect members of the CVD. The CVD helps prepare the communal development plan, mobilize village resources, and animate village development. Each CVD has a bureau with a president, vice-president, and so forth, and is elected for five years. Although they are elected, these officers are not considered part of the decentralized state. The CVDs’ authority is limited to development and they do not directly control the affairs of the village. Despite their elected status, the CVDs have less legitimacy than the previous village structures, which had grown from the self-organization of rural communities. CVDs now represent villages in higher-up structures instead of the preexisting youth, women’s, or socioprofessional associations. This has triggered conflicts for leadership between these associations and the CVDs. There also are frequent conflicts between the CVDs and the municipal councillors from the same village as to who can best represent the interests of the villagers. The fact that they remain an agency of the centralized state, despite their locally elected nature, makes the CVDs more a case of delegation than devolution. Thus, they end up partly complicating and weakening the local structure of governance, thereby undermining decentralized local institutions. The CVD problem hints at the extent to which decentralization has potentially reduced institutional accountability to associative structures. Local civil society associations and NGOs seem to be marginalized to the benefit of new institutional actors established to implement decentralization. Before decentralization, the providers of different local services (often nonstate actors) were by-and-large accountable to the local populations for which they existed. But citizens are more reticent to demand accountability from perceived sovereign structures, whether because of cultural norms or fear of reprisal.

58

Burkina Faso

Capacity

In general the fiscal, administrative, and political capacity of communes and regions is minimal. The lack of local revenue is the main cause of weak fiscal capacity, as discussed above regarding financial autonomy. In this section, we focus on administrative capacity. There is a significant imbalance between the administrative capacity and managerial competence of the central government, which is rather high by West African standards (Englebert 1996: 109), and the institutional weakness of the communes and regions. One factor of low capacity among communes, especially rural ones, is the lack of permanent salaried staff. Communal councils meet once every three months and delegate executive authority to the mayor and his or her deputies. Being unpaid, the latter frequently experience conflicts between their municipal obligations and their work. As a result, there is absenteeism at work among councillors in rural areas. In addition, although they are not salaried, municipal councillors are de facto local “big men” and must therefore allocate large amounts of their time to social functions. For example, they must attend funerals in their commune and, at each occasion, they are expected to make some financial donation (Sebahare and Sindogo 2004: 384). Despite the commitment of many councillors to public service, their lack of payment (while understandable in Burkina Faso’s budgetary environment) reduces their capacity and can create incentives for corruption. Low literacy levels and low levels of education among municipal councillors are also a serious problem. According to Richard Gerster (n.d.), “Many elected individuals are lacking a basic understanding of a community’s tasks as well as their own role. Technical capacities needed for project planning and implementation hardly exist.” Although the number seems high, Gerster reports four out of five municipal councillors cannot read or write. Even in a relatively large city like Koudougou, Pamphile Sebahara (2000) indicates that ten out of thirty municipal councillors were illiterate. CONAD’s Programme d’Action Triennal for 2008–2010 acknowledges that more than half the municipal councillors in rural communes cannot read or write and that their rather advanced age likely precludes them from making much progress in this (Republic of Burkina Faso 2009: 26). Thus, the most effective administration, when it exists, is performed by the secretaries general delegated to the communes by the MATD. These issues affect the quality of local governance. For example, the determination and distribution of land parcels has been one of the most important municipal activities, in part because it represents the greatest opportunity for local revenue, which is crucial both for the functioning of the communes and for the fulfillment of clientelistic obligations. To be eligible to get a parcel, one needs to purchase a card. In many communes,

Pierre Englebert and Nestorine Sangaré

59

however, more cards are sold than the parcels available because of mismanagement, incompetence, and, sometimes, corruption. More positively, municipal economic development planning has registered certain successes, with the design of 191 rural communal development plans as of 2009 and the initiation of the remaining 111. Three regions had also already completed their regional development plans by mid-2009 (Republic of Burkina Faso 2009: 17). However, lack of financing and administrative capacity prevents many communes from beginning to implement their plans.

The Political Economy Understanding the causes behind Burkina Faso’s limited decentralization requires an appreciation of its political economy, with particular attention to the political incentives facing the actors responsible for decentralization. In this section, we outline these incentives and how they have resulted in constraints on the process. Political Incentives and Constraints

Burkina Faso’s authorities embarked on decentralization in 1991 because donors wanted it. Of course, there was a continent-wide movement toward democracy and decentralization at the time, but it was primarily the donors’ agenda that exported these ideas to Africa where many governments responded out of political realism. In this context, it is not surprising that there initially was resistance among some powerholders. Many politicians in the ruling party were anxious about the potential of decentralization to create new loci of power and to strengthen the opposition. They dragged their feet accordingly in designing the necessary legal architecture. Similarly, Burkina Faso’s central administrative services also dragged their feet because they were suspicious of the capacity of local devolved authorities to carry out their task. According to Augustin Loada and René Otayek (1995: 142), staff at the Ministry of Territorial Administration experienced decentralized devolution as a “dispossession.” This situation partly explains why it took fifteen years from the adoption of decentralization in the constitution of 1991 to the first nationwide municipal elections of 2006. None of this is surprising, however, nor did it represent an insurmountable obstacle as subsequent developments have illustrated. Nonetheless, some administrative resistance continues. A more significant and more political reason for the slow pace of decentralization came instead from the upheaval that was unleashed on the national stage with the assassination of journalist Norbert Zongo in 1998, and the subsequent trop, c’est trop

60

Burkina Faso

(“enough is enough”) opposition movement that coalesced in response. The crisis seriously undermined the legitimacy of the Compaoré regime, which was widely believed responsible for the assassination, and put it on the defensive for a few years. Most major political initiatives were suspended while the government focused on staying in power and containing public outrage. As a result, progress on decentralization stalled until 2003. However, while initial resistance to decentralization eased, there is a more insidious constraint to decentralization that has so far largely succeeded in preventing it from being effective and democratic. This fundamental constraint is the neopatrimonial nature of the regime and the monopolistic domination of the ruling CDP party over all organs of the state. The Compaoré-CDP regime is an authoritarian system at its core, opposed to democratic competition and decentralization of power. It is unwilling to share power and authority with nonclients and to surrender control of the political process. By using communes and regions as instruments of patronage and penetration into the countryside, the CDP has reinforced this. Ironically, although the decentralization sequence was slowed by the Zongo crisis, decentralization in the end might have saved the government and restored its social foundations. National-Level Actors and National Politics

There is a certain symbiosis between politics at the national and decentralized levels. The domination of the 2006 local elections by the CDP mimics its domination of national politics.4 As a result, and with the understanding that decentralization does not represent a redistribution of power away from the national incumbents, it is not a particularly controversial national political topic. Simon Ouédraogo, the mayor of Ouagadougou, nicely illustrates the connection between national and decentralized politics. For a long time, he was the chairman of the ODP-MT, the ancestor to the CDP. He was appointed as the ODP-MT candidate for Ouagadougou mayor as a reward for political loyalty. Similar appointments of party barons and other important patrons have occurred throughout the country at different levels of authority. Such practices help explain why there is general support for decentralization within the party. The neopatrimonial nature of the regime, predicated on the handing out of parcels of sovereignty to big clients of the president, who are in turn patrons of their own followers, has found in decentralization a wonderful opportunity to create additional patronage opportunities. It is hard to win in an urban commune against a representative of the ruling regime. And in rural communes, loyal chiefs are in turn rewarded. Hence, there is little resistance to decentralization from within the ranks of power but, at the same time, there is relatively little democrati-

Pierre Englebert and Nestorine Sangaré

61

zation taking place through decentralization. It reproduces and spreads the domination of the ruling party through the country more than it opens new avenues for local governance. The influence of national political parties on local politics prevents the emergence of local parties that might better represent the concerns of the local inhabitants. This pattern undermines the potential of decentralization to promote local democracy. While there is limited political opposition to decentralization at the national level, there remains a certain though not critical amount of administrative opposition, mainly in the form of passive resistance. For example, the different ministries involved in the implementation of decentralization in the CONAD three-year plan failed to budget for their planned interventions, thereby guaranteeing that they would not be implemented. The plan’s appraisal report also notes that deconcentrated personnel worked with limited enthusiasm and had little commitment to the success of decentralization. The report notes a Jacobinist reluctance to adapt to the new institutional environment (Republic of Burkina Faso 2009: 24). As such the report also notes that decentralized ministries, like health and education, continued to display organizational structures (e.g., health districts and circumscriptions for basic education) that did not match the new territorial collectivities. Moreover, the deconcentrated structures had little local accountability, little local authority, and amounted to little more than relay stations for centralized decisionmaking (Republic of Burkina Faso 2009: 24). Finally, CONAD, together with the strategic compact for the implementation of decentralization (CSMOD), marked a significant yet ambiguous push by the national government toward decentralization. After endless conversations on the question of effective transfer of competence, the CSMOD finally set the momentum for the passing of the 2009 decrees on health care, education, water, and sports and culture. Yet at the same time, both CONAD and the CSMOD marked a recentralization of the decentralization process. By reinforcing the leading role of centralized structures (under the direct authority of the prime minister) over the local dynamics of communes and regions, they represent a certain dispossession of these local structures (albeit not an intentional one). As such, they embody the strong centralized tradition of the Burkinabe state and its propensity for top-down approaches. Decentralization has so far done little to erode this. Local-Level Actors and Local Politics

Aside from the deconcentrated and delegated state agencies mentioned above, the other local-level actors that communes interact with are customary authorities. Traditional chiefs are a complex institution in Burkina Faso, and chieftaincy varies across regions. It is common, however, for villages to have a political/administrative chief and a land chief, who typically come

62

Burkina Faso

from different lineages (Laurent 1995: 34–43). Some interesting dynamics with these indigenous authorities have unfolded since the beginning of decentralization. Historically, before the Compaoré regime, chiefs were frequently repressed by the central government, and some appear to have seen in decentralization a possibility to add a legal dimension to their traditional legitimacy to prevent this in the future (Ouédraogo 2003: 101). Thus, some traditional chiefs have run successfully for local office. While this may seem to provide some indigenous representation to the state, it can also result in problems since communes are in charge of people of different ethnic groups whereas the authority of chiefs might be more conscribed to certain communities or domains such as the distribution of land for agriculture and livestock. As a result, local elected officials who are also traditional chiefs have at times found themselves on shaky legal grounds (Ouédraogo 2003: 101). Since many chiefs are co-opted by the ruling party, it is not surprising that they tend to run on the government’s list, and their elections represent as much the local extension of national neopatrimonial politics as the representation of local interests. For example, the Larlé Naba, a top-ranking Mossi chief, was also elected municipal counselor of Ouagadougou in 1995, in addition to his post as member of the National Assembly (Loada and Otayek 1995: 137). The holding of multiple offices (as also occurs in France) is not uncommon in Burkina Faso. Sometimes, however, decentralization perturbs the established way of things at the local level and can undermine or reshuffle existing structures of authority and social relations. This is essentially due to the fact that decentralization empowers communes, which are aggregations of several villages, whereas indigenous authority typically resides at the village level. From this perspective, decentralization itself means in fact a certain degree of centralization. This displacement of authority to the communes involves a loss of control by village hierarchies and populations. This relocation of authority can sometimes clash with local social dynamics. In the west, for example, cultural hamlets formed by migrants from the Mossi Plateau are typically in a subordinate position in the indigenous villages from whom they borrow land. Yet the aggregation of their votes at the commune level dissolves this relationship and reduces accountability among these groups, which can trigger grievances from indigenous communities increasingly couched in a language of autochthony (Hilgers and Jacob 2008: 182–183). If such trends generalize, decentralization may end up being associated with increased social polarization. There is no widespread evidence that this is a large trend so far, but there is new evidence that decentralization has promoted the rise of localized political contention. Harsch has documented 207 reported instances of demonstrations, mass rallies, marches, sit-ins, strikes, and riots in more than

Pierre Englebert and Nestorine Sangaré

63

thirty of Burkina Faso’s forty-nine urban communes. Harsch argues that many of these events contest the “goals and process of decentralization” (2009: 265), particularly in the allocation of urban land titles, corruption, “rivalries among local elites,” and limited democratization. Some of these also pitched indigenous communities against perceived newcomers and challenged the results of the elections or subsequent manipulation in the installation of mayors and their deputies by municipal councillors. As we have seen, much politics (and occasional corruption) takes place at this stage, with the frequent intervention by national political figures and representatives of the CDP to adjudicate disputes among candidates and promote certain individuals. More worryingly, perhaps, Harsch’s (2009) data show that instances of contention have increased over time and peak in years following municipal elections (thirty-two instances in 2001 and forty-nine in 2007). Civil Society Actors and Identity Politics

Burkina Faso has a comparatively weak civil society and has had limited civil society involvement in decentralization (Ouédraogo 2003). But this is not to say that civil society is irrelevant. In fact, the president of the CND, Antoine Raogo Sawadogo, began his public career as president of a large agricultural CSO network called Naam. Nevertheless, it is quite possible that the decentralization process, which was initiated and implemented with little involvement by civil society organizations, may in fact be hurting the interests of CSOs and weakening local civil society. As discussed earlier, decentralization has allowed for the emergence of new political actors (e.g., CVDs and political parties) that might have less capacity than CSOs to promote local interests. These new actors become local institutions without necessarily benefiting from any particular social capital and social trust. There is particularly a certain degree of the competition between CSOs and political parties. The penetration into local development by the new decentralized and deconcentrated institutions of the state has pushed back NGOs. One might expect that decentralization would have multiplied their influence but, in some areas, the opposite seems to have happened. The penetration of political parties at the local level, through municipal elections, has brought politics to the communes and the villages and squeezed out associative life. NGOs shy away from participating in structures that replicate locally the national political game. As such, they might have lost some local influence (Sebahara and Sindogo 2004: 388). In addition to civil society, it is important to note the effects of decentralization on identity. Decentralization can reshape local public arenas in which struggles for power, resources, and influences are waged. Such reshuffles have positive or negative consequences regarding the goals of

64

Burkina Faso

decentralization. One important consequence is that decentralization, by changing the local institutional context, also contributes to reshaping and altering the salience of identity. Although ethnic politics is not as dominant in Burkina Faso as in other African countries, tensions do exist among some ethnic groups, particularly in the west, where recent migratory patterns have pitted indigenous communities like the Bobo, Dioula, and Senoufo against newcomers like the Fulbe (who are seminomadic) and the Mossi, many of whom have acquired land in the west in response to demographic pressures. The increased prerogatives of local authorities under decentralized rule have made the notion of who is indigenous and who is not more important than before, and thus have somewhat increased social polarization among ethnic groups. The Senoufo, in particular, have claimed autochthonous status in some districts. Sten Hagberg suggests that “through decentralized political institutions and structures, politicians have come to use local collective identities in rallying electoral campaigns, leading to situations in which first-comers (or ‘autochthons’) are opposed to latecomers (or ‘strangers’)” (2004: 52). This social dynamic challenges the idea that decentralized political structures are necessarily closer to the people. In multiethnic environments, where ethnicity can represent a powerful factor of mobilization, the decentralization of authority increases the stakes of social competition and allows latent divisions to surface. It can also lead to discourses of autochthony and allochthony (indigenous vs. nonindigenous) through which established elites try to reduce competition for office and resources by declaring some groups legally ineligible. Nigeria’s concept of what it calls “indigeneity” provides a worrisome example of what can lie ahead for countries that follow this trajectory. Hence, it is the liberal notion of citizenship, the very foundation of democracy, that can find itself challenged through decentralization.

Conclusion Despite the problems associated with decentralization, particularly the extent to which it is in fact a centralized process, there is nevertheless a remarkable national momentum in favor of it. For better or worse, decentralization is now fully part of the Burkinabe institutional and political landscape. Yet although it is now well established and heavily legislated, Burkina Faso’s decentralization is also rather limited in scope and effectiveness. The extent to which it promotes democracy, development, or state capacity is unclear. There is little doubt that it has contributed to placating donors and to reducing domestic political tensions, thereby contributing to domestic stability. But such stability might have come at the expense of genuine democratic decentralized governance.

Pierre Englebert and Nestorine Sangaré

65

It is somewhat unreasonable to expect good and democratic local governance in a political system characterized by authoritarianism, patronage, corruption, and factional politics. To a large extent, local governance reflects national politics. As Thomas Bierschenk and Jean-Pierre Olivier de Sardan once wrote in the context of Benin, decentralization brings politics into the village, but lets power seep out of it. While politics is indeed closer to the citizen than it used to be, it is unclear whether this corresponds to any improvement in actual governance in the aggregate. Finally, the link between decentralization and development is at best tenuous in Burkina Faso. Decentralization might actually have disempowered local village development associations, and the lack of local resources makes communal development plans more theoretical than practical. It thus is far from certain whether Burkina Faso’s decentralization balance sheet will come out in favor of democracy and development. An important dimension of Burkina Faso’s decentralization, and one of its main liabilities, remains the imbalance of power between deconcentrated and devolved units. The salience of deconcentration probably reflects a commitment problem by central authority. Decentralization is alright as long as it does not effectively displace power. In practice, the significant parallel structure of deconcentrated institutions is redundant and undermines the substance of devolution. The problem of the effective transfer of power is related to the sequencing of the administrative, fiscal, and political dimensions of decentralization in Burkina Faso. Tulia Falleti (2005) suggests that implementing political and fiscal decentralization first tends to increase the power of local governance structures while early administrative decentralization, which transfers administration and delivery of social services without changing the rule of selection of local authorities and without giving them fiscal autonomy, tends to have the opposite effect. In Burkina Faso, the lack of effective decentralization of power is linked to the effective limits on all three types of decentralization, none of which is complete. But it also derives from the fact that the government has tended to privilege administrative decentralization at the expense of the two other dimensions until recently. The principle of delegation of authority regarding provision of public goods was established as early as 1993. Yet the first nationwide municipal elections (which still do not fully provide for political decentralization) did not take place until 2006, and a serious (albeit partial) step toward fiscal decentralization was not achieved until the adoption of the March 2009 decrees. In addition, the decision to use administrative circumscriptions, such as the department and the region, as units of decentralization is unfortunate because it promotes a dual system that undermines the autonomy of the communes and regions. It probably was well intentioned to have deconcen-

66

Burkina Faso

trated and devolved units at the same level, for the state is supposed to provide support to the communes. But in practice, this has led to a lack of clarity with respect to what each institution is in charge of (reinforced by the lack of financial means of the communes and regions) and a dilution of communes as organs of local governance. At the same time, the most basic unit of social, political, and economic life—the village—finds itself largely excluded from decentralization. Communes include several villages in a top-down process, according to whether they belong to the same department. Contrast this with the experience of Mali where villages were given the opportunity to come together on a voluntary basis to form communes. Villages in Burkina Faso are not without institutions, but they are either traditional (chiefs) or largely deconcentrated (like the CVDs). Thus, the essential building block for development finds itself excluded from the decentralization process, one of the aims of which was to increase local ownership of development. Variations over time in the institutional and legal landscape are also somewhat problematic, as they reduce local ownership of the process and signal wavering government commitment and unpredictability, all of which reduce the incentives to participate. For example, provinces were first territorial collectivities, but then this designation was removed. Villages first had their delegates, then they had CVDs. Communes were first supposed to take over eleven domains of competence, then they effectively got none, and finally received four. It appears the instability of this process is ongoing rather than abating. The entrenched urban bias of Burkina Faso’s decentralization is another major problem. For practical purposes, effective decentralization is largely limited to Ouagadougou and Bobo-Dioulasso, whether in terms of institutional innovations or finances. Other urban communes, while not as advantaged as the two large cities, also benefit disproportionally from decentralization. Meanwhile, the 302 rural communes and their populations continue to be neglected. This bias largely deprives decentralization of its purpose of bringing government closer to the people because the people of Ouagadougou and Bobo-Dioulasso were already closest to government and its services and more influential as well. A few lessons arise from Burkina Faso’s decentralization experience. First, top-down decentralization is probably not an ideal process; neither is gradualism. A more radical transition to decentralized governance earlier in the process, as several donors unsuccessfully pushed for, might have had better effects. A more endogenous process, like Mali’s, also appears more capable of effectively transferring power and accountability to local communities. A second crucial lesson is that decentralization can be only as democratic as the regime that presides over it. The penetration of authoritarian neopatrimonialism into the villages and the hegemony of the CDP in municipal elections have emptied much of the process of its democratic potential. The law should allow for local candidates unrelated to national

Pierre Englebert and Nestorine Sangaré

67

parties or even ban explicit party affiliations in communal elections in order to increase local ownership of governance institutions. Mayors should be elected by universal suffrage rather than by municipal councillors, which is a process that is excessively liable to manipulation. In general, donors should think hard about the usefulness of their efforts toward decentralization if they are unable to encourage or constrain the national regime to democratize. Third, there might be a threshold of income below which decentralization is practically unfeasible. Burkina Faso tests these limits. It might be possible for a country to be too poor to decentralize. If there is insufficient creation of wealth in the villages to offer some basis for taxation, communes will remain underfinanced and at the mercy of the national government. Similarly, local villagers will not develop a sense of ownership of their commune if they do not participate significantly in its financing. At the national level, decentralization also represents a fiscal burden, particularly in the transition period when there might be considerable overlap among old and new agencies. In a country as poor as Burkina Faso, such fiscal burdens imply difficult trade-offs. Finally, it is important to keep in mind the relationship between decentralization and development. Only under certain conditions will decentralization promote development. In Burkina Faso before decentralization, there were provincial cadres de concertation technique provinciaux (technical exchange committees) that brought together all the actors in local development such as technical services from ministries, public administration, and CSOs. With the decline of the province, these committees have disappeared and governors thus far have not successfully recreated these institutions at the regional level. The governor of the Sahel region reported that he deplored the difficulties in bringing together these different actors. He mentioned existing conflicts between mayors and department prefects, whose roles in development are not adequately differentiated by law. In this case, decentralization might have undermined local state capacity because previous institutions were disassembled and new and potentially competing ones were created. In short, the problem lies in the fact that the new institutional architecture does not put local development—and local ownership of development—at its core. Consequently, ad hoc solutions are invented, and thus far these do not perform as well as the structures that existed beforehand.

Notes 1. The succession of small steps implied by gradualism reflects in part a desire to learn and adapt on the part of implementing authorities. However, it can also be a convenient excuse for a tepid rate of implementation and an occasional lack of commitment.

68

Burkina Faso

2. The following lists are not exhaustive and include only the most or important responsibilities of communes. 3. See Articles 236–239 of the Electoral Code of 2001 (in French: LOI N°0142001/An Portant Code Electoral). 4. The CDP’s domination is probably greater even at the communal level. With some 70 percent of the votes in 2006, the CDP won almost 90 percent of the mayorships.

4 Ethiopia: Ethnic Federalism and Centripetal Forces J. Tyler Dickovick and Tegegne Gebre-Egziabher

Ethiopia since 1991 is a crucial case study in African decentralization. The country is characterized by ethnic federalism in which the largest ethnicities (nations, nationalities, and peoples in the Ethiopian context) each have their own regional states with substantial autonomy, at least constitutionally. Emerging from civil conflict in the 1990s that toppled a MarxistLeninist dictatorship, this ethnic federalism is a unique and striking experiment on the African continent, a big push in the political, fiscal, and administrative dimensions that devolved power and authority to elected subnational officials. Decentralization in Ethiopia continued after the 1990s devolution to the regional states. In the early 2000s, a second big phase of decentralization was undertaken with the transfer of power and resources to the woredas (districts) in the so-called District Level Decentralization Program (DLDP). This resulted in an additional devolution of authority, substantial fiscal transfers, and a furthering of electoral accountability to local populations. Despite these achievements, significant constraints remain for subnational governments, most notably in three areas: limited tax bases and ownsource revenues; top-down partisan control by the dominant Ethiopian People’s Revolutionary Democratic Front (EPRDF) party; and low capacity at subnational levels. Together, these convey an image of a decentralization that is quite robust de jure and, while considerable in African comparative perspective, more limited de facto. In the terms used through this volume, the country has devolved significant authority, but SNGs are more circumscribed in the areas of autonomy, downward accountability, and capacity. It is in the formal balance of power between levels of government that Ethiopia has advanced farthest; we elaborate on the significant limitations to decentralization in greater detail below. 69

70

Ethiopia

The Origins of Decentralization Decentralization in Ethiopia began with the establishment of an ethnofederal state codified in the 1995 Constitution. Prior to this, in 1991 an agglomeration of armed resistance movements toppled the Marxist-Leninist regime known as the Derg (1974–1991). Power consolidated around the EPRDF, a coalition of movements led by the Tigray People’s Liberation Front (TPLF) and the Eritrean People’s Liberation Front (EPLF); the TPLF became the dominant member of the coalition after Eritrea claimed independence in 1993. The establishment of Ethiopia’s new regime began with the premise that the country has a multiethnic (or multinational) structure, which translated into strong constitutional guarantees for the respective ethnicities (or nations). Demography—particularly, ethnic demography—is the foundation for the ethnic federal regime and the devolution of power. Ethiopia is comprised of a multitude of nations.1 With the establishment of the federal regime in the 1990s, the five largest of these were each allotted their own regional state, the largest being Oromia (the state of the Oromo people) with over 27 million people and 37 percent of the population (Garcia and Rajkumar 2008: 20–22). The second-largest regional state is Amhara, with about 23 percent of the population. The regional states of Tigray, Somali, and Afar each correspond to a specific nationality while the four remaining states of Ethiopia are multiethnic in character: the small states of Gambella, Harari, and Benishangul-Gumuz (each of which represents less than 1 percent of the population) and the large state known as the Southern Nations, Nationalities, and People’s Region (SNNPR), which comprises about twenty percent of the population (see Table 4.1). Ethiopia’s experiment in ethnic federalism was intended to address the considerable tension between the country’s nationalities. Political divides were historically most acute between the Amhara peoples, the Oromo, and the Tigray. The tension dates to the early periods of state formation in present-day Ethiopia under Menelik II in the late nineteenth century, and persisted through the regime of Emperor Haile Selassie I (1930–1974).2 Of additional demographic significance for federalism and decentralization is the country’s sheer population size. In Africa, Ethiopia is now second only to Nigeria in population, and continues to grow rapidly. As with other large federal countries, Ethiopia’s population size gives rise to multiple layers of SNG, both politically autonomous or devolved units (the regional states and woredas) and deconcentrated units (the zones in most states, although the SNNPR has zones that are devolved units). Historical legacies—especially the centralizing efforts of the Haile Selassie and Derg regimes of the twentieth century—informed the debate about federalism and decentralization in Ethiopia in the 1990s. Ethiopia is

J. Tyler Dickovick and Tegegne Gebre-Egziabher

71

Table 4.1 Distribution of Population by Region, 2007 Region Tigray Afar Amhara Oromiya Somali Benishangul-Gumuz SNNPR Gambella Harari Addis Ababa Dire Dawa Special enumeration Total

Population 4,314,456 1,411,092 17,214,058 27,158,471 4,439,147 670,847 15,042,531 306,916 183,344 2,738,248 342,827 96,570 73,918,505

Percentage of Population 5.8 1.9 23.3 36.7 6.0 0.9 20.4 0.4 0.2 3.7 0.5 0.1 100.0

Source: Central Statistical Agency (2007). Note: SNNPR is Southern Nations, Nationalities, and People’s Region.

the only African country that was not successfully colonized by a European power; even the brief and brutal Italian occupation of 1936–1941 did not result in the construction of a meaningful colonial state. Rather, Ethiopia in the nineteenth and twentieth centuries was governed by indigenous state builders who crafted political structures that brought the country increasingly under centralized control, ending a long period of warlordism by local princes and elites. Up through the Haile Selassie era from 1930 to 1974, top-down centralized government coexisted with strong regional elites in a system often considered akin to feudalism. The period included some limited administrative decentralization enacted between 1942 and 1966, which created three subcentral levels of government mainly to facilitate central administration in performing the functions of local government at lower levels (Tegegne Teka 1985). The Derg instituted top-down governance, developing a penetrating state and a militarized Marxist-Leninist vanguard party. At the local level, the Derg established rural peasant associations and kebeles (urban dwellers associations), grassroots extensions of the formal government structure in rural and urban areas, as well as village development committees. The regime nationalized industry, collectivized agriculture, and forced villagization, the destructive consequences of which were epitomized by the 1980s famines. When the Derg came under pressure from insurgencies in the late 1980s, the regime attempted a hasty and haphazard decentralization to the embattled regions of Eritrea, Tigray, and the Ogaden, but these were rejected summarily by the newly emboldened rebel groups (Meheret 2007: 75).

72

Ethiopia

Contemporary Ethiopia decentralized on top of a unique set of preexisting conditions, relative to the rest of Africa: a feudal legacy, the lack of an externally imposed colonial state, and the decaying Marxist-Leninist party structures and village collectives. The 1995 Constitution passed by the EPRDF sought a complete break with the Derg, most notably with the rejection of centralized power in favor of decentralized governance under ethnic federalism. The creation of regional states, along with the constitution’s unprecedented right to secession (elaborated on below), was an attempt to manage national political divisions by accommodating the various nationalities. Notwithstanding the changes in ideology and in the creation of federalism, however, the EPRDF after 1995 replicated certain of the top-down characteristics of the Derg. As discussed below, power has not been dispersed as much as the Ethiopian constitution might imply, largely because of continued de facto centralized power in the hands of the dominant party. Decentralization began after 1991. With Prime Minister Meles Zenawi (1995–2012) at its head, the TPLF became the leading organization in regime transition, in the drafting of the new constitution, and in the EPRDF coalition government that followed. Meanwhile, under Isiaias Afwerki, the EPLF founded modern Eritrea, splitting off from Ethiopia in 1993. Initially, the TPLF sought regional decentralization as a means to hold together the peoples of Ethiopia and to begin the process of state reconstruction. However, the EPRDF quickly developed top-down patterns of governance—led by the TPLF—that pushed toward a more centralized federation (Keller 2002). The TPLF leaders in the coalition increasingly sidelined nationalist movements, most notably the Oromo Liberation Front, which boycotted the first national elections after the Derg’s fall. As the TPLF consolidated its power in the new government, it created parties and organizations under the EPRDF umbrella to channel demands from the different nations. These included the Oromo People’s Democratic Organization (OPDO) and the Amhara National Democratic Movement (ANDM), the two parties that won the most seats in the coalition in 2000. Ethiopia’s regional decentralization after 1991 was a big push or big bang decentralization, with a dramatic commitment to devolution instead of deconcentration or delegation. The changes included political decentralization with the election of subnational politicians; fiscal decentralization with the transfer of a significant portion of the national budget to the regions (starting in 1992), and the transfer of expenditures in major social service areas such as education, health care, and agriculture; and administrative decentralization with the transfer of planning and administrative responsibilities in those expenditure areas. Resources and responsibilities totaling about 40 percent of the government budget were transferred to the regions in short order (cf. Garcia and Rajkumar 2008: 27). Up to the present, the

J. Tyler Dickovick and Tegegne Gebre-Egziabher

73

share of total spending at this level is estimated at about 38 percent of total government expenditure, and about 45 percent or more of total nondiscretionary expenditure (Garcia and Rajkumar 2008: 31). The big push decentralization has led some regions to lag behind others in the implementation of decentralized functions. In particular, some regions (i.e., Gambella, Benishangul-Gumuz, Afar, and Somali) have shown capacity gaps in budget utilization, planning, and budgeting and implementation of development activities due to a relative lack of preparedness and institutional infrastructure in these regions.3 Woreda decentralization followed in 2001–2002 from the prior regional decentralization. It was also nearly simultaneous across the political, fiscal, and administrative domains in each regional state, but it was asymmetrical with the first adoption coming in the country’s three most populous states— Oromia, Amhara, and the SNNPR—along with Tigray (which may be fourth in population among the states, being roughly the same size as Somali). This decentralization was then followed in the remaining states (and two city administrations) in the subsequent years. In all states, finance roughly followed function, with revenues and expenditures being transferred in parallel along with administrative and planning authority.

Authority: The Legal Framework of Robust Federalism The current legal framework for decentralization—the basis of the legal authority for the various levels of government—was established in the 1995 Constitution. There are five major layers of government for which we outline the basis of authority: the federal government, regional states, zones, woredas, and kebeles. Federal Government

As noted above, several of the key structures of Ethiopia’s federal government are predicated on the notion of ethnic federalism. The national-level institutions of particular importance for decentralization are the structure of legislative-executive relations (including electoral rules for legislators) and the structure of the political party system. With respect to legislative-executive relations, Ethiopia’s legislature is uncommon in Africa because it is parliamentary, with the prime minister serving as the head of government and a ceremonial president selected as head of state. Parliament is bicameral, with a House of People’s Representatives (lower chamber) elected in single-seat constituencies from districts and the House of Federation (upper chamber) selected by the legislatures of the regional states. This arrangement gives a territorial basis for representation in both chambers. The lower chamber of

74

Ethiopia

the national parliament (the highest political organ) does not involve itself in regional and local affairs; it is limited to legislating laws of national importance and overseeing the national executives and national departments. The House of Federation, on the other hand, can be regarded as a “council of nationalities.” It has a mandate to determine the allocation of intergovernmental transfers to different regions, among other functions. In the process of allocating transfers, the House of Federation consults with regional governments and receives a final proposed allocation from a standing committee that is comprised of the heads of regional governments. The allocation formula is ultimately approved by members of the House of Federation. Regional States

Ethiopian federalism is based on eleven regional-level governments: nine regional governments (each with their own constitutions) and two city administrations for Addis Ababa and Dire Dawa that are accountable to the federal cabinet largely for demographic reasons (most notably their large population sizes, but their diversity in terms of nationalities and ethnicities). These regional governments have considerable constitutional protections. The most unique of these is Article 39, which guarantees Ethiopia’s regions a right to secede, albeit on completion of a substantial political process that includes supermajorities in regional legislative action and a subsequent referendum, followed by approval from the House of Federation. The constitution also instituted a reserve power clause that ensured residual powers for the regional states. Thus, there is a significant vertical division of authority between the center and the regions, and regional authority is robust on paper. These powers involve enacting state constitutions; establishing state administration; formulating and executing economic, social, and development policies; administering land and natural resources; levying and collecting taxes and duties; enacting and enforcing the state civil service; and establishing and administering the state police force. Regional governments also have significant fiscal authority, both in expenditure responsibility and revenue collection powers, with regions having their own independent revenues as well as shared revenues used jointly by the federal and regional government. Zones

Below the regional states and above the elected district level are administrative units known as zones. These zones are appointed by the regional governments in most cases, with the exception of one state—the SNNPR— where elections are held for zonal cabinets (see Tegegne Gebre-Egziabher and Kassahun 2007). In addition, three zones in the Amhara region and all

J. Tyler Dickovick and Tegegne Gebre-Egziabher

75

three zones in the Gambella region have elected councils. At this level, we thus can observe some degree of deconcentration of authority from the regions to the zones. Zones are responsible for coordinating among the woredas and monitoring woreda compliance with regional prerogatives; as such, “in most regions, they have become the arms of the regional government” (Garcia and Rajkumar 2008: 13). Woredas

Woredas are the essence of Ethiopia’s second wave of decentralization since 2002. These elected local governments number approximately 600 or more (with woredas periodically subdividing into multiple units).4 Woredas are not indicated in the federal constitution, but they have authority that emanates from each of the respective regional constitutions. Constitutional provisions vary modestly across regional states, but in each the woredas are the key autonomous self-governing units to which power and resources are devolved. While rules vary from region to region, members of woreda councils are elected by the citizens of the district, and these councils select their executive cabinets. Woredas are also well ensconced in ordinary law and are on solid institutional footing, having been given substantial powers nationwide with the DLDP beginning in 2001. In general, woreda administrations are responsible for economic development and social service plans. Woreda fiscal authority is also defined by each regional constitution. On the revenue side, woredas collect revenue on behalf of the regional government. The locally collected revenues are used to cover expenditures in the woreda, though woredas may also receive transfers from the regional government to cover expenditures necessary to fulfill their allocated budget. Because woredas act as delegated units collecting revenue on behalf of regional governments, they technically depend on regional government allocations for all of their expenditures (though a portion of this is raised and collected within and by the woreda). This is different from the fiscal authority of regional governments, in which regions have autonomy in the use of their own-source revenue while also being eligible to receive additional intergovernmental fiscal transfers. Woredas thus have relatively less fiscal autonomy than regional governments by virtue of their locally collected revenues, and how they are spent, are subject to regional authority. Woredas are also accountable for implementing directives from above (from the regional state or the zones). On the expenditure side, woredas are responsible for a range of local public goods, including local police, roads, utilities, water and sewerage, street lighting, and sanitation (Tegegne Gebre-Egziabher and Kassahun 2007). They are also responsible for a range of activities in the public services that have been decentralized since

76

Ethiopia

2001: education, health, and supporting water services. This includes hiring and paying teachers and health staff, coordinating primary health care, and constructing and maintaining local health clinics (Garcia and Rajkumar 2008: 16–18). Kebeles

Below the woredas are kebele administrations, the grassroots governance units, serving as local-level institutions in both rural and urban areas. Intended for a population numbering about 5,000, they are governed by approximately 100 (or more) council members. This equates to a number of seats estimated at around 3.6 million. The stunning number of seats means that there is a local councilor for about every twenty Ethiopians and, as discussed below, this has significant implications for politics given the partisan composition of council members. Along with elected local councils, the kebeles have executive cabinets that coordinate planning and administrative functions; these are selected by the members of the local councils. They respond to the woredas’ directives and rely on the woredas for their budgets (Tegegne Gebre-Egziabher and Kassahun 2007: 15). The EPRDF dominates politics across these various levels of government. It won 499 out of 547 seats in the House of People’s Representatives in the 2010 general election, with allied parties winning all but two of the remaining seats. This gives the EPRDF a voting bloc that controls greater than 99 percent of seats in the legislature as of 2014. In recent elections, the EPRDF also prevailed in 1,903 out of 1,904 regional council seats. It also dominated the 2008 subnational elections, running essentially without opposition at the woreda level and winning more than 3.5 million out of approximately 3.6 million kebele seats. In fact, the EPRDF dominance in the 2008 local elections was seen by some as a precursor to the 2010 general election. Given that the 2005 general election had witnessed some decline in EPRDF dominance, the 2008 local elections presented a preliminary opportunity for the EPRDF to reconsolidate its power (Aalen and Tronvoll 2008).

Autonomy: De Facto Limitations on SNG While de jure authority is well established, de facto subnational autonomy is more restricted. This can be judged by looking at three areas: political autonomy, fiscal autonomy, and administrative autonomy. We begin with political autonomy. As our discussion of authority suggests, political power in Ethiopia is more decentralized than in most African countries. The political system nominally guarantees considerable autonomy to the subnational

J. Tyler Dickovick and Tegegne Gebre-Egziabher

77

levels, but this formal autonomy is offset by the dominance of the EPRDF, which dominates politics at all levels. This dominance up and down the system may have the positive effect of facilitating communication vertically across different levels, but it also presents an unofficial mechanism through which national level leaders can exert influence over decisionmaking at subnational levels (see Dom and Lister 2010: 36; Vaughan 2011). Regional governments have fiscal authority, but the degree to which regional governments (and woredas) have fiscal autonomy is more circumscribed. The distinction here comes down to the sources of subnational revenues, particularly that between intergovernmental transfers and own-source revenues that SNGs collect on their own tax bases. Fiscal decentralization in Ethiopia is substantial by African standards, if one takes into account intergovernmental transfers (e.g., block grants and conditional special purpose grants) as well as own-source revenues. In addition, expenditures are quite decentralized, with regions becoming responsible for as much as 37 percent of total expenditures as early as 1994, and increasing to well over 40 percent thereafter (Tegegne Gebre-Egziabher and Kassahun 2007: 16–19; Garcia and Rajkumar 2008). Moreover, revenue distributions are formula based and thus largely transparent. Grants from the federal government to the regional states as reported were based for some years on four indicators, in declining order of importance: population (accounting for about half of the formula), then development indicators, regional tax effort, and finally poverty indicators (Tegegne Gebre-Egziabher and Kassahun 2007: 21). In recent years, however, the budget allocation has been based on assessment of the expenditure needs and revenue potential of each region. For transfers from the regional state to woreda level, the formulas used differ from one location to another, but some regions employ the same expenditure need approach. While this is considerable, the limits to the fiscal autonomy of Ethiopia’s SNGs begin with the limited tax base facing many regions and woredas, especially in rural areas. Along with the low levels of private sector economic activity in some areas, another reason is the assignment of lower-yielding and narrower tax bases to regional governments. This leads to a heavy reliance (by all levels of SNG) on intergovernmental fiscal transfers. In the 1990s, SNGs (under the aegis of the regions) collected an average of 16–17 percent of total government revenue, with the center raising the remaining 83–84 percent (Keller and Smith 2005: 273). For many observers, the central problem of Ethiopia’s fiscal federalism is that expenditures are so decentralized while revenue collection is centralized (Meheret 2007). Centralized revenue collection in turn gives the central government leverage over subnational expenditures. Transfers from the federal level to the regional states cover from 45 percent to 80 percent of the regions’ expenditure assignments. Such high dependency of regional governments on the federal government reduces the autonomy of the regional states and compro-

78

Ethiopia

mises their decisionmaking capability. At the woreda level the situation is even more imbalanced, given the woredas’ reliance on regional governments. As a rule, woreda finance passed on from the regional state is insufficient to meet expenditure requirements. Accordingly, the woredas may seek revenues from other sources, including foreign assistance or earmarked central government funds. The reliance on regional state and external funding compromises local autonomy because the financing arrangement requires woredas to respond to regional government directives or to depend on ad hoc special transfers that may be increasingly conditioned on certain actions set by the higher level (Dom and Lister 2010: 36). The levers of political party discipline and agreements of SNGs to adhere to national policy frameworks mean that this leaves autonomy significantly reduced (Vaughan 2011: 67). It should also be noted that the budgets transferred are generally used for recurrent expenditure, with nearly 91 percent of woreda budgets earmarked for administrative and operational expenditures (Meheret 2007). This leaves little room for discretionary capital budgeting at the woreda level. Most capital activities at the woreda level are financed from donors and aid agencies or from specific purpose grants from the center such as food security finance. Reliance on donors and aid agencies to undertake capital activities raises issues of sustainability. And reliance on specific purpose grants to undertake specific development activities limits the woredas’ abilities to address overall development outcomes; these grants reduce power to make discretionary expenditure decisions as a result of budget guidelines required by regional governments (CIDA 2005). Subnational units in Ethiopia are established as administratively autonomous and are empowered to approve budgets and to prepare and implement development plans. According to Serdar Yilmaz and Varsha Venugopal (2008), local governments can regulate land use, cooperatives, and community activities while municipal local government can regulate markets, sanitary services, and other municipal services. Local governments also have the autonomy to enforce regulation by sanctioning and punishing citizens in elected social courts and woreda courts; Woreda governments have discretionary power over tendering and administrations of small projects usually less than ETB100,000 (100,000 Ethiopian birr). For larger projects, however, the woreda defers to zonal and regional governments. As noted by Yonatan Fessha and Coel Kirkby (2008: 259), Ethiopia’s regions have a civil service that is separate from the national civil service. This enables the regions to hire and fire personnel, including the large number of teachers each has under its purview. Civil servants such as teachers, sector heads, health workers, and development agents are hired by woreda governments (though the practice varies from region to region). In Amhara and Oromia woredas play a central role in recruiting and hiring of civil servants, but in Afar the woredas send their needs to the regional government.

J. Tyler Dickovick and Tegegne Gebre-Egziabher

79

In Ethiopia, regional governments have had autonomy in planning and budgeting since the first phase of decentralization. As this autonomy was limited to regions, woredas suffered from inability to autonomously plan and budget until the implementation of the DLDP in 2002. Subsequent to the DLDP, regions devolved considerable portions of their experienced manpower (along with financial resources in the form of block grants) to woredas by scaling down the role and resources of the zones, which hitherto served as intermediate tiers of administration linking woredas to regions (Kassahun and Tegegne 2004). The significance of this is that woredas are responsible for their planning and implementation within the bounds of resources available to them without seeking authorization from higher levels of administration. This has marked significant improvement in the exercising of autonomy of planning and budgeting at the local level. Case studies in Tigray and Amhara woredas in 2004 have shown that the finance and economic development offices of woredas are responsible to prepare woreda plans (Kassahun and Tegegne 2004). These plans are based on sector plans prepared by different sectors (e.g., health, education). The finance and economic development offices prioritize sector plans and checks consistency with regional policy. The woredas thus prepare plans autonomously, though they utilize the regional policy framework to ensure consistency. Subnational autonomy in Ethiopia therefore presents a mixed bag and, in this area, Ethiopia’s massive decentralization looks more ambiguous. EPRDF dominance and the party’s structure call into question political autonomy while limitations on subnational fiscal autonomy—especially in the form of limited own-source revenues and the resulting need to follow central directives in many areas of expenditure—hinder the ability of SNGs to target public services to the specific needs of each area. Given limited own-source revenues, and the need to dedicate transfers for recurrent expenditure, decentralization results in less fiscal autonomy than the highly decentralized authority would suggest. That the influence of the center is greater de facto than it is de jure is a theme that reemerges in the discussion of accountability.

Accountability: Downward and Upward How decentralization works depends on patterns of accountability: who is accountable to whom? In principle there is a range of accountability mechanisms such as administrative processes (e.g., consultation with technical staff, report cards), formal political processes (e.g., voting in elections, consultation with elected officials), and use of civic media (e.g., local press and radio, public forums organized by nongovernmental organizations).

80

Ethiopia

At the various levels of SNG in Ethiopia, dual accountability exists. This takes the form of upward accountability of SNGs to higher levels of government and downward accountability of SNG officials to electors. All SNGs have popularly elected local councils that are nominally accountable to the electorate through elections every five years. There is also a system of horizontal accountability between branches of government. Woreda councils have standing committees to oversee sector activities. This practice is also evident at the federal level where the prime minister, as chief of the executive, and various sector ministers face the parliament, which assesses their performance on the basis of reports. The oversight functions of councils, however, are compromised because of conflict of interest, as cabinet members in all woredas are also members of the councils (Yilmaz and Venugopal 2008). At the woreda level, in addition to representative accountability, there is administrative and service delivery accountability (CIDA 2005). This accountability refers to woreda sector offices (e.g., health, agriculture, education) that are headed by appointed professional civil servants. The main problem here is the blurring of the distinction between the elected and appointed personnel. At the woreda level, those who run sector offices are both elected and appointed officials. Public scrutiny remains one of the most important means of ensuring accountability, but this social accountability is not a well-developed area in the Ethiopian decentralization schemes. The tradition of scrutinizing plans, budgets, and other documents is not instituted in the decentralization drive. A recent citizen report card survey conducted in seven regions in twenty-one rural woredas showed that 87 percent of the households have no information on the woreda-level government budget, 83 percent of the households have no information on strategic plan of the local governments, 85 percent have no information about the agenda of their local governments, and 84 percent are not informed about the decisions made by woreda council (Ministry of Capacity Building and Ministry of Works and Urban Development 2008). This is a clear indication of the lack of accountability and transparency on the part of woreda functionaries to their constituencies. Though budgets are published in Negarit Gazettas (official journals or gazettes), the levels of literacy and awareness of the public remain obstacles for the people to scrutinize and hold the government accountable. Mechanisms such as participatory budgeting also are not well instituted in the country. Political competitiveness is constrained at the local level as well. Party politics at the local level is critical and permeates many aspects of woreda administration, including appointments and recruitments to local offices (Meheret 2007). The fact that the EPRDF dominates the political scene narrows the political sphere, leaving little space for programmatic alternatives. Local officials are loyal to party leaders rather than to the populace, which

J. Tyler Dickovick and Tegegne Gebre-Egziabher

81

can reduce the incentive to address local development needs. There is a gulf between the constitutional provision that allows a multiparty system and the de facto exercise of this system since the ruling party controls the executive and legislative branches of the national and regional governments (Yilmaz and Venugopal 2008). There are still other systems of accountability in the country. These include a national capacity-building program; the system of Gemgema (ongoing evaluation) that is mainly used for intraparty accountability within the ruling party; and the formal grievance-processing mechanisms installed in different sectors or offices at various levels of government (CIDA 2005).

Capacity: Constraints and Challenges There is a clear capacity gap in managing decentralized responsibilities in Ethiopia. Apart from financial and budgetary problems, woredas lack competent and trained personnel. A first challenge is the availability of required number of staff in a particular situation. A study by Meheret (2007) indicated that many positions remain vacant in the woredas, though there are some variations among woredas; this has serious implications for local service delivery and local economic development. This is due to the lack of adequate budgets to hire new staff members and low remuneration that makes woreda employment undesirable. This means the woredas lack competitiveness in service delivery and economic development. The second aspect of the human resources challenge involves the capability of woreda officials in administrative, oversight, and planning functions. The administrative capacity of these officials is worrisome at best. Local government officials are popularly elected and owe their election largely to party allegiance; they are recruited and promoted by party structure. These individuals, however, are expected to lead the total development activities and policymaking process. One study showed that woreda personnel hold diploma and certificate qualifications, but many are former elementary or secondary school teachers with little or no formal training in public service delivery, good governance, or economic development (Meheret 2007). Another study also indicated the lack of capacity on the part of councilors at local level to review the planning, budgeting, and service delivery functions (Yilmaz and Venugopal 2008). An area related to planning capacity at the local level is the capacity to track public expenditure. One of the activities that woredas perform is to track public expenditures to ensure that they are made in conformity with approved budget plans. Woredas thus deploy internal auditors to examine the books of account of the sector offices, which are expected to provide a clear picture of expenses incurred in view of approved budget lines for

82

Ethiopia

specified activities (Kassahun and Tegegne 2004). Tracking public expenditures associated with the capital budget, however, is often too complex to be handled by the expertise available to the woredas since capital expenditure involves technical skills to determine the amount of money spent on a specific aspect of a project. Zonal experts (e.g., engineers and surveyors) assist the woredas in supervising whether activities are performed according to plans and standards, and determine whether the amount of money requested by private contractors for their services is commensurate with plans and budget (Kassahun and Tegegne 2004). The capacity gap of local government is therefore pervasive and compromises the functions of local government in service delivery, regulation, and oversight. A national survey in 2004 showed that about 56.3 percent of experts indicated that local government has poor capacity to manage the various decentralized responsibilities while another 17.5 percent believed that they had no capacity at all. About 22.5 percent suggested that the local government has limited capacity while only about 4 percent felt that the local government has some capacity (Tegegne Gebre-Egziabher et al. 2004). The central government has attempted to train local officials in centrally run formal institutions and also has organized various symposiums and seminars (Tegegne Gebre-Egziabher et al. 2004). These include governmentlaunched training programs for the elected officials in woredas, zones, and regions, the modules of which include agricultural development, basic management, financial management, integrated rural development, and ethics (Asmelash 2000). The National Capacity Building Program targets fourteen sectors, including civil service reform, tax reform, agricultural and industrial training, civil society, and higher education.

Political Economy of Intergovernmental Relations: The Dominant-Party State The reasons for decentralization in Ethiopia can be understood as a set of efforts to ensure governability and facilitate a process of state reform and democratization, combined with incentives to enhance partisan leverage. What was good for stability and for decentralization in the 1990s was also good for the party. Many of the central political incentives and constraints behind decentralization are found in the political party system, particularly in the dominance of the EPRDF. The power of the central party in the decentralized federation was such that the system originally considered a “holding-together” federation designed to accommodate Ethiopia’s peoples came to be understood by some as a top-down “putting-together” federation (Keller and Smith 2005: 269; cf. Stepan 1999). As noted above, there were incentives for the EPRDF to accommodate ethnoregional interests in the

J. Tyler Dickovick and Tegegne Gebre-Egziabher

83

first stage of regional state decentralization but, as the party consolidated power in political life over the period from 1995 to the present, it has had reason to move systematically toward assertions of central control. In light of its partisan control at all levels of the political system, the EPRDF has mixed motivation with respect to decentralization. In many instances, dominant parties do not have incentives to fragment or fracture their grip on central power through decentralization. While decentralizing could indeed result in such dispersals of power, the extension of political authority to subnational levels has also accommodated large numbers of EPRDF party activists and clearly has deepened the political linkages between the party and local populations. Put affirmatively, this could enable effective channeling of demands from local levels to increasingly higher levels of government (Meheret 2007: 89). But the structure of the party itself points in the direction of a greater degree of state control, clientelism, and “top-down governance in which decentralized units perform more as transmission belts for central prerogatives than for upward transmission of local demands” (Riedl and Dickovick 2013). The political goal for the EPRDF thus may be to continue to deepen the existing decentralization of governance, partly in the interest of improved service provision, but not necessarily in the interest of more bottom-up democracy. At the national level, the key institutional feature shaping decentralization is the structure of the EPRDF. Despite the features of the constitution that apparently empower the many nations of Ethiopia, several observers have found that this devolution of power has been significantly undercut by partisan realities (Chanie 2007; Keller 2002; Keller and Smith 2005; Kidane 2001). For several of these analysts, the near-single-party dominance of the EPRDF has gutted the idea of democratic federalism. In this vein, Edmond Keller (2002: 31) argues that Ethiopia’s structure was originally conceived not as a “coming-together” federation of peoples from the bottom-up, but rather as a top-down “holding-together” federation designed by the EPRDF to maintain some degree of unity in the multinational state. This lack of devolved power was exacerbated when the country shifted to become a “putting-together” federation led by a dominant political clique, in what might also be termed “federation from above” (cf. Keller and Smith 2005: 268). The decision to promote woreda decentralization was less motivated by questions of ethnic identity, and more by the motivation to extend governing authority down to local levels. Woreda decentralization has had several stated aims, including improvements in public service delivery and more participatory governance at the local level. It should be noted, however, that woreda decentralization also has had profoundly political consequences: as Green (2011) notes, the decentralization process, combined with the centralized structure of the EPRDF, has consolidated the party in power.

84

Ethiopia

At both the regional and woreda levels, the decentralization process thus has a mix of worthy objectives and political consequences that may help explain the motivation to decentralize. In this context, it is useful to consider the sequential relationship between these processes and the larger macropolitical environment, particularly with regard to considerations in the EPRDF. Regional decentralization took place simultaneously with the end of the insurgency against the Derg and the establishment of the current EPRDF regime. That is, regional decentralization went hand-in-hand with large-scale initiatives to stabilize the polity, create order, reconstruct the state (on an ethnic federal basis), and reform the economy (Riedl and Dickovick 2013). The process occurred as a party (TPLF) and the coalition it led (EPRDF) was attempting to establish power. By contrast, woreda decentralization took place in the context of a stable regime in which the coalition had a clear governing majority coming out of the 2000 elections. Whether one takes the stated aims as given or assumes a narrow political calculus by the TPLF, the difference in the macropolitical sequences meant that the motivation for decentralization turned from questions of basic stability with regional decentralization to questions of the political and economic reach of the state with woreda decentralization. The EPRDF has integrated vertically in recent years. At the state level, it has transitioned from being a loose coalition of regional parties—led by the TPLF and allied parties in Oromia (OPDO), Amhara (ANDM), and the SNNPR—to a more cohesive political force in which the major regions are now governed by EPRDF-allied parties. At the woreda and kebele level, concerns about Ethiopia’s dominant-party state are replicated and even heightened. In Ethiopia’s system, local-level administration is characterized by massive staffing and large numbers of elected positions. In 2008, local elections were held for approximately 3.6 million local administrative seats; the result was an overwhelming victory for the EPRDF. The EPRDF touts extensive state penetration as a way to ensure participatory democracy, but the large number of positions is also an advantage for the EPRDF, insofar as it is the only party capable of fielding so many candidates across the country (Aalen and Tronvoll 2008: 116). This has made Ethiopia virtually a one-party state—with remarkable reach—at the local level (Riedl and Dickovick 2013). As noted in the section Accountability: Downward and Upward, officials in Ethiopia’s SNGs (region and woreda) have dual accountability to local electorates and to higher levels of government. With respect to higher levels of government, there are several sources of accountability from above that shape the actions of SNGs. These include formal and legal accountability (as specified by law) and fiscal accountability (through some degree of earmarking and guidance on expenditures). For some, such as Paulos Chanie (2007), this is further rounded out by partisan accountability in the patron-client relations established in the EPRDF.

J. Tyler Dickovick and Tegegne Gebre-Egziabher

85

Bureaucratic politics at the ministerial level is an additional complicating factor, with some (though not all) ministries and central government officials resisting or otherwise hindering the decentralization process. This can be both a political incentive to retain power in some circumstances and a lack of capacity in other circumstances. Among the former, there were reports that as of 2005 the Ministry of Education, for instance, exhibited “some resistance to the notion of devolution” (Watson and Yohannes 2005: 15). Among the ministries that may support decentralization actively, yet have shown limited effectiveness in doing so, is the Ministry of Capacity Building, which David Watson and Lissane Yohannes (2005: 15) report is focused on national capacity building and on provision of direct technical assistance rather than on building subnational capacity in a self-reinforcing fashion. Similarly, as noted above, the Ministry of Finance and Economic Development has lagged in the provision of fiscal resources to subnational levels, even as the ministry remains an active player in the decentralization process. There is no clear evidence that the ministry is deliberately slowing the pace of decentralization; rather, we observe simply that the implementation of fiscal decentralization is likely to be challenging in resource-poor environments. A final noteworthy feature at the subnational level is the overlapping of geographic decentralization with ethnic decentralization. As noted above, identity-based and ethnic decentralization was primarily installed at the regional state level whereas the more local subnational entities—namely, zones, woredas, and kebeles—are merely geographical entities created on the basis of administrative expediency. However, there are some exceptions to the rule in which zones and woredas follow an ethnic logic. These are the multinational SNNPR state where zones and special woredas are created explicitly to accommodate ethnic diversity; the Amhara region where there are three ethnic zones, namely, the Awi zone dominated by the Agew people and the Oromiya zone dominated by the Oromo speakers; and the Gambella region, which has three ethnic zones.

Conclusion Decentralization in Ethiopia engenders a range of conflicting interpretations, with certain aspects of the decentralization process receiving considerable praise and other aspects being incomplete, lacking in implementation, or otherwise undermined. We thus conclude with an overview of where the process has advanced farthest and other areas where it has lagged. We find that authority is rather clearly established and secure in Ethiopia for SNGs, though some levels of SNG are more robust than others. The degree to which decentralization has met the objectives of autonomy

86

Ethiopia

and accountability is more ambiguous while capacity seems to be a factor that seriously constrains implementation. This can also be viewed through the lens of the three dimensions of decentralization: political, fiscal, and administrative (as shown in Table 4.2). Ethiopia’s decentralization process offers a range of potential lessons for other African countries, both in terms of what makes decentralization work and what hinders the process. Among the greatest achievements has been the experience of the ethnic-based decentralization. Being a multiethnic and multicultural country, Ethiopia has chosen what has been called ethnic-based decentralization, with subnational regions created on the basis of ethnolinguistic criteria. This form of decentralization is different from the more commonly designated decentralization based on geography, resource, or other criteria. Two views are forwarded on the potential outcome of ethnic-

Table 4.2 Overview of Decentralization in Ethiopia Aspects of Decentralization Political decentralization

Fiscal decentralization

Administrative decentralization

Achievements • Elections at the levels of regional state, woreda, and kebele • Federal system in formerly centralized state • Regional states and woredas have strong legal authority • Regional states have independent authority to recognize their own subnational units • Fiscal transfers: guaranteed and substantial • SNGs spend nearly half of all government revenues • Formula-based distribution of revenues • Decentralized planning and administration • Civil service employees such as teachers fall under purview of woredas

Constraints • Dominance of EPRDF • Central government control over SNG officials • Partisan powers to appoint candidates for local elections

• SNGs have limited own-source revenues, rely on centralgovernment transfers • Transfers allocated primarily to recurrent expenditures; scarce resources for capital budgeting • Capacity constraints

Note: SNG is subnational government. EPRDF is the Ethiopian People’s Revolutionary Democratic Front.

J. Tyler Dickovick and Tegegne Gebre-Egziabher

87

based decentralization. One view argues that ethnic decentralization will lead to dismemberment of the nation (Seyom and Yacob 1999) and leads to unbalanced resource distribution (Assefa 2006). The opposing view argues that ethnic decentralization, by providing autonomy and self-rule to the various ethnic groups in the country, restrains further disintegration (see Asmelash 2000: 143). Ethiopia has remained intact for nearly two decades under ethnic federalism. Ethnic heterogeneity is the rule rather than the exception in African countries (originally resulting from arbitrary drawings of borders by colonial powers), and observers of African politics will find the Ethiopian experience worthy of observing closely as an approach to resolve problems of nationality and ethnicity. The other great achievement is the establishment of a significant degree of political, fiscal, and administrative decentralization in a short period of time. This process was relatively orderly and comprehensive. Given that this occurred in sub-Saharan Africa’s second-largest country in terms of population, with a complex ethnic demography and a legacy of top-down (indeed Marxist-Leninist) authority, the Ethiopian experience suggests that comprehensive decentralization is attainable and can persist in such environments even under inauspicious conditions. It also matters that Ethiopia’s substantial decentralization (while rapid and comprehensive once initiated) came in two waves: regional decentralization followed by woreda (district) decentralization. The empowerment of woredas (and kebeles at the local level) followed the empowerment of the regions by several years, a sequence that may have facilitated the establishment of effective patterns of intergovernmental relations, especially in a postconflict context where regime change occurred and stability was at issue. It is noteworthy that regional decentralization was initiated in Ethiopia before the EPRDF grew to its present dominant role, and that woreda decentralization occurred when the EPRDF had already established firm control. For many observers, the presence and power of the EPRDF has undermined the decentralization process; yet it may also be said that decentralization has consolidated and flourished in tandem with the governing party developing stronger links down to local levels (cf. Meheret 2007: 89). Strong party linkages up and down the vertical layers may or may not hamper local democratization. National-regional-local party linkages can theoretically either transmit demands from the bottom up or imperatives from the top down. An optimistic perspective may hold that the EPRDF’s decentralization has effectively constructed local governance at the woreda and kebele levels, and has provided considerable fiscal support to the woredas. In addition, while decentralization in the dominant-party state seems to have favored the extension of EPRDF power, it is not clear how party politics interacts with patronage. The relative transparency of the fiscal trans-

88

Ethiopia

fer formula suggests that access to funds is not primarily conditioned on personalistic relations between subnational officials and the center. Finally, on the fiscal side, the key lessons are with regard to the devolution of tax authority relative to intergovernmental transfers. The difference between these two categories of revenue sources illustrates the ways that fiscal decentralization can be cast in different lights. On the one hand, Ethiopia is quite fiscally decentralized, with high levels of government expenditures taking place at the subnational level and intergovernmental transfers providing resources for this. On the other hand, Ethiopia’s fiscal decentralization is limited by the low levels of own-source revenue at subnational levels. The impacts of this are felt not only in inadequate resources (though there is some evidence of this), but in the expenditure autonomy of SNGs: intergovernmental transfers are shaped overwhelmingly by central prerogatives, even where they are not explicitly earmarked or otherwise formally mandated. Thus, while own-source taxation faces obvious constraints in many countries in Africa, it is worth noting that the difference between these resources and intergovernmental transfers may be significant. There is conflicting evidence on the extent of the expenditure autonomy of the devolved units of government (the regional states and woredas). On the one hand, the revenues transferred are sent to these subnational levels as block grants, implying considerable latitude in expenditure. On the other hand, there is reliance on these funds for recurrent expenditures (cf. Tegegne Gebre-Egziabher and Kassahun 2007). Moreover, there have been reports that the government uses its partisan leverage to shape the decisions of subnational officials (cf. Meheret 2007). It is clear that autonomy is partially restricted on the revenue side, due to questions of overall tax bases and the capacity for collections. While this is likely to be true across many African countries, the more intriguing questions for Ethiopia seem to be on the expenditure side given the relatively secure access to intergovernmental transfer revenues enjoyed by the regions and woredas. If Ethiopia is a leader among African countries in the proportion of revenues devolved to subnational levels (albeit not the extent of own-source revenues available to SNGs), then a leading question for further investigation would be to what extent the center exercises leverage on the spending side of the ledger, particularly as it pertains to conditional grants. It would be interesting to provide a fuller accounting of the nature of official mandates, earmarks, standards, and requirements for subnational compliance with central directives, but complete information on this remains surprisingly elusive at this stage. Finally, we can ask how Ethiopia’s decentralization experience provides leverage for an understanding of decentralization processes across Africa or more generally. Of particular interest are the timing and sequencing in decentralization. Ethiopia’s decentralization sequence cannot be seen

J. Tyler Dickovick and Tegegne Gebre-Egziabher

89

as gradualist, given its rapid implementation at the regional level after the fall of the Derg in 1992 and its similarly rapid extension to the woreda level in 2001–2003. Yet the process also cannot be seen as a one-shot shock therapy because of the two-stage sequence that led from regional decentralization in the 1990s to the DLDP in the 2000s. The process was also largely (and impressively) simultaneous across political, fiscal, and administrative aspects, but it remains ongoing and incomplete in its implementation and desired outcomes. Furthermore, the process covers the whole of the country more or less equally, but is also asymmetrical: woreda decentralization was first implemented in the four big regions (Amhara, Oromo, SNNPR, and Tigray) before being extended to the others, and woredas themselves are established (with modest variations in form) by the regional state constitutions rather than the federal constitution. Finally, this postconflict state has practiced substantial devolution (but under a dominant party that exercises a degree of top-down control) and an ethnic federalism that paradoxically both channels and possibly deepens ethnic divides (cf. Vaughan 2006). Together, this suggests a form of decentralization that is not necessarily sui generis, but fits uneasily with any particular doctrine of decentralization.

Notes 1. We use the term nationalities, and not ethnic groups, given the long historical construction of local states and the degree of linguistic diversity 2. Haile Selassie was seen by many non-Amharic elites as favoring the Amhara, though the emperor was himself of mixed descent, with Amhara and Oromo parentage. 3. This has prompted interregional cooperation in implementing decentralized functions; cooperation between the Tigray and Afar regions, the Amhara and Afar regions, and the Gambella region and SNNPR are cases in point. Currently, there are no formal or legal provisions for such interregional ventures. 4. Intended to range in population size from 100,000 to 120,000, it is clear that woredas fall outside this range (see Tegegne Gebre-Egziabher 2007).

5 Ghana: Decentralization in a Two-Party Democracy Joseph Ayee and J. Tyler Dickovick

Over the past two decades, Ghana has been one of the most stable democracies in Africa. The country also has undertaken decentralization efforts during this time, but these have been partial and halting. Decentralization has been established and has endured, but remains incomplete. In this chapter, we evaluate Ghana’s decentralization experience from 1992 to the present. Our main finding is that decentralization in Ghana has resulted in the formal transfer of authority to elected officials at subnational levels, but that several characteristics of decentralization—autonomy, accountability, and capacity—are subject to multiple constraints that prevent the full flourishing of the decentralization program. Achievements are strongest in the development of a robust legal framework, but there are also real limitations on subnational authority because transfers of resources and responsibilities are incomplete and sometimes muddled by conflicting legal provisions regarding civil service staff. The autonomy of subnational actors is hindered by a lack of revenue, particularly with respect to internally generated and own-source revenues, though fiscal transfers for local governments have been established. Furthermore, much of the accountability in place at local levels is upward to nationallevel actors rather than downward to local populations and civil society. Finally, capacity is an enduring challenge, in both fiscal and administrative terms. The central challenges to Ghana’s decentralization come from bureaucratic politics more than electoral politics. Importantly, both major political parties—the National Democratic Congress (NDC) and the New Patriotic Party (NPP)—generally favor decentralization, though incentives for each are mixed and may vary with the alternation of power at the national level since Ghana’s demographics and stabilizing two-party system ensure that the 91

92

Ghana

out-party will have some regional and local base. More problematic for decentralization are the bureaucracy and civil servants in sectoral and line ministries (e.g., health, education, and agriculture that have large field administrations), which can be reticent to concede policy authority to the local government units. Additionally, understanding the incomplete realization of decentralization’s promise must include an understanding of the intended career paths for members of the local government service, which for most civil servants are directed toward posts in major metropolitan areas.

The Institutional Backdrop and Political System, 1992–2012 Ghana suffered an economic decline through much of the period from its independence in 1957 to the 1980s. In recent years the country’s economy has stabilized, though growth has been modest. Since Ghana’s return to multiparty rule in 1993, the country has witnessed political stability through five successive national-level elections. The presidential elections of 1992 and 1996 were won by Jerry Rawlings’s NDC while the 2000 and 2004 elections were won by the NPP of John Kufuor. The December 2008 elections were won by NDC candidate John Atta Mills after a tight runoff with the NPP’s Nana Akufo-Addo. Following Atta Mills’s death in office, Vice President John Mahama of the NDC assumed the presidency; he later won the presidential election (along with a parliamentary majority for the NDC) in December 2012. These elections were noteworthy for the acceptance of the results by the defeated NPP and for the stability that followed, despite some acrimony between officials and supporters of the two major parties. (The NPP brought an election petition before the supreme court challenging the outcome, but it was dismissed in August 2013.) Ghana thus is distinctive in Africa for its relatively stable two-party system in a democracy that has endured for twenty years. Ghana’s unicameral parliament is composed of district-based constituencies, with members elected from 275 territorial constituencies. But the link between parliamentarians and subnational governance is weak. The country fits the pattern of African governance in which political contestation is first and foremost about the presidential election, with parliamentary elections having a secondary role (cf. Van de Walle 2003).1 Moreover, there is no upper chamber to represent the country’s ten administrative regions. The NDC and NPP combine to hold a large majority of the seats in the legislature, with the NDC holding 148 and the NPP 123 of the 275 total seats currently occupied; small parties and independents hold the remaining 4 seats. Two-party democracy distinguishes Ghana not only from less democratic systems in sub-Saharan Africa, but also from the continent’s other democracies in which party systems tend either to be more fragmented

Joseph Ayee and J. Tyler Dickovick

93

(e.g., as in Benin) or dominated by a single party (e.g., as in South Africa). Ghana’s two parties do have some regional basis, though both draw support from across the country. This maps onto the ethnicities that have significance for electoral politics. The Akan ethnicity is the largest and contains a number of subgroups, of which the Ashanti subgroup, centered in the city of Kumasi, has traditionally supported the national candidates of the NPP. The Ewe people, who are clustered in the eastern part of the country (in the Volta region), are another major ethnic group of significance for Ghana’s political history, as longtime president Rawlings (1979, 1981–2000) was of Ewe parentage. The Ewe also have featured prominently within the military. While each party has an ethnoregional core constituency, recent elections have demonstrated that victory depends on assembling a crossregional, multiethnic coalition. Ghana’s decentralization process has established multiple layers of subnational authorities, the most important of which are the local government units known as district assemblies (DAs), as well as metropolitan assemblies and municipal assemblies in larger cities and metropolitan areas, respectively. We refer to these together as the metropolitan, municipal, and district assemblies (MMDAs).

Patterns of Authority: Institutions of Decentralization The legal framework of decentralization in Ghana is extensive on paper, beginning with the 1992 Constitution and extending to acts of parliament, legislative instruments, bylaws, and standing orders. They cover areas such as the objectives of decentralization, finance, personnel, planning, budgeting, tendering, accountability, and central-local relations and the relationship between the DAs and other public services. Ghana has a four-level structure of government, divided into the following levels: 1. National government, including line ministries and deconcentrated entities; 2. Ten regions with regional coordinating councils (RCCs) responsible to the center; 3. One hundred seventy districts with elected MMDAs; and 4. Subdistrict structures: urban, zonal and town councils, and unit committees. The core units of decentralized governance, representing local government, are the MMDAs. Whether a locality is governed by a metropolitan, municipal, or district assembly depends on population size: metropolitan assemblies existing only in the country’s six largest cities, municipal

94

Ghana

assemblies in the next forty largest cities, and district assemblies throughout the remainder of the country.2 Each of these forms has its own subdistrict structures: metropolitan assemblies have town or area councils, municipal assemblies have zonal councils, and district assemblies have urban, town, or area councils. By law, the MMDAs have eighty-six authorized functions that empower them to provide deconcentrated, delegated, and devolved local public services.3 These range from provision of basic services in education, health, water supply, and sanitation to public safety and revenue collection. Devolved functions of the MMDAs include the mobilization and management of revenue as well as the construction and maintenance of feeder roads, streets, parks, cemeteries, crematoria, and other public utilities. A deconcentrated function of the MMDAs, as agents of the Ghana Highway Authority, is the maintenance of trunk roads lying within the boundaries of their jurisdiction while a delegated function is water provision throughout the entire district in consultation with the Ghana Water Company and Community Water and Sanitation Agency. MMDAs are primarily responsible for local-level policy and planning (based on, and informed by, national policies and programs). They constitute a hybrid form of decentralized authority, combining elected and appointed officials; 70 percent of MMDA members are elected by universal adult suffrage and 30 percent are nominated by the president. The district chief executive, the political head of the district, is nominated by the president and subject to approval by two-thirds of the MMDA members present and voting. The chief executive chairs an executive committee comprised of one-third of the MMDA members. The MMDAs are nominally nonpartisan, as candidates are not sponsored by political parties. The stated aim is to ensure consensus building and promote development, but evidence suggests presidential administrations routinely appoint copartisans. In terms of fiscal authority, MMDAs have several revenue sources, the most important of which is the District Assemblies Common Fund (DACF), a set of intergovernmental transfers. A portion of total government revenue is paid into the DACF and shared according to a revenue-sharing formula approved by parliament. Originally allocated 5 percent of general government revenue, recent reforms under the previous NPP government increased this fiscal transfer to 7.5 percent. On the budgeting side, the decentralized departments responsible for different sectors are intended to send their budgets through the MMDAs for transmission to the Ministry of Finance (in a process known as composite budgeting), but this shows a continued degree of line ministry control. The MMDAs are the sole taxing authority in the districts, and they prepare their own annual budgets for recurrent and capital expenditures. Further, the MMDAs are the sole district planning authorities. They are corporate bodies that can sue and be sued,

Joseph Ayee and J. Tyler Dickovick

95

and they have the power to award contracts not exceeding GHC2,500 (2,500 Ghanaian cedi, approx. US$1,250 at $0.50 = GHC1). The MMDAs also implement development plans and programs coordinated by the central government’s planning commission while also working alongside district-level policy, planning, and program implementation by deconcentrated line ministries (health, education, and agriculture) that continue to receive funding and directives from their national and regional offices. Officially, the responsibilities for the actions of line ministries and departments are placed under the MMDAs, but this coordination is incomplete. Similarly, while a local government service nominally contains all staff of the decentralized departments, this too competes with sector-specific rules on staffing that complicate MMDA control over local civil servants. Central government ministries have significant power over the revenues and spending of MMDAs. The Ministry of Local Government and Rural Development is charged with supporting decentralization and coordinating subnational action. It has a major role in approving specific budgets for the MMDAs, along with the Ministry of Finance and Economic Planning. Line ministries plan and implement sectoral development policies. They also monitor and evaluate sectorial plans, lead budget processes, and represent their sectors in acquiring and disbursing budgets, including all funds for services and investments at the lowest level. This is especially the case with the three ministries that have strong, vertically integrated deconcentrated departments: the Ghana Health Service, Ghana Education Service, and Ministry of Food and Agriculture. The RCCs are entities in Ghana’s ten administrative regions with delegated authority to coordinate and harmonize district-level development. They monitor, coordinate, and evaluate the performance of the MMDAs and report to the center. The RCCs are chaired by a presidential appointee (regional minister), and led administratively by coordinating directors and their planning units. The RCCs also bring together other key actors in Ghana’s decentralization matrix: chief executives of the MMDAs, representatives of traditional authorities, and the regional heads of various deconcentrated line ministries. This mix of devolved, deconcentrated, and delegated authorities has made aligning the RCCs into the decentralization process complex and somewhat contradictory. Many functions to be delegated to the RCCs continue to be performed by the regional offices of deconcentrated line ministries, which still obtain funding from the central government in their respective sectors. Below the MMDAs come several subdistrict structures known as councils (urban councils, zonal councils, or town councils) and unit committees.4 Some of the functions of these bodies include naming streets, promoting public health, and performing day-to-day administrative duties (Ayee 2003). The evidence suggests that subdistrict structures are not viable: they are too

96

Ghana

numerous and small, have too many members, and are not fully elected (see Ayee 2003). Moreover, their functions are unclear and they have virtually no personnel or financial resources to perform their functions (NCG and DEGE Consult 2007).

Limits of Fiscal Autonomy While our discussion of authority suggests that Ghana has witnessed devolution and deconcentration, there are major limitations on the autonomy of subnational governments. These are most noteworthy in the area of fiscal autonomy, though there are some de facto limitations on the political autonomy of MMDAs, which we discuss in the next section on accountability. On the fiscal side, the MMDAs have scant own-source revenues and rely on central funding, which comes with strings attached by the central government. Internally generated funds or own-source revenues of Ghana’s MMDAs include rates, lands, fees and fines, licenses, rent, and investment income, which constituted 18 percent of their revenues in 2007.5 Ownsource revenue is low for a number of reasons. First, tax bases of the MMDAs are static and limited. One fundamental constraint is the lack of a tax base for land in rural areas. A large proportion of the country’s lands are under either state control or customary authority (stool lands). The MMDAs receive only a portion of the revenue from the latter, with a portion accruing to the local customary authorities (the stools and traditional councils). While this empowers traditional authorities to engage in local development initiatives, it also implies a limitation on own-source revenue for the MMDAs. Similarly, databases for determining revenue potentials are weak, limiting tax capacity. Second, there are complex structures in terms of planning for revenue collection in most districts. Extracts from the auditor general’s report of 2007 reveal that over 200 revenue collectors misappropriated revenues generated.6 Third, mechanisms for collecting internal revenues are ineffective. For instance, revenue collections from property taxes are low because properties have not been properly valued. Revaluation is the responsibility of a central government agency (part of the Lands Commission) that lacks the capacity to fulfill this function efficiently. This problem is exacerbated by the poor logistics for revenue collection. Fourth, there is an absence of appropriate sanctions or punishment for tax defaulters. Most residents, especially those in the informal sector, do not feel obliged to pay taxes (Inanga and Osei-Wusu 2004). The 1992 Constitution created the DACF, into which a portion of total government revenue is paid (in quarterly installments) for disbursement to the MMDAs based on criteria to be approved by parliament. The NPP government in 2008 proposed to increase the proportion of total government

Joseph Ayee and J. Tyler Dickovick

97

expenditure allocated to the DACF from 5.0 percent to 7.5 percent. The DACF amounts to 37 percent of revenues for the MMDAs (or twice the amount raised by own-source revenues) while total grants provided by the central government, in the form of salaries, proceeds from the Debt Initiative for Heavily Indebted Poor Countries (HIPC), and donor support accounts for the remaining 45 percent. The DACF is thus the single largest source of revenue of the MMDAs. By providing a constitutionally guaranteed minimum share of government revenue, it offers some financial independence. However, the amount is insufficient for the broad range of responsibilities devolved to the MMDAs. Moreover, expenditure autonomy is limited, as central government directives to the MMDAs determine about 75 percent of expenditure via earmarks, which leads to a loss of operational autonomy. Several other challenges have emerged in the administration and management of the DACF, including failure of MMDAs to submit their budgets promptly for release of their share of the DACF, failure of MMDA chief executives to abide by procedures for disbursement (which has landed some of them in trouble after they left office), and delays in releasing disbursements (which sometimes has lagged as much as two quarters). In short, Ghana exhibits a modest degree of autonomy for subnational governments. Fiscal decentralization has relied on the DACF. In 2008, an election year, the NPP increased the DACF to 7.5 percent of general government revenues. MMDAs thus operate on a relatively stable budgetary margin, with relatively transparent criteria for distribution set by parliament. This suggests some fiscal autonomy. However, the center predominates in terms of funding flows, functional assignments, and reporting. Moreover, the predominance of intergovernmental transfers (DACF) over internally generated funds, with strong central government control over expenditure, limits SNG autonomy. MMDA control over financial matters is limited by the constitutional provision that the Ministry of Finance and Ministry of Local Government and Rural Development will determine the areas of expenditure regarding the use of the common fund by the MMDAs. This budgetary control may be seen as turning the local governments into something resembling delegated agencies more than fiscally autonomous local governments.

Upward Versus Downward Accountability After fiscal autonomy, the second major limitation to Ghana’s decentralization is the prevalence of upward accountability of SNGs to central government officials, as contrasted with limitations on downward accountability to citizens. While both upward and downward accountability are appropriate,

98

Ghana

the preponderance of upward mechanisms in Ghana has reinforced central control in the decentralization framework. The limitations on downward accountability begin with the electoral process. MMDAs are partially elected and partially appointed by the national president. Elections to the MMDAs are held every four years (and have been since 1994), at least six months apart from parliamentary elections.7 Of the MMDA members, 70 percent are elected by universal suffrage (by electoral wards within the district) while 30 percent are appointed by the president who nominally must consult with traditional authorities and interest groups in the district. Elected members of the MMDA are expected to conform to nonelectoral mechanisms of accountability as well. For instance, they are obligated to meet their constituents before and after every meeting; in practice, this provision has been largely unmet due to problems such as the payment of meager allowances. The main achievement in downward accountability is the political decentralization that gave rise to elected local government, but its success is tempered by an incomplete electoral process. The empowerment of the MMDAs is incomplete for several reasons. First, election of district chief executives and, eventually, all assembly members would be required to make the system fully representative and democratic. Second, the nonpartisan nature of the elections, in an environment where partisan superstructure has been constructed at the national level, is anomalous. Beyond the electoral process, MMDAs are further criticized for a lack of responsiveness to the public in the conduct of their affairs. While there are formal and informal procedures and opportunities for popular participation in the local policymaking process (through MMDA meetings and MMDA members meeting with the electorate), these have been irregular and inadequate, leaving voting as the sole accountability mechanism exercised by most people at the local level. Moreover, while district plans are subject to a public hearing before adoption, this has not occurred in many cases for several reasons, including the rushed planning cycle, weak citizen contributions (which has discouraged officials from seeking further inputs), and citizen apathy as a result of low MMDA effectiveness. Accountability also depends on the ability of the MMDAs to cooperate and work with civil society organizations. The legal framework of decentralization stipulates that the MMDAs should coordinate, integrate, and harmonize development programs promoted or carried out by CSOs as well as act in cooperation with them in the district. There are examples of CSO action in exacting accountability such as civic unions from twenty districts participating in a USAID-funded program in 2005, CSOs consulting MMDAs, and CSOs undertaking civic education and organizing the public to demand accountability through questioning officials of MMDAs about

Joseph Ayee and J. Tyler Dickovick

99

the awarding of contracts (Ayee 2006). Moreover, media outlets (particularly local FM radio stations) have helped to build local accountability by providing avenues for local people to question their MMDA officials and debate with them over issues.8 Thus, there has been some citizen participation in demanding accountability from MMDAs, but progress remains slow. A key concern is the quality and low levels of civic participation and CSO involvement. The culture of MMDAs remains opaque, with some MMDAs failing to provide key information on their operations such as meeting schedules, annual plans and budgets, and annual expenditures. Contrasting with the limited forms of downward accountability are the extensive forms of accountability upward toward appointed officials and central government ministries. Unelected officials play an important role in the MMDAs, the most important of which is the district chief executive who is nominated by the central government and elected by two-thirds majority of members present and voting, including the 30 percent of members appointed by the president. Though the chief executive can be removed by the MMDA members, this usually is difficult because a vote of no confidence must be supported by two-thirds of all MMDA members. Political incentives impede reform in this area: the NPP previously advocated electing the district chief executives in its party platform, but retained the spoils system after its national victory. Government appointees (30 percent of members of the MMDAs), may bring professional and technical skills and expertise, as well as representation for marginalized groups such as women, youth, and persons with disabilities, but in practice the majority of government appointees are selected because of loyalty to the central government and governing party. Moreover, there is evidence to show that the president does not consult with traditional authorities and other interest groups within the district; rather, the party in power selects the government appointees from a party list drawn up by the district chief executive and the regional minister.9 Deconcentrated officials serve to control and circumscribe devolution in Ghana. The two most obvious examples are laws applying to the major sectoral ministries: the Ghana Education Service Act and the Ghana Health Service and Teaching Hospitals Act.10 These two acts outlined arrangements for service delivery at the local level that did not necessarily coincide with the arrangements for MMDAs. Both acts are inimical to devolution because they sought to create deconcentrated semiautonomous and separate services whose district heads do not report to the MMDAs, but rather to their parent regional and national offices. Health, education, and agricultural officers stationed in districts across the country report to central ministries. Moreover, budget transfers from the central government to local governments are heavily earmarked and severely limited in their possible applications.

100

Ghana

The establishment of the Ghana Health Service was aimed at providing a structure for the implementation of approved national health policies. The Ghana Health Service and Teaching Hospitals Act decentralized delivery of health care services by creating both regional and district health committees, along with a district director of health services who would be answerable to the director-general of the Ghana Health Service for policy measures and the district chief executive for administrative purposes. The staff at the district level remained staff of the Ghana Health Service. However, the Local Government Act made provision for health officers of the DAs. It indicated that any person who discharged the duties of a medical officer or sanitary inspector for any area would be an officer of the DA in that area (Ayee 2003). The Education Service Act also had this anomaly and required the district director of education to report to the director-general of the Ghana Education Service through a regional superior. Thus, while DAs are responsible for the provision of health and education infrastructure, they have little access to information about financial, human, and other resources of these services. Tensions have arisen in the past when assemblies have tried to inspect or get more involved in the use of this infrastructure (e.g., as in Cape Coast in 2002). At the same time, the two laws made provision for popular participation in development at the local level through the creation of the district education oversight and district health committees. Since enactment of the Local Government Service Act (Act 656) of 2003, efforts have been made to move all staff at the district level to the new local government service and eventually to integrate these into the MMDAs. The complete separation of budgeting, funding, and accountability of district-level staff from their parent ministries or departments and transfer of these to the MMDAs’ jurisdiction will take longer. Implementing the unified local government service has presented various challenges, including persistent allegiance of civil servants to central ministries and lack of MMDA control over the tenure of officials working in their localities. The MMDAs often found themselves without key officers such as coordinating directors, planning and administrative staff, and finance officers since these responded to the Office of the Head of the Civil Service or the RCC. Implementation of the local government service has fallen short, notably in the significant areas of health and education discussed above. Some MMDA functions and responsibilities are also provided by other service delivery agencies, which leads to passing the buck between the MMDAs and the service agencies in the provision of some functions; namely, water, roads, education, electricity, cemeteries, and crematoria. The legislative instruments creating the MMDAs would need to be updated substantially to modify their functions vis-à-vis those of institutions such as the Ghana Education Service, the Ghana Health Service, the Ghana Highway

Joseph Ayee and J. Tyler Dickovick

101

Authority, and the Ghana Water Company. In short, the lack of an integrated intergovernmental fiscal framework has resulted in multiple funding, planning, budgeting, accounting, reporting, and auditing systems. The Civil Service Law of 1993 was intended to define the relationship between the center and the local levels of government. It identified the duties of central government as policy planning, monitoring, coordination, and evaluation, and assigned the responsibility for local-level development implementation to the district assemblies. The law also provided for the assignment of staff, setting out the technical and administrative support for the decentralization process. A decentralized development planning system exists with planning bodies at the district level and regional level (RCCs), and sector agencies, ministries, and the National Development Planning Commission at the national level. In practice, however, most MMDAs do not have trained planning officers and, therefore, rely on consultants to develop their plans. This has undermined interest in the plans by the public because of lack of ownership. In addition, the requirement for public hearings is not adhered to and, when hearings are held, there are few debates and limited public understanding. Finally, the tendering and procurement process has been questioned because of the lack of MMDA members who adhere to conflict of interest requirements. Contracts have been awarded to companies belonging to MMDA officials, thus creating conflicts of interest. In other cases, contracts have been given to those with partisan connections without following approved tendering procedures. Consequently, the tendering process has created avenues for corruption (Ayee 2008).

Capacity Constraints Decentralization in Ghana is hindered by capacity constraints, which have resulted in a low quality of public services provision.11 A first capacity challenge is the nonviability of subdistrict structures. The legal and institutional framework of decentralization has placed emphasis on the district to the detriment of the subdistrict structures (area, town, and zonal councils and unit committees), which are expected to perform functions such as the registration of births and deaths, naming of streets, and education of the people on their rights, privileges, and obligations. These subdistrict structures have a weaker legal and political foundation than the MMDAs, and face operational difficulties due to the large number of staff required and inadequate financial and human resources. This has limited the flow of community-based inputs into planning and implementation. The MMDAs have performed poorly in service provision (in areas such as sanitation). When natural disasters such as floods, bush fires, and epi-

102

Ghana

demics have occurred, the MMDAs have not been in a position to provide immediate assistance because of a lack of resources. This has resulted in lack of faith in decentralization and in the MMDAs, with citizens looking toward CSOs and NGOs to deliver instead of the MMDAs. Decentralization is dependent on the provision of information to the citizenry. Here, progress is limited, though the legal standing orders enjoin the MMDAs to provide information to the citizenry. Keeping rural citizens (particularly the rural poor) informed of their rights and opportunities may not directly improve life conditions, but it will make indirect contributions toward improving knowledge, choices, and participation. Few MMDAs have truly met the requirement of posting information on public notice boards, and local languages have rarely been used because technical documents such as minutes and contract awards are written in English and assembly discussions are held in English. Consequently, the nonliterate MMDA members sometimes have found themselves unable to follow discussions on issues. Finally, as noted above, there are limited internally generated funds and a continued dependence on intergovernmental transfers and support from development partners. This can be seen as a capacity constraint in the sense that the MMDAs have limited fiscal capacity. Thus, the MMDAs have weak managerial and administrative capacities, with low levels of training and rudimentary management systems as well as limited revenue authority. This challenge correlates with the social and economic structure of many low-income rural districts where constituencies are socially heterogeneous, dispersed across large areas, and dependent on weak transportation and communication networks. This makes public goods provision challenging; for instance, local governments have difficulty resolving free-rider issues, managing common pool resources, and enforcing decisions due to the high transaction costs associated with their limited local legitimacy. Indeed, some districts lack legitimacy due to arbitrary demarcations of local governance units, which has led to conflicts over boundaries and control. The set of capacity constraints facing MMDAs feeds into the capacity of nonstate actors as well; for instance, transaction costs associated with poor communication and transportation impede the ability of publics to organize, lobby, and energize local governments (cf. Olowu and Wunsch 2004). Weak local media and information outlets result in information asymmetries between the public and local government personnel. Beyond weak transportation and communication systems, constraints on participation by civil society are manifested in the high opportunity costs of political activity for poor members of the public. In many rural districts, some citizens cannot afford indirect costs of time away from labor that political participation may imply; this is added to the potential direct costs of involvement if participation involves physical displacement from a remote village to a district seat, for instance.

Joseph Ayee and J. Tyler Dickovick

103

There is evidence that the MMDAs have made efforts in the provision of public services. They have allocated most of their financial resources toward education, health, and local government and rural development programs. They have undertaken development projects such as the construction and maintenance of feeder roads, school classroom blocks, clinics, toilets, and markets as well as the provision of water and electricity. The issue of poverty reduction is central to decentralization, and a number of poverty reduction strategies have been formulated with district-level implementation in mind (Asante and Ayee 2008a, 2008b). However, MMDAs have not been able to introduce programs to improve economic productivity, create employment, and generate income; in spite of HIPC funds and the poverty alleviation fund, many micro-, small-, and medium-scale enterprises still have difficulty accessing credit (Asante and Ayee 2008a, 2008b).

Political Incentives and Constraints to Decentralization In Ghana, the stated aims of decentralization were (and remain) deepening democratization at the local level and promoting local development and service provision, but the impetus for decentralization in Ghana’s central government has modified over time with changes in partisan needs and the country’s macropolitical and macroeconomic environment. We thus do not assume that the stated objectives of decentralization are always the real objectives of decentralization. In this section, we assess the incentives to decentralize as well as incentives to limit decentralization. Incentives to Decentralize: Political Parties and Historical Sequences

In the 1980s the Provisional National Defense Council (PNDC) made the case that decentralization was part of a program to improve governance and bring it closer to the people, yet it clearly also engaged in a calculation about its own capacity to govern. Similarly, opposition parties have advocated strongly for decentralization in their platforms (e.g., as the NPP before 2000 and the NDC before 2008), only to backtrack from commitments after gaining national power. While stated objectives may mask political objectives, it still is worthwhile to consider the reasons given for decentralization. We place greater emphasis on the political imperatives that may have promoted decentralization as well as the political constraints that have hindered it in spite of stated commitments. This gap between stated objectives and actual practice may be seen as analogous to the de jure status of decen-

104

Ghana

tralization (i.e., what is on paper in officialdom) as contrasted with the de facto status of decentralization (i.e., what is the reality in practice). The legally stated objectives go part of the way toward informing an understanding of decentralization, but must be coupled with an understanding of the political incentives of parties, elected officials, and state actors and an understanding of local governance processes on the ground. Since independence, subsequent governments in Ghana have had mixed preferences regarding decentralization, sometimes viewing it as a necessary condition for development and political legitimacy while also worrying that decentralization would promote divisive tendencies (Ayee 1994, 2004, 2008). For instance, the first postcolonial government of the Convention People’s Party under Kwame Nkrumah (1957–1966) feared decentralization could compromise national rule and thus encouraged centralization in the nation’s capital, particularly the Office of the President. The government also at times has fragmented the decentralized units as a way of weakening them (Ayee 1994), and prioritized a unified civil service to enhance local administration. These weaknesses in decentralization historically, combined with inadequate financial and human resources, continued up to the passage of local government reforms in 1988 (Ayee 2004). Decentralization was initiated in relatively close proximity to the two other major state reforms in the country’s recent history: economic liberalization and political democratization. The economic liberalization process (the Economic Recovery Plan of the 1980s) preceded the decentralization effort, though the PNDC regime of the time did initiate a process of local governance reform with the aforementioned establishment of local institutions (e.g., Committees for the Defense of the Revolution and unit committees). Conversely, the decentralization and democratization reforms were more nearly simultaneous, with both being set in motion at the end of the 1980s and formalizing in the early 1990s with multiparty elections in 1992 and the new constitutional charter. During this period, Ghana undertook several efforts to extend local governance and some of these were later incorporated into the present-day system. These included the creation of local cells initially called Committees for the Defense of the Revolution, along with unit committees. In 1988 the PNDC established the district assembly system and followed this with the 1992 Constitution which, along with subsequent statutes and enabling legislation, constituted a new local governance framework. This was elaborated on in 1993 with the Local Government Act as well as with subsequent legislation. After the immediate crisis of the 1980s–1990s was resolved, decentralization was perpetuated through a path-dependent process. As the country democratized in the 1990s, the PNDC’s successor party emerged victorious in national elections under President Rawlings. Called the NDC, this party

Joseph Ayee and J. Tyler Dickovick

105

continued the decentralization process that Rawlings and the PNDC had initiated. Meanwhile, the emerging opposition party—the NPP—adopted a platform that favored increasing decentralization, a decision that can be explained by its initially greater success in local elections and strong support in key regions of the country (especially the Ashanti region, often considered the country’s heartland). As the two-party system consolidated around the NDC and the NPP, and as power alternated from one to the other in 2000 and then in 2008, the party in opposition at any given time routinely advocated for greater decentralization. This out-party preference for decentralization can be partly explained by rational electoral calculations, as opposition parties in an evenly matched two-party system have regularly maintained a strong subnational presence even when out of power nationally. As a result of Ghana’s historical sequence, decentralization has (in principle) received support from both major parties at different times. After the PNDC, subsequent governments of the NDC and the NPP did not see decentralization as a way of self-legitimization, but rather had a mix of motivations. Some of these were noble, such as adhering to the 1992 constitutional provision that stipulates decentralization, as well as the global enthusiasm for decentralization in the third wave of democratization, and the good governance agenda generally (Crook and Manor 1998; Olowu and Wunsch 2004). The NDC and NPP governments also saw decentralization as an administrative reform, frequently driven by donor agencies, to slim down an ineffective central administration (Ayee 1997, 2004, 2008). Much of the decisionmaking in decentralization is subject to political calculations. This can even be found in terms of the very number of SNGs created. As of 2010, Ghana had 170 districts,12 up from 110 just a decade earlier (see Table 5.1). The 1992 Constitution and the Local Government Act of 1993 stipulate criteria for the creation of districts (e.g., population, geographical contiguity, and economic viability), but the legal framework is superseded by the interests of governing parties to create new patronage positions by creating new districts. The fact that the two most recent waves of district creation occurred during election years is revealing. Although the constitution gives parliament the main role in the creation of districts, the Local Government Act gives the president a dominant role through an executive instrument. The result has been the creation of districts not as a technical exercise, but as a political choice made by the central government. Even though there are criteria for delimitation of districts, final decisions are determined by the pressure that vested interests can generate. Indicative of the problem is that the creation of districts has been conspicuous as general elections approach (in 2004, 2008, and 2012, as seen in Table 5.1). As a result, the creation of districts has not per se enhanced local capacity to perform responsibilities, and many new districts are not fiscally viable.

106

Ghana

Table 5.1 Creation of Districts in Ghana Region Ashanti Brong Ahafo Central Eastern Greater Accra Northern Upper Easta Upper Westa Volta Western Total

Region Ashanti Brong Ahafo Central Eastern Greater Accra Northern Upper Easta Upper Westa Volta Western Total

Districts Before 1988–1989

Created 1988–1989

Total Before 2004

Created 2004

10 8 8 9 3 7 4 3 8 5 65

8 5 4 6 2 6 2 2 4 6 45

18 13 12 15 5 13 6 5 12 11 110

3 6 1 2 1 8 2 — 3 2 28

Total Before 2008

Created 2008

Total Before 2012

Created 2012

6 3 4 4 4 2 1 4 3 4 32

27 22 17 21 10 23 9 9 18 17 170

3 5 3 5 6 6 4 2 5 7 46

21 19 13 17 6 21 8 5 15 13 138

Total in 2013 30 27 20 26 16 29 13 11 23 24 216

Source: Ayee 2012. Note: a. Before January 1983, the Upper East and Upper West regions were collectively known as the Upper Region. In all, the Upper Region had 7 districts, 4 from the Upper East area and the remaining 3 from the Upper West area.

Decentralization also establishes a pattern of relations between the state bureaucracy, governing party, and customary authority. Ghana’s traditional authorities (led by paramount chiefs and regional kings) have ambiguous roles in the formal processes of the MMDAs and are not fully incorporated into local government per se, yet retain strong de facto roles in many areas of local interest, including resolving land tenure and family law disputes. These authorities control some local resources, drawing tribute from the so-called stool lands under each chief (stools being the symbol of traditional authority). Despite the incompleteness of their integration into the formal decentralization process and weak consultation of the govern-

Joseph Ayee and J. Tyler Dickovick

107

ment with the chiefs on appointments, customary authorities also have representation at the regional and national levels in the houses of chiefs, making them active players in political discussions even where their decisionmaking authority is circumscribed by the predominance of the national executive and parliament. The largest and most politically powerful of Ghana’s subnational kingdoms is Asante (headquartered in the city of Kumasi in the relatively prosperous Ashanti region), led by the Asantehene. This region is the heartland of the Akan ethnic group and is strongly affiliated with the NPP party. At the origins of decentralization in the 1980s, this region was an area likely to generate opposition to Rawlings’s PNDC, which featured many prominent leaders from a minority ethnic group (the Ewe) in the Volta region. Accommodating the chiefs and kings was not a stated aim of decentralization, and indeed further empowering Ashanti chiefs would have been anathema to the Rawlings government, which may help explain its initial preference for formal local government as a counterweight to traditional authority. Nonetheless, this alternative source of governing authority clearly impacted decisions to decentralize, insofar as it was a key element of the PNDC effort to build its legitimacy. To the present, while ethnic conflict is insignificant in Ghana relative to other African countries, the continued salience of ethnicity in voting (between the NDC and the NPP) suggests that decentralization may have the benefit of empowering local actors in decisionmaking, thus providing the much-theorized stake in governance to those who are out of power nationally. Constraints to Decentralization: Partisan and Bureaucratic Politics

While political elites of both current parties have supported decentralization for the right reasons, there are also questions of partisan advantage that have hindered the furtherance of decentralization. Governing parties at the national level have limited incentives to decentralize further. The NPP made little progress on decentralization, despite the party being seen as the hope for more proactive decentralization at the last alternation of power in 2001. Party politics may also help account for the tendency of NPP governments in 2004 and 2008 to increase the number of districts and the proportion of government funds allocated to the DACF, as decentralization ensures multiple opportunities for district-level patronage and electoral seats. However, this logic is not simple and should not be overstated. In brief, the political benefits of decentralizing in Ghana’s two-party system can be ambiguous. For example, the NPP won the 2004 national election and lost the 2008 election in a close race. In general, governments should

108

Ghana

be expected in rational terms to favor more centrally controlled patronage if they expect national victory, but seek more opportunities for decentralized electoral contests if they expect to lose nationally. In Ghana’s democratic two-party system, where both parties have reasonable probabilities of winning and losing, the benefits of creating more local elections alongside more centrally controlled patronage positions can be ambiguous. These relate particularly to the question of presidential appointment. While an out-of-power party has reason to clamor for more decentralized governance and a broader distribution of power, the party holding the presidency has little incentive to alter arrangements that favor central appointment power. This pertains to the appointment of 30 percent of MMDA members, as well as the district chief executives. The transition from the NDC to the NPP in 2001 illustrates this constraint. Prior to the election, the NPP offered a platform that included amending the constitution to make local government elections partisan and freely elect chief executives (Ayee 2004: 149–150). However, Gordon Crawford (2008) shows that the NPP fully used its appointment power upon election, immediately revoking the mandates of all district chief executives and all DA appointees put in place by the NDC (even though the latter had a term intended to continue until the following year). This led to the conclusion that the “non-partisan nature of the DA system is effectively a myth, but one that serves to conceal mechanisms for ongoing central government and ruling party control” (Crawford 2008: 252). It is noteworthy that this took place even with the prodecentralization NPP, which did increase the DACF by 50 percent. The other principal political constraint to decentralization comes not from partisan politics, but instead from bureaucratic and administrative actors. Many central government employees—ranging from ministers to locally deconcentrated staff in rural areas—have reason to favor administrative procedures that allow the center to exert influence over local government. At the apex, for instance, the Ministry of Local Government and Rural Development collaborates with the Ministry of Finance to review district budgets for the use of the DACF monies, thereby reducing the spending autonomy of local governments. Similarly, the aforementioned structures in health, education, and agriculture ensure that administrative staff members see their career paths within a line ministry, even if they are intended to be local government service employees overseen by MMDAs. Indeed, many personnel resist posting to rural areas—known to many as “the bush”—and into local government service, preferring a “centralized career structure” (Crawford 2008: 253). Line ministry departments at local levels largely follow the policy prerogatives and administrative provisions laid out by their parent ministries, not the MMDAs. These together suggest a political impediment to decentralization from unelected officials.

Joseph Ayee and J. Tyler Dickovick

109

Conclusion Our analysis gives rise to several conclusions. We must emphasize that achievements in decentralization in Ghana have been paired with countervailing shortcomings. Since 1988 Ghana has developed an impressive legal framework of constitutional provisions, acts of parliament, and legislative instruments with the objective of transferring decisionmaking and resources closer to the citizenry. This includes a devolution of political power (with local elections) and the creation of an established legal framework for revenue and expenditure decentralization. There remain several contradictions in the legal framework that need to be corrected. In certain instances, the legal framework either will have to be harmonized or realigned, particularly with regard to deconcentrated officials of line ministries. As we have noted, there remains ambiguity in the degree to which these officials are accountable to the MMDAs as opposed to the national ministries. In concluding, we observe how these challenges can be understood analytically through some of the lenses developed in Chapter 1 of this volume: collective action and principal-agent dilemmas. The result of constraints on Ghana’s decentralization is that the MMDAs are often ineffective at organizing and sustaining collective action (Olowu and Wunsch 2004). This ineffectiveness relates to the authority and capacity challenges noted above. Several constraints on the authority (or perhaps autonomy) of the MMDAs hinder their ability to confront collective action problems at the local level. This goes beyond the inability to manage key services themselves (e.g., in health, agriculture and extension, education, roads). With respect to collective action, key limitations are on the MMDAs’ authority to pass regulatory bylaws to regulate natural resource usage and to reach beyond their jurisdictions to deal with broader problems. Lacking the authority to bargain and make binding agreements with other local government jurisdictions (e.g., neighboring districts), MMDAs are vulnerable to cross-boundary externalities that render programs and actions less effective. Improvement of the decentralization process would depend on central government action to clarify the vertical assignment of revenues and responsibilities across levels of government. Revenue assignments to the MMDAs require alignment with functions in a way that allows them sufficient room to adjust revenues to local needs. Ideally, the assignment of functions across the different tiers of government would be based on the principles of subsidiarity, economies of scale, and efficiency in service provision. This would necessitate a review of the functions of the RCCs to ensure that the central government responsibilities are delegated to them as far as practicable. It would also likely result in a reduction in the number of

110

Ghana

subdistrict structures and the number of personnel on each, and would put in place mechanisms for sharing of revenues collected from subdistricts. Steps to harmonize the legal framework for decentralization with sectoral legislation would also be necessary to promote devolution of authority from the central government to the MMDAs. Finally, full implementation of the local government service would provide MMDAs with a functioning administrative structure and adequate resources. Taken together, such a rationalization would clarify the assignment problem and give rise to subnational governments and administrative units capable of undertaking service tasks appropriate to their level. Ghanaian subnational government can also be examined through the lens of principal-agent dilemmas; in particular, the difficulties lie in shaping accountability for subnational personnel (see Olowu and Wunsch 2004). When viewed through this lens, local governments simply are not (for the most part) principals of local personnel, and local publics are not the principals of local governments. The predominance over decentralized government units by policies, decisions, and personnel at the center have shaped Ghanaian governance in such a way that many subnational actors view their principals as being central-level officials. Principal-agent challenges in Ghana emerge from several sources. The main challenges revolve around the orientation and outlook of subnational personnel. Most administrative personnel receive low pay and, hence, have low motivation and loyalty to their governance units. There are also few rewards for merit (e.g., in areas of exceptional service delivery or managerial performance). The result is an outward orientation of most personnel in among devolved and deconcentrated officials, meaning that many seek reappointment or promotion to the national or regional capital and to other large urban areas. Among senior administrative and executive personnel as well, career tracks and ambitions generally point toward higher office within the central government. It should also be noted that local legislative bodies have weak resources and powers vis-à-vis local executives; this means ambitious personnel seek executive authority and this system generates a self-sustaining equilibrium in which local councils and assemblies remain weak. With respect to the public serving as the principal of local government agents, another problem is representativeness. The composition of the MMDAs is rarely reflective of the populations they are governing. Elected representatives are usually drawn from professional classes, with men in the highly educated and elite occupations accounting for over 89 percent of members. The poor and women have been underrepresented groups. In spite of a government directive that 30 percent of government appointments would be reserved solely for women, the dominance of men in the work of the MMDAs is overwhelming (Ofei-Aboagye 2004). In 2002 the number of

Joseph Ayee and J. Tyler Dickovick

111

female candidates was only 965 out of the total number of 13,950, though this was nearly double the proportion in 1998 (Ayee 2008). The gender imbalance did not improve in the 2006 local government elections, in spite of vigorous advocacy and publicity work of gender-sensitive CSOs. Ghana’s decentralization reforms have resulted in some devolution of power and resources to subnational governments, but these have not led to energetic and effective local governments. Our findings echo those elsewhere in Africa (and in this volume), where shortcomings of devolved governments have included a range of problems such as persistent central control; weak professional and technical support by the center; weak administrative, professional, and technical capacity; dominance of local councils by local executives and administrative personnel; weaknesses in local public and civil society participation; and failure of local officials to ensure accountability of local service delivery personnel. There is a general political consensus in Ghana in favor of decentralization as a positive force for democracy and development; this holds among the executive, parliament, and subnational politicians. Decentralization enjoys both constitutional and legal guarantees and open bipartisan support. Given the slow progress with decentralization, in their 2008 election platforms the two major parties (the NDC and NPP) each promised to reform and deepen decentralization. Upon winning the elections, the NDC government set up a committee in 2009 to undertake district and regional forums to gather views on advancing the process of decentralization. However, political incentives to decentralize are modest. As with the NPP in 2001, the NDC now faces incentives to perpetuate a status quo in which the central government has considerable appointment power over officials at subnational levels. Moreover, significant resistance to full implementation of the decentralization agenda may come from ministerial and bureaucratic or administrative actors who resist local autonomy. Decentralization in Ghana requires substantial coordination and political consensus to advance because it involves several actors at the international, state, regional, and district levels with different responsibilities and perspectives. Questions of bureaucratic coordination are linked to political realities and are exacerbated by genuine capacity limitations at all levels. In Ghana (and elsewhere in Africa), decentralization is likely to succeed when the procedures and institutions in place ensure proper technical coordination of the process. However, recognizing the importance of coordination does not presume that decentralization is simply a matter of technical capacity and adequate information; political incentives to impede decentralization are also important. Decentralization receives political support from many key actors in Ghana, yet political obstacles remain. Understanding political impediments— whether they emerge from partisan incentives to retain appointment powers

112

Ghana

over subnational officials, ministerial objectives to set expenditure rules, or bureaucratic resistance by national civil servants to become part of a local government service—are crucial to understanding why gains in decentralization in countries such as Ghana have been more modest than ambitious legal frameworks would imply. Even Africa’s most enduring example of a stable democracy with a strong two-party system shows that the creation of a legal framework for decentralization does not necessarily guarantee democratic and effective local governance.

Notes 1. The constitution nominally guarantees a judiciary that is independent of both the executive and legislative arms of government, but judicial independence is restricted because there is no upper limit to the number of justices of the supreme court that can be appointed by the president. 2. The six largest cities are Accra, Cape Coast, Kumasi, Sekondi-Takoradi, Tema, and Tamale. 3. In formal terms, the responsibilities of MMDAs can be seen as: political and administrative authorities, planning authorities, development authorities, budgeting authorities, and rating authorities. 4. At a slightly larger level than these councils and unit committees, Ghana’s metropolitan assemblies in the six largest cities have submetropolitan district councils charged with the administration of self-help projects; recordkeeping on ratable properties; and the collection of rates, levies, and taxes. 5. This figure is according to the administrator of the District Assemblies Common Fund. 6. Reports can be found at www.ghaudit.org/gas/site/reports. 7. Moreover, passage of the Internal Audit Act (Act 658 of 2003) and the Procurement Act (Act 663 of 2003) has been seen as promoting financial accountability. 8. This is akin to the situation in Uganda where local radio stations hold phonein programs in which mayors and chairpersons are invited to respond to people’s questions and complaints. 9. The progovernment stance of most government appointees has brought them into conflict with the elected members, though some government appointees have constructively criticized government policies (Ayee 2008). 10. See Act 506 of 1995 and Act 525 of 1996, respectively. 11. These capacity challenges were documented in the 2007 Draft Comprehensive Decentralization Policy Framework. 12. The NDC government created 46 new districts and municipal assemblies in 2012, bringing the total to 216. The creation of new districts in an election year led the opposition to accuse the government of gerrymandering. See www.ghanaweb .com/GhanaHomePage/NewsArchive/artikel.php?ID=235800, accessed April 13, 2012.

6 Mali: Incentives and Challenges for Decentralization Susanna D. Wing and Brehima Kassibo

For nearly two decades Malian democracy, with decentralization as its defining element, was a model for Africa. After the overthrow of Moussa Traoré’s authoritarian regime in 1991, decentralized government was considered by many Malian citizens to be intrinsic to bringing governance closer to the people. It was grounded in the country’s vibrant civil society, engaged citizenry, and propensity for national dialogue. Malian decentralization, however, was also born out of conflict and fueled by rentseeking behavior. In this article, we examine the paradox of decentralization in Mali in which the imbalance between the democratic aspirations that many Malians associated with the process were derailed by institutional constraints and elite interests. In keeping with the themes of this volume, we find that decentralization has had minimal effects on democratizing politics or strengthening state performance at the local level. In Mali the shortcomings of the implementation of decentralization appears to have been one contributor to the recent crisis of regime stability and democracy at the national level. Nevertheless, the Estates General on Decentralization that convened in Bamako in October 2013 illustrates the way in which decentralization remains a centerpiece for rebuilding democracy. The impetus for contemporary decentralization dates back to the Tuareg insurgency in the latter days of the Traoré regime and demands for greater democracy across the country. The Tuareg demands for autonomy, first accepted in a cease-fire agreement in 1991, resulted in a general push toward decentralization in the 1992 Constitution. Political will to support decentralization was widespread from 1992 until 2005, when declining political will sent a signal to the Malian people that the government was not fully committed to any real transfer of power. Ultimately, decentralization 113

114

Mali

failed to live up to the expectations of many Malians and helped to rekindle violence in the north in 2006 and again in 2011, resulting in the most recent insurgency in the north by various groups including most prominently the National Liberation Movement of Azawad (MNLA) and Ansar Dine. It also contributed to the general malaise evident at the time of the March 2012 coup d’état that ousted President Amadou Toumani Touré (known as ATT) from power. Nevertheless, the Tuareg rebellion and democratization in the 1990s set in motion a remarkably thorough process with genuine local initiatives such as the creation of communes from below by the participating villages. Whatever its early successes, Malian decentralization was plagued by underlying structural weaknesses. It was relatively easy to make laws (e.g., no fewer than twenty-four decentralization-related laws and decrees have been promulgated since 19901), but far more challenging to implement them and finance devolution of power to local authorities. And a dearth of financial resources meant that few sought to join the civil service at the subnational level. Decentralization also multiplied layers of government and did little to reduce the effective control of the central government through deconcentrated agencies. Decentralization was supposed to increase transparency and accountability, but unfortunately the structures in place constrained such developments. Community involvement remained limited and a small number of individuals tended to take charge, increasing chances for corruption. Indeed, some political elites apparently pursued decentralization precisely for access to rents that would multiply as layers of bureaucracy increased (Coulibaly, Dickovick, and Thomson 2010). Malian decentralization, at the time of the March 2012 coup d’état led by Captain Amadou Sanogo, was hindered by long-standing economic and political realities. Prior to the coup, poverty was widespread. Decentralization was meant to lessen poverty, yet that same condition remained a primary challenge to decentralization. Moreover, additional constraints included a lack of financial and human resources; weak technical, administrative, and revenue-generating capacity; and the neopatrimonial nature of Malian politics where the boundary between public and private roles is not clearly established. In this chapter, we offer a historically informed assessment of how the democratic principles of decentralization in Mali have eroded over time, and how this contributed to the recent troubles that the country has experienced.

History of Decentralization The Mali Empire, which ruled over the region between the thirteenth and seventeenth centuries, was decentralized. Thus, contemporary decentralization has long-standing roots, relying on a culture of local deliberation and

Susanna D. Wing and Brehima Kassibo

115

checks and balances that are central to village governance (Rawson 2000: 266; Kassibo 2007). Yet centralized power has been the dominant characteristic of politics throughout Mali’s colonial and postcolonial history. The Soudan Francais, as Mali was known during French colonial rule, was divided into cantons from 1895 to 1959, each of which included villages ruled by a canton chief appointed by the governor. In 1917, the governor general of French West Africa argued rigorously for decentralized government in order to “[reconcile] the interests of the dominant power with those of the local people” (SNV and CEDELO 2004: 15). In practice, however, decentralization has always been slow to materialize. The commune is the political and administrative unit at the core of decentralization. At first, only urban communes existed. The communes of Bamako and Ségou were created in 1919 by colonial authorities, followed by an additional eleven communes between 1953 and 1958. Communal authorities were also appointed by the colonial administration until 1955 when mayors and councils began to be elected by universal suffrage (Seely 2001: 508). At independence in 1960, decentralization was established as a central feature of the new constitution. Article 41 established the general principle that “territorial collectivities” were to freely administer themselves by elected councils. These units included communes, villages, tribes, and nomadic groups (Seely 2001: 508). The 1960 Constitution also established a ladder of administrative units beginning with the arrondissement (subsectors made up of a few villages), and moving up to cercles (grouping several arrondissements) and regions (grouping several cercles). Although the constitution provided for “free self-administration” of these units, local officials were actually appointed by the central government, a practice that was retained by the Traoré regime after the 1968 coup that toppled President Modibo Keita. In 1974 the constitution was revised, and Articles 74 and 75 removed self-administration from the local government structures. In 1977, local authority was further undermined with the redrawing of subnational unit boundaries and appointing bureaucrats to control the redesigned regions, cercles, and arrondissements. In 1981–1982 regional and local development committees were created, but they had only consultative roles. In fact the Ministry of Territorial Administration retained effective control, particularly over the collection of local revenues. The core of local financing rested in the taxe de développement, régional et local (regional and local development tax, TDRL), which consolidated several local taxes into one. A percentage of the TDRL was to be retained at the local level and another fraction assigned to the corresponding cercles and arrondissements. The decree determining these percentages was never issued (Seely 2001: 509–510). By 1991, administrative decentralization existed in only nineteen urban communes—the twelve communes created by colonial government, the six communes of Bamako, and the commune of Bougouni (SNV and CEDELO 2004: 16). Until the democratization

116

Mali

movement gathered momentum in the early 1990s, Mali’s history of local governance was one of limited and weak deconcentration. Decentralization in Mali from the 1990s is linked closely to the history of the Malian state’s relationship with the Tuareg. The ineffectiveness of promised decentralization hastened, but did not create, the rebellion in the north. The Tuareg are a seminomadic people in northern Mali, Niger, Burkina Faso, Algeria, and Libya that represent no more than 10 percent of Mali’s population. In the principal cities of the north (Timbuktu, Gao, and Kidal), only in Kidal do the Tuareg make up a majority of the population. The Tuareg were once a dominant people in the region, but they suffered progressive marginalization and impoverishment throughout the independence era, culminating with dramatic droughts in the mid-1980s that all but wiped out their livelihoods and lifestyles. Many of the Tuareg sought refuge in Libya and were recruited into Muammar Qaddafi’s Islamic Legions. An estimated 3,000 fighters returned to Mali in the early 1990s and launched a quasi-separatist insurgency. Most of the conflict, from 1990 to 1995, was characterized by relatively low-intensity skirmishes and irregular rebel attacks on local military or government outposts. In the 1990s young, unemployed Tuareg intellectuals joined mercenaries returning from Libya (Humphreys and Ag Mohamed 2005: 255). They had failed to find employment in the local deconcentrated state structures that were mostly staffed by southerners or Songhais (sedentary northerners) (Humphreys and Ag Mohamed 2005: 274). The rebels wanted a free Tuareg state. After negotiations with the Bamako government, the Tamanrasset Agreement of 1991 offered the rebels decentralization, a high degree of autonomy, and special status for the three northern regions (Kidal, Gao, and Timbuktu) that constitute part of the traditional Tuareg territory known as Azawad. The agreement provided for local assemblies with legislative and executive branches to regulate all economic, social, and cultural issues specific to these regions. Implementation of the Tamanrasset Agreement, however, unraveled with the overthrow of Traoré. Further negotiations with the new regime led to the adoption of the National Pact in April 1992. The pact also provided for special status for the north, including local, regional, and interregional assemblies responsible for agriculture, livestock, water, urbanism, housing, environment, industry, transport, communication, health, education, culture, and tourism. Application of the pact was slow and deficient due, in part, to the government’s weak capacity and hollow goodwill. The 1992 Constitution did not include recognition of the north’s special status and this contributed to scattered violence that continued into the 1990s. The concessions initially granted to the Tuareg were further reflected in the decentralization reforms that began to take shape in 1992. In these ways, the insurgency spurred decentralization.

Susanna D. Wing and Brehima Kassibo

117

Implementing Decentralization in the Third Republic Decentralization was a core element of democratization and a central theme of President Alpha Oumar Konaré’s administration (Wing 2008a; SNV and CEDELO 2004: 19). Not surprisingly, democratization unleashed a widespread backlash against centralized government that had historically extracted resources from rural populations while not providing public services in return. Simultaneously, the dominant donor discourse of economic liberalization and privatization promoted a decentralized conception of public authority and resource allocation. Many Malians expected democratization to result in a redistribution of resources, and decentralization was presented by the regime as the ideal method by which greater equity could be reached across the country (Kassibo 2007: 1). A National Conference was held in Bamako in 1991 and delegates from government, the opposition, and many groups in civil society affirmed a moral commitment to decentralization. Most Malians live in rural areas and feel particularly distant from the urban centers that had historically garnered all political influence and the bulk of development resources. Farmers at the 1991 Rural Estates General, for example, demanded more equitable treatment and representation of rural interests in the political sphere. Thus, villages figure prominently in the decentralization framework. The 1995 Code des Collectivités Territoriales (Local Government Law) obliges communal councils “to take advice from village, pastoral grouping and/or neighborhood councils concerned before deliberating on certain important issues regarding land use, financial matters and other affairs” (Djiré 2004: 6). Despite a remarkable degree of national consensus and goodwill from the new authorities, decentralization reforms unfolded haltingly. First, in conventional centralizing fashion, a number of administrative structures were set up to oversee the process. The first was the Decentralization Mission created in 1992 with Ousmane Sy as director of a twelve-member board. Conflicts of interest arose with the Ministry of Territorial Administration that hosted the mission. The mission was moved to the Office of the Prime Minister where it was renamed the Mission de Décentralisation et des Réformes Institutionnelles (Decentralization and Institutional Reform Mission, MDRI). The mission had only a consultative role, but it can be credited with important and original contributions to Malian decentralization. Most crucially, it established the principle that villages were allowed to group themselves together in a bottom-up process to set up communes and decide on the seat of these communes among themselves. Initially, the MDRI stipulated a constraint that communes had to be economically viable, in addition to being geographically contiguous and encompassing between 10,000 and 25,000 inhabitants (Seely 2001: 513). Not unexpectedly, social

118

Mali

solidarity between villages soon became the driving force behind the new communes, to the detriment of fiscal viability. Mali’s governmental structure is noted for its complex and interlocking levels. The administrative structure includes four levels of government: the national administration, regions, cercles, and communes. The latter three form the local government authorities. There are eight regions, plus the Bamako District (which is not a district in the sense of others, but more a distinct administrative region), 49 cercles, and 703 communes. Each level of government has distinct legal prerogatives and competencies, and is administered independently through elected executive bodies. Communes elect councils of eleven to forty-five members for five-year terms, depending on the size of their population. Council members in turn elect mayors and executive committees by secret ballot. Members of the cercle and regional councils are selected from representatives of the lower administrative structures. Thus, members of communal councils are selected by their peers to participate in cercle councils. Cercles, in turn, choose the members of regional councils. Cercles and regions also have executive committees.2 In the areas administered by these regional collectivities, such as education and health care, the law provides for a complete transfer of responsibility from the central government (Seely 2001: 513). Creating communes was a complex process, and elections for municipal councils were phased in gradually. The first communal elections were held in urban municipalities in 1998, and in the nearly 700 rural communes in 1999. Meanwhile, the country had already undergone two rounds of presidential and legislative elections (1992 and 1997). In February 2000, the MDRI was dissolved and a Ministère de l’Administration Territoriale et des Collectivités Locales (Ministry of Territorial Administration and Local Collectives, MATCL) was created, with Ousmane Sy as minister, an appointment that communicated the government’s commitment to decentralization. The government also established groupes d’étude et de mobilisation aux niveaux régional et local that were designed to increase the role of civil society activists in developing infrastructure, education programs, and mobilization strategies for decentralization (Rawson 2000: 278). However, the lack of resources and competence of civil society participants forced the government to bring civil servants back into the process. In addition, the government also set up another layer of centralized supervision in 2000 with the Direction Nationale des Collectivités Territoriales, suggesting that old habits die hard (SNV and CEDELO 2004: 20). With its new—and somewhat unwieldy—institutional apparatus, the government continued to create the legal infrastructure to finalize decentralization. A few new communes were created in the north as a result of Tuareg pressure (to reach a total of 703) and, by 2004, elections were held for the first time in all communes across the country. In that year, over 10,000 local government councilors were elected.3

Susanna D. Wing and Brehima Kassibo

119

At the national level, responsibility for decentralization and local administration rests with the MATCL and the Haut Conseil des Collectivités Territoriales (High Council of Local Government Authorities, HCCT). The HCCT is made up of representatives from all regions and is designed as a consultative body for matters of subnational government. Unlike the government and the National Assembly, the HCCT cannot be dissolved, so it provides a limited check on executive power (Seely 2001: 511; SNV and CEDELO 2004: 20). Prior to the overthrow of ATT in March 2013, there had been ongoing negotiations to replace the HCCT with a second legislative house (Sénat), as proposed by the Daba Diawara Commission that was appointed by ATT. This proposal was to be voted on by referendum in April 2012, but the vote was aborted following Captain Sanogo’s seizure of power. Interestingly, if the referendum had passed, not only would the HCCT have been replaced but a large number of seats (a number that remained unspecified, but reported to be as high as 33 percent) would have been appointed by the president of the republic, further eroding limited representation and autonomy of political interests outside the center.

Challenges to Decentralization: Authority, Autonomy, Accountability, and Capacity Authority

By law, authority over multiple domains is transferred from the central administration to communes, cercles, and regions. This transfer has not been effective. Running parallel to decentralized authorities, there are newly created deconcentrated agencies of the state, represented by governors in regions, prefects in cercles, and subprefects in communes, that retain considerable de jure—and even more de facto—authority. The multiple layers of authority are presented in Table 6.1. Some of these institutional arrangements are of dubious use aside from maintaining control over communes and preventing individual initiatives. For example, coordination structures between the different decentralized and deconcentrated layers of government, referred to as local steering committees, are chaired by the prefects and include the commune councilors, district council members, and representatives from the chambers of agriculture and commerce and from NGOs (SNV and CEDELO 2004: 27). The local steering committees approved the Centres de Conseils Communaux (Center for Communal Councils; CCC) programs, which previously served as donor-sponsored advisory structures for decentralized development projects before the CCCs were phased out in 2004. The regional steering committees define, coordinate, monitor, and evaluate technical support to the local steering committees and communes. In a country plagued with weak

120

Mali

Table 6.1 Supervisory Authorities for Decentralized Structures in Mali Level

Elected Bodies

Supervisory Authority

Steering Committees

State

Supreme Council of Local Government Authorities

Ministry of Territorial Administration and Local Collectives (MATCL)

National steering committee

Region (8)

Regional Assembly

High commissioner governor of the region

Regional steering committee

District (cercle) (49)

District council

Prefect

Local steering committee

Bamako Districta (6 communes)

District council; commune council

Governor of district

Commune (703)

Commune council

Subprefect

Source: SNV and CEDELO 2004: 26. Note: a. The Bamako District has a special status. It is similar to a region and led by a district governor. It has a central mayor and a district council, but also includes six communes, each with its own council and mayor. Unlike cercles, however, prefects do not supervise the Bamako communes.

capacity and severely limited human capital, it is hard to see much benefit in this type of structure other than creating mechanisms to retain control. The Local Government Law of 1995 transfers to the communes responsibility for local budgets and accounts; preschools, primary schools, and literacy teachers; centres de santé communautaires (community health centers, CSCOMs); transportation, roads, and communication; urban and rural waterworks; local markets; sports and cultural events; and environmental protection and planning. Cercles inherit responsibility for secondary schools, health facilities, roads, and communications and rural waterworks. Regions have authority over secondary, technical, and professional schools and special education as well as over regional hospitals and support of vulnerable populations (see also Lippman and Lewis 2008: 5). Act 96-056 charges municipal governments with local economic, social, and cultural development planning and the management of their own affairs. Each commune is supposed to administer the census, municipal police, sanitation projects, archives, and official documentation for the commune. At the same time, deconcentrated layers of government retain authority over significant issues. Prefects continue to exercise competencies over services related to health, waterworks, land, conservation of nature, and some administrative services. At the regional level, the governor oversees the regional management offices for health, education, waterworks, taxes, budget, treasury, conservation of nature, and some other administrative services such as programs, statistics,

Susanna D. Wing and Brehima Kassibo

121

computing, development, and population (RLD 2008). Thus, there is a significant overlap in terms of domains of competencies between decentralized and deconcentrated authorities. Moreover, in exercising their own prerogatives, decentralized authorities often run into constraints that limit their effective authority. For example, although the municipal council draws up its budget, the prefect must approve the budget before it is implemented (UNCDF 2007: 6–7). The communes were given some prerogatives that were ultimately reassigned to other levels of governance. For example, voting cards and polling stations were to be controlled by mayors, but the electoral code was subsequently amended to return these powers to the central government, on the rationale that mayors (as members of political parties) could not be neutral enforcers of the electoral code (SNV and CEDELO 2004: 38). Many of the local government civil servants are in fact appointed or delegated by the central government and its deconcentrated agents, including commune secretaries general who, in some cases, are the only literate workers in a commune. Not surprisingly, these structural factors arguably slowed the actual transfer of competencies to communes. It was not until June 2002 that health, education, and water services were transferred to the communes by presidential decree. Even then, this transfer was made without appropriate financial resources. Similarly, the transfer of the management of natural resources to communes was delayed. The Netherlands Development Organization (Stichting Nederlandse Vrijwilligers; SNV) notes that, when three communes in the Bourem and Menaka districts wanted to restrict the types of wildlife that could be hunted, the prefect rejected the plan and hunters simply obtained permits from a higher administrative level (SNV and CEDELO 2004: 45; see also Ribot 2001). Finally, communal financial resources are insufficient to translate their competencies into effective authority. First, the income of the communes is deposited into the central tax office account. As a result, it is difficult for communes to remove funds because the state often has already used them for other purposes. Second, municipalities are supposed to derive their income from “their own resources (fiscal revenue from duties and taxes transferred by the state; receipts and outputs from their estate, such as sale or rental of land and immoveable assets, charges on the use of the forest, wildlife or fishing resources, etc.) and external resources provided by the state or local government partners, authorized loans, gifts and bequests” (Djiré 2004: 6). As Moussa Djiré (2004: 6) argues, there should have been a transfer of resources corresponding to the transfer of authority, in order to cover the costs of running local governments. The devolution of authority was slow and, once transferred, competition with deconcentrated agencies and a lack of financial resources undermined the effective authority of local government structures.

122

Mali

Autonomy

Decentralization was further challenged by the restrictions on political, fiscal, and administrative autonomy at the local level. To some extent, direct universal suffrage guarantees the political autonomy of communes. Communal councilors are elected to a five-year term by proportional representation with closed lists and without preferential voting or panachage (the option to vote across lists or strike out candidates). Indirect elections then follow, as noted above, for positions at the cercle and regional levels. Bamako has a special status with six urban municipalities, each with its own council that appoints district council representatives from among its members. The district council appoints the mayor and his or her deputies. The authority of the district council is the same as those of the regional council. Although there is local autonomy in selecting communal councilors and, through them, the rest of decentralized authorities, this political autonomy is restricted by the relative lack of the administrative autonomy. Deconcentrated supervisory authorities (see Table 6.1) check the legality of measures taken by the elected bodies. These deconcentrated authorities also provide support and advice to the elected bodies. Subprefects, prefects, and high commissioners formerly held most of the powers that have been transferred to local governments and they often have substantial experience. The continued existence of these supervisory offices, however, undermines the effective administrative autonomy of decentralized structures and promotes significant confusion over the division of labor between decentralized and deconcentrated bodies. For example, in some cases councillors believed they replaced all aspects of the former administration and therefore the prefect (who remained as a supervisory authority) should be removed as well (SNV and CEDELO 2004: 26). The HCCT and the Fonds National d’Appui aux Collectivités Territoriales (National Fund for Territorial Collectivities, FNACT) are examples of institutional multiplication in the context of decentralization. The HCCT provides oversight to all levels of subnational government. The FNACT, under the aegis of the Agence Nationale d’Investissement pour les Collectivités Territoriales (National Investment Agency for Local Government, ANICT) was created in 2007 to replace the CCCs. It focuses on regional economic development by financing investments for regions. Although some of these structures are no doubt helpful, they also undermine the administrative autonomy of communes, cercles, and regions, which have to jump through many hoops with limited capacity. These structures also reflect a tradition of state control and a propensity to address policy issues by creating consultative organs. While these facilitate dialogue and can promote consensus, crucial values of Malian democracy (Wing 2008a), they hamper local decisionmaking and create avenues for elites to maintain control.

Susanna D. Wing and Brehima Kassibo

123

Finally, the administrative autonomy of communes is negatively affected by their lack of control over personnel. There are two categories of civil servants at the local level. On the one hand, there are those named by the public administration in Bamako and appointed locally such as governors, prefects, subprefects, and technical personnel. On the other hand, there are those civil servants who work directly for local governments. These include, for example, the secretary general (who runs the daily operations of communes), régisseur (manager), personnel employed at the CSCOM, administrative secretaries, and teachers in community schools. A branch of the civil service dedicated to local government recruits these employees. While appointed by central authorities, their salaries depend on the communes that typically are short of funds, in no small part because the government fails to transfer what it owes them (see the subsection Fiscal Autonomy). With the salaries of these civil servants often several months behind, it is difficult to find qualified recruits. Importantly, the chieftaincy—the historical locus of real decentralized power in Mali—has been absent from decentralization structures. During the recent Estates General on Decentralization, participants supported increasing the role of traditional authorities within the communes (RFI 2013). Traditional authorities represent legitimate authority in many communities and, in practice, village chiefs retain considerable power and are often the first level of authority to which citizens flock in time of need. In recent public opinion surveys in Mali, support for chiefs was clear: 82 percent of Malians stated that they have a great deal of trust for traditional authorities whereas only 50 percent said the same about police and only 43 percent trusted courts a great deal (Afrobarometer 2009). Of course it may be precisely the official distance between traditional authorities and the state that preserves the legitimacy of chiefs. The village chief is chosen by village councillors who are, in turn, elected by the village assembly; the prefect then confirms the chief within three months (hence, deconcentrated authorities also supervise traditional village authorities, illustrating the continued reach of the central state throughout the territory). Formally, the mandate of the village chief and his or her councillors is for five years and can be renewed an unlimited number of times. In practice, however, it is permanent and usually ends with the chief’s death. The resilience of chiefly power at the village level further squeezes communes between the authority of chiefs and that of the prefects, thus contributing to their limited autonomy. Fiscal Autonomy

The gap between law and practice is the greatest with respect to fiscal autonomy. Legally, communal finances are composed of resources transferred from the central government, those raised locally and loans. The

124

Mali

state allocates funds to local government with respect to size of the population of the commune, the ability to mobilize resources, and the extent to which facilities and infrastructure exist. These transfers are derived from income from patents and licenses, 80 percent of the TDRL, and taxes on livestock and firearms, salaries, motorcycles and bicycles, and delivery of the carte professionelle (professional identity card) (Sy 2010). These transfers are then expected to be complemented by local resources. Local taxes are on transport vehicles; small boats and carts; games installed in public venues; nightclubs, restaurants with entertainment, sales from drinks and small restaurants; publicity in public spaces; construction authorizations; cereal mills; mining and construction materials; and garbage collection. In addition, communes collect fees from the authorization of events and income from patrimoine (historical resources).4 Finally, communes can be loaned money directly (because they have legal standing) and they have the authority to “act in cooperation” with nongovernmental organizations in their region (Seely 2001: 500). This authority to deal with foreign NGOs is important for communes and opens up individual avenues for revenue acquisition. Those who are authorized to negotiate with donors become the local gatekeepers of these flows (and, hence, controllers of the rent) from which significant local power can be obtained. In general, it is exceedingly difficult for communes to be financially self-sufficient. First, the government is typically remiss at making transfers. Second, local governments struggle to collect taxes. The TDRL currently represents 75 percent of the communes’ operating funds, and tax compliance is highly variable among communes and over time. Average compliance rates tend to be low, at 20–40 percent, but some mayors have reported compliance of 70–80 percent (Tefft et al. 2003: 6). In the area around Sikasso, there is great fluctuation in farmers’ cotton income and, therefore, high variability in tax compliance from year to year. Some mayors believe that in order to improve tax compliance, there needs to be civic education and regular communication between councilors and the citizens. In one example, a village did not pay taxes because no one from the village was elected to the communal council; therefore, the community did not believe that the municipal government would respond to their interests (SNV and CEDELO 2004). Chéibane Coulibaly, J. Tyler Dickovick, and James T. Thomson (2010) found Sikasso city councilors with personal interests in private sanitation companies, an area that communes are involved. Many believed that mayors, prefects, treasurers, and other government officials collude to divert public funds. Local officials complain, with little irony, of the unwillingness of citizens to pay taxes (Coulibaly, Dickovick, and Thomson 2010: 35–36). In the Mopti region, mayors argued that “people are too poor to pay” (Tefft et al. 2003: 6). Each of these factors contributes to low tax compliance.

Susanna D. Wing and Brehima Kassibo

125

Decentralized communities are rarely in effective control of the fiscal tools given to them by law. For example, communes find it exceedingly difficult to receive their portion of commission levied on wood fuel by the Rural Management Service, even though they are entitled to raise taxes on village forests belonging to the patrimoine. This is largely due to the failure of the state to transfer forest conservation legally to the communes (see Kassibo 2009). The last option for communes to obtain funds is the ANICT that makes money available for capital investments and can be accessed only if a commune can meet the required 20 percent local contribution. Because tax compliance is so low, it is hard for many communes to come up with these initial funds. For example, 75 percent of the communes in the Kolondieba District of the cotton zone could not use their 2001 ANICT drawing rights for infrastructure and equipment investments (Tefft et al. 2003: 6). Regionally, communes in the Sikasso region have been the most successful in terms of mobilizing resources and realizing projects (CFA22 billion for 2,596 projects). Ségou follows with CFA21 billion and 1,879 projects, slightly ahead of Mopti (CFA20 billion and 1,739 projects). The region of Kidal in northern Mali has had the least amount of money (CFA3 billion) and only 229 projects (ANICT 2010: 8.) and this may have contributed to dissatisfaction with decentralization. The original explosion and then cooling off of decentralization fervor is visible in annual drawings of funds by communes that went from CFA579 million in 2001 to CFA28 billion in 2007, and then dropped to CFA22 billion in 2009. In the Sikasso region, village associations provided funds to cover the commune contribution, but this solution is limited to instances where the ANICT investment for a project is in the association’s village. The members of the association are often communal council representatives. In some cases, an association will pay expenses such as salaries or lodging of teachers or health workers, or an annual contribution to CSCOM (Tefft et al. 2003: 7). Accountability

Decentralization did not significantly address underlying differences in political power. The election of commune councillors, for example, seen as one of the more promising mechanisms for political accountability, produced slight rather than substantive advances. Largely, communal elections (held in 1998, 1999, 2004, and 2009) are considered to have been competitive, free, and fair. Nevertheless, voter turnout was low and there have been incidents of fraud, purchased votes, and sale of the placement of names at the top of the electoral lists. Voter participation in 1998 in urban municipal elections was 33 percent on average, and 43 percent in the first rural communal elections in 1999. In 2004, the national average was again

126

Mali

43 percent. In every case, Timbuktu (both the urban municipality and the region) had the highest rates of participation (1998: 50 percent, 1999: 46 percent, and 2004: 54 percent) (Wing 2008a: 97).5 One reason for this appears to be decentralization’s important role as a response to conflict in the north in the 1990s and the importance that the population there places on local government as a result. By 2004, the frustration with decentralization was widespread, as political interest in it subsided. Nevertheless, the relatively high level of voter participation in the Timbuktu region raises questions about the support in this area for an independent Azawad.6 Citizens have a right to information on budgets at every level of government (Kimenyi and Meagher 2004: 174), but across the country people are rarely aware of their full rights or of the mechanisms in place to promote transparency. Civil society organizations and donors such as Le Programme d’Appui aux Collectivités Territoriales/Deutsche Gesellschaft für Technische Zusammenarbeit (Program for the support of local government/German Organization for Technical Cooperation, PACT/GTZ) have promoted civic education programs to help communities demand accountability and transparency in local government. Amid these modest advances, activities such as fraud and the manipulation of electoral lists reproduced a system that favors local elites. Specifically, traditional authorities have dominated communal elections. This trend appears to exclude women, youth, and disadvantaged classes (in 2004 only 6.7 percent of counselors were women, in 2009 8.6 percent; MPFEF 2009), and marks payback by village elites who were often left out of the decentralization reforms. On the one hand, chiefs seem to benefit from the elections by gaining local political legitimacy. On the other hand, their election to municipal councils embed them in local patterns of accountability and corruption. Political parties that might have promoted greater equity soon conflicted with centralized bases of power that were less amenable to dispersing influence. For years, established political parties dominated communal elections, although this domination is slightly less than at the national level, suggesting continued local pluralism. The Alliance for Democracy in Mali (ADEMA), the ruling party until 2002, and since then a partner in the presidential majority, gathered about 28 percent and 29 percent of the votes in 2004 and 2009, respectively. In comparison, it received about 33 percent in each of the corresponding legislative elections around the same dates. There was an erosion of the importance of political parties in Mali after the election of ATT as an independent in 2002. He gathered the support of ADEMA personalities, but did not join the party itself. Instead, he organized a loose coalition around his person, the Alliance pour la démocratie et le progress (Alliance for Democracy and Progress). For a time this created broad support for his regime and afforded him a high degree of autonomy,

Susanna D. Wing and Brehima Kassibo

127

but ultimately it undermined accountability by marginalizing political parties and causing a blatant return to a more personalized type of politics. Decentralization could not address deeper divisions and power bases. Following the overthrow of ATT in 2012, many Malians blamed a politics of consensus as having corroded democracy in the country. Consensus is inherent to Malian culture, but its pursuit has been at a significant cost. Consensus and democracy are difficult bedfellows. Consensus requires unanimity and compromise that can be a useful, if time consuming, method of getting things done. This process, however, carries great risks in a country in which personal politics dominate and access to the state virtually guarantees power and material resources. Under ATT, consensus gagged all opposition and kept the National Assembly under the yoke of executive power. Accountability at the local level was undermined by the fact that political parties had little autonomy and no realistic platforms. This extended from the national level to local elections where personal politics continued to dominate. Capacity

In a country as poor as Mali, it is no surprise that technical and administrative capacity is low. Examples can include the low literacy level of councillors and municipal secretaries. In some instances, for example, only one municipal employee can read and write. Low pay also promotes high staff turnover (SNV and CEDELO 2004). As noted earlier, the communes’ lack of fiscal resources compounds the problem. As a result, there is strong reliance on external actors to meet basic needs of communities. This dependency toward donors was originally structured within the CCCs, but these were phased out in 2004. For years, donors operated though the FNACT. In a country as vast as Mali, with a predominant desertlike environment, geography constrains isolated communes. The ongoing crisis and annexation in northern Mali has resulted in further isolation of the north and the instability in the south led to a freeze on many donor funds and the loss of hundreds of billions of CFA francs.7 With presidential and legislative elections now complete, donor funds are once again forthcoming. One of the strengths of the Malian political system is its civil society, which is made up of national and local organizations. Although the technical and administrative capacity of local organizations depends on the level of support that they receive from external sources, there are many capable organizations. Civil society and external donors were successful in pressuring the Malian government to move forward with decentralization, and they have been active in many aspects of decentralization. International NGOs work with local NGOs to support health care and education initiatives, serve as CCCs, and provide much-needed guidance and support to communes (Tefft et al. 2003:4).

128

Mali

In fact, effective capacity at the local level sometimes lies more with civil society actors, such as service-providing NGOs, than with the communes themselves. International NGOs provide resources to these local actors; however, their funds can be difficult for communes to access. Indeed, even when the CCCs were in place, the richest communes benefited most from their subsidies and loans, in part because the municipalities were required to pay a deposit of 10 percent on the amounts requested. Access to loans thus correlated with the wealth of communes, reinforcing existing inequalities. Communes that have managed to become sister cities with communities abroad also have a significant financial and material advantage. By the comparative standards of West Africa, Mali has a strong civil society that has been integral to decentralization as well as for democracy and development in general. As noted above, the decentralization process has included the groupes d’étude et de mobilisation aux niveaux régional et local that used civil society to help design the communes and to promote civic education across the country. Village associations provide resources for decentralization. All inhabitants of a village are members of the association, and these associations were once used by the state to govern local communities. With democratization and decentralization, these organizations have provided important infrastructure and helped disseminate information to communities. They also remain in the control of local elites who can use their influence to undermine decentralization. Kafos (multivillage associations united by proximity and blood ties) have been important sites for returning some power to the community level. These institutions date back to the Ghana Empire of the eighth century, and were designed to address community problems that expanded beyond the village level. They were strengthened by the French colonial regime and disbanded by Modibo Keita. Currently, they have been reinvigorated as contemporary development associations (Coulibaly, Dickovick, and Thomson 2010: 5–6). Whereas communes struggle to mobilize local revenue through taxes, it should be noted that “kafos and local development associations have demonstrated an ability to develop local infrastructure facilities, which they finance with adequate funds voluntarily mobilized by their members” (Coulibaly, Dickovick, and Thomson 2010: 46). In addition, many Malians draw on indigenous institutional resources for development needs, rather than waiting for the state to finance projects. According to Coulibaly, Dickovick, and Thomson, “Organization occurs through hometown associations, youth associations, women’s groups, socio-professional groups, and migrant societies, among other forms. Malians use these with consistent economy to develop infrastructure facilities (school classrooms, clinics, wells, mosques) in the communities that residents desire” (2010: 40). International donors have operated within these shifting tendencies. In terms of local service provision, several international donors work with

Susanna D. Wing and Brehima Kassibo

129

local civil society organizations to operate community schools and health centers (CSCOMs). CARE, World Education, and Save the Children have all supported community schools through funding and training. Interestingly, the communes have constituted themselves into lobbying associations that are now considered civil society actors. The Association des Municipalités du Mali (Association of the Municipalities of Mali, AMM) and the Association des Collectivités, Cercles et Régions du Mali (Association of Collectives, Cercles and Regions of Mali, ACCRM) are two groups that represent local government. AMM is organized by cercles and is supported by various European organizations. Unfortunately, AMM does not function well and its impact is seen only during the annual Day of the Commune (Sy 2010).

Political Incentives and Constraints to Decentralization Given that decentralization was a principal goal of the 1991 transitional government and came as a response to popular demands for good governance, elite political resistance to decentralization was negligible, despite the lack of capacity and financial resources that constrain the process. Indeed, regional elites were among the main supporters of decentralization because it gave them more opportunities to gain power in regional and local administration and obtain greater access to aid money (Seely 2001: 513). This cuts both ways, however, as it creates opportunities for development as well as for rent seeking (see Coulibaly, Dickovick, and Thomson 2010). Under centralized rule, rural communities and geographically remote areas experienced arbitrary collection of taxes and few benefits from money taken by the state. The driving force behind decentralization was to give communities a stake in local government and to encourage people to vote and pay taxes as a way to participate in the future of their communes. At the end of the day, local support for decentralization varies as a function of the benefits received. Communes where investments in infrastructure and equipment are lacking tend to exhibit more lukewarm support (Tefft et al. 2003: 3). Those where economically stimulating or socially needed investments were made in the wake of community planning meetings are more likely to believe in the benefits of decentralization. Decentralization has had its organized opponents and has revealed competing blocs of power. Employees of central administrative structures have often been reticent to give up control. Moving the Decentralization Mission from the Ministry of Territorial Administration, as well as attempts to keep civil servants who had served in Mali’s Second Republic (1978– 1991) away from the new decentralized organs, illustrated the will of Third Republic (1991–2012) advocates to maintain control of these structures. In

130

Mali

addition, deconcentrated administrators of the state, such as governors and prefects, have posed challenges to mayors whose new authority they resist (Lippman and Lewis 2008: 9–10). Decentralization has been marked by broad popular support. Some have contended that insofar as decentralization was initiated as a response to demands from the Tuareg of the north, it was not a goal equally shared by southern communities.8 The pervasiveness of decentralization in Malian political discourse weakens such a claim. Decentralization was, for example, strongly supported by the National Conference. Interestingly, at times the very institutions involved in decentralized public services are the most apprehensive about it. Indeed, decentralization may threaten already effectively decentralized public services provided by NGOs, associations, and cooperatives. For instance, CSCOM workers, who are paid by the communes, came into conflict with Associations for Community Health, which are local private community cooperatives that sign agreements with the state to be local providers of community health. Local representatives of these associations resisted the transfer of local health authority to CSCOM because it undermined their power. As outlined above, during President Konaré’s term in office (from 1992 to 2002), decentralization was a key aspect of democratization. Mali’s progress on decentralization helped establish the country as a model for democracy on the continent. Konaré benefited from the fact that Mali was at the forefront of African democratization. Nevertheless, there were fundamental issues concerning the dominance of his political party, ADEMA, and local elections were delayed because of the difficulty in establishing infrastructure throughout the country. Complications related to the establishment of legitimate voting rolls hindered the decentralization process by raising doubts about the transparency of elections. Decentralization was also a political strategy of the government to buttress regime legitimacy. It became a showpiece for donors and part of what has garnered Mali such praise as a model democracy in Africa. Yet decentralization has not been successful in reducing poverty, one of its key goals. It was meant partly to ease the burden on the state by involving civil society in the creation of community schools and community health centers and transferring responsibility for these services to local government. Unfortunately, the lack of resources in municipalities has meant that these communities struggle to meet the basic needs of their populations. Prior to the overthrow of ATT’s regime, the main ongoing decentralization-related initiative was the proposal by the Daba Diawara Commission that the HCCT be replaced with an upper legislative house or senate. The proponents of this change included the pro-ATT coalition of parties. Opponents resisted the proposal that a large portion of the senators be appointed by the president of the republic, and argued that it would be too expensive

Susanna D. Wing and Brehima Kassibo

131

for the state to take charge of senators’ expenses. Many viewed the proposal as a consolidation of executive power. Political incentives at the subnational level must also be considered. The postcolonial strategy of traditional authorities has always been to seek to maintain their influence and control over land irrespective of the political system. They have historically controlled access to land in their communities and have struggled to maintain this power, which is the foundation of their political status. Given the authority legally granted to communes over land issues (e.g., roads, waterworks, markets, environmental protection), many chiefs have sought election to communal councils and subsequent selection as mayor. In other cases, they have nurtured deep personal ties with the mayors and council members. According to Susanna Wing, “In places where local notables have not been elected they maintain veto power over community decisions because of the legitimacy they hold in the community. It would be unwise for local officials to act counter to notables who often have the ability to mobilize the community to block any reforms of which they do not approve” (2008b: 30). Thus, elected officials can do little without the support of traditional authorities. This situation creates the potential for local elites to capture decentralization. There are several strategies that a local chief might use to perpetuate his political influence. One is to ensure that he or she has a son in each political party running for local election. In this case, no matter which party is elected, the chief will have a direct link to the party in power.9 Another example involves an eighteen-village commune near the border of Côte d’Ivoire where immigrant families are the majority and have considerable economic power. The local chief made certain that, even though by virtue of demography the mayor would be elected from the immigrant families, the chef lieu du commune (municipal seat of government) would be located in one of the five villages still dominated by local families,10 which would continue to symbolically project their power. Yet it is a matter of perspective whether chiefly control of decentralized institutions represents a hijack of power or traditional legitimacy. The power of chiefs in villages, unequal though it may be, is embedded in historical practice, reciprocal relations, and legitimate authority. The state, even through its decentralized manifestations, remains the outgrowth of the colonial enterprise and continues to be seen by many as parasitic. If chiefs take over decentralized state institutions, they might hamper their functioning as anticipated by the laws and modern conceptions of accountability, but they might also embed these institutions in local social practices and confer greater legitimacy on them where they might become the locus of real decentralized institutions of collective choice. Intervillage conflicts represent another subnational problem largely triggered by decentralization. The main issues typically lie with the drawing of communal borders and the location of the municipal seat. In the com-

132

Mali

mune of Sanankoroba, for example, news of decentralization “intensified tensions between villages, different social and ethnic groups, and indigenous inhabitants and incomers over where the seat of the mayor should be and who should be given decentralized powers” (Djire 2004: 11). In some cases, commissions of arbitration have had to intervene in determining the boundaries of communes (Kassibo 2007). The location of the municipal seat is important because it is seen as the repository of real power, reminding people of the chefs de canton or chefs d’arrondissement from the predecentralization days, with whom mayors are often assimilated in people’s minds. The term papa commandant is often used to describe mayors and used to be applied to the chefs de canton and d’arrondissement. Thus, the village that hosts the seat of municipal government is typically seen as aggregating powers and it can dominate other villages in the commune (Kassibo 2007). Unequal representation of villages on municipal councils has also become a point of conflict. In Sanankoroba, Moussa Djiré found that municipal elections resulted in a municipal council that was viewed as unrepresentative by some villages. He writes, “The 17 councilors on the municipal council represent six political parties and only six of the 26 villages in the municipality, with 10 councilors from the village of Sanakoroba” (Djire 2004: 13). This disparity in representation of villages within communes challenged the ability of decentralization to address local governance concerns.

Conclusion Decentralization has fallen short of its promise as a force for significant change. Across Mali, political elites recognized its importance as a way to bring power closer to the people. Yet as Ousmane Sy (2010) notes, the state lost its “vision,” and in the process its “inspiration.” Indeed, decentralization has faced numerous hurdles that ultimately contributed to the overthrow of ATT in March 2012. The challenges faced by communes, including weak technical and administrative capacity, financial constraints, and a lack of transparency and accountability, have had ramifications for society as a whole. Communes could not meet the basic needs of the local populations and suffered from a lack of organization and basic structures, compounded by weak technical assistance from the central government and donors. A structure for the professional training of elected officials and local government civil servants was never created (PACT/GTZ 2009). The deconcentrated services of the state lacked adequate preparation for the transfer of responsibility, and there was a great deal of reticence to transfer responsibilities and power or to let go of their former ways of doing business (Diallo 2010: 1). More-

Susanna D. Wing and Brehima Kassibo

133

over, political parties saw decentralization as an opportunity to expand their reach into villages across the country, but politics of consensus instead meant the continued political dominance of ADEMA and the pro-ATT coalition. For these political actors, decentralization provided another layer of government to which they had access. Although decentralization can promote local accountability, politics in Mali remained heavily influenced by patron-client relations, which decentralization reproduces locally and thereby creates additional opportunities for corruption. Finally, Malian decentralization has been characterized by donor dependency. In terms of investment in the 2001–2005 period (which followed the installation of the first rural communes), 76.9 percent of local government investments, which amounted to US$7.7 million, were financed by donors.11 The National Budget financed 12.8 percent and communes contributed 10.3 percent. For the period to 2009, the total financial contribution for decentralization, according to ANICT, amounted to $25.6 million, supporting no fewer than 11,792 projects. In a country as impoverished as Mali, it is not surprising that decentralization would rely heavily on external actors. The latter years, however, have seen some significant progress in the collection of the TDRL, which improved from 2007 to 2009 from approximately $6 million to $18 million. As a result, some communities have become less dependent on external aid (Diallo 2010). This may be the result of civic education or increasing legitimacy of the local government after multiple election cycles. The lack of political will ensured that the decentralization process fell short of achieving important goals. Mamoutou Diallo notes that decentralization was a priority until 2005, with the national government actively promoting the administrative and technical independence of local governments (Diallo 2010: 2). Widespread donor support brought technical assistance and resources. Beginning in 2006, however, decentralization became but one aspect of a broader development plan in which the overall institutional development of the state figured prominently. In addition to this switch in emphasis, Mamoutou Diallo (2010) also notes the institutional weakness of organizations that are integral to decentralization, including the HCCT, AMM, and ACCRM. The lack of continued political will contributed to the widespread malaise that was apparent at the time of the March 2012 coup d’état. The delicate balance between deconcentrated state services and communes left too much control in the hands of the former at the expense of the latter. The obstacles to decentralization restrained the process and contributed to the fall of ATT’s regime.12 Challenges included the need to build capacity (administrative, fiscal, and technical), the lack of resources and skills that undermined decentralization’s achievements to date, and the shift away from decentralization by the political elite. These threatened the process as

134

Mali

a whole. Finally, the dependence on external actors is particularly overwhelming because of Mali’s relative poverty. Donors not only provide resources that the state is unable or unwilling to provide; they are also a symbol of the state shirking its responsibility for providing resources to the people. Elites with connections to state institutions, at all levels, often displayed wealth far beyond the means of most Malians, further underscoring this divide. For years, Mali was ahead of several countries of West Africa with respect to decentralization and local governance. Nonetheless, compared to other countries, Mali’s decentralization suffered from weak local elections characterized by fraud, the purchase of votes, and weak voter participation; a generalized lack of transfer of power, resources, and property to the decentralized collectivities; strong dependence on external aid; and a low level of participation of women as elected representatives in the subnational and national government structures. What is Mali’s future? Will Mali reestablish a secular, democratic country with an intact territorial integrity? The presidential election took place in 2013, with Ibrahim Boubacar Keita of the Rally for Mali (Rassemblement pour le Mali) being elected. Mali has legal codes in place to continue to pursue decentralization. And the Estates General on Decentralization marked the government’s keen interest in reinvigorating decentralization across the country, with a focus on the transfer of capacity and resources to the local level (RFI 2013). The MNLA was skeptical and refused to participate in the three-day conference. President Keita opened the Estates General with calls for increased local autonomy. While government supporters argued that decentralization “encouraged Tuareg ambitions of independence,” some Tuareg groups believed the reforms did not go far enough in establishing regional autonomy (Diallo 2013). The failures of decentralization will have to be directly addressed if the country is to have any hope of anchoring and sustaining meaningful democratic practices and structures. Reconciliation and durable peace with the Tuareg is a foundational element to stable democracy in Mali. But as the country moves slowly toward national reconciliation, it remains unclear where the demands of the Tuareg fit into this picture.

Notes Susanna Wing would like to thank Cameron Sells for research assistance on this project. 1. For a complete list of decentralization laws and decrees, see Wing and Kassibo (2010). 2. If nothing else, decentralization has thus created a vast amount (10,000+) of new positions of public authority. 3. ADEMA won 28 percent of the votes; the main opposition parties, Union pour la République et la Démocratie (Union for the Republic and Democracy, URD)

Susanna D. Wing and Brehima Kassibo

135

and Rassemblement pour le Mali (Rally for Mali; RPM) received 14 percent and 13 percent, respectively. A new round of municipal elections was held successfully in 2009. ADEMA (now renamed ADEMA-Pasj) and URD were the only two parties presenting candidates in all 703 communes. They obtained respectively 29.5 percent and 17.9 percent of the votes. RPM received 7.2 percent and all other parties received less than 5 percent. See Wing (2008a). 4. For most people at the grassroots level, decentralization therefore translates broadly in a vast amount of new taxes. 5. Accurate data on 2009 communal elections is not available. Informal observer tallies from the US embassy ranged from 15 to 35 percent turnout in Bamako and perhaps 40 percent in the interior. See “Local Elections in Mali,” unclassified cable, April 27, 2009, http://wikileaks.org/cable/2009/04/09BAMAKO259.html. 6. The argument here is that disillusioned individuals would be less likely to participate in elections. However, electoral participation in the north has been relatively high. If individuals widely supported an independent Azawad, then it is logical that participating in elections would be a low priority. 7. See Adam Thiam, “Comment la crise plombe l’économie du Mali,” Le Républicain, June 14, 2012, http://www.malijet.com/a_la_une_du_mali/45500-comment _la_crise_plombe.html, accessed November 25, 2013. 8. Chéibane Coulibaly, cited in Kassibo (2007). 9. Ousmane Traoré, deputy to the National Assembly, interviewed by Susanna Wing, Bamako, Mali, January 10, 2007. 10. Ousmane Traoré, deputy to the National Assembly, interviewed by Susanna Wing, Bamako, Mali, January 10, 2007. 11. The main decentralization partners are the European Union, the French Agency for Development, the Direction du Développement et de la Coopération (Agency for Development and Cooperation) of Switzerland, the Kantonale Fachgruppe Wellenberg (Wellenberg Cantonal Department), the SNV, and the Canadian Agency for International Development. 12. It is important to note that this is just one of a multiplicity of factors, including the collapse of Qaddafi’s regime, the movement of arms into Mali, and a surge in activities of the al-Qaeda in Islamic Maghreb.

7 Mozambique: Decentralization in a Centralist Setting Beatrice Reaud and Bernhard Weimer

Prior to the first multiparty election in 1994, Mozambique was governed by a highly centralized government. That government was the product of a socialist experiment by the Frente de Libertação de Moçambique (Liberation Front of Mozambique, FRELIMO) and the inheritance of Portuguese colonial institutions. The 1994 election marked the end of a sixteen-year conflict and brought stability by creating space for the conversion of a guerilla group, Resistência Nacional Moçambicana (National Mozambican Resistance, RENAMO), into an opposition political party. The 1990 Constitution, followed by the passage of decentralization laws in 1997 and 2003, set the framework for decentralization to provincial, district, and municipal levels. Among others, these laws created thirty-three autarquias (politically autonomous municipalities), which were augmented by another ten municipalities in 2008. Significant economic development, sanitation, and public health responsibilities were devolved to these autarquias, along with some fiscal decentralization of fee and taxation authorities. Three subsequent municipal elections and increasing utilization of fiscal authorities by municipalities have consolidated gains, despite a wide variety of capacities, infrastructure, population, and access to resources. Despite these successes, our main finding in this chapter is that Mozambican decentralization is compromised by its bifurcated, or asymmetric, nature. The main challenge to decentralization is the centralizing tendency of a dominant-party state that dictates a gradual approach to decentralization (gradualismo), focusing on devolution to municipalities and deconcentration to districts of responsibilities without sufficient resources. The result is a parallel system of decentralized subnational units: the deconcentrated, centrally controlled provinces and districts, known as Órgãos Locais do Estado (Local Organs of the State, OLEs), and forty-three municipalities with 137

138

Mozambique

devolved authority, autonomy, and elected leaders. While Mozambican decentralization resulted in a carefully engineered, multi-institutional setup of these three subnational units of government, the overall decentralization objective shifted from postwar peace through democratic decentralization and poverty reduction to consolidating the dominant-party system at the local level. Several factors undercut subnational autonomy; namely, FRELIMO’s dominance of the Mozambican state, the weakness of municipal revenue bases, a lack of coordination of sector ministries at the district level, and a bifurcated hiring system. Despite the dispersion of participatory governance processes across autarquias, accountability to voters is constrained by a combination of FRELIMO party politics that limit performance incentives, a lack of viable political alternatives, the limited authority and resources of consultative provincial and district bodies, and voter apathy. Lastly, the lack of qualified public personnel at all levels, district infrastructure disparities, and the existence of a small and Maputo-based civil society all limit capacity. Before detailing these aspects of decentralization in current politics, we offer a brief survey of the political background and history of decentralization in Mozambique.

Political Background and Structure Mozambique’s colonial institutions shaped present-day decentralization, the political dynamic of the dominant political party (the former liberation movement, FRELIMO), and the centralizing tendency of today’s national government. From 1900 to 1962, the Portuguese established a highly centralized government administration based in the capital Lourenço Marques (later renamed Maputo) and a few urban areas and towns along the coast and in the economically or strategically relevant hinterland. Like today’s fortythree municipalities, these enjoyed some degree of self-government autonomy through the câmara municipal (municipal council). The indigenous population lived outside this system in rural areas governed by district administrators and postos (subordinated administrative units) headed by a chefe do posto, all accountable to the capital. This geographical and horizontal administrative division continues to structure the administrative and territorial division of Mozambique today. The bifurcated system of local government and administration was reaffirmed by the decentralization reforms in the past twenty years, leading to a separation between autarquias in urban areas with a certain degree of administrative and fiscal autonomy and deconcentrated, subordinated units of the central administration for the rural areas. After a ten-year war for independence, Portuguese colonial control ended in 1975 and was followed by a period of single-party dominance by FRELIMO, which did little to change the government’s structure from its

Beatrice Reaud and Bernhard Weimer

139

Portuguese roots. With 11 provinces, 128 district governments, and 393 postos administrativos (administrative posts) and conselhos executivos (local executive councils, the central government replicated itself throughout the country, with almost all ministries represented at all subnational levels. The president appoints provincial governors who, in turn, approve the district administrators that are appointed by the Ministry of State Administration. Between 1977 and 1992, Mozambicans lived through a harrowing war waged between the FRELIMO government and the resistance movement that was to become RENAMO. RENAMO became a legitimate political party after signing the Rome Peace Accords with the FRELIMO government on October 4, 1992. It is estimated that destruction from the war totaled $15 billion, that 900,000 Mozambicans were killed, and that some 4.5 million people were displaced as refugees in neighboring countries, with Tete and Zambezia provinces being the most affected (Abrahamsson and Nilsson 1995: 66). Political liberalization began with the peace process between RENAMO and FRELIMO and the National Assembly’s approval of Mozambique’s first multiparty constitution in November 1990. The president elected in Mozambique’s first multiparty elections in October 1994, Joaquim Chissano, won by 53.3 percent of the vote, with FRELIMO winning 129 seats in the National Assembly. The RENAMO candidate, Afonso Dhlakama, garnered 33.7 percent of the vote, with his party winning 112 seats. Mozambique, however, remains a de facto one-party state; FRELIMO effectively has controlled the state and all resources, including the public administration and public enterprises, since independence. FRELIMO “represents the ‘foundational’ political force turned into a dominant party whose changing leadership and elites have been shaping the politics and policies informing the distribution of wealth and power in Mozambique” (Weimer 2012a: 3).

The History of Decentralization Decentralization in Mozambique was driven by three primary factors: the need for postwar economic and political stabilization; the subsequent need for economic recovery and shift from a central, planned economy to a market-based economy; and FRELIMO’s attempt to consolidate its system of government and control at the local level. The war exacted such a crippling effect on the Mozambican economy that the government adopted political and economic liberalization reforms both to bring an end to the war and to stabilize the economy. FRELIMO proposed and passed a constitution which, while locking into place conditions that were favorable to the party’s continued dominance, also opened up political competition to the new opposition (RENAMO) and created a space for democratization of local government (Manning 2005).

140

Mozambique

In early 1994, the FRELIMO parliament envisaged a devolution model with elected assemblies and mayors for urban and rural administrative units. But with the surprisingly good electoral performance of RENAMO, notably in rural areas, the FRELIMO leadership had second thoughts about extending democratic decentralization to the rural areas. This led to an amendment of the 1990 Constitution that cemented parallel systems of local government, or asymmetric decentralization: urban autarquias (povoações) referred to as poder local (local power) on the one hand, and the aforementioned deconcentrated OLEs on the other. This legal framework is the basis for the modern-day bifurcated process of decentralization that combines deconcentration of central state planning, budgeting, and service delivery functions to provinces and districts with devolution of fiscal and political authorities to autarquias. Specific legislation was enacted in 1997 for thirty-three municipalities (with an additional ten added in 2008) and for the deconcentrated units in 2003. The first multiparty municipal elections were scheduled for 1996, but postponed three times due to RENAMO-FRELIMO politics, notably regarding the implications of the 1996 constitutional amendment (Weimer and Fandrych 1999). Prior to decentralization, the government issued regulations that classified the country’s main urban areas based on several criteria that included population density, social and technical infrastructure, and economic potential. There was some controversy about the selection and designation of 10 out of a total of 128 vilas (district centers), which, together with the 23 cities and towns, made up the first 33 municipalities. This controversy contributed to a boycott by RENAMO of the first elections because party members felt that they were excluded from negotiations on the selection, and the result was only a 15 percent voter turnout (Cuereneia 2001). When ten additional population centers were similarly elevated to municipal status in 2008, bringing the total to forty-three, some political analysts speculated that most of the cities chosen were “FRELIMO towns” selected to stack the deck in favor of the dominant government prior to the 2009 mayoral elections. RENAMO has participated in all municipal elections since 2003 and the voter participation has increased from 15 percent (1997) to 46 percent (2008). Its best result was in 2003, when the party won five out of thirty-three mayoral contests.

Authority: Structure, Institutions, and Legal Framework The two main legal frameworks of decentralization are the Pacote Autarquico of 1997 and the Lei de Órgãos Locais do Estado (Law on Local Bodies of the State) of 2003. The resulting constitutional and legal framework gives Mozambique four main levels of government (see Tables 7.1 and 7.2):

Beatrice Reaud and Bernhard Weimer

141

1. Central government (including the president, prime minister, Council of Ministers, National Assembly, and line and sectoral ministries);1 2. Eleven provincial governments (consisting of delegations of central government ministries and directorates); 3. One hundred twenty-eight district governments (consisting of district-level central government delegations that provide services to localities, villages, and administrative postos); and 4. Forty-three autarquias (which are politically autonomous municipalities).

Table 7.1 Degree of Decentralization and Capacities by Subnational Units Subnational Unit

Degree of Decentralization

Capacities

Municipality (43) (autarquia)

High (Devolution)

• Municipal elections (1999, 2003, 2008) for mayor and municipal assembly. • Only subnational government that can charge and keep fees and taxes. • Responsible for policing, local infrastructure (drainage, markets, public health, and sanitation), and management of urban space. • Funded primarily through central government transfers with varying degrees of reliance on own-source revenue.

Provinces (11)

Medium (Partial devolution with deconcentration)

• Governors, appointed by central government, recommend appointment of district administrators to the Ministry for State Administration. • Provincial assembly elections (2009). Assemblies are consultative bodies. • Provinces collect taxes and deliver services as arms of central government.

Districts (128)

Low (Partial deconcentration)

• Participatory institutions and consultative councils are appointed, not elected. • Districts are accountable to provincial and central government; collect taxes and provide services on behalf of center.

142 Table 7.2 Time Line of Decentralization in Mozambique Mid-1980s

Election of local party officials; provincial governors and ministers appointed.

1990

National Assembly approves first multiparty constitution.

1992

Rome Peace Accord, end of internal war.

1994

Decentralization legislation (Law 3/94) passed; first multiparty presidential and National Assembly elections held.

1995

First local government elections postponed.

1996

Constitutional amendment introduces two types of local government: devolution and autonomy to urban municipalities, deconcentration to rural districts.

1997

Laws 2/97, 7/97, 9/97, 10/97, and 11/97 (Pacote Autarquico) designated the 33 original urban municipalities.

1998

First municipal elections (RENAMO boycott; FRELIMO wins all 33 mayoral elections. First guidelines for district planning published. First district development plan and introduction of PPFD process.

2003

Law 8/2003 (OLEs) passed. Decree 63/2003 creates the representative of the state. Second municipal elections held; 5 of the 33 municipalities gained RENAMO mayors.

2005

Constitution reconfirms parallel system (OLEs vs. autarquias), introduces provincial assemblies.

2006

OIIL funding approved for districts. Regulation for transfer of additional competencies to municipalities approved.

2008

Finance law amended to provide additional revenue authorities to autarquias. FCA and FIIL transfers reduced from 3 percent to 1.5 percent of central government own-source revenue. Law 3/2008 designates an additional 10 municipalities. Third municipal elections held: FRELIMO wins 42 of the 43 mayoral seats.

2009

First Provincial Assembly elections held.

2010

Armando Guebuza government reaffirms stated commitment to decentralization.

Notes: RENAMO is Resistência Nacional Moçambicana (National Mozambican Resistance). FRELIMO is Frente de Libertação de Moçambique (Liberation Front of Mozambique). PPFD is Programa de Planificação e Financiamento Descentralizado (Program of Decentralized Planning and Finance). OLEs are Órgãos Locais do Estado (Local Organs of the State). OIIL is Orçamento de Investimento e Iniciativas Locais (Budget for Investment and Local Initiatives). FCA is Fundo de Compensação Autárquica (Municipal Compensation Fund). FIIL is Fundo de Investimentos de Iniciativa Local (Investment Fund for Local Initiatives).

Beatrice Reaud and Bernhard Weimer

143

The 11 provincial and 128 district governments are deconcentrated entities that represent the central state. Provincial governors are appointed by and report to the president, with executives that are deconcentrated and called delegations of line ministries, such as the Ministry of Health or Education, at the local level. The central government replicates agency functions at the provincial level, and selected line ministries are grouped into district services at the local level.2 District administrator appointments are made with input from provincial governors. The personnel at the municipal, provincial, and district OLEs are civil servants who are usually recruited, remunerated, and promoted according to central government regulations through the Ministry of Public Service. They are part of the regular body of civil servants. Direct recruitment of personnel by deconcentrated units on a contractual basis outside the civil service complements the human resource allocation provided by the government. Municipalities are also entitled to employ public servants directly contracted by the municipality. Two key differences between OLEs and municipalities are that members of the vereadores (municipal town councils) change with the political mandate of the mayor, and municipal elected officials are paid from own-source revenue. Unlike municipalities, provincial and district governments have few revenue sources of their own, collecting only a nominal amount of taxes and fees on behalf of the central government. While provincial and district governments are not politically decentralized, recent bodies have been created to promote a degree of downward accountability in these deconcentrated units. Since 2008, provincial governments have been monitored by elected provincial assemblies while district administrations have the duty to consult with the consultative councils, nonelected institutions of participation and community consultation. Consultative council members are chosen by the community and confirmed by the district administrator. They provide links between the district government and the public and are charged with participating in the elaboration of district plans and budgets, monitoring their implementation, and participating in the decisionmaking concerning the distribution of the Fundo de Desenvolvimento Distrital (District Development Fund, FDD). The combined sanctioning effect of these bodies remains weak, and they are often seen as FRELIMO-biased, rubberstamping decisionmakers by district administrators. Cases of patronage (e.g., privileged access to credits from the district fund) have been reported (Forquilha and Orre 2011). Beneficiaries were predominantly party members, the reimbursement rate of the credit was low, the funds were often used for consumption rather than for investment, and the district government had enormous difficulties in performing the functions of a bank (Forquilha 2009). While this modality has been extended to municipal provincial capitals, the average repayment rate is low (below 10 percent) and a recent study found that the selection of bene-

144

Mozambique

ficiaries is biased in favor of the ruling party (Forquilha and Orre 2011; Weimer 2012a: 7). The forty-three autarquias are politically decentralized, with mayordominant governments advised by municipal assemblies, both of which are elected to five-year mandates. In the three municipal elections held since 1998, FRELIMO has won the vast majority of mayoral seats and majorities in almost all municipal assemblies. The Pacote Autarquico and subsequent amendments assigned twenty-nine competencies across nine areas that include health, education, sanitation, and transportation as well as the authority to charge fees and taxes that include a municipal income tax, municipal property tax, economic activity tax, and market and trash fees (Waty 2000). Despite own-source revenue authorities, autarquias depend for at least 50 percent of their annual budgets on transfers from the central government, in the form of a general purpose grant known as the Fundo de Compensação Autárquica (Municipal Compensation Fund, FCA) and the Fundo de Investimentos de Iniciativa Local (Investment Fund for Local Initiatives, FIIL), a conditioned grant for investments and capital expenditures. One of the reasons for this is the low degree of utilization of their own tax base (Weimer 2012b), resulting in a low per capita yield of ownsource revenue, amounting to between $3 and $20 per capita per year, depending on the size of the municipality. However, only approximately 1.5 percent of the total domestic revenue is transferred to the municipalities. Lastly, Decree 63/2003, passed after RENAMO won five municipalities in the second municipal elections of 2003, created an office called the representative of the state at the municipal level that is directly responsible to the provincial and central government. These representatives are meant to increase the FRELIMO-dominated central government’s oversight of local affairs and are thus a significant policy change in the direction of recentralizing power (Chiziane 2011). Conflicts between the municipal mayor and the (nominated) representative of the state of the parallel town administration persist. This is particularly true for the five out of thirtythree municipalities that were ruled by RENAMO mayors during the 2003– 2008 mandate. Through its representative of the state, the central government attempted to limit those autarquias’ scope, territory, tax base, and authority (Chamaite 2010). In terms of powers and resources, the autarquias and deconcentrated units (provinces and districts) are charged with providing basic services, alleviating poverty, and promoting development in response to local needs. Line ministries primarily deliver services such as health and education at provincial and district levels through the sector funds (i.e., health), specialized funds (i.e., road fund), and public utilities. District administrations are also responsible for services such as solid waste management services, which in the case of the municipalities are provided by the municipal gov-

Beatrice Reaud and Bernhard Weimer

145

ernment. Districts submit their annual district plan to the provincial governments that are responsible for both the provincial and district plans and the budgets. These plans and budgets are funded through the annual budget proposed by the Council of Ministers and approved by parliament. This means that the central government has the final control over spending at the district level. In 2010, the donor-supported Programa de Planificação e Financiamento Descentralizado (Program of Decentralized Planning and Finance, PPFD) was turned into a national program, aimed at improving infrastructure and investment at district level. However, districts are subject to double subordination, in that they are accountable to both central government line ministries and provincial governments. Districts generally do not have the resources to adequately provide the services and investment planned in their annual district plan. In terms of resources, provinces receive a greater share of the budget than districts. However, in 2006 the government created a fund called the Orçamento de Investimento e Iniciativas Locais (Budget for Investment and Local Initiatives, OIIL), capitalized by MTn7 million (7 million meticals, US$300,000) per district. Initially available for public infrastructure development and investment, the FDD, managed by the district, was later converted to a credit facility for small-scale investments, turning the district into a kind of local state bank. The reimbursement of the FDD loans has been, on average, only around 10 percent of the disbursed loans. It is seen as part of FRELIMO’s clientelist system aimed at entrenching its rule in rural areas and securing continued electoral support (Orre and Forquilha 2011). A major challenge to transparency and effectiveness of the FDD is its sustainability. Since provinces and districts have received funds from a central budget line, districts have no legally enshrined claim to those allocations, which can be discontinued at any time by executive decision.

Autonomy In a centralist and unitary state politically dominated by one party, the political autonomy accorded to municipalities, provinces, and districts is constrained. Municipalities have held three municipal assembly and mayoral elections since 1998, with FRELIMO winning thirty-three of thirty-three mayoral seats in 1998, twenty-eight of thirty-three seats in 2003, and fortytwo of forty-three seats in 2008. The creation of the representative of the state after the 2003 elections constrained municipalities by imposing an additional layer of government favorable to the dominant party. The first elections for provincial assemblies were held on October 28, 2009, in which FRELIMO won the overwhelming majority of seats (703 seats), followed by RENAMO with 83 seats and two other parties with 26 seats com-

146

Mozambique

bined (Government of Mozambique 2010). The stated purpose of provincial assemblies is to monitor the activities of their respective provincial government, but final approval of provincial plans and budgets remains with the central government through the provincial governor and representatives of central ministries. The control of mayors and municipal affairs by central FRELIMO party organs has increased in recent years. The central party bodies in Maputo have the final say over the selection of candidates for mayor in municipal elections, and the mayors are increasingly accountable to them once elected. For example, mayors in three municipalities (Cuamba, Quelimane, and Pemba) were forced by the party to step down by 2013, prompting by-elections. Municipalities are underresourced, partially due to insufficient efforts in organizing and tapping their own tax base and partially due to the fact that central state funding via transfers, based on population and size, is legally capped at 1.5 percent of the central government budget for all municipalities. Most municipalities do not have the ability to cover recurrent expenditures. On average, own-source tax revenue contributes only 10 percent to the budget while nonfiscal revenue (license and market fees) is about 28 percent. Thus, transfers received from the central government are often used to finance salaries instead of investment and services. However, recent studies on the tax potential of municipalities have indicated that locally generated revenue could increase dramatically if property and property transaction taxes were collected (Weimer 2012b). Municipalities have limited access to revenues, even though about 80 percent of central government revenue is generated from economic activities within municipalities, with the bulk generated from Maputo, Beira, and Nampula (Ilal 2008). Central government transfers do not make up for the loss of this potential municipal income. In 2008, the National Assembly approved a new law on municipal finances, which provided the municipalities with additional sources of own-revenue (i.e., a vehicle tax, a property transaction tax, and a charge for public investment improvement) but caps transfers (FCA and FIIL) by reducing the percentage share of national revenue from 3.0 percent to 1.5 percent. This severely limited the fiscal autonomy of the municipalities, especially when more municipalities were created, meaning each received less. The 2009 state budget demonstrates this point. According to the Government of Mozambique (2008), the FIIL and FCA combined totaled MTn837 million ($27.9 million), which pales in comparison to total provincial and district expenditures at MTn19.5 billion (approx. $641.9 million). A study of twelve municipalities found that they covered only half their operating expenses through own-source revenue, making them unsustainable. Few municipalities collect property and property transaction taxes (Weimer 2012a: 8, 2012b). Provincial and district governments have no fiscal autonomy and little control over their budget, though expenditures in some major sectors have

Beatrice Reaud and Bernhard Weimer

147

been deconcentrated. By 2008, planning and personnel in the education sector were largely transferred to provincial governments. The Ministry of Energy’s Fundo de Energia (Energy Fund), which promotes supply of electricity, solar energy, and fuel in districts, is present in ninety districts (Government of Mozambique 2009). Starting in 2008, the government also deconcentrated funds to construct rural health centers at the district level and provide local salaries and expenses. However, one study showed that the flow of funding to local health units was still fraught with institutional challenges preventing the regular flow of funds and thus jeopardizing the capacity of local units to effectively deliver health services (DFID 2012). Provincial, district, and municipal governments are also subject to a series of constraints on their administrative autonomy. A significant challenge to subnational hiring is the requirement that job applications be screened by the Ministry of Public Service and the Tribunal Administrativo (Administrative Tribunal) to confirm that the application is consistent with legal requirements. Public personnel can be hired under two rubrics: fora do quadro (outside the framework, i.e., employees contracted temporarily by the state) and dentro do quadro (within the framework). Those who are hired within the framework qualify for retirement benefits while those who are outside the quadro do not. As a general (yet unofficial) rule, those within the quadro, and increasingly also those outside, are supposed to be members of FRELIMO. 3 These recruitment procedures, screenings, and lack of financial resources delay local hiring. Although this system constrains provincial and district hiring, those employees are considered to be central government employees, with personnel spending highest at the provincial level. The existence of several central government agencies involved in personnel policy results in cumbersome, lengthy hiring procedures.4

Accountability Downward accountability to voters in autarquias is limited by a combination of FRELIMO party dominance, lack of viable alternatives, and voter apathy. However, the formation of a new party, the Movimento Democrático de Moçambique (Democratic Movement of Mozambique, MDM), and participatory planning and monitoring processes in the majority of municipalities provide some evidence of growing accountability. Political Accountability

Mozambican political parties have proliferated since 1994, with roughly twenty parties and citizens’ groups competing in local elections. However, FRELIMO remains the dominant party, with RENAMO being a distant sec-

148

Mozambique

ond and progressively weaker opposition party. This weakness at the subnational level was most evident in the 2008 elections when RENAMO lost all forty-three mayoral seats. RENAMO did poorly in the subsequent national elections in October 2009, with their presidential candidate Afonso Dhlakama winning only 14 percent of the vote to FRELIMO candidate President Armando Guebuza’s 75 percent. FRELIMO won the majority of the 248 seats in the National Assembly, and 703 out of 812 seats in the provincial assemblies. The MDM won only 3.93 percent, corresponding to eight seats. The lack of an effective opposition reduces political accountability because it cannot provide an effective check against the dominant party. Political party competition escalates during election campaigns, and occasionally has turned violent in the past. The provision of basic services has been politicized, especially in municipalities run by RENAMO. In Nacala, for example, water sources were tagged to the political parties, with the representative of the state providing in the same municipal area FRELIMO water, and the municipal council doing the same on behalf of RENAMO. Another aspect that undermines the democratic legitimacy of assembly members and mayors is that candidates are normally handpicked by the parties’ central and provincial decisionmaking bodies. This makes them less accountable to the electorate than to the central bodies of their parties, which may use them to pursue policies and programs defined at the national rather than local level. Since the last municipal elections (2008), cases of interference by party officials in municipal human resource management have been reported. Some technical staff members who are members of opposition parties have been dismissed, both in FRELIMO-run municipalities and in Beira, which is run by the MDM. Party dominance can also reduce political accountability by not rewarding performance, as shown by the case of Eneas Comiche, the former FRELIMO mayor of Mozambique’s largest city and capital, Maputo. In 2005, Maputo introduced a citizen scorecard to gauge performance and Comiche initiated a presidência aberta (open governance initiative) in which he held periodic, rotating public hearings in each of Maputo’s neighborhoods to listen to what residents thought about municipal services. This eventually developed into Maputo’s first participatory budget. Comiche’s administration cut municipal jobs and increased property tax receipts.5 However, Comiche challenged party interests. A FRELIMO deputy in the municipal assembly indicated that “in fact there were some errors in (dealing with) the [electoral] base,” speaking of Comiche’s administration (Moquivalaka 2008). Of the three candidates put forth in the internal vote for the FRELIMO 2008 mayoral nomination, Comiche, despite improvements in municipal services, came in second with only 32 percent of the vote (Nhamirre 2008). Despite a lack of maneuverable space in the dominant party, an opening in the opposition landscape propelled a high-performing mayor to suc-

Beatrice Reaud and Bernhard Weimer

149

cess as an independent candidate and traction on a national platform. Daviz Simango, the popular RENAMO mayor of Beira, successfully cleaned up trash collection in his first mandate. Despite this, intraparty rivalries reported in the press resulted in the RENAMO party leadership not renominating him for a second term. So, Simango ran as an independent and won the 2008 mayoral contest with over 60 percent of the vote. He ran on a results-based platform with Ele promete e faz (“He promises and delivers”) as his campaign slogan and was so successful that it gained enough momentum to create another (national) party, the MDM, and Simango won almost 10 percent of the vote for president in 2009. In Quelimane City, the capital of Zambezia Province, the MDM candidate, Manuel de Araújo, won the mayoral contest in the 2012 municipal by-elections, despite strong campaign support by FRELIMO heavyweights from central government for their candidate. These examples demonstrate how the mechanisms of political accountability operate in nuanced ways in this decentralized context, to the detriment of FRELIMO’s interest in dominating and controlling local governments. Voter turnout for municipal elections is on the rise, steadily increasing from 15 percent of eligible voters in 1998 to 24 percent in 2003, and to 42 percent in the 2008 polls. The ability of municipal assemblies to hold the mayor accountable is weak. Cases of municipal assemblies refusing to approve the budget or the conta de gerência (municipal accounts) are rare exceptions. In recent years, there were cases in the provincial capitals of Manica Province (Chimoio) and Cabo Delgado Province (Pemba), in which the FRELIMO majority in the assemblies did not approve budgets proposed by the FRELIMO mayor. This led, in both cases, to interventions by the party (even the head of state in the case of Manica) to resolve the conflicts. In general, however, municipal assemblies do not challenge the mayors and their executive appointee. This is often due to lack of capacity and knowledge of assembly members in matters related to budgets and accounts. Mayoral discretion over selection of their executive officers allows them to appoint members of the assembly to executive positions. Fiscal and Administrative Accountability

Subnational governments are held fiscally and administratively accountable in two ways: through central government monitoring procedures and through participatory governance bodies. Provincial, district, and municipal governments are subject to the same reporting requirements in terms of their budgets. The Ministry of Finance is charged with reviewing budget requests from subnational entities and receiving quarterly reports from all three levels of subnational government. The municipalities often have difficulties in the timely and correct production of their budget reports on the

150

Mozambique

conta da gerência because detailed paperwork (i.e., current expenditures, bank transactions, asset register) is required by both the Ministry of Finance and the aforementioned administrative tribunal, which in Mozambique performs the function of auditor general. Some municipalities, like Maputo and Beira, post their quarterly financial reports at city hall, and all of them are obliged by law to publish their bank deposits on a daily basis— although most of them do not abide by this practice. Participatory processes evolved through three distinct periods. In the first mandate (1998–2003), municipalities were occupied with reorganizing their administrative infrastructure from one with ill-equipped offices and personnel with little education to a modern state bureaucracy, with its own statutes, more experienced staff, and offices with computers. The second mandate (2003–2008) saw a consolidation of these processes (i.e., revisions of regulations, reorganizations) and inclusion of community authorities in consultations. By the end of the second mandate, several municipalities had adopted participatory processes for commenting on strategic plans and municipal budgets. At least thirty-three of the forty-three autarquias now have some sort of consultation or hearing process (Nguenha 2009: 15). With the exception of Maputo, CSOs outside of religious organizations are rather weak. Lack of interest, technical knowledge, or a fear of reprisals has led citizens and CSOs to be passive in engaging local leaders and denouncing irregularities. Only recently, encouraged by President Guebuza’s approach of the “Open and Inclusive Presidency” have residents and CSOs denounced abuses of power, which, in some cases, has led to the dismissal and disciplinary measures of such leaders. Even in the municipalities, citizens are careful to avoid confronting, even engaging, the mayors or deputies in a formal way. The lack of formal complaint procedures and codes of conduct for municipal leaders results in a clear preference for addressing and possibly resolving contentious issues informally.6 At the provincial level, Nampula Province, has a national reputation for national and local NGOs with a focus on local governance (D. Jackson 2007). Nampula CSOs have also been in the lead in working together with the provincial government to institute Provincial Poverty Observatories (Macuane et al. 2012). Some proactive involvement of provincial CSOs in local government and decentralization monitoring has also been reported in other provinces such as Niassa (Åkesson and Nilsson 2006). During the past ten years the media, both print and electronic, at national and local levels have also taken an increasing interest in municipal and local governance affairs. Features and critical reports on municipal and district governance have become quite common, notably in television programs that examine in-depth achievements and challenges of decentralization, notably during periods of elections. Local media (especially, community radio) are an exception to this since many of them are controlled, and

Beatrice Reaud and Bernhard Weimer

151

some even financed, by local leaders to whom the staff feel loyal. In highly contested municipalities, such as Nacala City, the local community radio station was simply closed down when the RENAMO mayor took over the reins of the city from his FRELIMO predecessor after the 2003 elections.

Capacity Capacity is a challenge for all actors involved in Mozambique’s decentralization. CSOs face various challenges. Among them are low financial management capacity, insufficient internal governance, and high donor dependence, especially outside of Maputo. Key findings of a recent study on Mozambican civil society and citizen participation indicate that it is weak for two reasons: first, organizations lack financial and organizational capacity; and second, citizens in practice have little economic and political influence (Francisco et al. 2008). According to Robert Mattes and Carlos Shenga (2007), civil society participation in Mozambique is half that of Tanzania, Senegal, and Nigeria and only a quarter of that in neighboring Malawi.7 Poverty, lack of qualified staff, and lack of infrastructure are challenges that all subprovincial governments face. Provincial variation in economic development favors the south, posing particular challenges to development in the population-rich northern and central provinces. Northern Mozambique has the lowest Human Development Index average (0.33), with the lowest life expectancy rate at birth (forty-three years) and literacy rate (34.6 percent) of the three regions (UNDP 2005). Maputo Province has the highest averages, with a literacy rate almost three times that of the northern Mozambican provinces combined at 94.8 percent (UNDP 2005). Infrastructure access remains a critical capacity deficit in Mozambican districts. Northern provinces tend to have greater needs relative to their resources,8 with about one-fourth of district centers still lacking electricity. Lack of personnel capacity in public administration plagues all subnational levels of government. The central government employs over 170,000 employees, of whom only 40 percent have more than a basic education (AIM 2010). Existing human resource policies and legislation limit the ability of municipalities to hire, fire, and compensate staff, leaving them with low-skilled workers and few qualified senior staff, although the number of trained and qualified staff is increasing. Considerable internal conflict potential exists between the executive (mayor and town councillors) on the one hand and the technical staff in the municipal departments, notably the departmental heads. The former, in a position of power but often with little technical capacity, have a limited mandate of five years while the latter are long-term employees of the municipal administration with considerably more expertise. This often

152

Mozambique

results in situations in which the latter will commit an irregularity ordered by his “boss,” fearing for his job otherwise.9 Despite significant financial constraints, including the suboptimal utilization of their own-source bases and administrative capacity, overall performance in service delivery has increased over the past ten years.10 Progress has been made in the following areas: • There has been gradual enhancement of revenue collection, mostly fees, and improved financial management; • Participatory planning methods have been introduced in a few municipalities; • Solid waste management, sanitation, drainage, and road construction and maintenance have seen some progress in selected urban areas of many municipalities; • The administrative apparatus has been reorganized for improved performance; and • In some municipalities urban upgrading of informal settlements and allocation of serviced plots in extension areas have been accomplished, together with improved infrastructure, especially municipal markets (World Bank and Associação Nacional dos Municipios de Moçambique 2009).

Political Economy and Political Incentives Given Mozambique’s vast territory, low population density, and rudimentary technical infrastructure, public service provision is largely limited to the capital, the forty-three municipalities, and a few district centers. Political power and economic resources continue to be highly centralized. In 2007, only about 3 percent of total public sector expenditures were spent at the district level and around 1 percent at municipalities. While the state budget has seen an increase in domestic tax revenue and a decrease in foreign aid, aid inflows have made Mozambique one of the world’s most aid-dependent countries (de Renzio and Hanlon 2007). Less than 10 percent of the Mozambican active population pays taxes or holds a tax identification number (AllAfrica 2010b). Generation of own-source revenue has increased in recent years from around 14 percent to 20 percent of gross domestic product (GDP), and is expected to further increase, given the massive foreign direct investment (FDI) in extractive industries (coal, gas, mineral resources). It is estimated that foreign direct investment in extractive industries in the next ten years will amount to $90 billion as compared to a GDP of $12.3 billion in 2010 (EIU 2012).

Beatrice Reaud and Bernhard Weimer

153

Yet regarding municipal revenue, overall domestic tax revenue remains below the existing potential (Weimer 2012b). This is in large part due to few profitable local private businesses and investments, especially small- to medium-sized enterprises; weak capacity, and little effort to tax businesses and parastatals; rent-seeking behavior by the (local) political elite; and generous tax holidays for large-scale projects and foreign investment in strategic economic sectors.11 Some members of FRELIMO have considerable stakes in mining and energy ventures, as well as in the tourism, real estate, and service sector of the economy, and they often benefit from generous subsidies and tax exemptions. President Guebuza’s mandates have coincided with recentralizing tendencies, despite political rhetoric that stresses decentralization, thus echoing the former socialist model when the party was fused with the state. As one observer shares, “You must see the FRELIMO state as an employer, the only employer, who guarantees the daily bread for us and our families.”12 The FRELIMO party and the central government are the main actors in decentralization. Without a national decentralization policy in place, a gradual policy shift from municipalization to deconcentration with a focus on districts has been observed since Guebuza won the 2005 and 2009 elections. This approach favors the consolidating central power of party and government at the local level via the OLEs.13 Municipalities are much less frequently visited than districts, with the exception of the provincial capitals and Maputo. FRELIMO: Proponent and Opponent in One?

Mozambican decentralization is primarily a FRELIMO homegrown process, as can be gauged from Law 3/1994 passed by the then one-party National Assembly. At the time, it reflected a genuine recognition by the party leadership that the vast, thinly populated Mozambican territory could not be centrally governed, and that decentralization provided an opportunity for postconflict stability by accommodating the political opposition.14 Yet various factions within FRELIMO hold different assumptions and objectives concerning decentralization. Proponents of a FRELIMO-dominant state are poised to maintain central and political power and control of people (voters), resources (rents, taxes), and territory, especially given the trajectory toward an extractive economy in which Mozambique finds itself. Only a minority of the party entertains a vision of more decentralized, pluralistic, and democratic political institutions at subnational levels. These contradictions may be seen as the primary reason for the government’s gradualist approach to decentralization. While it gradually introduces and tests municipal capacity to enhance service delivery, it simultaneously

154

Mozambique

maintains and strengthens the strong central state via its deconcentrated local branches (OLEs) at provincial, district, and subdistrict levels, in which the municipalities are territorially embedded. Seen from this angle, decentralization appears politically useful and necessary only if it strengthens the political party that claims control over all levels of government, the country, and its resources. Within this overall strategic framework, tactical approaches to shaping intergovernmental relations may vary from benign neglect, via cooptation and power sharing (in the case of allies), to usurpation in the case of rivals (Boone 2003). This depends on factors such as the social capital, ethnicity, and resource endowments of local elites with sufficient leverage and capability to challenge central party decisions. National elites have had a tendency to marginalize local elites, the co-optation of individual regional strong-men notwithstanding (Weimer, Macuane, and Buur 2012). Few opportunities exist for members of the local elite and mayors to become part of the central power structure. Other political parties represented at the national level have had little influence over the decentralization process. RENAMO systematically opposed decentralization by almost always voting against relevant legislation and boycotting the first municipal elections in 1998. Only in 2003, when its candidates for mayor won five out of thirty-three municipalities, did decentralization and local governance start to matter politically, at least in the official discourse.15 The Dominance of FRELIMO and the Executive

Concerning the executive, there is no clear functional and operational responsibility for decentralization. The Ministry for State Administration has national directorates for municipal development and for local administration, but the driving force in planning has been the Ministry of Planning and Development. It is the host to the donor-supported National Decentralized Planning and Financing Program.16 The Ministry of Finance is also a key actor, exercising financial oversight over municipalities. It receives and approves the municipal budget (after approval by the local executive and legislative), and transfers the recurrent and investment funds to the muncipalities. The ministry’s inspector geral das finanças (inspector general) examines the municipalities’ accounts, asset registers, and internal control functions, but reports are not made public. Municipalities are charged (by Decree 33/2006) with providing several services that further transfer competencies from sectoral agencies of the OLEs accompanied by adequate resources and sufficient municipal capacity. At present, hardly any of the municipalities meet these conditions and the decree lacks operationalization, leaving sectoral decentralization at its infant stage and sometimes neglected or resisted by the line ministries.17

Beatrice Reaud and Bernhard Weimer

155

The Ministries of Health and Education best exemplify the principle of gradualismo in the transfer of sectoral competencies and resources (human and financial) to the municipalities. Similarly, the first experiences of decentralizing financing road construction and maintenance through the Road Fund have been fraught with difficulties such as insufficient financial management and procurement capacity by the local governments. Concerning local water supplies, the picture is mixed. Major towns have benefited enormously from the deconcentrated water management arrangements with the national Water Investment Fund, which provides financing for investment and maintenance of hardware and reticulation systems, while the management of operations is in the hand of local companies. The public water supply systems in smaller towns and district centers, however, face major challenges concerning management capacity, often lacking cost efficiency and thus sustainability (Uandela 2012). Lastly, the Grupo Interministerial de Desenvolvimento Autárquico (Interministerial Group of Municipal Development), a coordinating body encompassing all key ministries, was set up a few years ago. So far it has produced little impact in effectively shifting functions of line ministries to municipalities or in coordinating and monitoring their responsibilities.18

Conclusion Mozambican decentralization has created a subnational governance structure that divides local governments into deconcentrated state organs (OLEs) and elected municipalities (autarquias), with the latter having decidedly more authority, autonomy, accountability, and capacity. Capacity and resource constraints exist at all levels with improved service delivery, accountability, and governance remaining formidable challenges. FRELIMO has extended its clientelist practices to the local level. However, the introduction of provincial assemblies and consultative councils at the district level brought some degree of accountability. The same is true for some municipal assemblies. Own-source revenue authorities at the municipal level are beginning to be exploited, with market revenues increasing and a culture of fee and taxation payment gaining ground. Municipalities are in need of increased investment into financial management systems and tools (i.e., registers, property value assessments) to enable them to tap into and effectively administer a hitherto insignificant part of their tax base; namely, property and property transaction taxes. The planning and budgeting cycle seem adequate and increasingly produce the required outputs (budget plans, activity plans, conta de gerência), though often with difficulties in quality and timeliness. Scarce resources (of smaller municipalities) and considerable levels of recurrent expenditures (salaries and goods and services)

156

Mozambique

notwithstanding, public infrastructure and services are increasingly produced, making the municipalities quite effective in implementing their projects and providing services. In comparison, the deconcentrated units face major challenges, though they are generally much better endowed with central government support. Lack of harmonization of planning, budgeting, and execution at district level, where multiple actors intervene simultaneously with little coordination, poses a continued challenge. Frequent changes in the delivery modality of the local investment funds, with widely perceived opaque and corrupt practices surrounding the access to and distribution of the funds also persists. Despite considerable progress with participation in district planning, the deconcentrated units lack the capacity, information, and resources to do better, especially with regard to monitoring the execution of performance plans jointly defined by central government and donors. Given their considerable differences in the institutional frameworks, the rights of their citizens, their resource endowments on the one hand, and the lack of comparative studies on the other, it is difficult to compare OLEs and municipalities with regard to key criteria for successful decentralization such as efficiency in resource use, responsiveness to taxpayers and the electorate, and the production and distribution of basic services. However, most of the evidence suggests that devolution to municipalities has fared somewhat better than deconcentration. If factors such as the higher degree of (formal) democratic legitimacy by regular local government elections or the positive correlation between taxation and state building suggested by Deborah Bräutigam and colleauges (2008) are taken into consideration, this assessment may be even more positive. The municipal template for local government provides a more fertile institutional base for the emergence of citizens, taxpayers, and voters with an awareness of their rights and duties. A study of five municipalities confirms this view: on average 72 percent of respondents would even pay higher taxes to get better services from their local governments (de Brito, Pereira, and Forquilha 2007). Focus group interviews that we conducted in four municipalities—Beira, Maputo, Dondo, and Matola—also confirm this finding.19 In conclusion, the deconcentrated OLEs have not functioned effectively in service provision and governance. From a political economy perspective, however, one of the motivations for deconcentration may have been to consolidate FRELIMO’s state power in Mozambique and minimize the constituencies of the opposition. This suggests, in keeping with a theme of this volume, that the intent behind decentralization in Mozambique may have been more about regime maintenance and strengthening the central state than about strengthening local-level institutions. While there is evidence that this is increasingly the case in the municipalities as well, autar-

Beatrice Reaud and Bernhard Weimer

157

quias have shown resilience not only in surviving but also in delivering public services and promoting democratic local governance, despite a playing field tilted to their disadvantage.

Notes 1. Mozambique has a semipresidential system in which the president is elected through direct universal suffrage to a five-year mandate and appoints the prime minister (who is accountable to the president) and the Council of Ministers. The National Assembly is elected by proportional representation under closed party lists. 2. For example, the line ministries of education, youth and sports and science and technology are grouped to the District Service for Education, Youth and Technology. 3. This is officially denied by the Minister of Public Service, Victoria Diogo (AllAfrica 2010a). 4. The Estatuto Geral de Funcionarios e Agentes do Estado (General Statutes for Public Servants and Agents of the State), Law 44/2009, foresees the periodic performance evaluation of the civil service and municipal staff for promotions, transition into other careers, and salary scales. At provincial and district levels, the permanent secretary and the district secretary, respectively, are charged with this task. At the district level, Decree 5/2006 conferred on district administrators the ability to recruit their own personnel and set salaries. 5. Eneas Comiche, former mayor of Maputo, interviewed by Beatrice Reaud, Maputo, December 12, 2008. 6. An exception is the activities of the Gabinete Central de Combate à Corrupção (Anti-Corruption Office) in Maputo, Beira, and Nampula. 7. Other important actors are research institutes, to include the renowned independent Instituto de Estudos Sociais e Económicos (Institute for Social and Economic Research) and Centro de Integridade Publica (Center for Public Integrity). 8. A case in point is Nampula Province. It is third in tax revenue, but has one of the highest incidences of poverty, illiteracy, and endemic diseases. 9. Many of the irregularities reported by the inspector geral das finanças (inspector general of finances) and Tribunal Adminstrativo have their origin in this situation. Cases of embezzlement and corruption have also been reported in the media in which both the mayor and head of Department of Administration and Finances drained the municipal treasury of public funds for their own private purposes (see www.canalmoz .co.mz/hoje/20150-pio-matos-renuncia-ao-cargo-e-nega-acusacoes-de-gestao -danosa.html). A study commissioned by USAID’s Projecto de Governação Municipal Democrática showed that there is a considerable risk of corruption in the municipalities, especially in public tenders for infrastructure, allocation of plots of land, and granting of construction licenses (Nuvunga, Mosse, and Varela 2007). 10. Technical and management capacity, notably at the intermediate level of municipal technical staff, has increased considerably due to various training programs run by the Instituto de Formação em Administração Pública e Autárquica (Training Institute for Public and Municipal Administration) and on-the-the job training, often supported by donor agencies. Most of the municipal administrations are equipped with computers and some technical departments boast of considerable equipment for civil engineering and public works. 11. Data provided by the Autoridade Tributária de Moçambique (National Tax Authority).

158

Mozambique

12. Senior lecturer at one of Mozambique’s public universities who is also a member of a CSO, personal communication with the authors, February 20, 2010. 13. In the president’s rhetoric and working visits to districts, the key messages are deconcentration, local participatory planning and participation, management of the OIIL, and local economic development for combating poverty. 14. One of the most dynamic driving forces for decentralization was the former minister for state administration, Aguiar Mazula. 15. In 2008, the party and Dhlakama ousted their confirmed incumbent candidate for Beira, Daviz Simango, from the party’s election campaign. Simango founded the MDM, which won Beira an overwhelming majority in the national elections. The party is still too young to be gauged on its position concerning decentralization. 16. The donors were the United Nations Capital Development Fund (UNCDF), the World Bank, the Netherlands, Ireland, Norway, Germany, and Switzerland. 17. So far, only Maputo City was selected to assume responsibility for primary education. 18. Other important actors at the national level include: the Tribunal Administrativo, the Autoridade Tributária de Moçambique (National Tax Authority), and the Associação Nacional de Municípios Moçambicanos (National Association of Mozambican Municipalities). 19. Our findings are the result of assessments gathered from over 100 participants in eleven focus groups in the four cases mentioned above. See Reaud (2012).

8 Nigeria: Issues of Capacity and Accountability in Decentralization Dele Olowu and James S. Wunsch

With an estimated population of 174 million, Nigeria is Africa’s most populous country and the largest federal government on a continent whose population totals over one billion. The present structure, functions, finance, and accountability arrangements governing decentralization were created by military rulers in 1976. Succeeding governments, both military and civilian, have subscribed to and supported the logic of a three-tier federation. The performance of these institutions has differed over time and space, influenced by varying priorities of rulers at the national and state government levels over this period. The Nigerian experience poses both negative and positive lessons. Negatively, the comprehensive or big bang approach to devolutionary decentralization was ill advised without an effective effort at creating strong lateral and vertical accountability structures. This was further aggravated by the near-complete dependence of the local government system on federal revenue-sharing arrangements—financed almost exclusively from oil revenues—for all three levels of government, without requiring performance benchmarks in terms of development or revenue matching from local governments. On the other hand, the local government system has significantly decentralized the Nigerian system of government and public finance while standardizing the system of local government in this huge country. Local governments spend 21 percent of total public expenditures or 5 percent of GDP, high by African and developing country standards (Olowu and Wunsch 2004). This decentralized structure has helped to increase possible entry points into the political system. Local government authorities (LGAs) have successfully served as recruiting grounds for state and federal level politicians (two of the post-1979 presidents and several governors served in local 159

160

Nigeria

governments). It has also functioned as an important training ground for citizens and organized civil society groups, although not so much for private sector actors or the media. Second, the system has enhanced the capacity of local government compared with what it was before the 1976 reforms. Finally, important intergovernmental organs have been created during the military period for mediating relations among the three levels of government. What remains to be done is to create effective organs to enable state governments, civil society actors, and donors to assist LGAs in improving their capacity to provide citizens with quality public services, and to develop effective accountability structures to incentivize better performance at all subnational government levels.

Background In 1976, the federal military government, as a part of its preparations for transition from military rule to democracy, embarked on a reform of local government. The government gave two reasons for this: (1) to return to democracy starting from the grassroots; and (2) to ensure that the country’s oil revenues were utilized to build the infrastructure and services essential for economic modernization (Nigeria 1976). This was pursued through three strategies. First, a homogeneous structure of local government was created across Nigeria, replacing the premilitary systems of region-based local governments with a national system of local government. Second, the reform articulated clearly the responsibilities of this third of tier government, distinguishing it from the federal and state governments. In the meantime, the state governments had also been reconfigured from three to six units and later to twelve (just before the civil war) and to nineteen units after the war was concluded. Third, huge financial resources from the national federation (which came mainly from oil) were made available to these new local governments. Their expenditure profile within the public sector moved from only 3 percent before the reform to 10 percent and, subsequently, to the present 23 percent. They were also allowed to recruit and manage their own staff. Several staff and devolved services were sent to the local governments over a relatively short period of time. (See Table 8.1.) In total, 299 local governments were created. This has been increased (301 in 1976; 449 in 1987; 500, then 589 in 1991; 593 in 1995) to the present total of 774 since 1996. More states were also created bringing their present total to thirty-six. However, only a few states and local governments have been able to translate these large infusions of federal cash transfers into governmental infrastructure; thus, unfortunately, infrastructure and service delivery have barely improved. Rather, the reforms have fueled an increasing demand for additional local government councils—as a way of

Dele Olowu and James S. Wunsch

161

Table 8.1 Summary of Local Government Reforms in Nigeria, 1976–2005 Date

Reforms and Changes

1976

Reform of local government system; uniform local government system adopted for the whole country. Abolition of the Ministry of Local Government. All local government matters moved to new Department of Local Government; all other functions reassigned to other state ministries. Introduction of directorate system into the local government service. Creation of the office of local government auditor. Federal government assumed legislative competence over local governments. Empowerment of local government chairmen to appoint supervisory councillors instead of being elected. Empowerment of local government chairmen to appoint secretaries. Announcement of local government autonomy with respect to operational and financial matters. Local government budgets were to be approved by local government councils, no longer by state governments. Transfer of primary education and primary health care to local governments. Adoption of a presidential or strong-mayor system universally in all local governments. Local government revenue shares in national intergovernmental revenues rose to 20 percent in 1992. Constitution of 1999 enters into effect. Establishment, structure, composition, finance, and functions of local government by state government law. Value-added tax contributions to local governments. Inclusion of local governments in work of anticorruption body.

1988

1989

1990

1991

1999

2004 2005

Sources: Constitution of the Federal Republic of Nigeria 1999; Olowu and Wunsch 2004.

“sharing the national cake.” Some state governments created new local governments as a result. The state of Lagos, for instance, created an additional thirty-seven local governments in 2004, which generated controversy that remains unresolved to this date. By 2005, 30 percent of total public service employees worked in Nigeria’s local government. This is a high figure for African countries, although low when compared to Organisation for Economic Co-operation and Development, Latin American, and Asian countries.

Fiscal Decentralization Government introduced a formula for sharing federal revenue among the three governmental levels in 1981. The statutory allocation to local government was increased from 3.0 percent in 1976 to 10.0 percent in 1981, 15.0

162

Nigeria

percent in 1990, 20.0 percent in 1992, and 20.6 percent in 2002. During the same period, state government allocations from the federation account dropped from 30.0 percent to 26.0 percent while that of the federal government dropped from 55.0 percent in 1981 to 48.5 percent in June 2002. In 1988, there were further steps to strengthen the local councils when direct payment of federal allocations were made to local governments instead of passing them through state governments. As part of these financial reforms, local government was expected to receive 10 percent of internal revenue of state governments. Most state governments have failed to comply with this. When the value-added tax was introduced in 1994, 25 percent was allocated to local government. The local government share of VAT represented 10.7 percent of its total revenue for that year and 18.3 percent in 2005. This large infusion of financial resources into local governments has produced both positive and negative consequences. On the positive side, it first has attracted a crop of strengthened political and administrative leadership into local government (Adamolekun 1984). Second, it has enabled local governments to undertake a number of services, including construction and maintenance of roads, streets, street lighting, drains, parks, gardens, open spaces; naming of roads and streets; and numbering of houses. Others include provision and maintenance of public conveniences and sewage and refuse disposal; maintenance and regulation of slaughterhouses, markets, and motor parks; and establishment and maintenance of cemeteries and burial grounds. On the negative side, accountability has been weakened because local governments rely heavily on federal transfers with minimal input from the local citizenry. Second, local government officials often are not paid for months (Khemani 2005). A third problem is that each local government is forced to have exactly the same structures, personnel, and pay systems as others, even when these are not appropriate to their respective local contexts. Finally, the transfers from the national treasury seem to discourage and displace local revenue mobilization efforts.

Authority LGA Internal Functional Authority

LGAs have only modest authority over personnel. Their personnel systems are divided into two levels. All managerial and professional personnel are employed by the Local Government Service Commission, an independent organ run and managed at the state level. This structure engages, deploys, disciplines, and discharges these personnel. Thus, LGAs have little authority over them, but it does protect staff from the fear of political victimiza-

Dele Olowu and James S. Wunsch

163

tion, a real issue in Nigerian governance. Rotation among LGA assignments typically occurs every three to four years, so there is some instability for LGAs regarding this cadre. Still, the rotation system means that even poorer and more rural areas have access to more skilled and experienced personnel. Employees below the managerial grade, primarily less-skilled and subordinate personnel, are engaged by LGAs. While LGAs do have full authority over these personnel, before 1999 there was a tendency to overstaff and the high cost of eliminating such positions and discharging redundant personnel limit effective LGA power. Thus, overall, LGA power over personnel is minimal. Revenue authority for LGAs is limited to a few, low-yielding and highcost sources stipulated by the 1999 Constitution, and by state preemption of property taxes in some states. LGAs also lack the ability to influence their share of federal allotments since these are distributed according to factors such as population and size. LGA revenue efforts and policy performance have no bearing on their shares. Budgeting and expenditure authority varies among states. Some states, such as Kaduna, Lagos, and Jigawa, treat LGA budgeting and expenditures as strictly LGA decisions. Others require state approval for LGA budgets and expenditures above minimal levels. LGAs are all required to perform audits of their expenditures. States are normally responsible for supervising audits, but they vary in their diligence. Generally, LGA audit performance is regarded as weak. LGA Sector Authority

According to the 1999 Constitution, LGA policy and administrative authority is limited and could be better regarded as delegated than devolved. Most responsibilities are under states or concurrent state and federal authority. Education, for example, has been removed from LGA control and, while frequently deconcentrated in some measure to the local level, is governed by independent boards at state, LGA, and village levels. Economic development is a shared state-local function, but subject to state delegation. International agencies have confirmed that state and federal agencies dominate economic development planning, and that LGAs have little input in planning or implementation of projects (IFPRI 2009). There are variations across states and localities. Security is a federal responsibility. However, some control over this function has effectively devolved to the state and LGA levels through the emergence of community “militias” such as the notorious “Bakassi Boys” of Onitsha (Harnishfeger 2003). Such militias have also developed in the states of Rivers and Bauchi. These are semiprivate bodies, sometimes authorized

164

Nigeria

by state legislation (Barkan, Gboyega, and Stevens 2001) because federal policing has lacked competence and integrity in protecting lives and property. A downside to this is that these militias are managed in opaque ways and the influence of state and LGA personnel over them is unclear. They have worked with politicians and played a political role in some areas, intensifying tension and fear among some ethnic communities (Harnishfeger 2003). Where active, the militias have largely preempted the authority of the Nigerian police. LGA power is the clearest over markets and local roads and, while shared with states, is important in primary health care (PHC). However, LGAs are constrained by low resources remaining from federally mandated expenditures and the federal salary adjustment of 2000 that hit state and LGA budgets. Another limitation on the effectiveness of LGAs is weak accountability regarding their use of resources and weak incentives to mobilize additional locally generated resources. Overall with the lack of effective structures of accountability (laterally and vertically) for LGAs as well as the perverse incentives created by Nigeria’s fiscal federalism (which as a common pool resource creates no incentives for enhanced local revenue effort or more effective LGA services or local economic development), it is not clear that enhancing LGA power would lead to improved performance. It is significant that some local government chairmen have nonetheless distinguished themselves such as in the states of Lagos, Jigawa, and Kaduna (Barkan, Gboyega, and Stevens 2001). Even in the state of Rivers, where systematic local and state government corruption has been documented, one local government (Khana) distinguished itself with high-quality performance (Human Rights Watch 2007a).

Autonomy Since Nigeria is a federal system with real powers exercised by both state and federal levels, analyzing the autonomy of its 774 LGAs requires attention to all three levels of government. It requires examination of the 1999 Constitution, fiscal federalism as Nigeria practices it, and state statutes and administrative practices regarding LGAs. Needless to say, the fact that there are thirty-six states in Nigeria, with varying approaches to their LGAs, means that generalizations are limited. Additionally, only a few states have been studied, so there are large gaps in knowledge. Nonetheless, one can make several important observations. In some respects, decentralization in Nigeria can be regarded as robust and vital, and as an effective foundation for LGA autonomy. The three levels of governance have constitutional status, the states have or share with the Federal Government of Nigeria substantial authority and responsibility,

Dele Olowu and James S. Wunsch

165

and LGAs are entrenched in the constitution with some (though not a great deal) of autonomy. More to the point, states and LGAs have constitutional rights to substantial shares of the federation account, amounts that have increased dramatically since the 1970s. The Federal Constitutional Court has also ruled in several cases in favor of subnational governments against the Federal Government of Nigeria, protecting their autonomy and enhancing their share of federal monies (Suberu 2009). Finally, with the exception of the robust centralizing trend under several periods of military government, Nigeria has a long and strong federal tradition and history, dating to the early 1900s. Even General Ibrahim Babangida (though primarily for political reasons) substantially enhanced the autonomy of local government during his rule (Barkan, Gboyega, and Stevens 2001). Otherwise, the military rulers’ suspension of state and LGA autonomy and rule through military officers (governors) can be regarded as a period of exceptional subnational governance ineptitude, incompetence, and corruption. It was in part in reaction to the military’s abuses that the 1998–1999 transition revitalized federalism. Therefore, federalism and local government are well accepted, entrenched, and institutionalized in Nigeria (Suberu 2009; Barkan, Gboyega, and Stevens 2001; Gboyega 2003). The current system provides for at-large election of the LGA executive, (the chairman), election by wards of the members of the LGA legislative body (the council), and for appointment of a cabinet (department heads) by the chair. In these respects, de jure autonomy is secure even though the major political parties play the key role in determining who runs for elections at local and state levels. However, there are several operational aspects of Nigerian governance that substantially reduce the effective autonomy of local governments, though not so much that of states. First, Nigeria’s determination at the federal level of all government employees’ salaries, allowances, and conditions of separation has dramatically eroded the effective autonomy of LGAs. When the salary adjustment of 2000 was instituted, its 2.5-fold increase in civil service salaries, along with the existing generous conditions for separation, largely paralyzed LGAs’ programmatic and capital budgets. They also made LGA personnel reductions and restructuring virtually impossible. The fivefold increases in salaries for political personnel worsened a terrible situation. Second, Nigerian fiscal federalism impedes development of an LGAfocused political life and a public base for genuine autonomous action by LGAs. For instance, the preemption of resources for primary school costs reduced resources available to LGAs and some, in effect, had zero allocations remaining (Barkan, Gboyega, and Stevens 2001; Gboyega 2003). Third, LGA politics revolves around the pursuit of the spoils of political office (and the opportunities they bring), public employment, and determining which ethnic community will get each (Aiyede 2009).

166

Nigeria

Thus, there is usually little to no money for programmatic budgets, and capital projects have dried up since the 2000 salary adjustments. New facilities frequently cannot be staffed or operated even if completed. This means most LGA decisions become decisions over who will get what are effectively private goods such as local jobs or contracts for LGA purchases and infrastructure maintenance (Ostrom 1990). Consequently, there is little incentive for local residents or civil society to mobilize to assert and support local autonomy in order to make choices regarding public goods such as which ones to provide, how to provide them, how to pay for them, and how to enhance their quality. The politics of individual or ethnic community spoils dominate LGA governance, and as long as the spoils come from federation accounts (not local taxes), their allocation affects only a few rank-and-file Nigerians. Since the local political class already controls this, there are few inducements for it or the broader public to assert local autonomy regarding policy or programs. Finally, LGAs in some states lack effective autonomy to make choices in most functional and sector areas: over their revenues and personnel, frequently over budgeting and expenditures, and over most development projects. Overall, the Nigerian constitutional structure, its history, and its very size and diversity provide a positive framework for LGA (and state) autonomy. However, the practices of Nigerian fiscal federalism undermine the incentives needed for its politicians and publics to use this structure to develop and sustain real political and programmatic autonomy (Kalu 2008). Also, state dominance in functional or sectoral areas further limits LGA autonomy. However, with 36 states and 774 LGAs, the level of autonomy held by individual LGAs varies a great deal, varies among various sectors, and varies with informal and cooperative mechanisms and procedures that may have developed over the past decade.

Accountability As noted in Chapter 1 of this volume, accountability can be best understood as having three dimensions (Diamond 2008): • downwardly, to the electorate and civil society; • upwardly, to superior levels of government; and • laterally, to various checking and internally regulating mechanisms at the specific level of government. Nigerian LGAs generally rank poorly on electoral, civil society, and general public accountability. Elections have been marred by low turnouts (20 percent in the 2007 national election) and by what are generally

Dele Olowu and James S. Wunsch

167

accepted to be patterns of fraud, violence, and intimidation (Rawlence and Albin-Lackey 2007; Joseph and Kew 2008). Elections are primarily personal contests over who will have access to LGA resources. Parties have no real programmatic differences, but are coalitions formed around personalities for the purpose of pursuing personal power and the spoils of office. Once in office, party membership reportedly plays little role (Barkan, Gboyega, and Stevens 2001). In spite of Nigeria’s rich civil society at all levels of governance, particularly at the community level, what research there is on civil society at local levels of government has indicated an absence of public knowledge regarding LGA budgets, weak civil society involvement, and rare media activity in most of Nigeria’s LGAs (Khemani 2005; Dibie 2003; Alemiki and Chukwuma 2004). Both local and national media cover some local issues, but national issues receive greater attention. Even in the Delta state, where local civil society is arguably the most politically active in Nigeria, youth groups target the petroleum companies and the federal government, and generally not LGAs (Anugwom and Anugwom 2009; Ikelegbe 2001a, 2001b, 2005). E. E. O. Alemika and I. C. Chukwuma (2004) reported some attempts by civil society organizations to work with LGAs, but that these were brushed off by LGA politicians. Ethnically based civil society has been active, however, in the campaign for new LGAs, the location of their headquarters in their communities, and other distributional benefits (Ukiwo 2006; Alapiki 2005). In some instances, community-based organizations and CSOs have become more active in education and health projects at the local level. Overall, accountability of LGAs to the general public can only be regarded as weak (Crook 2003, Human Rights Watch 2007a), but occasionally effective where the community and LGA leadership promotes it (Olowu et al. 2000; Olowu et al. 2010). Lateral accountability is weakened by the dominant role played by LGA chairmen, by generally weak and ineffective LGA councils and audit officers, the latter which are understaffed and lack enforcement powers, and by poor systems of account management (Barkan, Gboyega, and Stevens 2001; Gboyega 2003; Agundu and Waleru 2008). Even though LGA councils are authorized to approve local council budgets, councillors have been preoccupied with personal and constituency benefits rather than with programmatic assessment or evaluation of budgets or programs (Human Rights Watch 2007a). In a sense, this is only a little different at the national or state government levels (Barkan, Gboyega, and Stevens 2001). Furthermore, the lateral accountability at times provided by strong cadres of professional civil servants and professional organization is also absent (Dibie 2003). Tangible evidence of this can be seen in research on LGA PHC, which consistently has shown poor performance, fund leakage, and arrears in salary payments and the like (Khemani 2005; Olowu and Wunsch

168

Nigeria

1995; Olowu et.al. 2000). The authority of LGA councils, PHC staff concerns, and public discontent has not translated into better performance. An exception to this pattern can be seen in the deconcentrated system of primary education. This is autonomous of LGAs and research on several dimensions of primary education has indicated that, when decisionmaking and management is actually deconcentrated to professional personnel at the school level, performance is significantly enhanced (Ikoya 2007, 2008). Lateral accountability among the educational civil service professional seems to be a likely explanation for this stronger performance. Also important is the role of parent-teacher associations (Geshberg and Winkler 2004). Under the 1999 Constitution, state governments play the dominant role in establishing the statutory and regulatory framework within which Nigeria’s LGAs operate. Thus, there is substantial variation among Nigeria’s thirty-six states in the roles that state governments play. In some cases, state governments closely supervise and must approve LGA budgets and expenditures. In others, state governments take a largely hands-off approach to these functions (Barkan, Gboyega, and Stevens 2001). However, the research available on the role of state governments does suggest that one critical function—audit—is weak to ineffective across most state governments. State auditors are in general poorly staffed, funded, and equipped, and are running months to years behind in tracking state expenditures. They also lack any enforcement powers. This weakness means that a crucial element of upward accountability is generally ineffective (Barkan, Gboyega, and Stevens 2001; Agundu and Waleru 2008). As noted under the Capacity section, state-level intergovernmental relations mechanisms range from weak to completely ineffective. Underlying the lack of accountability of Nigeria’s LGAs, as in so many aspects of its decentralization, are the dynamics of the federal account allocations. First, with little revenue raised locally, a local tax burden to stimulate public motivation to involve itself in its raising and spending is absent. Generally, some 90 percent or more of local revenues are federal transfers. Second, as several scholars have argued, the driving force of LGA politics is the capture of shares of the federal cake via LGA employment, political appointment, or election (Suberu 2009; Aiyede 2009). Since virtually all LGA revenue now goes to meet personnel rather than operating or capital expenses, there are few public goods over which the general populace might be concerned. Public goods are limited and poor in quality, and this is unlikely to change. Furthermore, according to several observers, the competition over LGA employment and contracts and over who is eligible for them has translated itself into an ongoing conflict between LGA indigenes and settlers. Insofar as the public is concerned with LGA functions, it seems to be largely over who gets to share in these resources. At times, this translates itself into serious communal violence (Ukiwo 2006; Alapiki 2005).

Dele Olowu and James S. Wunsch

169

Dishonest and fraudulent elections, absence of local public goods to stimulate public involvement in LGA affairs, weak linkages with local civil society, ineffective levels of councillor and state government supervision, politicization of LGA elections by state governors (where state governors involve themselves in LGA elections in order to elect LGA chairs who will support them and their factions in state races), and dominance of the LGA chairmen over other LGA actors combine to leave LGAs virtually unaccountable to anyone except local political elites and to those attentive to their share of local spoils (Crook 2003). Overall, the evidence seems to suggest that the opportunities for improving LGA accountability lie in strengthened local political processes, enhanced federal and state oversight for transferred resources, and more effective media and civic engagement (Lessman and Merkwardt 2010).

Capacity Capacity in Nigeria’s LGAs can be grouped under the categories of fiscal resources, personnel, and capacity of local council assemblies. All are hindered by a number of reinforcing factors. Fiscal Resources

LGAs have minimally remunerative tax sources. These include primarily tenement (property) rates, fees, and user charges. Tenement rates have largely been preempted by state governments, although in a few states they constitute the greatest revenue source for local governments. Together these provide an average of only 10 percent of LGA revenues, although intensively urbanized areas such as Lagos do mobilize much higher locally generated revenues. The remainder of LGA revenues comes from their shares of the federation account, VAT, and state revenues, the last of which is not reliably transferred to them by the states. While the federal monies, at 21 percent of the federal account, are not inconsiderable, the large salary budgets that the frequently overstaffed LGAs carry nearly exhaust these monies. Since Nigeria’s fiscal federalism pays no attention to LGA revenue efforts or program performance, these allocations provide no incentive for LGAs to enhance their capacity and performance. Furthermore, the limited local revenue sources that LGAs may tap into give them no incentive to develop their local economies. Overall, these reinforce rather passive LGA governance. As a result, there is neither a fiscal carrot nor stick pushing LGA governments to enhance revenue capacity or performance. This, along with high personnel costs, leaves LGAs paradoxically fiscally weak, even though they have access to large revenues compared to most local governments in Africa.

170

Nigeria

Where LGAs mobilize more local resources, as in Lagos, the story is significantly different. Personnel

The reform of LGAs resulted in improved local government personnel, though professional and political personnel in many LGAs still are not well qualified (Dibie 2003; Gboyega 2003; Freinkman 2007). The cause is that local government service has historically had lower status among civil service personnel. Also, local amenities, particularly in the more rural and less developed areas, are much inferior to those offered in federal or state capitals. Opportunities for promotion at the local level are relatively limited compared to the state and federal services. Opportunities for further education, spousal employment, and so forth are also limited in most LGAs, though not in state capitals or urban centers where basic infrastructure exists. Professional and managerial local government personnel are rotated among LGAs every few years, causing turnover in critical managerial and professional positions. Furthermore, the rapid growth of the number of LGAs in the 1990s introduced great personnel turmoil and thinned out personnel in the many subdivided LGAs. It is unclear how much they have recovered from this. Perhaps because of all these factors, research on both account management and local revenue has found the performance of LGAs to be low (Adedokun 2006; Agundu and Waleru 2008). In general, sources argue that LGAs are overstaffed with frequently ineffective personnel, which they inherited from the era of military rule. High severance costs have made personnel rationalization and restructuring difficult to impossible for LGAs. Finally, the political imperative of spreading the spoils of the federal cake to local indigenes leads to overstaffing and irrational staffing decisions as well, with nonindigenes frequently being excluded from local service (Alapiki 2005; Suberu 2009; Gboyega 2003). However, several factors have helped to mitigate these problems. First, there is a commitment to training local government personnel throughout the country. At the beginning of local government reforms, training was localized to three regional university centers: in Zaria for the northern states, at Nsukka for the southeastern states, and at Ile-Ife for the southwestern states. Over time other institutions, including the national training center, the Administrative Staff College of Nigeria, and a number of other public and private institutions, have become involved in training local government staff. Another positive element of the reform has been harmonization of salaries of local government with those at the state and federal levels. This ensures that similar conditions are enjoyed by staff of the same qualifications at all levels of government, although this also poses challenges for small, rural local governments in paying high salaries. Finally,

Dele Olowu and James S. Wunsch

171

the creation of local government service commissions has ensured that local government staff is protected from victimization by local government politicians. Similar arrangements also exist at state and federal levels. Local Council Assemblies

Research on local councils suggests these are weak and ineffective. Politically they are far less powerful than the LGA executive (the chairman), and personally they generally have less education and familiarity with government and administration. The spoils that the chairman controls are immense, and many councillors are vulnerable to the chairman’s ability to offer appointive positions, influence contracts, and offer LGA jobs to them or their family members. Furthermore, they lack staff resources to strengthen them as autonomous, legislative institutions. Information on the 2003 and 2007 local elections suggests that they were fraught with fraud, violence, corruption, and intimidation, largely over control of the spoils available to LGA chairs (Olowu and Wunsch 1995; Wunsch and Olowu 1996; Barkan, Gboyega, and Stevens 2001; Joseph and Kew 2008; Rawlence and Albin-Lackey 2007). In this, they may not be any different from the elections to state and federal offices, both legislative and executive. Little is known of the capacity of the LGA “cabinets” to develop policy and lead management of their sector responsibilities. Again, research dating to the 1990s is discouraging, showing them generally to be of low professional competence (Wunsch and Olowu 1996). Overall, the capacity of LGAs is generally regarded as weak, although this is actually not peculiar to LGAs (Barkan, Gboyega, and Stevens 2001, Olowu and Erero 2009). However, much of the research on LGA personnel is dated, and more recent research that has been spotty in its coverage suggests some LGAs have made considerable progress (Olowu et al. 2010).

Political Economy of Decentralization Decentralization in Nigeria is a de jure fact. While political forces during the First Republic centralized government at the regional level, and military rulers centralized government during the civil war until the 1980s, the bulk of Nigeria’s history has been as a federal system. Today, decentralization to states is a legal and institutional reality, and sustains de facto political existence as well. However, its administrative and programmatic dimensions, at least to LGAs, are weaker. The forces behind Nigeria’s decentralization include its size and large population, and its regional, ethnic, and religious diversity. These create administrative complications and sociopolitical centrifugal focus whose

172

Nigeria

impact cannot be underestimated (Suberu 2003, 2009). However, from a political-economic perspective, the more powerful force for federalism and local-level governance is the powerful distributive dynamic that has developed around Nigeria’s oil income. Consensus among scholars and other commentators on Nigeria is that Nigeria’s political class at all levels is a rentier class, with its large subnational elements depending on oil account allocations to states and LGAs for direct and indirect income. Furthermore, along with state and LGA elected and appointed political officeholders, there are some 1,160,000 civil servants and nonpolice personnel employed by state and local governments, all dependent on a steady flow of federal oil rents (Adegoroye 2006). These resources are the “fuel” that sustains what many regard as the still strong patrimonial patron-clientage dimensions of the Nigerian polity (Alapiki 2005; Joseph and Kew 2008). As such, subnational governance is powerfully institutionalized in Nigeria. Since LGAs are seen as a source of incomes to local residents, there is a strong gravitational pull to establish more of them, to locate LGA headquarters in their communities, and to resist their loss. Rather than enhancing grassroots democracy or local development, this seems to have driven much of LGA politics in the preand post-1999 era, and at times has led to considerable violence (Ukiwo 2006; Diprose and Ukiwo 2008; Alapiki 2005). Thus, there is no pressure to reduce the numbers, budgets, employment, or nominal roles of LGAs. Indeed, such are the distributional processes in Nigeria that it was only the constitutional entrenchment in 1999 of 36 states and 774 LGAs that slowed what seemed to be an inevitable proliferation of ever more of each (Aiyede 2009; Alapiki 2005). In 2006, some 45 percent of total government expenditures were made by state and local governments, which totaled approximately 20 percent of gross domestic income (GDI). Ninety percent of these revenues came from federal accounts. The political class shows neither desire nor intention to eliminate or redirect to public benefits, such as propoor policies, the institutional structure that “feeds” it (Crook 2003). There are probably justified capacity concerns held at state and federal levels that incline some to minimize LGA de facto authority over programs and policies. However, the experience of reversing the devolved responsibilities for primary health and basic education in the late 1990s did not ensure more effective performance in either of them, much less democracy. Of course, there is reluctance of federal ministerial personnel to allow greater decentralization. This may be explained in part as the normal reluctance of bureaucracies to cede authority and control over resources to other entities, the possible loss of opportunities to capture rents in project funding and implementation, even though this does not deny the fact that there are real concerns regarding the probity of LGA officials. Similarly, while policy of the Federal Government of Nigeria emphasizes decentralization to

Dele Olowu and James S. Wunsch

173

enhance agricultural and rural development, considerable effective control over these remains at federal and state ministerial levels, particularly in project planning and implementation. This too has not been conducive to remarkable improvements in performance (IFPRI 2008, 2009). It is clear from the research available that neither state governments nor LGAs have generally been effective agents of democratization or development. Both the 2003 and the 2007 state and national elections were riddled with fraud and violence. Human Rights Watch (2007b), Amnesty International (2007), and other such organizations were highly critical of both. The emergence of the Peoples Democratic Party as the victorious party in twenty-eight of the thirty-six states in 2007 also does not augur well for a positive role played by subnational governments in developing democracy. The fact that thirty out of the thirty-six governors elected in 2003 had been investigated for corruption before their terms were up is equally worrisome for their performance. What evidence exists on LGAs as agents of democracy and development is also discouraging, with 2007 LGA elections badly tarnished by fraud and intimidation, and with substantial concerns raised about their chairs’ tendency toward corruption (Diamond 2008; Aiyede 2009; Suberu 2009). Research on the partially deconcentrated and delegated educational system (over which the central government retains substantial control) also shows mixed performance, which suggests these problems are deeply rooted (Geo-Jaja 2006; Ikoya 2007, 2008). In spite of these performance issues, there are immense pressures sustaining continuation of a three-tier system of governance in Nigeria. In this respect, decentralization is self-reinforcing since the political interests of a large proportion of Nigeria’s multiethnic elite directly benefits from the resources it spreads across the 36 states and 774 LGAs. Indeed, a leading Nigerian scholar of its federalism estimated that 90 percent of Nigeria’s federal monies go to 1 percent of the population (Suberu 2009).

The Decentralization Sequence The effectiveness of decentralization as a developmental or democratizing reform is not clear, particularly regarding the LGAs. Much of the reason for this may lie in the sequence decentralization followed. The contemporary institutional structure and the fiscal basis of the present local government system date to the military regime of Olusegun Obasanjo and Shehu Musa Yar’Adua. In this sense, all three steps (political, administrative, and fiscal) of decentralization, were accomplished in one fell swoop. However, the reality is that nominal legalistic changes do not translate into realities on the ground, and that in the dynamic and changing Nigerian political landscape one cannot translate this complex process into a sequence such as

174

Nigeria

political to administrative to fiscal. In Nigeria the process has been disjunctive and erratic, with movement in one direction, then reversals, then movement in another. Political autonomy at the local level was accomplished early as was administrative and some fiscal decentralization. However, the key issue for Nigerian decentralization is that the alternation of military and civilian leaders seriously compromised the structure of accountability. This failure has little to do with the sequencing—whatever it might have been— of these three elements of local government. The legal and institutional structures and fiscal bases for decentralization were established in 1976, and substantial functions were added rapidly. Elections of varying quality have occurred at LGAs several times, with civilian rule interrupted by its suspension and replacement with direct rule by military officers. Political autonomy was part of these elections, and one might think all the ingredients for successful decentralization were then in place. However, that did not develop. What did was a local political life focused on capturing spoils: jobs, buildings, contracts, and, optimally, an LGA itself (Suberu 2009; Aiyede 2009). Some sources argue this fit with Babangida’s motivations in enhancing LGAs and increasing the number of states, which were entirely Machiavellian: to feed the political classes, to stimulate subnational conflict over state and LGA creation, and thereby deflect attention from his endless democratic transition (Kraxberger 2004; Suberu 1997). This might explain the unfortunate direction that decentralization took in this era. Taking anyone’s rhetoric about the supposed purposes of decentralization at face value is liable to be misleading and probably naïve as well. Multiple actors with many different agendas—some no doubt hidden—have been involved in this process over its thirty-five years. If one sought a single driving factor in decentralization throughout its history, both to states and LGAs, it would probably be trying to cope with Nigeria’s fissiparous ethnic, religious, and regional tendencies. Many involved in Nigeria’s governance have supported decentralization because they genuinely believed it would improve services (i.e., Minister of Health Ransom Kuti regarding primary health care) or provide a stronger foundation for democracy (probably Obasanjo and the elder Yar’Adua in their earlier military days). Improved service delivery and (re)building democracy from the grassroots were articulated as objectives of Nigeria’s local government reform in 1976 (Nigeria 1976). It is not clear that these have been supported consistently by other military and civilian leaders since that time. For Babangida, for example, encouraging the emergence of viable local or state political communities as democratic and development agents was not part of the agenda. Instead what emerged, encouraged by the substantial federal monies that flowed to states and localities, was an LGA politics focused primarily on patron-clientage at the local level, and one seen by the states under civil rule as primarily an arena to mobilize political support for state elections.

Dele Olowu and James S. Wunsch

175

This was sustained, as Rotimi Suberu and others argue, by federal revenue transfers that obviated any need to raise local revenues and because there were—and are—so few local resources available after fixed personnel costs (Suberu 2009; Aiyede 2009). Thus, local politics is about capturing private goods (jobs, rents, contracts) instead of making choices over collective goods. This is certainly not atypical of local politics anywhere. The key factor here may be that the sequence followed in Nigeria offered immense, locally cost-free resources to localities before they had to make any tough choices such as how to collect and allocate local revenues. As a result, working institutions to make such choices and hold political leaders accountable never developed. Thus, there is little sense of a local civic politics, and neither local politicians nor local publics expect much honest governance or accountability from the LGAs. Thus, local politics in Nigeria does not deal much with programmatic or developmental choices, quality of services, or distribution of tax burdens—except in a few rare cases. The LGA structure essentially established in 1976, though amended under Babangida, is still largely followed. Decentralization to the LGAs has been stagnant or gradually eroding since 1999, as most observers agree that states have expanded their roles. However, this has not led to improvement in development or democracy (Barkan, Gboyega, and Stevens 2001; Gboyega 2003). While states could—and probably must, if LGAs are to improve—play a stronger role in enhancing vertical accountably and in supporting measures to enhance horizontal accountability within LGAs, they have not done so thus far. Effective decentralization in Nigeria can occur only with two complementary components: strong states and strong LGAs. Most observers believe that the shortfalls at both levels are because of the corruption and rent-seeking behavior that plague all levels of Nigerian governance (Olajumoke 2011). Nigeria front-loaded a sequence that neglected requirements and incentives for local communities and governments to make real decisions regarding locally taxed resources and local priorities. It is this sequencing that is critical in understanding the problems that Nigeria’s LGAs face today. Political, administrative, and fiscal decentralization were rapidly implemented, but have been erratically supported and followed through since. These steps, in their haste, may have precluded growth of effective local democracy, civil society, and public accountability because they were taken without establishing a nexus between raising revenue and making choices. Absent this to incentivize local publics to check, punish, and reward local officials for imprudent and irresponsible behavior regarding local resources and priorities, the flood of resources and autonomy that came to local political elites seems to have facilitated their capture of and lasting grasp on local government (Crook 2003). Weak state capacity to exercise vertical accountability, probably caused by similar factors, may have furthered these weaknesses.

176

Nigeria

Intergovernmental Relations While the national level set the stage for decentralization in Nigeria through the constitutional structure and its reliable 21 percent fiscal transfers, it is the subnational arena that is particularly important to Nigerian decentralization. The 1999 Constitution places primary authority over local government in the states. Local government responsibilities are limited, shared with the states, and require enabling legislation before they can be undertaken. Indeed, the 1999 Constitution effectively rolls back much of the authority LGAs had gained during the mid-1970s to1990s, moving the LGAs more toward delegation than devolution. Joel D. Barkan, Alex Gboyega, and Mike Stevens (2001) found that some states were taking a largely hands-off approach to their LGAs while others were closely supervising them, particularly their budgets and expenditures. While ministries in the state sector also have substantial authority over LGA programs, it is not clear how closely they have exercised it. PHC, at least, seems to have largely continued in its pre-1999 guise, with LGAlevel personnel apparently in charge of most aspects (Khemani 2005). Some evidence suggests that development projects are dominated by state and federal government (IFPRI 2008, 2009). State assembly members also do not appear active in LGA affairs except insofar as they offer resources for personal and parochial goals or affect factional and partisan conflict (Lewis 2009; Dibie and Bradley 2008; Smith 2007). Federal personnel are even less engaged in LGA affairs. This is in sharp contrast to their close involvement under military rule. Several sources have reported gubernatorial political involvement in LGAs. This has been observed regarding elections and the effort by state governors and political parties to get support from LGA chairs. Reports of extensive fraud, ballot box stuffing, closed polls, violence, and intimidation in elections suggest this might have more purposes than get-out-the-vote efforts (Rawlence and Albin-Lackey 2007; Joseph and Kew 2008; Human Rights Watch 2007a). Most of the governors also raid LGA accounts and deny them of their federal financing, which is LGAs’ lifeline (Olowu et al. 2010). Lev Freinkman (2007) reports extremely weak mechanisms of intergovernmental relations between states and the federal government. He found a dearth of good information regarding programs as well as weak analytical capabilities at all levels to assess what information there was. Freinkman also reports that information is not shared across levels of government, there is little clarity about who is responsible for what, and policy is not well coordinated; consequently, the three levels of government engage in uncoordinated and at times duplicative activities. Other researchers have found that much information was collected and passed upward, to both states and federal government, but little analysis was done with it and no feedback regarding it was returned to decentralized govern-

Dele Olowu and James S. Wunsch

177

ments (Olowu et al. 2000). A consistent complaint aired by PHC personnel at LGAs in the mid-1990s and in 2000 was of the utter absence of state support for or invidious involvement in their programs (Olowu and Wunsch 2004). While federal frameworks have been established in both PHC and primary education, it is not clear how well they have been followed to maintain and strengthen performance in these two sectors. As in so many things, the pursuit of federal monies seems to be the driving force in federal, state, and local politics. It is access to these resources that probably drove the fraudulent and at times violent elections of 2003 and 2007 as well as the growth of Nigeria from three to four regions, and then from twelve to ten, twenty-one, thirty, and finally thirtysix states (Kalu 2008; Mustapha 2009). Similar pressures have pushed the number of LGAs to 774 (Alapiki 2005; Suberu 2009; Aiyede 2009). While a few state governors were hailed as progressive new-men who were part of a new generation in 2001, by the 2007 elections thirty of thirty-six governors had been investigated for corruption and four were charged with money-laundering about US$400 million. Many of these state governors have successfully used the courts to grant ever wider powers to states, challenging such federal initiatives as the Fiscal Responsibility and Public Procurement Acts aimed at reducing malfeasance in governance. Unfortunately, the record of LGA chairs has been similar, even if on a lower scale (Barkan, Gboyega, and Stevens 2001; Suberu 2009; Aiyede 2009; Alapiki 2005). Nevertheless, even these are counterbalanced by a few states and local governments that differ from the above generalizations: some state and local governments have bonded effectively to achieve more impressive developmental outcomes than others (Olowu et al. 2010). In sum regarding intergovernmental relations, over the past thirty years a system of local government has been created that bears great potential if more robust accountability and management systems can be instituted. At the initial stages, local government reform was sponsored and managed by the federal military government. Civilian democracy (1999 to present) has reinforced the role of states while diminishing that of the federal government in local governments. Unfortunately, except for a few cases, the states neither have the capacity nor the motivation to support and assist local governments in building their capacity for development or democracy. The challenge for intergovernmental relations in Nigeria therefore must be to enhance the capacity and incentives for the states to fulfill this critical role.

Conclusion In spite of the many issues and challenges that we discussed, Nigerian decentralization has accomplished a great deal since its current framework was established some thirty-five years ago. From a country only a few

178

Nigeria

years removed from a disastrous civil war and whose future as one entity was in doubt, Nigeria has consolidated itself as a highly multiethnic federation under a strong, though challenged, government. Nigeria’s powerful fissiparous pressures have been contained; it has conducted three national elections, flawed as they have been; and it has successfully transitioned executive power twice. Its National Assembly has made strides as a serious legislative body and check on the executive branch (Lewis 2009). The country has consolidated a 36-state federation, to which many real functions and powers have been delegated, and has instituted 774 local government units. The local governments have elected executives and legislatures in place, and functioning state-led civil service structures that are deploying improved personnel throughout. Nigeria has established an institutional framework with the potential to consolidate democracy and provide basic services and development for its nearly 150 million people (Suberu 2009; Aiyede 2009; Kalu 2008; Mustapha 2009). Yet there are serious obstacles to reaching this potential, ones clearly seen at the local level of governance, and as we discussed in this chapter. Despite their problems, Nigeria’s LGAs have made impressive strides in their de jure and, at times, de facto autonomy and authority. This capacity, while still weak in a number of areas, is far ahead of where it was ten years ago. Some local governments have indeed performed very well, though primarily as a result of strong local leadership (Olowu 1990; Barkan, Gboyega, and Stevens 2001; Human Rights Watch 2007a). Sadly, however, the vast majority have fallen far short of the hopes once held. The greatest impediments to their progress toward becoming institutions able to make collective choices, manage conflicts, and provide public goods have been clear weaknesses in accountability; overdependence on federal account allocations for fiscal resources; and, less prominently, structural defects and rigidity, and constitutional ambiguity over their authority. The most powerful factor causing the first two of these has been the size, availability, and allocation policies of Nigeria’s immense oil rents. Many Nigerians and scholars describe oil as “the curse of Nigeria” (Diamond 2008; Collier 2007). While these monies make many good things possible for Nigeria, their availability introduces a multitude of situations that encourage opportunistic behavior and allow evasion of the consequences of those behaviors, so rent-seeking behavior and moral hazard are rampant in politics and governance (Karl 1998; Ostrom 1990, 2005). Elinor Ostrom (2005) has described the singular challenge of effective governance as managing human tendencies toward these behaviors. Any time that actors control disproportionate political or economic resources, as well as advantages in information, it must be expected that some will engage in opportunistic behavior—behavior that give them advantages at the cost of others. When they are not held accountable for these

Dele Olowu and James S. Wunsch

179

actions, they face moral hazard and can only be expected to continue in these actions. When this dynamic continues and others see it, their willingness to restrain themselves from such behavior for moral or ethical reasons is reduced: no one wishes to be played for a sucker and, furthermore, simple survival may require them to behave this way. This is the dynamic that causes the tragedy of the commons as well as erodes investment in public goods (Ostrom 1990; Olson 1965). The public treasury is such a commons while honest and efficient governance is such a public good. When everyone else is raiding the commons for their own benefit and when no one—or only a few—are investing in public goods, not only is it personally disadvantageous to restrain oneself and invest, it is futile. The virtuous behavior of a few—or even many—cannot negate the opportunistic behavior of others and, once unleashed, these patterns become stubborn, vicious circles. Oil rents create these dynamics in Nigeria, partially because of their immense size but even more because of how they are allocated. States and LGAs receive their shares of federal accounts automatically, regardless of policies or performance. These resource flows have grown to where they are now some 95 percent of federal accounts, and the temptation that encourages opportunistic behavior by state and local officials and local civil society has grown apace. There essentially is no consequence for opportunistic behavior beyond an occasional and usually ineffective criminal investigation since these monies are allocated without regard to local policies, programs, or performance; since state and federal audit functions are underfunded and toothless; and since few among these actors have an interest in accountability for these funds. Furthermore, because generally few local resources are raised, there is no revenue-expenditure nexus to stimulate local concern over distribution of a commons regularly replenished without the efforts of local dwellers. Just as seen regarding commons such as ocean fisheries, there is instead a scramble to capture the resources for the private consumption of individuals (Ostrom 1990). This corrupts public institutions both from within (among public officials as they pursue their interests and rents) and from without (as local publics tend to organize only around getting their share of the commons). Thus weakened, local institutions are ineffective in providing public goods—the first of which is honest and effective governance itself! The desirable competition among politicians for the electorate’s favor by producing public goods is displaced by a competition among them to capture a share of oil rents to use to purchase votes from that same electorate (Collier 2007). Democracy and local development, two of the supposed goals of decentralization, decay as a result. Stability, the third, is not far behind, as these goods are never evenly distributed (Ukiwo 2003). The critical lesson for Nigeria and other African states pursuing decentralization is that building accountability—in all dimensions—must be the

180

Nigeria

first step in that process. Forcing local governments to raise a significant proportion of their funds from their own populations is necessary—though surely not sufficient—to mobilize a potential local check on opportunistic behavior and moral hazard. Monies allocated from a central government must carry strong requirements for local cost sharing from locally raised revenues to strengthen accountability downward. Close attention to benchmarks set for local policies, programs, and performance must also be part of central revenue sharing. When LGAs do not perform, measures such as slashing shared revenues must be followed. There need to be strict performance and fiscal standards and audits to help assure them. These standards must apply to state governments as well as LGAs to ensure that these dynamics do not continue at that level and contaminate local government from above. Strong supervisory frameworks regarding LGA functions, both internal ones such as revenue, budget, and audit as well as those for services such as education, primary health care, agricultural extension, and general development, will be necessary along with strong local public accountability. In Nigeria, the states must set and assert these standards and guidelines, with the federal government assuring that the states follow through. Stronger horizontal structures are also needed and can be partially developed through judicious use of deconcentration and maintaining the states’ civil service structures and the role of professionals in delivering local services. This is suggested by research on primary education (Ikoya 2007, 2008). The LGA councils also need strengthening, with enhanced authority, staff resources, and training. LGA chairmen may currently be overly powerful vis-à-vis the local councils. However, there is ample evidence of current extensive corruption among local councillors, so until Nigeria’s federal transfer problems are dealt with, this is a problematic situation regardless (Human Rights Watch 2007a). The current ban on additional LGAs (and states) should be sustained to reduce the instability this causes LGA governance and administration as well as the divisive distributional dynamics that it stimulates among ethnically based associations. Because each new LGA automatically gets a roughly equal share of federation accounts, there is an opportunistic incentive to add LGAs relentlessly, that being a quick and easy way to capture a significant share of the federal fiscal commons without having to raise any local revenue from the public or provide any public goods. The larger the rents available, the more difficult it is to control their effects, particularly on electoral democracies, given the vast flow of resources to use in patronage politics (Collier 2007). Another lesson in Nigerian decentralization is the desirability of more flexibility in LGA structures and clarity in constitutional powers. Currently, all Nigerian local governance is provided through identically designed local institutions of comparable population sizes and, usually, of similar area.

Dele Olowu and James S. Wunsch

181

However, this has meant that highly urbanized and rural areas have identical structures for very difficult governance challenges. It has also meant that more populous areas sharing common and related economic and social problems, such as metropolitan areas, are subdivided into multiple local governance units. Exploring and managing positive and negative spillovers (externalities) is blocked by this. The possibility of options for alternate local governance institutions with varying authority and autonomy is desirable for Nigeria and other African states that are decentralizing. Finally, the 1999 Constitution is ambiguous as to what LGA powers actually are in most sectors. The primary lesson of Nigerian decentralization is that effective accountability must be established before the other three dimensions (authority, autonomy, and capacity) can be expected to lead to effective local democracy or development. Poorly performing local governments also seem unlikely to lead to state stability. All three dimensions of accountability (downward, lateral, and upward) need to be in place because they complement and reinforce one another. As access to resources drives human behavior, that access must be possible only through structures that incentivize others to pay attention to and demand responsible instead of opportunistic behavior by LGA officials. This is never a simple process anywhere, as all public officials—in less-developed countries as well as developed states—with access to transferred resources will exhibit opportunistic behavior (Kersting et al. 2009). The particular challenge facing Nigeria is the great volume of oil money—essentially huge rents—entering the political system. Thus, chances for opportunistic and corrupt behavior will continue to create moral hazard at all levels, including inevitably LGAs. State and federal politicians will no doubt continue to look to the localities and local political elites to deliver votes, and will use these rents to suborn democratic checks at all levels (Diamond 2008). Paul Collier (2007) found that the unrestrained electoral competition for votes typical of natural resource–based democracies could be restrained only by effective systems of checks and balances. This is why a free media, an honest and independent judiciary, and a strong and public goods–oriented civil society working with broad-based intergovernmental organs, among other checks, are so crucial at all levels but particularly for developing effective decentralization in Nigeria (Lessman and Merkwardt 2010).

9 South Africa: Decentralization and the Apartheid Legacy Louis A. Picard and Thomas Mogale

In this chapter, we examine the state of decentralized governance at the provincial and local government levels in South Africa. Emphasizing the institutional setting (fiscal and administrative) and the politics of decentralization, we also argue that historical factors have greatly affected South Africa’s political culture and its definition of local governance. Decentralized governance in the country emerged from the conflict resolution process (1990–1994) that ended its system of apartheid. After the establishment of nonracial government, the devolution arrangements that came from the negotiations were, at least in part, abandoned by the ruling African National Congress (ANC) in its quest to ensure government control over the development planning processes in rural and urban South Africa, and to manage political processes at the subnational level. The result has been a set of contradictory processes that reflect a governance theory that emphasizes decentralization and local control, but political arrangements that are both highly centralized and often inefficient.

Overview The institutional framework for decentralization in South Africa is based on three things: (1) the 100-year history of white local government, which goes back to the 1890s in the Cape Colony; (2) the urban base of challenges to apartheid, which focused resistance in large part on segregated local government structures in the 1980s; and (3) a series of constitutional and legislative arrangements that evolved out of the negotiations between the National Party and the ANC between 1990 and 1994. This legacy has affected the administrative, fiscal, and political interface in local govern183

184

South Africa

ment ever since. Many of these mechanisms are well established (and institutionalized), but some are more transient or fluid. The time line for negotiations on local government was defined by civil conflict (the antiapartheid struggle), the four years of negotiations, and the constitutional changes that followed. As a result, there was a reverse sequencing of decentralized governance because it was linked to democratization and preceded state reconstruction. As democracy became more entrenched, SNGs came more under centralized control. Thus, as a component of the negotiation process, decentralization preceded rather than followed democratization. To put it another way, as political stabilization increased, there was increased centralization in South African government. The sequencing of decentralization was complex. The timing of decentralization, relative to other major governance changes, was based on the 1990s negotiation process. Political decentralization came first (as a result of the 1990–1994 negotiations), but was preceded by a set of administrative and fiscal practices that had been introduced much earlier and conformed to the late apartheid policies of P. W. Botha in the 1980s. These fiscal and administrative procedures, at least in part, remain in place. The goal in reforming local government was the integration of black and white units of local government and the creation of entities that incorporated all segments of the population into devolved political units. In the negotiations the National Party looked toward small units to be created; the ANC looked toward larger units and, in large cities, to the creation of metropolitan entities (which now number eight). As was the case with most elements of the negotiations, the ANC was able to get its way, resulting in municipalities that in practice were also more deconcentrated than devolved. The aim of the ANC-dominated government after 1994 was to increase the administrative capacity of local authorities to deliver services (Atkinson 2003), not to increase citizens’ participation in their government. There were two elements to the concerns of the ANC-led government regarding decentralization: one related to ideology and political control, and the other to administrative competence. ANC leaders believed in a planned process of economic development, even within the adapted framework after 1996 of market principles. This meant that the party did not wish to see deviation from its policies, which has occurred at various times in Natal and the Western Cape and even occasionally in opposition-led local governments in medium and small municipalities outside of the Western Cape (most recently in Gauteng). Second, SNGs were perceived by both the ANC leadership and the public to be administratively weak and corrupt. This led to increased central government control of and strings placed on the fiscal authority of subnational authorities after 2000 (Freund 2006). Several political, social, and economic factors, which all help to explain the decision to decentralize, have also created limits on it. Politically, oppo-

Louis A. Picard and Thomas Mogale

185

sition to the ANC plays itself out at the subnational level. Opposition from both the left and the right is geographically delineated. Politics remains intimately defined by regional political culture in South Africa. The social dimensions of local government continue to be determined by racially defined distinctions. Residency, service delivery, size of local government units, and taxation issues all link up (if only obliquely) to race and to residential patterns that, except for a narrow elite, remain largely segregated.

Authority Authority over SNGs is defined constitutionally, by the nature of the ANC’s political culture of centralized control, and through historical patterns of state relationships with SNGs in South Africa. The balance between national and local government, inherited at the time of majority rule from the National Party and from earlier regimes, has remained constant if deracialized. South Africa is a unitary state, with SNGs playing a largely deconcentrated, social service delivery role, though with elements of opposition located in certain geographical areas. This deconcentrated centralism has contributed to the stability of the South African political system in the sixteen years since the beginning of nonracial government, but has meant an increasingly depoliticized and in some cases alienated electorate (Bond 2000). Local government has faced significant challenges in South Africa. While the country has made great strides in economic development, social equity has worsened and local governments have failed to address issues of social service delivery. Municipalities (the generic term for local governments) are almost entirely financed by transfer income, with the exception that the large metropolitan areas such as Johannesburg and Cape Town collect as much as 97 percent of their revenue. In the past five years, the South African government has admitted that municipalities have come under increased levels of stress and carry unacceptable levels of unfunded mandates (Schulz-Herzenberg 2007: 138). In 2006 (and beyond), South Africa has gone through a significant number of public demonstrations and violent reactions to the ineffectiveness of local authorities in the delivery of social services (Atkinson 2007). Bureaucratic inefficiency, political arrogance, corruption, and resistance to tax collection have all been blamed for the local government crisis of 2006– 2007 (Cuthbertson 2008). There continue to be significant numbers of public protests against local governments, though the intensity and amount of demonstrations have leveled off in the past few years. Authority outside of the national sphere is both deconcentrated and delegated. National programs in education, health, housing, water, and land

186

South Africa

use all define deconcentrated authority in South Africa’s deconcentrated departments and subnational authorities. The South African state has also owned a wide number of public enterprises and other parastatal bodies, which carry major social and economic responsibilities for the country. Large sections of the economy have been controlled by these autonomous bodies under delegated authority.

Autonomy Subnational autonomy reflects the extent to which local and regional governments can make political, administrative, and fiscal decisions with some degree of independence from central government. The autonomy of SNGs in South Africa is limited. These limits are often exacerbated by a poor understanding of the principles of separation of the party and government by members of the ruling party. Ironically, this has changed little either before or after the end of the apartheid regime. South Africa, legally and politically, is a unitary state with high levels of centralized policy control, though mediated by uneven implementation caused by inefficiency, corruption, and limited fiscal resources. Patterns of control over subnational institutions are defined by electoral policy and administrative practice. Local government elections are controlled by the country’s political parties and, except for the Western Cape, in effect by the ruling ANC. With the exception of areas where there are pockets of opposition, such as Natal, the Northern Cape, and the Western Cape, the ANC in any given election and at all levels of government is likely to receive between two-thirds and three-quarters of the popular vote. Political loyalties infiltrate local government services, beyond those offices that are scheduled as political. De facto political appointments penetrate deeply into the administrative system and patronage appointments are widespread. Political corruption and patronage are linked to patterns of appointment and political loyalty throughout South Africa (Lodge 2002: 129–152). SNGs in South Africa have access to equitable share revenue. Compared with many developing countries, and especially other countries in Africa, local government revenue is both extensive and relatively secure. However, public expectations of government services were high in 1994 and have increased in the years that followed. Virtually all access to revenue is through transfer income except in the country’s largest cities. South Africa’s revenues are inadequate to cover decentralized responsibilities and little money is available for developmental activities. The country suffers from a classic unfunded mandate problem that limits local autonomy. South Africa’s revenue problem is twofold. First, the primary national taxes on individuals are a national income tax and a value-added tax. Sub-

Louis A. Picard and Thomas Mogale

187

national authorities are restricted in the extent to which they can raise taxes on their own. Rural local governments (but not district authorities) actually have considerable legal leeway for raising additional revenue, including property rates and user charges for electricity. The main constraint is the lack of a tax base and the poverty of most rural local residents. Secondly, South Africa also suffers from a form of tax payment resistance inherited from the late apartheid period and, despite attempts to encourage or enforce tax payment (the so-called Masakhane campaign), SNG access to revenue is limited and may be decreasing on a per capita basis. Taxes on property and services are collected at high levels (95 percent of capacity) in former white areas, but are low in historically black areas and virtually nonexistent in former homeland areas. Control over expenditures and spending patterns is centralized and uncertain, and likely to remain so. Observers suggest that significant amounts of government money (according to press accounts) disappear through ghost workers, corruption, and administrative waste, particularly at provincial levels (Wittenberg 2004; Picard 2005: 331–332). This has led to an increasing skepticism about the efficacy of government at the local level. Administrative control over public sector employees is limited and intermittent. In principle, such personnel matters as recruitment, payment, discipline, and retention are controlled by national policies formulated and reviewed by the national Public Service Commission, and implemented by SNG administrative heads. However, in practice the supervision of public sector employees and administrators is both formalistic and often based on patronage relationships. Many positions, particularly in the areas of the former homelands, are considered sinecures. Local government administrators formally fall outside of the civil service, but are heavily regulated by the South African government, which is committed to creating a unified public service to incorporate all levels of government. Even the government admits that local government has suffered from political interference in hiring and patronage as well as from corruption. There also exists a quasi-official associational framework for decentralization in the South African Local Government Association (SALGA), which might be an advocate for local government autonomy and capacity and is largely active at the national level. SALGA is a quasi-NGO that is mandated to support local government in terms of research, policy strategies, technical development, and information dissemination. At the local levels, it is largely moribund in most areas of the country outside of the metropolitan areas. The main weakness with SALGA at the regional level is that it is thinly spread and poorly resourced. Even when it participates in provincial- and national-level debates about local government, observers suggest that it lacks technical skills to make a meaningful impact on

188

South Africa

debates around local finance, integrated development planning, and so forth; hence, national and provincial departments outwit localities or impose their views. In summary, only limited powers and resources have been devolved to SNGs in South Africa. Most authority and the bulk of fiscal resources have been deconcentrated or delegated to departments and public authorities. Moreover, the shortfalls of money at the local government level have been significant and show signs of increasing rather than decreasing over time. While the delegated sector appears to be shrinking as a component of the public sector (Bond 2002), a culture of oligarchy seems to be evolving in South Africa, one that combines political and bureaucratic inefficiency with political arrogance and disdain for media scrutiny and public openness (Cuthbertson 2008). An increasingly sluggish, openly elitist, and ever more closed system of political leadership within the ANC have been cause for concern among journalists and academics alike. The inability of opposition parties to be relevant at the national level appears to also have limited their ability to challenge state authority at local government levels (with the exception of the Democratic Alliance in the Western Cape).

Political Accountability Local authorities in principle are devolved bodies, but with limited powers. They are elected on a partisan basis in regularized national elections. The electoral process, considered free and fair by international standards, serves as the fundamental mechanism of political accountability and thus far it has helped ensure political stability in South Africa. Elections and responsible government are rooted in political party majorities and party-based elections at the local government level. Local governments have elected political heads based on majority control of the legislative body. In theory, local government civil services are nonpartisan and selected by merit, but in practice the senior administrators have been loyal to the governing political party, the ANC, since 1994. Political party systems are historically strong and founded on a fused government style of party discipline that is common in continental Europe. A strong ideology and an orientation toward development planning are important components of this process. Party discipline has been particularly strong within the ANC. Private member bills and maverick legislators at any level of government are uncommon. There have been a number of political splits and realignments between and among South Africa’s political parties, including a split within the ANC by a splinter group seen to be aligned with former president Thabo Mbeki that formed a new political

Louis A. Picard and Thomas Mogale

189

party (Congress of the People) in 2008 before the last elections. The party made some inroads into ANC support at the provincial and local government levels and is now the third-largest party in South Africa. The Democratic Alliance, which remains the primary opposition, has increased its support in the 2012 local government elections. Further party realignment is expected in the next national and local elections in 2014. In ANC-dominant areas, the government functions as a virtual oneparty system, with local representatives, key civil society leaders, and senior (and often middle-level) administrators all functioning under the umbrella of the political party movement. This is in part a legacy of the organization’s projected image as a broad-based mass movement in its struggle against apartheid. In several areas where there is significant opposition to the ANC (the Western Cape, the Northern Cape, Gauteng, and Kwa-Zulu Natal), opposition parties are able to approach 40 percent of the votes. Political representation in those areas takes on a more neutral, multiparty form and administrators are more sensitive to political neutrality, though in the Western Cape there is a built-in loyalty among local government bureaucrats to the Democratic Alliance. Where opposition is stronger at the local government level, patterns of neutrality are similar to that of provincial levels. At the municipal level, in ANC-dominant provinces there continues to be a considerable amount of loyalty to national party leadership, particularly in larger cities. In smaller cities and rural areas, there is considerable informal loyalty to provincial party leadership. However, the links between the grassroots and the national party leadership remain strong. The ANC continues to aspire to and function as a mass mobilizing party, though it is less issue-driven than patronage-based and uses ethnic and racial loyalties to continue to solidify its support. Local government elections are fought less on the basis of national issues (and still less on the basis of regional and local issues) than on the basis of historical bonds of loyalty to the antiapartheid movement and the ANC’s historic role. Though public opinion polls often show dissatisfaction with government policies within the electorate, this does not translate into support for opposition movements, even those with origins in the antiapartheid struggle. A dominant feature of the South African electoral system is the adoption and entrenchment of the hybrid (proportional and ward) system. This by its very nature is inclusive in its coverage of electoral voices in the country and, hence, guarantees the participation and representation of minority and disadvantaged groups including women. The formation of the Independent Electoral Commission as a statutory body with levels of independence provides avenues for all stakeholders to participate unhindered. The manner in which it is constituted and the nomination process for office-

190

South Africa

bearers is designed to guarantee transparency and, therefore, impartiality in the manner that it operates. Its mandate is clear and adequately funded to execute its functions. There are no legal partisan restrictions that impact on electoral participation in South Africa. Nor does the ANC dominant-party monopoly result from electoral legislation or administrative practice. That said, there are a number of what are referred to in South Africa as “no go” areas where minority parties are not able to access communities or campaign freely. Further, while the written press is considered free, the media (except for private radio stations and one private television station) show evidence of bias toward the ANC government. South African commentators tend to mockingly refer to the South African Broadcast Corporation (known by its acronym SABC) as the “SANC.” South Africa has a rich history of citizen participation and political activism in terms of civic and local government processes. Consequently, the local government system puts citizen access and participation at the center of its electoral, planning, and budgetary processes. Several pieces of legislation and policies between 1998 and 2007 brought this into effect, calling for active participation and input from the public in the local government and planning process. Meetings and report-back sessions are a part of South Africa’s political culture. Civil society groups in the late 1980s and into the 1990s were strong since their mobilization base was opposition to apartheid and their unification under the banner of the United Democratic Front was at the forefront of the liberation struggle. As part of the legacy of the antiapartheid struggle, most civil society leaders in South Africa function under the umbrella of the ANC, though via their historical affiliation with the populist United Democratic Front that disbanded in 1993 and is now under the South African National NGO Coalition umbrella. After 1994, two things happened to weaken NGOs as a subnational accountability mechanism. First, many strong civil society leaders moved into national or local politics. This undermined the civil society groups significantly, particularly when donor funding was redirected toward the new national government in the post-1994 environment. Second, many civil society groups became subservient to political discipline by the ANC. Nonetheless, when there have been outbursts of populist challenge to government policies at the national and subnational levels, the origins of these challenges can be found in trade unions, civil society groups, and other NGOs (Atkinson 2007). Civil society, through its political tools of public meetings, hearings, debates, transparency rules, and monitoring mechanisms, can challenge authority, particularly at the local level. After opposing the recognition and use of traditional authorities for thirteen years, the ANC in 2007 decided to recognize and give them a grassroots

Louis A. Picard and Thomas Mogale

191

development role. The view was that traditional leaders represent important social networks, provide access to voters, and can mobilize citizen behavior in rural South Africa (Lehman 2007). Given that currently traditional authorities stand to gain from decentralization, they have begun to defend what is called “cooperative governance” in South Africa. To advocates of traditionalism, recognition of the importance of traditional authorities is one of the best discoveries identified in South Africa, and it is suggested that the South African model of traditional administration may be applicable to other countries (Lehman 2007: 302–303). To critics, domestic and international, traditional authorities represent hierarchical authority, corrupt institutions, and mismanagement. Other popular accountability organizations are said to be available to subnational actors and civil society to control local administration and civil servants. Metropolitan, district, and local organizations all have political, administrative, and fiscal mechanisms that are designed to underpin decentralization. These tend to work better in larger metropolitan areas and in the formerly segregated white areas. There are well-developed ratepayers associations in the northern suburbs of Johannesburg, for example, and neighborhood watch associations. These urban social groups and associations would be familiar to residents of North American middle-class neighborhoods. The role that civil society groups continue to play in South Africa is important, though the intensity level is lower than immediately after the first nonracial elections. In the electoral process, the role played by the media, watchdog groups, and civil society representatives is significant. The capacity of the press, local-level organizations, and national and regional civil society groups and interests to monitor and pressure SNGs is largely independent of government. The access of citizens’ groups to SNGs—civic unions, town hall meetings, and referenda intersecting with district authorities—is essential to democratic local governance. Such access can be described as only nominal in the South African context, though it must be said that well-organized, mostly white, ratepayers can influence both the revenue and expenditure process. The ability of NGOs to both advocate to local government and to ensure service delivery is seen by their advocates as important mechanisms of governance in South Africa. The combination of relatively efficient electoral processes and the watchdog role played by civil society groups are two parts of the South African SNG system that provide a modicum of legitimacy for local governance despite the excesses of corruption at local levels of government. Despite all the legislation, policies, and participation mechanisms in place, however, public participation is not yet effectively achieved in South Africa, and what occurs is often largely symbolic. A major challenge highlighted by local government representatives and officials is the costly nature of public participation. A majority of South Africa’s local govern-

192

South Africa

ments are cash strapped and have a limited tax base. Thus, limited resources can be committed to public participation. This is despite the fact that government, through the National Policy Framework on Public Participation, already calls for municipalities to institutionalize public participation. Exactly how municipalities will achieve effective participation without the necessary resources is uncertain. In sum, local government in South Africa fits into a system of cooperative governance that reflects the holistic view of the ruling party, the ANC. Local-level political elites tend to be ANC activists and members of local civil society institutions. Many elites are able to demonstrate their leadership credentials through their history in the liberation struggle. The associations of local government and other supportive civil society institutions concerned about subnational governance, and the existence of traditional parallel local-level governments, are parts of a governing party network at all levels of government—national, provincial, and local. The rituals of participation are provided for by political and government participation, but are often largely symbolic and merely visual. They also reflect the lingering characteristics of the ANC as an aging dominant party. Thus, accountability of local governments to local publics is not clear.

Fiscal and Administrative Accountability The South African system of government relies on principles and institutions to promote and safeguard transparency at municipal level. These include the constitutional and legal framework, the electoral process, and the role played by the Independent Electoral Commission as well as the role played by the media, watchdog groups, and civil society representation. The constitutional and legal architecture in South Africa is designed to promote and safeguard fundamental freedoms and human rights. Further, a plethora of legislation exists, including that used to set up institutions such as that of the public protector. Electoral law provides for mechanisms to address transparency in the electoral process. Virtually all urban centers now function with executive mayors that are supported by a city manager and functional staff. Smaller municipalities function with executive administrators, called municipal managers (formerly, town clerks). While the structure and process of local government administration is in the English style, the post-1994 municipal structures are continental and incorporate rural and urban areas into the same municipality. Municipalities are graded according to size and authority. Formally, South Africa’s government departments function on the basis of deconcentrated authority. In practice, there is a great deal of ambiguity in the relationship between political and administrative leadership at all levels

Louis A. Picard and Thomas Mogale

193

of government, and often confusion about roles and responsibilities within government units. In practice, South African government at all levels— national, provincial, and local—operates on the basis of cooperative governance, which suggests a centrality-defined mission and strategy. Critics of South Africa’s government also suggest that the system at all levels is characterized by high levels of patronage, rent seeking, and embezzlement (Picard 2005: 245–280). There have been, and continue to be, calls within parliament and from outside interests for greater transparency at all levels of government. Subnational nongovernmental actors, groups, and media organizations also function to monitor subnational spending. The South African press, both open and competent, acts as a first level of fiscal monitoring throughout the country. There are a number of associations and civil society groups, such as the Institute for Democracy in Africa (IDASA), which monitor fiscal expenditure. Identification of corruption and malfeasance of funds is common in national and even local newspapers. Administrative accountability mechanisms exist, in theory, at all levels of government. These focus on the control of bureaucratic behavior at the national, provincial, and local government level. These mechanisms, available to the central government, are intended to monitor and control administrative behavior. They are common to national administrators, but are also designed to monitor local government officials. In practice, they tend to be formal and unsystematic in their application. Local government employees, as noted above, are not politically neutral but are usually bonded to party interests. Bureaucratic accessibility and accountability mechanisms for both citizens and civic groups are weak. There is a suspicion of civil society groups among local-level officials. Administrative and organizational development strategies and policies are guided by centralized directives and by sporadic monitoring exercises. Administrative control mechanisms: merit rewards, discipline, dismissal, and transfer are widely used and are important symbolically. A number of fiscal mechanisms are available on paper that are intended to ensure fiscal accountability in South Africa. These include internal and external auditing requirements that meet international standards. It is fair to say that fiscal accountability in South Africa compares favorably with that in the rest of Africa and throughout the developing world. The central government uses budget rules and reporting requirements to monitor or control provincial and local government spending. Fiscally, South Africa operates in a highly centralized manner. Evidence suggests that there is occasional divergence between proposed and actual expenditure and revenue patterns. That said, South African fiscal monitoring is generally well managed at the national and provincial levels (IDASA 2008). South Africa has had a history of revenue collections

194

South Africa

coming in above budgetary predictions. There is some evidence, however, that at the national level parliamentary fiscal-monitoring mechanisms vis-àvis the executive are weak. Neither is the National Treasury always forthcoming in terms of amounts or explanations when mistaken estimations occur (IDASA 2008). Some observers have also questioned the capability of the South African Revenue Service to collect taxes. Fiscal information becomes murky at the local government level outside of the major urban administrations. Over 50 percent of local government revenue comes from user fees and other cost recovery mechanisms. The rest comes from property taxes and revenue sharing. Critics suggest that this is a negative in terms of expanding social services (Hoffman 2007). While formally fiscal management is quite rigid, fiscal mechanisms may tie the hands of subnational officials who try to implement policy. The administrative cadres in South Africa also remain potentially powerful control tools, a legacy of the apartheid period. Though they are seldom used, there is occasional evidence of praetorian mobilization efforts at the grassroots in townships and rural areas. The ruling ANC plays an overt role in this control and mobilization process. There are not always strict concerns for legal norms with regard to the ANC’s role in this.

Capacity Issues Deconcentrated and devolved authorities have a primary responsibility to deliver services. As has been noted, there are serious concerns about the ability of many, particularly rural, bodies to accomplish this responsibility. These grow from shortfalls in personnel skills and fiscal resources, compounded by often ineffective mechanisms of accountability. An effectively single-party system at many localities also contributes to stagnation in personnel and policy, as does excessive patronage and political appointments to subnational civil service posts. The service delivery protests since 2006 suggest that municipalities are institutionally weak, lack capacity in human resource terms, and lack the finances to meet these obligations. A related challenge is the inexorable movement of people into central government or, more lucratively, into the private sector. The ineffectiveness of programs to strengthen SNGs illustrates the limits to capacity building at this level regarding technical and administrative skills for delivering social services and local development. Capacity building and professional development for deconcentrated and devolved staff has focused on short courses and bridging training. It has historically suffered as a result of the en masse inclusion of virtually all homeland administrators, including local governmental officials, in the administrative system after 1984. Affirmative action policies have resulted in decreasing standards at all levels of government,

Louis A. Picard and Thomas Mogale

195

keeping in mind that, outside of townships and homelands, these institutions were white reserves before 1994. As noted above, the revenue-generating capacity of subnational governments is limited by legislation, local poverty, and practice. That said, the problem in South Africa is less resource availability (though that is a real problem in urban slums and rural areas) than administrative systems, which limit the ability of subnational governments to implement programs. A further problem, to some critics, is the use of user fees to deliver services, which limits the extent to which social services can be delivered to the some 40 percent of South Africa’s population that is poor and unemployed. This is combined with a pattern of almost continual loss of skills at the local government level, outside of the metropolitan areas, since the late 1990s. This has led, according to one observer, to a general failure of local government in both service delivery and local development (Atkinson 2007). Further problems include poor definitions of administrative responsibility and the functions of employees (administrative, fiscal, and sector) attached to local authorities. Employees at the local level also remain weak and poorly paid. Subdistrict structures (rural local governments, areas, and wards) are mobilizing units and they have little bureaucratic support. Cities, towns, and district administrations have some administrative support (significantly in metropolitan areas), but there are little salaried staff in rural local councils. Government departments do have technical specialists available to support decentralization. The Department of Provincial and Local Government (now renamed the Department of Cooperative Governance and Traditional Affairs) is the primary vehicle providing administrative support to all subnational governments, including the newly strengthened traditional administration. After several years of ambiguity with regard to traditional authorities, the government now sees a strategic governance role for traditional leaders in the area of rural development. The Treasury Department and the several commissions that support intergovernmental relations support SNG fiscally and also are in the process of setting up regulatory and operational framework. A focus in the past two years has been on the government’s policy of 2009 to strengthen local government capacity (Republic of South Africa 2009). Capacity building is coordinated by the Public Administration Leadership and Management Academy, but delivered by a myriad of university and institute educational and training programs. These programs are the mother lode of capacity-building systems, but are only partly taken advantage of by the government of South Africa. A hidden resource for South Africa’s SNG is the army of consultants and academic advisers who staff the investigating commissions, evaluation teams, and assessment missions that seem to proliferate throughout South

196

South Africa

Africa year after year. The country has significant internal technical capacity and, most importantly, local governments have access to a wide array of technical consultants, many of whom are former government officials and academics. Often not well understood, it is these consultants who both develop policy choices, monitor implementation, and clean-up operations gone wrong or jump into the breach after local leaders are exposed in scandals by the national and regional press. Additionally, there is a sector-specific technical presence at subnational level. The nature of cadre organization is such that there are separate cadres attached to local-level departments such as water, health, agriculture, education, natural resources, or attached to central departments that are both significant in number and in place. These include social development (education, water, health) and economic development microcredit-level activities (agriculture, animal husbandry, rural production, microloans). Academics and political opposition leaders are prone to criticize South African capacity but, when compared to those found in much of the rest of Africa, they are efficient and effective. In this light, it should be remembered that capacity perceptions are in part contextual.

Political Context, Incentives, and Constraints on Decentralization A significant degree of recentralization followed the final Constitution of 1996. This represented the interests of the second president of South Africa, Mbeki, and his cabinet and administrative chiefs. The corporatist thrust of the cooperative governance policy illustrates the stated goal of national leadership and is seen on the part of the ANC as an essential component of regime maintenance. Decentralization as stability rather than democracy was, and is, the primary regime focus. National government is meant to ensure that local government operates within an enabling legislative and policy framework and is structured and capacitated in a way that best enables it to promote the development of citizens, local communities, and the nation. The national government is thus mandated to provide an overall legislative framework for local government within the general legal framework set out in the constitution. In line with constitutional precepts, South African national government has provided a framework for municipal capacity building and support of municipalities: Section 154(1) of the constitution tasks both national and provincial governments with supporting and strengthening the capacity of municipalities to manage their own affairs, exercise their powers, and perform their functions. The central government after 1996 was intended to provide administrative supervision over both provincial and local government authority.

Louis A. Picard and Thomas Mogale

197

What evolved instead was the imposition of (or attempt to impose) political will over SNGs against a backdrop of weak administrative supervision and in the context of significant (though sporadic) nepotism, patronage, and corruption (Picard 2005). It should be noted that this is less widespread in the technical areas of government as opposed to the enforcement, regulatory, and service delivery sectors. There are a number of reasons why political stakeholders and decisionmakers oppose decentralization, and there are technical and political constraints compelling the ANC that impede decentralization in South Africa. The incentives and political motivations of decisionmakers are threefold. First, there is a genuine feeling that development management is a national project and requires centralized guidance. This is a Keynesian assumption and is common throughout the developing world. Secondly, there is the memory that federalism, decentralization, and local government were tools of apartheid. From the beginning of the struggle in 1912, there was a commitment to a unitary government in South Africa. It is not lost on the ANC leadership that those who advocate federalism and decentralization come from the white, upper-middle-class business sector and tend to vote for the opposition Democratic Alliance, a political movement that is not considered legitimate by many in the ANC. Historically, decentralization has been associated with a particular brand of elites, formerly white political parties and interest groups. Despite practicing heavy-handed authoritarian and centralized government for the better part of 100 years, the National Party on the eve of its having to give up power advocated an extreme brand of decentralized government. The cynicism of this turnabout was not lost on the liberation leadership. Thirdly, there are technocratic reasons to retain central control. There are limited professional skills in the country and these skills need to be used most effectively at the national level. In the short run, these concerns for standards are reason enough for central government officials, political and administrative, to retain power at the central level. Most development planners, union leaders, and other stakeholders in the camp of the ANC tend to oppose decentralized governance. That said, there is a group of urban and regional development planners who see that they, and their constituents in local government, have much to gain from strong regional and local government. For academics, consultant practitioners, and other intellectuals who support decentralization, this support comes from what might be called noble motives such as improving democratic governance rather than instrumental, ulterior, narrow, political, or personal concerns. The problem is these advocates tend to be apolitical and, in South Africa, most are members of minority populations (Oldfield 2002). Those who stand to lose from decentralized governance represent the dominant political elite. They see their constituents as being the majority

198

South Africa

African population who have little affinity for decentralized governance and are happy to gain increased levels of social service delivery from a centralized system. This is the group within the ANC that has never been won over or outmaneuvered in the 100 years that they have led the organization. The incentives of civil society organizations with respect to decentralization are ambiguous. Strengthened local government could weaken their role in primary service delivery, a function that they have had since the 1980s. There is a fear that funding could dry up for CSOs if local government increased its capacity to perform local-level service provision. There is also concern, and sometimes friction, between nonprofit groups and private sector contractors who are better at responding to tenders and implementing government contracts. Thus, civil society and local government relations are also ambiguous. There is an element of collegial comradeship reflecting the ties from the liberation struggle. However, there can be conflicts where financial interests overlap. Nonstate local actors (e.g., NGOs, religious groups) in theory support decentralization since it is a part of the rhetoric of liberation. There are tensions, however, and the incentives of local civil society and local government do not always coincide. Nonliberatory civil society groups (with no links to the ruling ANC) tend to support decentralization initiatives more strongly than those that are a part of the movement. Subnational leadership has developed an economic and, critics would suggest, a patronage relationship to local government; political leaders are broadly representative of local populations and most local-level administrators are embedded in the community. The local government networks in South Africa represent minor patronage networks and can be economically and socially captured by local political and traditional elites. This relationship defines the nature of the support for local government in terms of projects, local-level contracts, and a search for grant money that will filter down to the grassroots level. In the end, decentralization has taken place in only some sectors, and largely through deconcentration. This is because decentralization is seen to provide stability rather than the risk of political schism that some fear devolution represents. This is reflected in the reality that South Africa is a unitary state, with a level of deconcentrated and delegated decentralization characterized by responsibilities that are distinctive, interdependent, and interrelated. With 283 municipalities in South Africa, the challenge that remains is the consolidation and building of institutional capacity for these structures. While the national sphere provides the fulcrum and strategic leadership for the other spheres, South Africa’s leaders perceive that they require a sound mechanism for the coordination of actions when implementing policy or legislation affecting the material interest of other governments. That said,

Louis A. Picard and Thomas Mogale

199

political elites perceive that the evolution of a stable and a relatively wellfunctioning intergovernmental system has the potential to promote democracy and enhance services. Administrative and fiscal decentralization, monitored and controlled by central government, is more acceptable than political devolution of authority. This leaves policy decisions in the hands of the political elite, at the national and to some extent at regional levels. This sequence is driven by the nature of the ruling political party in South Africa, the ANC. Given this reality, it is not likely that enhanced decentralization will be legislated any time soon. Opponents in the central bureaucracy do not so much resist decentralization, but rather are indifferent to it and are not likely to be overly concerned about local-level empowerment.

Intergovernmental Relations and Processes The concept of cooperative governance is crucial to the understanding of intergovernmental relations in South Africa. Section 154 of the constitution calls for the establishment of the principle of cooperative governance in South Africa’s local government relationships. In short, the national and provincial governments are called on to support and strengthen the capacity of municipalities to manage their own affairs, to exercise their powers, and to perform their functions. Draft national or provincial legislation that affects the status, institutions, powers, or functions of local government must be published for public comment. At the same time, cooperative governance has been resisted by opposition parties that see it as a mechanism to impose ANC policies on SNG. The Intergovernmental Relations Act of 2005 was enacted to put into effect the principles identified in the constitution. The act provided for setting up structures and institutions to foster intergovernmental relations and promote cooperative governance. It called for coordination of policy efforts and legislative arrangements between national, provincial, and local government, and all organs of the state within government. It promoted coherence of government and the effective provision of services, monitoring and implementation of policy and legislation, and the realization of national priorities. Accordingly, the most important areas where coherence is sought include service delivery, public accountability, coordination and integration, effective implementation, and dispute resolution. However, cooperative governance has worked to strip South Africa of the richness of dialogue, debate, and difference that is at the heart of democratic governance. This is because national priorities are the focal point of cooperative governance and reflect the philosophical legacy of prefectoralism. This was at the foundation of the supervisory mechanisms established by colonial authorities, whether Dutch, French, British, or Portuguese, including

200

South Africa

those imposed by the Union of South Africa and National Party governments. Since majority rule, there has been a neoprefectoralism in South Africa in which executive and political leaders have political and policy authority over administrators, law enforcement agents, and technical operatives in a geographic area. Magistrates on the judicial side, and provincial premiers, executive majors, and municipal managers on the political side, are at the center of the new prefectoralism that constitutes SNG in South Africa. Regarding fiscal relations, it is a fundamental assumption of the government of South Africa that SNGs have the right to secure credit, borrow, and enter into contracts and engage in the normal manner of financial management that is defined throughout the world. However, the provincial governments play a strong supervisory role in this and other local administration. In Section 139 of the constitution, the provincial government is empowered to take steps to ensure that the executive obligations of a municipality are fulfilled to achieve service delivery objectives. Section 105 of the Municipal System Act enjoins the provincial members of the Executive Committee (MEC) for local government in a province to establish mechanisms, processes, and procedures to monitor how municipalities manage their own affairs, exercise their powers, and perform their functions, and to monitor the development of local government capacity and assess the support needed by municipalities. In fulfilling this role, the MEC for local government relies heavily on fiscal analyses of the municipality and extensive data generated from municipal in-year reports and annual reports that allow quantitative monitoring and evaluation. Accordingly under Section 136 of the Municipal Finance Management Act (MFMA), if the MEC for local government in a province becomes aware that there is a serious financial problem in a municipality, they must promptly consult the mayor to determine the facts, assess the seriousness of the situation and the municipality’s response, and determine whether the situation requires an intervention in terms of Section 139. If the municipality has failed to approve its budget or any revenue-raising measures necessary to give effect to the budget, the provincial executive must intervene in the municipality in accordance with Section 26 of the MFMA. If the municipality is in serious or persistent material breach of its obligations to provide basic services or to meet its financial commitments, or admits that it is unable to meet its obligations or financial commitments, the provincial executive must intervene in the municipality in accordance with Section 139 of the constitution. The national executive must intervene (according to Section 100 of the constitution) in cases where conditions for a provincial intervention in a municipality have been met, but the provincial executive does not become involved. These are the conditions set by legislation. The procedures, both constitutional and legislative, are impressive on paper. The policy and adminis-

Louis A. Picard and Thomas Mogale

201

trative framework and procedures and the constitution are among the best in the world, according to external observers (Roberts 2002; Nkoana 2007). The problem lies with the human resource conditions in South Africa, especially at the SNG levels. Administrative weakness, bureaucratic corruption, party nepotism, and hierarchical rigidity all combine to lead to a less than optimal result in fiscal management and monitoring practices. Nonetheless, the extent to which South African fiscal systems do work, at least some of the time, should not be underestimated. In sum, the national government’s responsibilities with regard to fiscal provisions for local government include managing the system of intergovernmental relations, situating local government’s roles and responsibilities within the national tax revenue structure, and passing legislation to determine local government’s equitable share of revenue raised nationally and to address a range of other financially related topics such as municipal financial management and budgetary reform processes addressed by the MFMA (IDASA 2008). Local governments are entitled to and do receive assistance from the national and provincial governments in order for them to fulfill their constitutional functions. This includes a duty for both provincial and national governments to monitor and, where necessary, intervene in local governments where and when they fail to fulfill their functions. Initiatives such as the Masakhane (build together) tax collection efforts, Project Consolidate, and various other capacity-building programs of the Department of Cooperative Governance and Traditional Leadership and the National Treasury have been designed to remedy these deficiencies. This interdependent dynamic implies corelationships built on the foundation of cooperative governance. While each sphere is distinctive, in principle all spheres have a relationship of relative equality. However, breakdowns occur in intergovernmental relations over issues of administrative weakness, corruption, and public exposure of patronage and nepotism. These are all external to the intergovernmental system and represent the weakness of internal monitoring mechanisms. These threats mean that opposition to decentralization is self-reinforcing and leads to further resistance. Political support for decentralization, among political actors, faces an uphill battle since it is often discredited within the political ruling class. Legislation and accompanying regulations are in place to allow for further decentralization. However, the discourse on decentralization does not match the reality of where political will lies when it comes to devolved governance, and the support for decentralization does not empower actors who represent local interests to promote decentralization. Neither is decentralization legislation waiting to be implemented or to be translated into practice on the ground. It is not that national proponents lose interest in

202

South Africa

decentralization, but rather that decentralization is threatening for many within the ruling coalition.

Conclusion Efforts at decentralization were most successful at the cusp of nonracial government when issues of local government were central to the negotiations’ debate. At that point the ANC, for strategic reasons, supported strong urban local government while the ruling National Party supported a federal South Africa. The debate tainted support for decentralized governance, however, and following the election of the ANC to national government, focus turned to support for national government initiatives. At the same time, many civil society groups that supported local government development moved out of civil society and into the national government under the umbrella of the ANC. In South Africa, all sectors of society and all political parties give lip service to decentralization. However, it is clear that the ANC remains suspicious of local governance and sees it as a threat to its overall mission of economic management. The opposition Democratic Alliance, which remains a libertarian party, is a strong supporter of decentralization, as are most of the other opposition groups whose support is often regional in nature. There is more support in the white, mixed race, and Asian communities than in the majority African community. The trade unions and other pro-ANC groups share a suspicion of autonomous local government, looking to local and provincial authorities to deliver social services initiated by the national government. The business and agricultural sectors tend to support regionalism. Not surprisingly, local government leaders from all political parties support decentralization to a greater degree than national politicians. There is significant support for decentralization in the Northern Cape, Natal, and especially the Western Cape. The lessons of South Africa are significant for other African countries in several ways. First, the fact that a relatively developed South Africa should have problems implementing local governance reforms, fiscal decentralization, and professionalization of personnel should not be lost on other countries (and international donors) as they implement programs. This is not a short-term task, nor is it easy. Governance is complex and should not be considered a quick-fix need. The fact that donors are working on democratic governance in South Africa twenty years after Nelson Mandela walked out of prison is, or should be, sobering information for all who work in the local governance arena. Secondly, while two models of local governance are often discussed— democratic local government versus deconcentration (either through geo-

Louis A. Picard and Thomas Mogale

203

graphical or functional administration)—there is a third choice. This choice is between a security focus often involving an area prefect, magistrate, and police and a centralized service delivery system that uses deconcentrated departments and local governance units. It is the latter option that has been taken by South Africa’s government, a policy euphemistically labeled cooperative governance. Deconcentrated development and the recognition of the importance of traditional authorities are two practices identified in South Africa that may be applicable to other countries. The effective use of consultants is a third. South African policymakers, though they have not been able to fully implement a professionalization of the public sector especially at the subnational level, have been willing to bring in well-trained consultants to tackle policy analysis, monitoring, and assessment assignments. Some of the worst practices in South Africa also pose a warning. Though they may be less than in other African countries, the levels of patronage, nepotism, and corruption are unacceptably high. The subnational public sector is weak, poorly trained, and expensive. The public has lost its trust of the public sector, resisted efforts to collect revenue, and created a small but growing “economy of affection” in the country (Hyden 1983). The private sector and nonprofit organizations often seem to operate despite the government structures in South Africa rather than in partnership with them. South Africa, though an emerging economy and a democratic country, has many lessons that can be seen as cautionary tales for other countries. Parallels can be drawn with Nigeria, Kenya, and Ghana, all countries that have economic potential but are challenged by weak government structures. South Africa is not there yet, but may be moving that way given the fragility of its subnational political institutions. This would be unfortunate, not only for South Africa but also for the rest of the continent. These lessons are transferable because they are not unique to South Africa and are almost universal patterns of government in emerging economies. What is unique in South Africa is the continued bifurcation of governance and service delivery patterns in the so-called developed versus the underdeveloped parts of the country. Race continues to be a factor in South African governance at all levels, but particularly at subnational levels because people of different races, except for a small upper-middle-class elite, tend to be segregated geographically from each other. The most salient issues of subnational governance continue to be: first, developing and sustaining a political culture that supports democratic governance and tolerance of differences; and, second, developing a professionalized, subnational public sector. These are the two most salient issues for both democratic governance and decentralized governance. Though there is recognition of the importance of democratic institutions and fiscal auton-

204

South Africa

omy, donors and national-level planners tend to neglect the importance of a commitment to long-term education and professional training for efficient and effective intermediate and local government systems. These remain matters of unfinished business in South Africa.

10 Tanzania: Devolution Under Centralized Governance Per Tidemand and Nazar Sola

Tanzania has pursued a largely devolutionary form of decentralization for the past twenty years by transferring powers, functions, and resources to elected multifunctional local governments.1 The local government reforms have been supported on the mainland since 2000 through a Local Government Reform Program (LGRP) managed by a dedicated reform unit (the LGRP team). The LGRP originally aimed to reform the legal framework, local finances, and human resources management systems, and to enhance participation and promote good governance at the local level. In later years, the program has also included components to support sector coordination and strengthening of deconcentrated agencies. With assistance from development partners, the government introduced a development block grant for local governments in 2004, known as the Local Government Development Grant system (LGDG). In addition, several of the service sectors undertook reforms that decentralized in numerous ways—the education, health, water, roads, and agricultural sectors were transformed in the past decade with more emphasis on decentralized (although not necessarily devolved) governance. The current reforms began in 1996, within the framework of wider public service reforms and the broader liberalization of the Tanzanian economy and polity. However, local administration had undergone significant changes prior to 1996, with reasonably well-established democratic local governments in the early postcolonial period, the subsequent abolishment in 1972 of local governments, and their reintroduction in 1982. In this chapter, we focus on the reforms from 1994 to the present. Our main findings are that Tanzania has had a strong legal framework to support devolution, but this has been compromised by several factors: deconcentrated structures that counteract the autonomy of subnational governments; HRM systems that are supposed to be devolved, but are contradicted by sectoral provisions that centralize power; 205

206

Tanzania

and the centralizing impetus of the governing Chama Cha Mapinduzi (Revolutionary Party, CCM) and certain elements of the bureaucracy.

Historical Background The extent of devolution over the past fifty years can be described as a pendulum movement: the early introduction of local governments around independence swung the pendulum toward devolution.2 The pendulum swung back toward deconcentration when local governments were abolished in the 1970s, and back again toward devolution in the late 1990s. In most recent years, the government has taken steps toward recentralization, in particular with regard to HRM in local governments. The earliest experiences with elected local governments were in the late colonial period when elected councils were gradually introduced, with elected local governments strongest in the urban areas. During the colonial period, native authorities were established in the rural areas. They collected local taxes and were responsible for limited services such as primary education, sanitation, dispensaries, and village roads. The appointed district commissioner nominally controlled these authorities, but chiefs continued to exercise substantial executive and judicial powers right up to independence. In the late colonial period, the colonial administration sought to further modernize and democratize the authorities, but it was never effectively implemented in the rural areas. The gradual democratization of local administration was also tempered by the colonial state’s attempts to limit African control of these institutions by reserving European and Asian seats. Nevertheless, a basic architecture of autonomous and partially democratic local governments was largely in place at independence in 1961. Locally collected revenues from rural and urban authorities amounted to 17 percent of total public revenue in 1961. In 1962, soon after national independence was declared, the new de facto one-party state under the Tanganyika African National Union (TANU) undertook significant changes in the local administration by: 1. Replacing administrative officers who previously headed the provinces and districts with political appointees (regional and area commissioners); 2. Removing all executive and judicial powers from the traditional chiefs; and 3. Extending modern district councils throughout the country in place of the native authority councils. Philip Mawhood (1983) notes that the impact was mixed. On the one hand, rural administrations lost substantial lower-level authority through the removal of the chiefs, and the removal of more senior administrative

Per Tidemand and Nazar Sola

207

officers led to the loss of administrative capacity. On the other hand, the initial years of independence also witnessed a short period of strengthened elected local governments. In particular, local authorities in the wealthier cash crop–producing areas saw their revenue increase substantially, just as new and qualified staff members were brought on board. Central government transfers were also introduced to rural authorities. The capabilities of local authorities declined gradually in the late 1960s, primarily due to declining revenues and increasing demands for services. Local authorities’ own tax collections peaked in 1965–1966, but these powers were largely removed in 1969. In May 1972, the government published its policy of decentralization, based on strengthened deconcentrated administrations at regional and district levels. Elected local governments were abolished. The reforms transferred senior central government staff members to the deconcentrated regions and conferred ministerial status on regional commissioners. Partly elected district development committees were established at the regional and district levels, made up of former local government councillors, but these bodies lacked downward accountability, as the institutions were dominated by central government and party-appointed members. Rural development featured particularly high on the government’s agenda in this period. Villagization was a key component for the modernization of the rural areas. Urban local authorities were reintroduced in 1978 and strengthened by further legislation in 1980, though these had less autonomy than their predecessors since the region retained substantial influence and local revenueraising powers were curtailed. CCM3 included the reintroduction of local governments in its 1980 election manifesto, and in 1982 introduced a comprehensive system of local governments at the district and village levels in rural areas and at municipal and city levels in urban areas. Elected councils were empowered to enact bylaws, collect revenues, determine local budgets and plans, and so forth. These structures were given direct responsibility for service delivery in the areas of primary education, primary health, local water supply, local roads, and agriculture extension. In urban areas, they also became responsible for urban services such as solid waste removal and streetlighting. However, strong regional administrations remained; they undertook a large number of development activities directly and continued to control most local funding. Thus, elected local governments in effect had no substantial resources or effective service mandates.

Devolution and Local Government Reform After 1994 CCM included in its Election Manifesto of 1995 the promise of a substantive local government reform program. The decisive step was taken in June

208

Tanzania

1996, when the government announced it would restructure and downsize regional administration. In this period a painful civil service reform was undertaken, with substantial retrenchments and a comprehensive reorganization that created several new executive agencies, privatized several stateowned enterprises, and began a devolved decentralization. The most immediate changes were found in the deconcentrated agencies at the regional level (per the Regional Administration Act of 1997), where retrenchments were undertaken and a substantial amount of staff was transferred to local government authorities. This act abolished the implementing role of the regions, which had then been primarily responsible for local development projects in key sectors such as education, health, roads, water, and agriculture. For instance, some regions had water departments with over 100 staff reduced to 1 engineer!4 A government policy paper in October 1998 outlined the vision of a future reformed public service at national and local levels, and spelled out how decentralization of government would take shape in four main policy areas. The first three were political devolution to create a holistic local government system, financial decentralization to local government councils, and administrative decentralization of personnel to local councils. The fourth was the principle that the central government would have overriding constitutional powers, with the minister responsible for local government coordinating central-local relations and sectoral initiatives relating to local governments.

Authority: Legal Framework for Local Government The existence of local governments is entrenched in the constitution, which holds that “there shall be established local government authorities in each region, district, urban area and village in the United Republic” (Chapter 8, Section 145). Decentralized bodies in Tanzania now include twenty-one deconcentrated regional administrations. Below these is the higher level of elected local government, with local governments being cities, municipalities, towns, or districts, depending on population size. Villages serve as lower-level local governments in rural areas, and at the most local level are mitaa (subvillages) and vitongoji (streets). Central government ministries of relevance to local governments are of two types: sector or line ministries (e.g., health, education) and intersectoral ministries (e.g., finance, regional and local government.) Theoretically, the 1982 legislation made LGAs responsible for delivery of core services such as primary education, health, and local roads. After local government reform in 1998, the sector ministries were nominally limited to providing guidance on sectoral issues, setting relevant sector policies

Per Tidemand and Nazar Sola

209

and guidelines, determining sector-specific service delivery standards, and monitoring performance. The Prime Minister’s Office–Regional Administration and Local Government (PMO-RALG) was supposed to coordinate and support the reform process. However, in reality, sector ministries have remained heavily involved at all levels in these sectors while the intersectoral or coordinating ministries have exercised tight control over the allocation and usage of local government inputs, especially fiscal and human resources. The twenty-one deconcentrated regional secretariats are headed by centrally appointed regional administrative secretaries. The sector ministries (e.g., health, education) are represented in the regional secretariats. These secretariats retain certain oversight and monitoring responsibilities regarding LGAs. Under the Local Government Laws of 1982 and the Local Goverment Reform Policy of 1998, LGAs have wide-ranging, but vaguely formulated, functions. These include: maintenance of peace, order, and good government; promoting social welfare, economic well-being, socioeconomic development, agriculture, trade, commerce, and industry; providing health and education, and fighting poverty (all from 1982) and the provision of basic public services, with a special emphasis on sectors that target poverty reduction (from 1998). The latter include primary education, primary health, agriculture extension services, local water supply, and roads. The Local Government Act is even more vague when it comes to primary-level local governments such as villages; Section 142 outlines the general functions of villages as acts necessary for “economic and social development,” “welfare and well-being,” and planning and coordinating village-level activities and partnerships with other villages. Table 10.1 outlines the division of tasks and responsibilities between LGAs, central government, and other stakeholders within key sectors. In summary, the decentralization policy framework for Tanzania has a reasonably clear assignment of major functions to local governments. The key questions therefore relate to the practical implementation of the reforms. Limitations on Authority

For more than ten years, the Tanzanian LGRP has sought to implement devolution. The reform program claims to transfer authority to LGAs, to increase their relative autonomy, to enhance both downward (to local residents) and upward accountability (to central government or Parliament), and finally to enhance the overall capacity of LGAs and a national system for decentralized governance. Progress in implementing reform elements has been uneven and can best be understood by a discussion of its individual key elements.

210

Tanzania

Table 10.1 Division of Responsibilities According to Local Government and Sector Legislation Service

Main Provider

Comments and Legal Issues

Primary education

LGAs

• Decentralized by Education Act (2004), but there are parallel centralized procedures for management of teachers (Teachers Service Commission).

Secondary education Central government

• LGAs play a role in construction of secondary schools.

Primary and preventive health

LGAs

• Unclear roles of standing LGA committees versus decentralized facilities.

Hospitals

LGAs • National Health Service Bill (2004) states LGAs (district hospitals) are responsible for all health facilities up to district hospitals, but health boards operate in parallel to LGA structures.

Water supply, sewerage, and sanitation

Urban areas: autonomous authorities; Rural areas: mainly LGAs

• Capital investments largely managed by central government. • Water boards and water users’ associations manage water supplies parallel to LGAs. • Regional consultancy units in parallel to the regional administration to support LGAs.

Solid waste

LGAs

• Local governments exhibit capacity problems.

Roads

LGAs

• Problems with financing and technical capacities. • Legal issues with drafting of Roads Act.

Agricultural extension

LGAs

• Slow transfer of extension staff to LGAs. • LGA capacity is limited by the unresolved division of work between the private and public sector. • Privatization and use of public funds managed through farmers’ groups raise issues regarding legal basis for procurement and financial management.

Sources: Based on Steffensen, Tidemand, and Mwaipope (2004); Tidemand, Olsen, and Sola (2007). Note: LGAs are local government authorities.

Implementation of Legal Reforms

In the late 1990s, and especially with the 1997 Regional Act, substantive retrenchments were undertaken at the deconcentrated regional level and many staff were transferred to LGAs. The regions were no longer supposed to play a major role in the implementation of capital projects and direct delivery of services, but rather to facilitate and guide local authorities in doing so. In 1999, local government legislation provided an enabling

Per Tidemand and Nazar Sola

211

framework for decentralization. The LGRP sought initially to undertake the reforms in a geographically phased manner, starting with approximately a third of the LGAs. Two sets of regulations were subsequently issued to enable greater fiscal and human resource autonomy in the first group of thirty-eight councils, as discussed below. Yet little progress has been made toward further legal empowerment of LGAs. Since 2000, the LGRP has sought to harmonize key sector laws with local government legislation. Studies have also been undertaken of the regional administration and the local government legislation that recommended amendments for furthering the devolution process (see Shivji 2006). However, virtually none of the recommendations has been implemented to date. On the contrary, some new laws and amendments have weakened the LGR process. For instance the Public Service Act (discussed below) sought to create a unified public service and, in the process, centralized the legal arrangements for local government staff. Amendments to the local government laws in 2006 (Act No. 8 of 2006) strengthened the role of central government by requiring LGAs to consult with district commissioners and by central government appointment of up to three councillors in each LGA. Further recentralization was avoided primarily due to lobbying from the LGRP team. Implementation of Human Resources Management Reforms

At the onset of reforms in 1998, the human resources system at the local level was both fragmented and centrally controlled by six different appointing authorities. The LGRP (PMO-RALG 1998: vii) sought a radical change regarding decentralized local government personnel management, as it states: “The councils (city, municipal, town and district) will be fully responsible for planning, recruiting, rewarding, promoting, disciplining, development, and firing of (all) their personnel.” The policy anticipated an immediate and radical transformation in which each individual LGA would employ its entire staff. Only the council director (the equivalent of a county or city manager) “in the interim may be posted by (Central) Government.” Subsequent regulations issued in late 2000 (for the first batch of thirty-eight reforming councils) held that the council director was to be recruited and appointed by the minister, but recruitment and appointment of other staff was largely decentralized to the councils (PMO-RALG 1998: vii). The selected thirty-eight LGAs started to work according to these regulations while remaining LGAs worked under the centralized HRM system. In 2002, a new Public Service Act created a unified public service. The act stipulates that staff management is to be decentralized to the deconcentrated regions in addition to devolved LGAs. The 2002 Act left the President’s Office–Public Service Management (PO-PSM) in charge of staff

212

Tanzania

management for local government. In 2003, PO-PSM issued regulations that defined the council director as the appointing authority and gave the central government the power to transfer staff across ministries, regions, and LGAs. This maintains a dual system of HRM in local governments where central government at its discretion can overrule local management. Furthermore, the health and education sectors have explicitly been exempted from decentralized recruitment procedures; large numbers of health staff and teachers have in recent years been centrally deployed to LGAs. Other aspects of the LGA employer functions have not been reformed. In particular, budgets, salary schedules, and career paths remain centralized. Budget transparency has remained unchanged. Staff salaries are almost entirely paid from central government transfers where the central government allocates funds according to filled posts rather than based on formulas. The reform policy does not explicitly address this issue and the introduction of formula based grants as proposed within the finance component of the LGRP (see below) has evidently been made without agreement between the Ministry of Finance and PO-PSM. Specifically, it has been made clear during implementation that the developed formula-based grants will not apply to salary components. Budget control has likewise remained entirely centralized; local governments are consulted during restructuring exercises, but all decisions on staff budgets and numbers of approved staff are ultimately done by PO-PSM. Career management has been partially devolved, but career progress for senior staff continues to depend on their ability to move into central government ministries at some point in their careers. Pay policy remains centralized, except that LGAs are allowed to establish local incentive schemes. In practice, LGAs are not able to afford this—except for select staff categories in the wealthier LGAs. Actual HRM practices in LGAs thus have been a mix of decentralized staff management and centralized transfers and postings. Consequently, LGA staffs have dual allegiances; they have to satisfy both local and central government. Furthermore, senior staff members are aware that their career prospects depend largely on the satisfaction of the latter. Several LGAs have invested in capacity building of their staff, then subsequently seen staff transferred to other LGAs or to the central government. This frustrates local capacity-building efforts otherwise encouraged by the new system of providing LGAs with grants for this. Transfers are undertaken without much consultation with LGAs and with late replacements of staff. This is perceived by LGAs as the most frustrating aspect of current practices.5 Available data does not allow for a strict comparison of effectiveness of centrally deployed staff compared to local recruitment. However, both regional and district officials argue that teachers who were locally recruited by LGAs were far more likely to continue work within their posting areas

Per Tidemand and Nazar Sola

213

(especially in the rural areas) than teachers who were centrally deployed by the central government and sent to particular districts.6 In addition, central deployment of staff has generally failed to address geographic inequalities of staffing levels in local governments. There remain significant and persistent problems in attraction and retention of staff (especially, senior staff) in districts considered remote or marginalized (PMO-RALG 2005a). Urban LGAs tend to be better staffed than rural and remote LGAs, even for agricultural extension staff, which suggests the current system of staff allocations is neither transparent nor need based (see Tidemand, Olsen, and Sola 2007). The LGRP has tried to rectify this by having a more rational allocation of personal emoluments through a formula-based recurrent grant system. However, the impact of these changes has been limited as the government procedures for approval of individual LGA staff structures remains highly centralized. In practice, money follows staff rather than the other way around. This limits the impact of the reform of the recurrent grant system. Reviews of the LGRP in 2001 and 2004 found that restructuring exercises in LGAs are prepared without a hard budget constraint and without substantial LGA incentives for retrenchments.7 There appear to be no significant efforts to reform the current system.

Fiscal Reforms and the Question of Autonomy The LGRP aimed to strengthen local government finance in three main areas: the introduction of formula-based fiscal transfers; strengthening local governments’ own revenue collections through reformed local tax systems; and improving local financial management. The stated objectives of the local government fiscal reforms were to ensure that LGAs have adequate funding to deliver services, that they are granted autonomy in budget allocation, and that they use financial resources prudently. In addition, key sectors, such as education and health, have increased and transformed levels of financing at local level. However, the conclusion regarding fiscal autonomy is that LGAs rely heavily on central government transfers that are highly earmarked, leaving little true fiscal autonomy. Reform of Fiscal Transfers: Toward Greater Autonomy?

Beginning in 2000, reform sought to enable the first thirty-eight devolved LGAs to receive some block grants rather than earmarked subventions. This strategy was never realized; all LGAs were in practice made to follow the same national regulations for LGA finance. Some efforts were made to introduce grants based on national minimum standards, but standards were

214

Tanzania

set so high that they proved too expensive to apply nationally and subsequently fiscal transfers remained unchanged. After 2004 more reforms of central government fiscal transfers to LGAs were undertaken, with the establishment of formulas for fiscal transfers to LGAs for recurrent expenditures in the areas of education, health, roads, agriculture, water, and administration. The introduction of the reformed fiscal transfers would share fiscal resources more transparently and fairly by implementing a need-based formula and enhancing LGAs’ autonomy. The results of the reform have been modest. Although formulabased allocations in principle were agreed to for all the sectors in 2004 and endorsed by the cabinet, in practice they have not been applied.8 Recurrent transfers are predominantly composed of personal emoluments and, since staff recruitment and deployment essentially remain centralized functions, it has not been possible to apply the formula-based allocations. As a consequence, financial allocations to LGAs are unequal and not transparent. Allocations in education range from TSh5,000 to TSh20,000 (5,000 Tanzanian shillings to 20,000 Tanzanian shillings) per capita. Most of the recurrent fiscal transfers are earmarked personal emoluments (approximately 88 percent),9 and LGAs have no autonomy regarding reallocation of these funds toward other charges. The reforms have been more successful in transforming capital transfers for development. Up to 2004, the development grants to LGAs were minuscule.10 Most of the development funds in LGAs were provided through discrete donor-funded projects, mainly through various area-based programs, but also in the form of some sector support programs. In 2004, the government and international development partners agreed on establishing the Local Government Development Grant. Under this arrangement, all LGAs receive a discretionary development grant of approximately US$1.50 per capita (a total of some TSh50 billion) if they fulfill basic minimum conditions regarding the quality of their development plans, financial management, degree of local transparency, and procurement systems. In addition to the core LGDG funding, various sectors have started to transfer funds along similar principles as the LGDG system. Table 10.2 provides an overview of the gradual introduction of the LGDG system and the increase of sectorspecific allocations. In addition to formula-based development transfers, LGAs continue to receive project financing (e.g., from bilateral donors) as well as funds from sector ministries. Because of the classification of these funds in the government budget and accounts, it is difficult to establish exactly how much is spent at LGA level. To some extent, the LGDG has been successful in introducing a system for development funding that provides LGAs with more autonomy. However, two key issues are yet to be addressed. First, the Prime Minister’s

215

Per Tidemand and Nazar Sola

Table 10.2 LGDG Grant Allocation FY 2004/05 to FY 2011/12 (budgeted amounts, billions of Tanzanian shillings) Fiscal Year

Core LGDG

2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12

6.0 67.1 76.8 71.4 78.4 118.6 175.2 208.4

Agriculture Rural Water LGDG LGDG 0.0 0.0 4.7 51.4 54.4 53.3 55.6 42.1

Health Sector LGDG

Other Sectors

0.0 0.0 0.0 0.0 0.0 29.2 20.0 10.8

0.0 0.0 0.0 3.6 5.0 18.0 0.0 0.0

0.0 0.0 0.0 69.6 62.4 65.6 63.2 142.2

Total LGDG System 6.0 67.1 81.5 196.2 200.2 284.6 314.1 403.4

Sources: Unpublished PMO-RALG data provided to the authors. Notes: This table reflects budgeted amounts for each fiscal year. It is not easy to consistently ascertain the amounts that were actually disbursed from year to year. In certain years, the disbursement of certain sectoral grant windows was considerably less than the budgeted amount. Incomplete and delayed disbursement of sectoral grant windows seems to have been particularly common during the first year(s) of a new grant window. LGDG is Local Government Development Grant system.

Office has issued various instructions to LGAs that have effectively undermined their autonomy by directing financial resources to secondary education only. Second, the LGDG system remains largely donor funded, rather than being fully integrated into the national budget and funded through government resources. Although LGA budgets have increased in absolute terms, their relative share of public expenditure has remained relatively stable. The recent increase in FY 2006/07 (see Table 10.3) is explained by an increase in the number of teachers and their salaries, which constitute a large share of LGAs’ recurrent budgets. LGAs’ share of development funding is somewhat lower—it was estimated in 2007 as 17 percent of total development budget (PMO-RALG 2007).

Table 10.3 Local Government Authorities’ Share of Recurrent Budget Fiscal Year 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07

Total Recurrent Expenditure (trillions of Tanzanian shillings) 1,253.1 1,527.8 1,834.1 2,252.3 2,875.6 3,142.3

Local Government Share (percentage) 18.7 19.0 17.7 17.0 18.6 24.3

Source: PMO-RALG 2007. Note: Data has not been analyzed or published by PMO-RALG since 2007.

216

Tanzania

LGAs’ Own Revenue Mobilization

LGAs collect some TSh60 billion from local taxes (mainly service and production levies rather than, e.g., property taxes). This represents only 7 percent of total LGA expenditures and thus indicates a high reliance on central government fiscal transfers. Revenue in rural LGAs has declined over recent years whereas urban LGAs have witnessed some growth (see Table 10.4). Urban LGAs today collect more than four times as much as rural LGAs, and it is also in urban LGAs that the most significant potential for further growth is found. The main reasons for decline in revenue collections are the abolishment of a range of nuisance taxes in 2004, inappropriate tax designs, and poor collection systems.11 LGAs’ Expenditure Patterns

Recent data reveals several tendencies in local government spending (see PMO-RALG 2007). First, a large share (78.5 percent) of local spending is recurrent spending. Second, most recurrent spending (56.6 percent of all local spending, or almost three-quarters of local recurrent spending) is spent on personal emoluments. Third, spending is heavily concentrated within just two sectors: three-quarters of recurrent spending and two-thirds of all local spending is aimed at primary education and basic health services. It is clear that these local expenditure trends are driven primarily by the nature of the intergovernmental fiscal transfer system, rather than by local priorities and choices. This limits the spending discretion of local authorities between sectors and spending categories. The LGDG was intended to provide discretionary development funding to LGAs, but there is evidence of increasing central government interference in local prioritization. For example, an increasing share of the LGDG budgets is now funding secondary classroom construction since LGAs have received instructions to do so by the prime minster, regional commissioners, and district commissioners since 2005—in spite of the fact that secondary education in not a legal mandate of LGAs.12 The conclusion regarding autonomy is thus clear: local governments since 2000 have not been significantly empowered as autonomous institutions. The central government maintains significant and, in some areas, increasing con-

Table 10.4 Trends in Local Government Authorities’ Own Revenue Collections in Tanzania (billions of Tanzanian shillings)

Urban Rural

2001

2002

2003

2005

2006

2007

TSh per capita, 2007 (approx.)

23.113 28.086

25.569 22.774

28.656 29.083

23.728 19.142

28.139 21.151

36.271 27.113

4,830 1,050

Source: Unpublished PMO-RALG data provided to the authors.

Per Tidemand and Nazar Sola

217

trol over local government staff and budgets—the extent to which local governments can make budget or staffing decisions has not changed significantly. On the contrary, legal reforms have constrained local autonomy in human resources while fiscal realities have limited local fiscal autonomy.

Citizen Participation and Accountability Local government reforms have affected accountability in several areas: (1) Electoral participation; (2) direct participation by participatory planning and participation in meetings; and (3) financial accountability of LGAs, as measured by audits. Electoral Participation

Elections for local government councils are held in distinctly different ways for the higher-level councils (e.g., districts and municipalities) and the lower-level councils (villages, vitongoji, and mitaa, etc.), respectively.13 The elections for village councils (vitongoji and mitaa)—often referred to as “grassroots elections”—are managed by the PMO-RALG and the respective district council directors throughout the country. More than 554,790 posts were filled in 2004 (village council members, vitongoji chairpersons, mitaa chairpersons, and council members). However, the elections for the district (as well as municipal and city) councillors are held simultaneously with the national elections for parliament and the president, and are supervised by the National Electoral Commission. The latest grassroots elections were held in November 2009. They were the fourth rounds of elections under multiparty democracy following the 1994, 1999, and 2004 elections. The national and district council elections are held approximately one year later than the grassroots elections. Survey data indicate some improvements in electoral participation, but voter registration for grassroots elections in previous years has reportedly been so poor that it “it is difficult to ascertain how many people registered and voted” (Chaligha 2008: 54). However, detailed studies of the grassroots elections in ten councils undertaken by REDET indicated that voter turnout was 70 percent. The ruling party dominates the grassroots elections even more than other elections (see Table 10.5). The substantive dominance by CCM in grassroots elections, compared to parliamentary or presidential elections, indicates that the electoral system may overly favor the ruling party since only CCM has the required machinery for organizing this large number of candidates. A large number of seats go uncontested, with only the CCM forwarding candidates. In Tanzania, only party members can stand for elections to parliament and district and municipal councils as well as for the posts in villages, mitaa, and subvillages.

218

Tanzania

Table 10.5 Dominance of Chama Cha Mapinduzi (CCM) in Elections 1994–1995

1999–2000

2004–2005

61.8 80.2

71.7 87.5

80.3 88.8 92.8 97.2

Presidential election (percent of votes) Parliament (percent of seats) Local council (municipal and district) Grassroots elections (percent of seats)

94.6

Source: National Electoral Committee, www.nec.or.tz; Chaligha (2008). Summary calculations of the local council election data 2005 were kindly provided by Ole Therkildsen and Torsten Geeland at DIIS.

Direct Participation

The REPOA local government surveys provide a fairly consistent picture of increased citizen participation in local affairs, whether through local government institutions at the grassroots level or through other community institutions. As shown in Table 10.6, citizen involvement has also increased by participation in various sector-specific user committees such as school committees, water committees, public work committees, and farmers’ associations. The most convincing indicator of increased participation in local government affairs may be the number of respondents reporting involvement in preparation of village or ward plans (REPOA 2007).14 It appears from the 2003 and 2006 surveys that the increase in citizen participation is due in particular to increased participation by women and youth.

Table 10.6 Indicators of Community Participation, 2003 and 2006 (percentage) Respondents Who Reported That They or a Household Member Was Involved in: Member of village/ward leadership Participation in full council meetings School committee member Water management committee Preparation of village/ward plans TASAF project committee Public works committee Primary cooperatives/society/farmers association Agricultural/livestock extension contact group Source: Tidemand and Msami (2010). Note: TASAF is Tanzanian Social Action Fund.

Total 2003

Total 2006

17.3 24.2 28.2 13.3 19.7 1.9 8.8 8.7 2.9

22.9 28.1 35.8 23.2 35.0 13.7 19.1 12.1 6.4

Change Between the Two Surveys 32 16 27 74 78 621 117 39 121

Per Tidemand and Nazar Sola

219

Financial Accountability of LGAs

A key indicator of financial accountability in LGAs is the annual report from National Audit Office/Controller and Auditor General. This auditing of subnational accounts can be seen as one of the key mechanisms of upward accountability by which the center exercises control over local governments. Since 1999 the number of LGAs with adverse audit opinions has fallen sharply from 45 percent to 0 percent in the latest audits. This is a significant indicator of strengthened financial management capabilities in LGAs, though local accountability will require stronger citizen involvement in the scrutiny of budgets and accounts. The Verdict on Accountability

Regarding local-level accountability and participation, the picture is more complex, but we can summarize the main trends. There is significant electoral participation in lower-level (grassroots) LGA elections, but also limited competition since CCM is so dominant. There too is evidence of an increase in direct citizen participation in local institutions, but increases have been more significant in user groups than in local government institutions, which suggests deconcentration in sectoral reforms (in particular, education) has advanced more than strengthening of local government. Significant citizen involvement has occurred in preparation of village and ward plans, but there is limited evidence on whether this has translated into actual district plans and budgets, which appear to reflect national rather than local priorities. There are problems with regard to village and mitaa adherence to the required number of meetings and quorums. There is evidence of increasing trust in local government institutions, local government reform initiatives, and LGA political leaders—especially at lowest levels—but lack of trust in local government council staff. There is increasing citizen access to information. However, only a small minority (13 percent) has seen a local government budget posted. Finally, there is a positive trend in the quality of LGA accounts and a perceived decrease of corruption in LGAs.

Political Economy The political economy of decentralization reforms in Tanzania (i.e., the incentives and motivations of key stakeholders) can be examined by looking at each of several main actors in the political system: CCM and the president and executive branch, the bureaucracy and its subdivisions, and civil society.

220

Tanzania

The Political Economy of Policymaking

Since its independence, Tanzania has pursued a distinct path of nation building, which—with the exception of Zanzibar—has resulted in a peaceful society, political stability, and relatively sound macroeconomic performance. Since the introduction of multiparty politics in 1994, Tanzania has significantly strengthened democratic rule and domestic accountability. However, the political and economic transitions have been incomplete with a continuation of insufficient structural accountability and responsible governance. It is generally recognized that power is primarily exercised by the leading elements of CCM and the executive with only weak parliamentary oversight. The opposition parties are a small minority (except in Zanzibar) and have had decreasing representation since 1995. It is also clear that party and state structures remain closely intertwined. At the same time, CCM is influenced by public opinion. Recent work by Anne Mette Kjær and Ole Therkildsen (2013) has analyzed the decisions regarding landmark policies and concluded that, although CCM and the executive ultimately determine policies, many key decisions have been significantly influenced by debates stimulated by party competitive elections since 1994. In the introduction of universal primary education (2001), the abolishment of the development levy in 2003 (comparable to a rural personal income tax), and the introduction of a fertilizer subsidy (gradually in the period from 2003 to 2009), the presidential and executive branch made decisions mainly because of broad popular support for the policies and their assessment of the political opposition. CCM and the President and Executive Branch

The reintroduction of local governments in the 1980s was based on internal discussions in the CCM, yet there is some evidence that the executive branch resisted the devolution being pushed by some elements of the party. The clearest examples of executive resistance were the amendments that strengthened central oversight through deconcentrated regional administration. Beyond this, two other instances stand out. The first is the abolishment of several local government taxes around 2004. This was announced by the Ministry of Finance, apparently without consultation with PMO-RALG. As pointed out by Kjær and Therkildsen (2013), the driving forces for the decision were found within CCM, which feared that the opposition would make use of the unpopularity of these taxes. The second is the aforementioned resistance to devolution of local government staff: the LGRP is clear in outlining a complete devolution of local government staff but, as discussed earlier, in practice the reforms have not been implemented. Interviews with the prime minister in 2007 indicated that the CCM leadership found it unre-

Per Tidemand and Nazar Sola

221

alistic to have each LGA employ its own staff to any significant degree; the prime minister was also largely unaware of several of the piloted aspects of decentralized HRM (e.g., the establishment of local employment boards in LGAs).15 However, in general the president and executive branch have not been much involved in the operational aspects of the local government reforms. For instance, there is no evidence of their involvement in LGRP technical design or the introduction of the LGDG. PMO-RALG staff made such decisions in close dialogue with development partners. The Bureaucracy and Its Subdivisions

The government bureaucracy has not articulated a common stand on the local government reforms. On the one hand, devolution is official government policy; on the other hand, it is just one of many, often contradictory, reform initiatives—and has been challenged by some parts of the bureaucracy. The ministry responsible for local governments has been the lead institution in advocating for the reforms. The office has for some years been under the President’s Office and in recent years the Prime Minister’s Office (PMORALG). However, the permanent PMO-RALG staff and the contracted LGRP team have not always had common positions on policy reform issues. In general, the LGRP leadership team has pursued a strict constructionist interpretation of the LGRP and (particularly in the early years of reform) worked on a fairly narrow set of issues while PMO-RALG as a whole has supposedly worked on several broader aspects of the local development agenda. The LGRP team’s main objective was to operationalize the rather ambitious local government reform policy whereas the PMO-RALG had to implement several wider and often contradictory local development agendas. The Prime Minister’s Office was given the mandate to manage several sector development programs with significant central government control that only partially dovetailed with LGRP ambitions of building semiautonomous local government institutions. Another source of tension between LGRP and PMO-RALG arose from the continued use of transfers of LGA staff by PMO-RALG in spite of the declared intention to leave HRM in local control. The PO-PSM has since 1999 pursued a reform strategy with a focus on the devolution of staff to the LGAs. Rather than work jointly with PMORALG on how to address the issue and reconcile conflicting policy objectives, PO-PSM has pursued its reform policy in relative isolation. The LGRP tried to include PO-PSM in a consultative task force after 2001, but PO-PSM showed limited interest and the task force failed to significantly influence policy. Since PO-PSM in the end had much more clout, its interpretation of the policies was reflected in the final legislation governing public servants.

222

Tanzania

The Ministry of Finance has been included more actively in the LGRP since 2001, and around 2005 some progress apparently had been achieved with the introduction of a formula-based grant system. The ministry also worked actively to support capacity building for LGA financial management. However, the ministry ultimately resisted implementation of the formula-based grant system and much of the PO-PSM increasingly resisted elements of the reform that would grant local governments increased autonomy. The sector ministries have interpreted decentralization within their sectors in various ways. The result has often been a mix of devolution and various forms of deconcentration. Thus, all major key sectors (education, heath, agriculture, water, and roads as well as forestry and fishery) have administratively transferred their staff to district and municipal local governments. However, certain areas of activity are deconcentrated de facto. For instance, most of the funding for key sectors is in the form of various conditional grants where LGAs have limited autonomy for determining local priorities. The recurrent grants are mainly earmarked payment of staff salaries with limited or no local discretion in decisions on personnel budgets and conditions. The development funds are frequently in the form of project-based financial support, although increasing funds are also channeled through the LGDG. Another area where departments pursue sectorspecific decentralization strategies has been in relation to the introduction of local-level institutional arrangements (e.g., school committees, health facility management committees, beach management units, wildlife management authorities). These become problematic when their relationship to local government structures (e.g., the village government) is unclear or competitive. Civil Society

In recent years, CSOs have engaged in advocacy work related to government reforms. Most of the CSO policy and advocacy work is coordinated through a policy forum that has established a dedicated working group on local government reforms.16 This group has sought to interact with the government to influence the local government reforms, and it has been formally consulted as part of wider stakeholder consultation regarding the LGRP. The group has been particularly active in the debate related to the Constituency Development Fund (later renamed the Constituency Development Catalyst Fund); it undertook analytical work on the proposed bill and dedicated a website to debate its relevance and implications. The position of CSOs has been clear (see Mshana 2008): the best way to provide additional development funds at the local level is through increased LGDG transfers and a strengthening of the local government planning and service delivery

Per Tidemand and Nazar Sola

223

system. CSOs argue that the introduction of the Constituency Development Fund would jeopardize local planning and budgeting and confuse the roles of members of parliament. Rather than work on national policy issues, the attention of the MPs would be directed toward local service delivery issues that in principle are the responsibilities of local governments. NGOs rightly saw the MPs’ interest in the fund as closely related to the commercialization of politics. Despite CSO objections, the Constituency Development Fund was passed into law in 2009. In addition to these national-level activities, several NGOs have recently emphasized initiatives for local-level policy advocacy and accountability work, by disseminating popular versions of government policies and through public expenditure tracking that seek to empower communities and enable them to hold their local leaders accountable. However, such approaches have significant limitations since outreach is limited to the few participating NGO and CSO members, and the identification of accountability problems does not lead to continuous mass action. In addition, some of these exercises become small projects with their own life, primarily driven by NGO members’ access to allowances.

Conclusion Our analysis concludes that LGAs have been strengthened in the form of increased staff, skills development, and some increased resources. However, their relative autonomy has not significantly changed in the past ten years despite a deliberate official policy of decentralization by devolution. To further analyze the politics of reform, it is useful to distinguish between areas where the reforms have progressed and areas where they have had limited progress or even led to centralization. Table 10.7 offers an overview of progress across several issue areas. Several conclusions thus emerge from our examination of decentralization in Tanzania. 1. The local government structures: Like many other anglophone African countries, Tanzania has a distinct system of local governance for urban and rural areas reflecting the British model of local government. The benefits of this are the supposed local responsiveness to the specific urban or rural nature of the locality. However, with rapidly increasing urbanization, there is some evidence that the system also has become more problematic. Thus, between 2005 and 2010, a total of almost 100 townships were created—these supposedly will later be turned into full town councils with corresponding administrative cost increases. Another problem is that some of the remaining rural districts will be too small to survive when their urban

224 Table 10.7 Assessment of Relative Progress of Devolution in Tanzania

Issue Area

Significant Progress

Legal reforms

Finance reform

Improved financial management in local governments. Introduction of LGDG block grants for development.

Human resources

Increased number of staff and qualification of staff in LGAs.

Local accountability and participation

Increase in citizen knowledge of local government and increased participation.

Sector and reform coordination

Moderate or No Progress

Backlash Against Devolution or Centralization

Continued official commitment to the stated policy of decentralization of devolution.

No introduction of major local government decentralization aspects in the constitution, no major amendment of local government legislation, continued conflicts between sector legislation and local government legislation.

Development of local government own-source revenues. Abolishment of several taxes.

No adherence to formula-based allocations for personal emoluments (recurrent budget). Abolition of several local government taxes without alternative tax reforms. No progress in devolution of HRM autonomy. Public Service Act centralizes certain functions.

No reform of village and mitaa elections— low electoral attendance and little effective competition. Weak or no effective follow-up on participatory plans.

Significant increase in various sector-specific local committees with unclear relations to overall local government structure.

Reform coordination included as special component of LGRP.

No effective legal harmonization. Introduction of the Constituency Development Fund.

Note: LGDGs are Local Government Development Grant systems. LGAs are local government authorities.

Per Tidemand and Nazar Sola

225

component is carved out as an independent unit. Planning rural-urban linkages may also become more difficult. That said, the abolition of city and town government in 1972 (and the bad taste that it left with many policy elites) may make an integrated town and countryside council unlikely. A special feature of the Tanzanian system is the hierarchy of institutions that forms the districts down to the wards, villages, and vitongoji. The villages are constituted as full local governments while still being community-level structures with a population of approximately 2,000–3,000 people. 2. Legal framework for decentralization of functions: There are at least twenty major ongoing reforms in Tanzania, and each has its own policy paper.17 The LGRP attempted over ten years to consolidate scattered local government legislation into one comprehensive document and to pursue legal harmonization—but with limited progress. Other countries (most notably, Uganda) have a clear outline of the local government system and reforms in the constitution. Such an authoritative document is useful when many different interests and policies have to be reconciled. In the absence of such a legal provision in the constitution or an otherwise strong document, it has proved difficult in Tanzania to achieve the desired legal harmonization. 3. LGA finance: The financial resources of the LGAs have increased significantly over the past ten years. The increased funding has been accompanied by increasing capacity in financial management and for fiscal discipline. This is a significant achievement. Central government financing has also moved toward a system of formula-based grants (like those used in South Africa) that can give LGAs some autonomy in planning and budgeting. The LGDG is a particularly good practice: it is entirely nonearmarked and provides LGAs with incentives to improve their performance. In this manner it has helped improve local-level financial management. However, the system also has a weakness: the intended reform of LGA recurrent financing has failed because rules regarding devolution of personnel management functions were never implemented. LGAs are also increasingly dependent on fiscal transfers rather than own-source revenue. Although one should not overstate this as a problem (since it is impossible to raise a majority of funds locally), it is important to ensure that LGAs can adjust services at the margin by raising some local taxes. In rural areas, this has primarily happened through various user fees and contributions that are not well captured in local statistics, but play a significant role. 4. Human resources management: This has proved to be perhaps the most important obstacle to progress in decentralization by devolution and has proved central to the debate about local government in developing countries. It is an overlooked, but critical, issue in many countries, and the vision of decentralization of personnel in Tanzania is quite radical. Yet in

226

Tanzania

hindsight, it appears that the idea of a decentralized personnel system was never fully understood or shared by the key Tanzanian policymakers, in particular the senior leadership of CCM. A more stepwise and pragmatic approach to reform with initial emphasis on practical arrangements (and a concern for equitable conditions of oversight of local staff rather than total LGA control of personnel management) might have been more effective. 5. Downward accountability and participation: Tanzanian reforms have been accompanied by efforts to increase local-level participation and accountability. Participatory planning during the early days of reforms was supported by multiple approaches for citizen engagement. From around 2001, PMO-RALG successfully harmonized the various approaches into one common strategy for participatory planning. To some extent, this came to be regarded as good practice. However, the approach was later criticized for being overly expensive, not sufficiently linked to budget allocations, and overly focused on the easy social sectors with insufficient attention paid to productive sectors. Part of the initial problem was participatory planning producing wish lists without adequate connection to budget processes; this was overcome when the LGDG was introduced, but other weaknesses remained. NGOs have in recent years tried to strengthen citizen follow-up and monitoring (e.g., efforts for expenditure tracking). The results from these exercises are yet to be fully documented, but preliminary reports indicate that rapid scale-up of public expenditure tracking by civil society tends to weaken their quality and that effective local accountability continues to be constrained by the limited autonomy of the LGAs themselves. 6. Party politics and decentralization: Tanzania has been characterized by a stable political environment. However, the stability is largely based on the hegemony of the ruling (and one-time sole) party, CCM. As discussed above, there is some evidence that the local elections at the lowest levels (village, mitaa, and vitongoji) inhibits independent candidates and overly favors CCM, and that this local power base is subsequently used for electoral mobilization in national and presidential elections. There is also evidence that the elected local government structures are less popular than supposedly more neutral project committees. If independent candidates were allowed to contest local government elections, it is possible that local government structures would take on a relatively more local character and possibly serve as a broader alternative power base for those opposing CCM. 7. Reform coordination and local government oversight: A final key lesson relates to the institutional arrangements for reform coordination. Since the reforms were officially launched around 1999, it has been the ministry in charge of local governments that has led the process. This is a common feature of local government reforms in other countries, but in Tanzania this institutional anchor, located within the office of the executive, has been problematic when reforms move beyond local government

Per Tidemand and Nazar Sola

227

reforms (primarily concerned with building the capacities of LGAs) to wider decentralization reforms encompassing wider public sector reorganization: transfers of powers from ministries, devolution of public servants, wider fiscal reforms, wider legal reforms, and so forth. PMO-RALG has not had the required institutional influence to deal with those wider aspects of reform, especially those that are involved with devolved governance and intergovernmental relationships. In conclusion, local government has been a high-profile governance reform supported by significant resources during the past decade in Tanzania, and our findings show that change is real, but local government reform remains stronger on paper than in practice. As noted above, there have been several main achievements, most notably an extensive legal framework and technically strengthened system of local government institutions, plus improvements in financial transfers to the local governments. On the other hand, LGAs have significant limitations on their autonomy due to low ownsource revenue, a dominant political party (CCM), and an incomplete devolution of human resources. These have limited downward accountability and likely have stunted participation in ways that prevent Tanzania’s devolution from contributing more fully to democratic deepening at the local level. Tanzania has undertaken local government reforms that have helped to increase local capacity, but these have lagged in implementation as have efforts to increase autonomy for local governments.

Notes 1. The structure and history of the local government system differs significantly within Tanzania between mainland Tanzania and Zanzibar. In this chapter, we focus on mainland Tanzania. 2. This section draws on Steffensen, Tidemand, and Mwaipope (2004); Mawhood (1983). 3. CCM has been the ruling party in Tanzania since independence. CCM means “the Revolutionary Party,” as TANU was renamed in 1977 when it merged with the Afro-Sharazi Party in Zanzibar. 4. In regions such as Mbeya, Iringa, and Ruvuma during the years preceding the reform, this included up to ten water engineers, various technical officers, and also large numbers of support staff (including vehicle workshops). 5. Field visits as part of the LGRP evaluation in 2007. 6. These findings are from a range of interviews conducted by the author during the LGRP evaluation in 2007. 7. The PMO-RALG conducts regular reviews and evaluations of the LGRP. 8. See PMO-RALG (2007). 9. This is an average figure for the largest five sectors over the past three years. In the education sector, only 18 percent of LGA expenditures are for nonsalaries while in the health sector it is 28 percent (see, e.g., PMO-RALG 2007).

228

Tanzania

10. See, for example, analysis in PricewaterhouseCoopers (2004). 11. For a detailed discussion, see Fjeldstad et al. (2010). 12. Note that secondary education was later made a mandate of the LGAs. 13. The discussion of the grassroots elections draws primarily on Chaligha (2008). 14. Note that informants in REPOA (2007) were asked about their individual behavior whereas the REPOA local government surveys asked about whether “you or a household member is involved in . . .” 15. Prime Minster Edward Lowassa, interviewed by the author, February 2007 during the LGRP evaluation. 16. See www.policyforum-tz.org/groups/lgwg. 17. See a tentative list of policy reform papers in Tanzania, see www.hakikazi .org/policies.htm.

11 Uganda: Decentralization Reforms, Reversals, and an Uncertain Future Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

Uganda has received considerable attention in the decentralization literature as a developing country that embraced genuine and significant decentralization with an unusual level of enthusiasm.1 Following an extended period of internal conflict after gaining independence from Great Britain, the country began a period of greater stability with the rise to power of Yoweri Museveni and the National Resistance Movement (NRM) in 1986. Museveni quickly moved to establish a functioning state and adopted a major program of public sector reform. From the early stages of the NRM government, there was an emphasis on citizen engagement and developing local capacities to drive economic development and political legitimacy.2 Decentralization was seen as critical for democratization, service delivery, and fighting poverty. By the late 1990s a strong legal framework for decentralization was in place, and local governments quickly became among the most empowered and best financed in Africa. Ugandan decentralization was unusually homegrown and initially substantial for a low-income, postconflict country. There was genuine progress with systems, processes, and outcomes expected from reform. The nature, pace, and trajectory of reform, however, seems to have been too ambitious to take deep root, and there was probably too much emphasis on formal system development and not enough on building local accountability and governance. As political incentives shifted over time in reaction to delayed concerns among actors who initially did not realize that their power was threatened by decentralization and documentation of deficiencies in local government performance, Uganda began to face creeping recentralization. This experience thus illustrates a phenomenon observed in other countries in this volume: developing a more capable local state is far from an automatic result of decentralization—this requires carefully conceived efforts to nurture capacity and behavioral changes among all actors involved. 229

230

Uganda

Socioeconomic and Political Context Uganda achieved independence in 1962 from Britain, which had held Uganda as a protectorate since the late nineteenth century.3 After decades of shifting systems and conflicts, the National Resistance Army (NRA) led by Museveni seized power and established the “nonparty” NRM system in 1986.4 The ensuing development of a constitutional and legal framework paved the way for elections, and Museveni was elected to the presidency. Although Museveni and the NRM brought Uganda some political stability and economic growth (which has ranged from 4 percent to 7 percent over the past decade), over time concerns about the concentration of power emerged. A 2005 referendum initiated multiparty elections but presidential term limits were also eliminated, allowing Museveni to run for a third term. The first multiparty elections in 2006 resulted in a controversial victory for Museveni, and he retained the office again in 2011. The strong powers of the president and the dominance of the parliament by the National Resistance Movement Organization, a party whose members also dominate ministerial positions, have led to further calls for reform, but Museveni and his party have managed to hold on to power. To do so, they have taken steps (see below) that have weakened decentralization, although not necessarily for that specific purpose. Public opinion regarding the current Ugandan government is mixed and below average for the region. According to an Afrobarometer (2008) survey, 56 percent of respondents had a lot of trust in the president compared to the 65 percent average in nineteen African countries surveyed (see Table 11.1). Trust in local government councils and traditional leaders in Uganda were at or closer to the regional averages (53 percent vs. 55 percent and 61 percent vs. 61 percent). The public sector accounts for about 22 percent of GDP, but tax revenues are only about 12 percent of GDP, so the state is highly dependent on external

Table 11.1 Trust in Public Institutions in Uganda (percentage expresses a lot of trust) Public Institution President Parliament National electoral commission Police Courts Local government (councils) Traditional leaders Source: Afrobarometer 2009.

Uganda

Average of African Countries Surveyed

56 52 40 37 51 53 61

65 59 57 51 59 55 61

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

231

assistance. There have been substantial efforts to build state competence and target development and poverty reduction, but overall government capacity remains modest relative to the considerable resources expended.

A Brief History of Decentralization During Uganda’s early colonial period, appointed district commissioners were the key subnational officials.5 Over time formal local institutions were created and revised, alternatively strengthened and weakened. The traditional kingdoms maintained a separate identity until their 1967 abolition and absorption into the district administration system. During the war that brought Museveni to power, the NRA mobilized popular support with village Resistance Councils. The NRM government built on this approach by creating a hierarchy of elected Resistance Councils in villages, parishes, sub-counties, and districts with authority to plan, make decisions, administer local justice, and provide services. The NRM Resistance Councils provided the basis for the local government system that emerged in the 1990s. In this proreform environment, a robust formal decentralization framework emerged. The Local Government (Resistance Council) Statute of 1993, the 1995 Constitution, and the Local Government Act of 1997 provide considerable detail on powers and responsibilities of local governments, with districts being the main fiscal authorities. Regional governments are permitted by the constitution and an amendment to it, as well as in the Regional Governments Bill of 2009, but they do not currently operate. Traditional kingdoms were relegitimized, but with intentionally limited (largely cultural) functions. A number of institutional innovations were adopted to manage decentralization. These included the now defunct Decentralization Secretariat (attached to the Ministry of Local Government), which was intended to coordinate national decentralization actors, and the Local Government Finance Commission (LGFC), a broad-based independent body (defined by the constitution and legislation) that reports to the president. As a neutral body, the LGFC at times helps to moderate the effects of central institutional self-interested behavior and power struggles over control of the decentralization agenda. By the turn of the twenty-first century, decentralization was threatened by various forces and concerns were increasingly raised about local government performance. The Fiscal Decentralization Strategy (FDS) adopted in 2002 was framed as an effort to help local governments perform better; increase revenue generation; more objectively and transparently allocate resources and target the neediest areas; increase local autonomy in utilization of funding; and strengthen downward accountability, financial management, reporting, monitoring, and auditing. Although the FDS accomplished some of these goals, it also partially recentralized the intergovernmental system.

232

Uganda

In addition to the centralizing features of the FDS, there has been a politically motivated proliferation of new districts, from 39 in 1994 to 111 in 2012, as well as other local units.6 Some analysts see this growth as a threat to the viability and effectiveness of local governments (see below). Recent steps to recentralize the appointment of senior local government officials, weaken the local revenue base, and create regional governments also pose concerns for local governments.

Structure and Institutions of Decentralization Uganda’s subnational government activity functions through a unified system of elected authorities at the district level, which has primary local planning and budgeting responsibility, and four lower levels with more modest roles. District or city, subcounty, and village councils are elected by universal adult suffrage, with county and parish levels elected indirectly or appointed ex officio. The district or city councils and subcounty (municipal divisions and town councils in urban areas) are corporate bodies and accounting levels. Other levels— county councils, parishes (wards in urban areas), and villages (cells in urban areas)—are administrative units.7 At the national level, the Ministry of Local Government (MLG) is charged with coordinating, advocating for, mentoring and supporting, and inspecting and monitoring local governments. The Ministry of Finance, Planning, and Economic Development (MFPED) plays a key role in designing and executing intergovernmental transfers and financial oversight. The Ministry of Public Service and the National Planning Authority also have important functions related to civil service and planning system procedures. The usual range of sectoral ministries (e.g., education, health) ensure compliance with national policies and standards, and each is responsible for inspection, monitoring, technical advice, support, supervision, and training in their areas of expertise. Other key central agencies include the Office of the Auditor General, the Inspector General of Government, and the fiscal advisory body LGFC (discussed above). The central government maintains a presence and exercises oversight at the local level through the resident district commissioner, who is appointed by the president to coordinate administration of government services in a district and advise the district chairperson on national matters. The resident district commissioner also assumes other functions assigned by the president or prescribed by parliament (1995 Constitution, Article 203; Local Government Act, Section 71). Two local government membership organizations, the Uganda Local Governments Association and the Urban Authorities Association of Uganda, represent the general interests of local governments in national policy discussions. They have been influential in some policy debates.

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

233

Regarding the nonfunctioning regional governments noted above, the 2009 law provides that two or more neighboring districts may form a regional government to perform functions allowed by the constitution. Areas associated with several kingdoms (Buganda, Bunyoro, Busoga, Acholi, and Lango) are specifically designated as regional governments. The regional tier, however, is not universally accepted.8 There are concerns that they would mix administrative and service functions with cultural and traditional functions, would increase operating costs, would require the revision of the intergovernmental fiscal transfer system, and that implementation would be selective and ad hoc.

Assessment of Decentralization Decentralization is expected to improve delivery of public functions, as it has in Uganda, although unevenly. If decentralization is to be effective, a number of key dimensions—authority, autonomy, accountability, and capacity—need to be developed. Authority

Authority is divided between central line ministries and local governments as provided for in the constitution and the Local Government Act. The role of ministries is restricted to the standard setting and oversight functions noted above. Local governments have major responsibilities, although exact functions depend on higher-level decisions, local capacity, and rural versus urban location. Local governments are legally responsible for both infrastructure services (e.g., roads and water) and social services (e.g., health and education). Their functions, however, have been partially usurped by central agencies. Local governments receive a substantial share of annual national revenues, although it has been shrinking, from more than 25 percent in FY 2003/04 to about 20 percent in FY 2011/12 and 18 percent in the FY 2012/13 budget. They have become more dependent on transfers, which have been on average nearly 95 percent of local revenues in recent years (less in urban areas). The constitution allows three types of grants: unconditional, conditional, and equalization. Grants were originally designed mainly to cover the recurrent budget. A separate development transfer was piloted with donor funding and later mainstreamed under the Local Government Development Program. This was closely linked to a participatory planning process and compliance and performance-based conditions and was later subsumed under a broader capital budgeting process. Now there are recurrent and capital transfers, both of which can be conditional or unconditional. Over time, the balance has dramatically shifted between unconditional and conditional transfers, with the latter now dominating. The uncon-

234

Uganda

ditional grant declined from 17 percent of total transfers in FY 2000/01 to less than 11 percent in FY2010/11 and as low as 4 percent in FY 2012/13. Under the present system, the MFPED allocates funds through the Medium Term Expenditure Framework and budget envelopes to sector or line ministries. Ministerial sector working groups—with input from Uganda’s local government associations—make decisions on where and how funds allocated to the sector will be spent, often attaching conditions on local governments’ use of their allocations. Local governments have some own-source revenue. The major source (averaging 70 percent of local revenues outside Kampala) had been the graduated personal tax (GPT). This was suspended before 2006 elections. Local governments were compensated, but not sufficiently in most cases. Local service and hotel taxes were instituted, but yields are uneven across local governments. Property tax is also assigned to local governments, but few use it effectively. Even in large urban areas, administrative and political constraints undermine collections. Finally, local governments may use various local fees, licenses, and other revenues that are rarely productive. Beyond fiscal powers, local governments have authority to plan and budget, manage staff, and legislate, among others. Lower levels play a smaller role than districts, but they are involved in service delivery and are entitled to revenues. The constitution requires districts to indicate how conditional and equalization grants are to be shared with lower levels, but they are substantially retained by districts or municipalities. Subcounties are also entitled to a share of district own-source revenues, but transfers are commonly delayed or do not occur. Collectively, this situation constrains the involvement of lower levels in planning and service delivery and inhibits development of their capacity. Autonomy

The constitution and relevant legislation endow local governments with a significant level of autonomy, compared to many developing countries. In practice, however, substantial restrictions have evolved or been placed on local discretion since the original decentralization reforms. First, there is considerable control over local government expenditure functions exercised by central government agencies. These take the form of various types of service standards, conditions, and operational guidelines. As noted above, the transfers subject to these conditions constitute a large and growing share of local government funds. Second, even unconditional transfers often end up being used to pay for basic (essentially nondiscretionary) administrative and salary costs. With unconditional transfers being used for critical fixed costs that the local government would otherwise be unable to finance, there is relatively little room for truly discretionary local government spending. Third, own-source revenues are constrained both by central interference and local constraints. One local challenge is per-

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

235

vasive poverty, particularly in rural areas. Citizens may also be unwilling to pay local taxes if they are dissatisfied with local service delivery. Finally, the national government has recently placed new restrictions on local government hiring and procurement processes (as noted above and explained below). Thus, local governments do not have control over some important basic operational functions that are typically seen to be the right of local governments. Accountability

There are multiple channels of local government accountability in Uganda— downward, upward, and horizontal. Downward accountability occurs first and foremost through the electoral process by which citizens choose their local government council members and hold the councils accountable to them. The constitution provides for directly elected district chairpersons. There is some evidence that local elections have been more competitive and fairer than national elections, and that elections are the most effective of the various local accountability mechanisms.9 There are, however, concerns about the integrity of the electoral process and the inadequacy of the information needed by citizens to evaluate council performance. Lower local council (parish and village) elections have not been conducted since 2001 because the Forum for Democratic Change, an opposition party, petitioned that elections should not be held until the legal framework had been amended to reflect the multiparty system. Legislation for organizing these elections under a multiparty dispensation has been put in place, but the government maintains that elections have not been held due to a lack of funds. In addition, a 2006 amendment to the Local Government Act provided for the election at the village level only of the council chairperson, who is then empowered to appoint the rest of the council. The government has also devised means beyond elections to enhance downward accountability, often with external support. First, citizens are encouraged to demand local accountability, and there are provisions for citizen petitions and appeals. Second, the government periodically provides data and information on fiscal transfers and other local fiscal matters to the press, and districts and subcounties are required to display financial reports on notice boards. Council meetings and minutes are public. Third, monitoring committees are constituted at lower levels of local government. Fourth, local governments gather information about citizen needs through consultations and participatory mechanisms, allowing citizens to voice their views through local radios, and use of informal mechanisms for dialogue and feedback. Finally, it is the express responsibility of councillors to monitor service delivery and ensure accountability. These accountability measures, however, have faced a number of constraints. First, there are broad concerns about the capacity of all actors (see below). Second, public information about fiscal matters is primarily used by

236

Uganda

elites. Others may not know it exists or how to use it, feel insufficiently empowered to use it, or feel anxious about the consequences of using it. Third, citizens’ sense of connection to local governments is somewhat undermined by the local governments’ financial dependence on the center, service delivery deficiencies, and corruption. Fourth, critics allege that local councillors do not often use information provided in participatory processes or provide significant feedback to constituents. Finally, critics charge that civic education has been too limited, such that many citizens do not sufficiently understand their rights and responsibilities. Collectively, these factors undermine downward accountability (see Smoke, Muhumuza, and Ssewankambo 2010). Results from the Afrobarometer (2008) survey shed some light on citizen perceptions of local government. Ugandans seemed more likely than average (for the sample of nineteen African countries surveyed) to understand that local councilors should be accountable to the people (54 percent Uganda vs. 37 percent average). There was, however, only modest satisfaction with local governance quality, and it was generally worse than the African average. Levels of dissatisfaction on specific questions are shown in Table 11.2. Satisfaction with service delivery and revenue collection performance also fared poorly, with only one service or revenue (keeping the community clean) being rated as fairly or very well managed by half of the respondents, and all ratings were around or below the multiple-country averages. Some results are more difficult to interpret. For example 60 percent of Ugandans surveyed approved of the job their councillors were doing and 69 percent believed they were qualified, but only 39 percent thought councillors often or always listened to people, and only 33 percent believed they were honest with funds. Some of this greater dissatisfaction with governance and performance, of course, may

Table 11.2 Citizen Dissatisfaction with Local Government (percentage dissatisfied) Category Making council’s work program known Providing information about budget Allowing participation Consulting others Providing ways to handle complaints Using revenues for public services Difficult for citizens to have voice heard between elections (% saying somewhat or very difficult) Little or nothing one can do about local government Source: Afrobarometer 2009.

Uganda

Average of African Country Sample

55 63 62 60 62 65

48 55 54 45 49 49

65 64

59 58

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

237

result from the relatively greater awareness of Ugandans about the expected accountability of local governments to them. Downward accountability can be somewhat complicated by the role of civil society organizations and traditional authorities and their relationship with local governments and local citizens. The involvement of civil society is evident in local government processes, especially in regard to planning and budgeting as well as implementation of programs financed by the national Poverty Action Fund. Civil society groups have been involved in sensitizing and organizing the local communities to participate in decisionmaking, planning, demanding and monitoring services, and pressing for accountability. All of these processes help further the accountability goals of decentralization, but the role of CSOs is primarily limited to advocacy. Civil society inputs are not legally binding and, thus, depend on the goodwill of local leaders. As a result of this situation, relations between CSOs and local authorities tend to be harmonious only in areas where CSOs supplement development efforts of local authorities; for example, in service delivery (water, HIV/AIDS services, microfinance, education, health, etc.). On the other hand, advocacy CSOs that monitor service delivery, or mobilize to demand services and accountability, have faced some hostility from local leaders and their critical inputs are more frequently ignored. Unless their role is legally recognized and institutionalized, they cannot have a serious impact. Some analysts argue that Ugandan CSOs are weak because they have been created largely to serve selfish economic interests of local elites or narrow interest groups. They thrive in Kampala and urban areas, but are disconnected from smaller community-based organizations in rural areas. Many CSOs depend on donor resources and, therefore, may primarily serve their funders’ interests. The role of CSOs with broader constituencies and legitimacy is not pronounced (Smoke, Muhumuza, and Ssewankambo 2010). The Afrobarometer (2009) survey found that citizens had a higher level of trust (61 percent) in traditional authorities than in any level of government. Sixty-six percent of respondents believed that few or no traditional leaders were involved in corruption and 59 percent thought they should play a greater role in community governance. At the same time, only 49 percent of respondents thought that traditional leaders listened to local people. Although this seems low, it is better than the 39 percent who responded to the same question about local councillors. Upward accountability for procedural and sectoral service delivery standards is also important, and has become even more so. The general role of the central agencies is to guide, inspect, monitor, and ensure compliance with legal provisions under their respective jurisdictions. Upward reporting mechanisms include: financial and physical progress reports (general and sectoral), routine inspections, annual local government performance assessments, and external audits.

238

Uganda

Local governments often complain of inadequate coordination of reporting mechanisms, as manifested in multiple missions, reviews, inspections, and workshops (Smoke, Muhumuza, and Ssewankambo 2010). In some sectors, there are also parallel guidelines that undermine the power, discretion and downward accountability of local governments. These bureaucratic controls also suffer to various degrees from inadequate capacity and corruption, although this has lessened over time. Horizontal accountability—between elected local councils and local staff who execute the local budgets—is also central to local government performance, but often problematic in practice. The legal framework provides a clear division of roles—councillors set policies and oversee technical staff who implement them. Elected councillors possess sufficient formal authority to hold the technical staff accountable. Council meetings are held and technical staff members provide information to councillors during sector committee meetings. The councillors in turn provide feedback to technical staff after monitoring and inspection activities. The effectiveness of their interactions, however, is constrained if information is not easily understood by the councillors. Councillors also sometimes venture beyond their oversight role into implementation, leading to role confusion and conflict. Finally, changes in the hiring and firing powers of senior council staff have obvious implications for horizontal accountability. Capacity

Capacity is a major concern in Uganda, as it is in all developing countries and in most decentralization efforts. Capacity is needed at the national level to develop, support, and monitor local governments. Local governments require both technical capacity to meet their functions and governance capacity to work effectively with their citizens. Citizens themselves need to develop capacity to hold local governments accountable. Although there has been progress in all of these areas and efforts have been designed to deal with deficiencies, capacity remains a great challenge. In the early years of decentralization, much of the national technical support was provided by donor-funded external advisers and consultants or local consultants paid through parallel mechanisms (e.g., the Decentralization Secretariat) rather than by central government staff. Some argue that this has held back capacity development among the latter. Over time as systems of upward accountability have been developed and centrally based civil servants have been trained to perform their tasks, the situation has improved considerably. Nonetheless, there continue to be major staff shortages and skill gaps in central agencies. Several donors have devoted considerable resources to capacity building for ministries and other key partners such as the Local Government Finance Commission.

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

239

Technical capacity in local authorities has also improved over time. There are, however, considerable disparities among districts, and most suffer capacity challenges. Remote or hard-to-reach and newly created districts, as well as those in northern and eastern Uganda affected by insurgency, have the most serious constraints. The lack of local capacity is one of the key factors that led to recentralization. Some aspects of local capacity relate to staffing while others relate to the elected councillors or citizens. As suggested above, some councillors are not well equipped to meet their responsibilities, and demand for good governance is relatively weak in many areas. There are many related reasons for the lack of civic competence—people fear authority, they lack basic information and awareness of their rights, and they may be too poor to pay much attention. Collectively, these challenges limit the extent to which councillors exercise authority and to which people exercise civic rights and responsibilities. Participation in local planning and budgeting processes is often limited, so citizens are not as involved as they could be in determining local development needs, and there are not many documented instances where constituents have held their local leaders truly accountable. Downward accountability is limited in part because both the central government and most of the involved donors have heavily emphasized building institutional and technical structures, procedures, and capacity of local government. There has been insufficient attention to supporting the development of civil society capacity and the governance sensibilities required to make it effective, resulting in a limited linkage between development of local administration and local democracy.

Political Economy Important political economy considerations underlie the shape and performance of the intergovernmental system reviewed above. Certain initial drivers motivated decentralization, but changing conditions at national and local government levels later led to important shifts. Beyond the governmental dynamics, civil society and international development agencies have also played a role in how decentralization has been shaped. Initial Drivers and Dynamics

The significant political change ushered in by the rise to power of the NRM created an opportunity to bring Uganda back from a long period of conflict and underdevelopment. There is little doubt that decentralization was a conscious and central part of the NRM strategy to develop Uganda.10 There was broad support for this from parliament and other key actors. Although some critics may have seen decentralization as a way for the NRM to consolidate power, supporters attributed the approach to a genuine desire to better the country.

240

Uganda

Although Uganda is now a multiparty state, the NRM dominated in the early days and it remains powerful. Overt opposition to local government was limited beyond proponents of federalism based on traditional structures, and decentralization was not overtly contentious within the NRM. A key motivation was the core role that the Resistance Councils played in mobilizing public support. That success and momentum for change created broad consensus for a local government system and defined the general shape it would take. It is also clear that this was a homegrown desire, as demonstrated by the government’s efforts to hold international development agencies largely at bay early on as the new system was developed. The government engaged only a few external agencies to help with the initial design of decentralization. Also important was the timing (whether intended or not), which coincided with the constitutional review process, thus facilitating the integration of the decentralization policy into the constitution.11 The creation of a modern local government system may also have been a way to sideline traditional kingdoms and preclude a resurgence of support for federalism, but the NRA provided mixed signals. Many NRA fighters were from tribal groups associated with prominent kingdoms, and Museveni reportedly was willing to restore the kingdoms after an NRA victory, but this did not happen. In 1993 the kingdoms were permitted to reconstitute, but as parallel institutions and with the understanding that they were to serve a cultural rather than political function. Of course, the kingdoms by their nature have political meaning, but they were given limited funding and no power to tax, so they clearly were not intended to be major partners in formal governance and development. Three factors about the emergence of the local government system in the 1990s are important. First, even if the NRM commitment to decentralization was genuine, it also seems to have been based on an incomplete appreciation of some significant implications. The local government system created by the NRM government, which had high governance and functional expectations placed on it, was very different from the largely apolitical mobilization and conflict mediation role of the Resistance Councils system that in great part inspired the decentralization initiative. Second, early influential relationships with particular development partners that had limited international experience in decentralization led to the design of an ambitious local government system that could be considered highly unusual in a developing, postconflict country. The considerable functions, resources, and levels of autonomy rapidly given to local governments under the initial reform program were in hindsight not commensurate with the capacity and accountability levels of the local councils. Third, the responsibility and resources for developing local governments were concentrated at the MLG and the Decentralization Secretariat established under it with external assistance to coordinate decentralization reforms. The secretariat, however, paid better salaries than the MLG and performed functions that the MLG should have had, such that the secretariat became a key driver of

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

241

detailed design and implementation of reform and limited MLG capacity was developed. In addition, many local systems and procedures were developed entirely by the MLG and secretariat with minimal consultation with other key agencies. Since the MLG was not sustainably powerful and the secretariat was effectively a parallel mechanism, the stage was set for the emergence of future concerns and challenges from other agencies. Shifting Conditions and Dynamics

As major policies were articulated over time and central agencies became stronger, influential ministries awakened to their potential roles in service delivery and local development, with the realization that relevant powers and resources had been assigned to local governments. As ministries moved beyond developing basic systems and capacities, their new reforms increasingly began to conflict with decentralization. Public financial management (PFM) reformers in the MFPED, for example, started to pay attention to local government budgeting, which had previously been managed by the MLG and secretariat. In the early reform period, the MLG developed a new local PFM system and local officials were trained in its use. Only a few years later, the MFPED determined that the new MLG system was not adequately linked to national systems and replaced it. Local officials then had to be retrained. Around the same time, major sectoral ministries began to develop new approaches to service delivery, primarily under the Poverty Reduction Strategy Papers and sector-wide approaches (SWAps) promoted by donors to support the Millennium Development Goals. These reforms had centralizing elements that clashed with decentralization policy and the local government systems and procedures developed by the MLG. As other national ministries were independently beginning to carve out a role for themselves in local functions, research (largely financed by the donors) began to document serious problems with local government expenditure management and service delivery, raising questions about whether too much responsibility had been decentralized too quickly. These concerns were initiated by the results of early Public Expenditure Tracking Surveys conducted by the World Bank and other related research. This process was piloted in the education sector in Uganda in 1996 and was later deepened in that sector and expanded to other sectors.12 The results of the studies raised serious alarms. Education allocations over a five-year period, for example, often did not reach intended schools that received on average only 13 percent of allocations, with the bulk captured by local officials. Within local governments, moreover, there were great variations in the realization of allocations, suggesting that certain schools had differential political power. In the health sector, allocations to service providers were reduced considerably when the districts were given authority to form their own budgets. The emergence of these performance concerns bolstered decisions by other government agencies to deepen PFM reforms and SWAps that had nontrivial

242

Uganda

recentralizing effects.13 The convergence of centralizing interests and the findings of these studies resulted in policies, systems, procedures, and programs that undermined local powers provided for in the legal framework. In the late 1990s under the Poverty Action Fund, for example, the transfer system was made more conditional, also a feature of the PFM and SWAp reforms that began to emerge around the same time. The Decentralization Development Partners Group was concerned enough to commission a study to consider how to fortify local discretion. This resulted in the FDS, which introduced new budgeting and financial management processes. These reforms did lead to better information and monitoring, but they also imposed a budget template that limited the budgetary discretion of local governments.14 The FDS provided for some local flexibility, but this has been substantially stifled in practice, and its stated intention to ease restrictions after local government capacity improved was never implemented. In addition to the bureaucratic dynamics surrounding intergovernmental relations, shifts in political dynamics also played a key role in reshaping decentralization. First, in part due to the desire of the incumbents to keep control in a new multiparty environment, the central government recentralized power to recruit and control top local government staff and to manage procurement.15 Specifically, local governments lost the power to recruit chief administrative officers (CAOs), deputy CAOs, and municipal town clerks to the Public Service Commission. The justification was to ensure recruitment quality, to protect hiring from local political interference, and to facilitate CAO transfers to other districts without local reinterviewing. However, critics charge that these changes soften local accountability, confuse reporting and accountability between local leaders and the central government, and limit control by councils over their performance (Muhumuza 2008b). Second, political incentives have increased local government fiscal dependence. Some dynamics are local and have to do with citizens’ unwillingness to pay taxes to local governments that are not seen as credible service providers. Given conditional transfer growth and weak own-source revenues, local governments have limited ability to respond directly to locally expressed needs. Available resources are commonly captured by local elites and used for patronage.16 There is, however, some evidence that local governments do a better job of delivering services for which they have full responsibility relative to services that involve overlapping responsibilities of central and local governments.17 Other political dynamics involve national calculations. For example, prior to the 2006 national elections in which the NRM faced opposition from newly legalized political parties, Museveni suspended the GPT. The tax had long been criticized and was unpopular, but it was the main own-source revenue for many local governments and it was not adequately replaced.18 The central government has also failed to encourage better use of the property tax, which is poorly administered and raises far less revenue than, for example, in neighboring Kenya. Central government unwillingness to support property tax reform is likely due to the political sensitivity surrounding landholding and use.

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

243

Third, new local governments have proliferated in the past two decades. Although this is officially justified in terms of improving local citizen voice and better targeting service delivery, some analysts see it as a source of political patronage designed to help the NRM to win increasingly tough electoral battles.19 Creating a new district brings resources to a specific location and is visible, which benefits the ruling regime. In addition, since district governments have little influence on national policy formation, Museveni and the NRM can create new districts even in opposition areas with little real risk of generating serious negative consequences for national-level politics and policies. Critics charge that creating new local governments has distorted local structures; diluted already inadequate human resources; caused performance gaps in key services; substantially increased administrative costs and overheads; and placed an additional, significant supervision burden on the central government (World Bank 2013). In short, creating so many new local governments may worsen the problems that it was supposedly intended to address. Finally, the issue of federalism seems to be resurfacing. The traditional kingdoms favor federalism over the regional governments, and some have publicly clashed with the central government. The kabaka (king) of Buganda, Ronald Mutebi, has promoted a federal structure built around the kingdoms.20 This concerns the government not only because of the political threat, but also because of the recent discovery of oil in territory associated with a number of kingdoms. The potential volatility of the situation is evident from riots sparked in September 2009 by the government’s refusal to allow Mutebi to travel in his kingdom, allegedly for his own safety. How the situation will play out remains to be seen, but it could be an important factor in future decentralization debates. The Key Impacts of External Development Partners

The external development partners have played a role of considerable importance in Uganda’s decentralization. As noted above, the early NRM government used external assistance for decentralization selectively. This kept reform motivation high and prevented an onslaught of development partners bearing diverse ideas and programs—the curse of decentralization in many developing countries. On the other hand, some of the early donors were not experienced with decentralization in developing countries. This may help to explain why such an ambitious reform was so rapidly undertaken in a postconflict, low-capacity country. In addition, the establishment of the Decentralization Secretariat as an externally financed parallel mechanism attached to a relatively weak ministry repeated the consequential mistakes of several other developing countries. While the larger system was developing in this way, the government opened the door to other donors, and the resultant efforts were fragmented. After a while, some points of cooperation began to emerge. When the World Bank expanded its support to local government reform, for example, it built on revised versions of systems and procedures piloted by the United Nations Capital Development

244

Uganda

Fund and other donors. Eventually in 2002, development partners established a coordination mechanism that evolved to become the current Decentralization Development Partner Group, which works with government coordinating mechanisms and aims toward harmonized programming of donor support. The donor group was central in supporting the MLG’s Decentralization Policy Strategic Framework to coordinate efforts for deepening decentralization. The framework was the basis for local government sectoral investment plans that outline strategies, priorities, and budgets for service delivery; political, administrative, and fiscal reform; good governance; and local economic development. Current arrangements for coordinating policies and external support involve a set of working groups, including the Decentralisation Management Technical Working Group (DMTWG). Opinions about the overall effectiveness of the donor coordination initiatives vary, but they did improve on the status quo and generated productive efforts such as the joint decision to stop financing the parallel Decentralization Secretariat in favor of direct support to the MLG. On the other hand, the partners focused coordinated efforts largely on developing technocratic mechanisms and insufficiently on nurturing local governance. Much less successful than the coordination of decentralization efforts targeted at the MLG has been the coordination of decentralization with other aspects of public sector reform that affect local governments. The DMTWG is under the broader Public Sector Management Working Group, which should enhance linkages between decentralization and other public sector reforms. Issues affecting decentralization are also supposed to be discussed between donors and the government. This should allow better linkages to sector initiatives, although policy and donor fragmentation and inconsistency have persisted, and it is not clear that the various parties really want this to change. At some level, development partners correctly understood that decentralization had initially gone too far and local governments were not performing assigned functions effectively. Some recentralizing efforts undoubtedly were designed with good intentions to deal with this problem. Yet there can be little doubt that some initiatives were driven largely by ministerial power plays reinforced by donor interests or immediate service delivery concerns. Donors can more easily pursue poverty reduction goals through sectoral ministries than through more diffuse budget entities that complicate monitoring efforts. Some analysts even argue that development partners took advantage of Uganda’s aid dependence to “hijack” policymaking roles through conditionalities associated with debt relief, the Millennium Development Goals, and the poverty reduction strategy.21 Unfortunately, development partners never conducted broad-based institutional and capacity assessments that might have allowed them to assist the government to improve service delivery without significantly undermining fragile decentralization and local governance reforms, even as some continued to provide considerable support to these reforms. In addition, the way that develop-

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

245

ment partners pursued reforms—whether intentionally or not—supported the political objectives of an already dominant national political party, and reinforced the often problematic competition among central agencies.

Conclusion Decentralization in Uganda has received considerable international attention. Local governments were given major powers and autonomy under a strong legal framework, and reforms were initially implemented with enthusiasm by the government. Decentralization has promoted popular participation in periodic elections and governance processes, and there have been steps to improve inclusion of previously marginalized or neglected groups. Political officials and staff were also empowered, and incrementally learned, to manage local affairs and to deliver local services more effectively even if unevenly. As reform proceeded, central agencies that supported (or at least did not overtly oppose) decentralization came to recognize its implications for their roles and control over resources. Documentation of inadequate service outcomes also emerged, which in retrospect might have been expected given the unusually substantial responsibilities devolved to local governments insufficiently prepared to manage them. These factors, as well as shifts in the political landscape (principally multiparty elections) led to a number of recentralizing policies such as increasingly conditional grants, extensive creation of new districts, and more central control over local staffing and procurement. These measures collectively undermine the incentives and ability of local governments to improve performance and of their constituents to engage productively with them. Moreover, there are no obvious incentives for the central government, individual ministries, or donors to reverse their recentralizing behavior. The situation may be further complicated by the reemergence of talk about federalism from traditional kingdoms and the likely frame of mind of the NRM in a new era of multiparty government. Despite the challenges, there undoubtedly are possible ways forward if the right information can be assembled and the appropriate actors are engaged productively. Surprisingly little knowledge about decentralization in Uganda is systematically evidence based, and some issues are not well explored. Equally important, much of the more recent attention to reform has focused on service delivery, with limited concern for how to enhance the performance of local governments as semiautonomous entities or the potential benefits of doing so. A few issues are particularly central if Uganda is to improve its local government system and advance the original goals of decentralization. First, local government-citizen dynamics remain neglected. Intergovernmental system reforms and donor support have focused heavily on technical matters. The limited attention to local accountability and initially overambitious

246

Uganda

technical reforms together contributed to compromise performance. A few development partners put greater emphasis on civic engagement, but they tended not to link those efforts well to local government incentives or only worked in limited areas. It would be useful to understand better how citizens perceive local governments and interact with them because the local government system is unlikely to recapture its diminished credibility and improve its performance without greater attention to developing local accountability. Second, there has been no methodical attention given to understanding the details of how broader public sector reforms and poverty reduction initiatives support or weaken decentralization. PFM and SWAp reforms have been beneficial in some ways, but also have been synergistically problematic in others. The FDS reinforced the legitimate role of the central government in strengthening inspectorates, regulations, and accountability mechanisms to ensure good local financial practices. Similarly, SWAps have improved some aspects of service delivery in sectors that were too substantially and rapidly decentralized. At the same time, these reforms bluntly restrict local discretion, a core requirement for realizing the potential benefits of decentralization. Local governments have no incentive to develop capacity if they lack sufficient programmable resources and autonomy to make meaningful decisions on behalf of their constituents. On this front, it would be useful to more fully document the performance of FDS and sectoral transfers and to consider possible strategic processes for gradually and appropriately relaxing budgeting restrictions. A more modest, but also useful, goal would be to work on assisting local governments to make better use of the discretion they have (within the restrictions imposed by the FDS and conditional transfers) and may not yet effectively use. Third, the sectoral fragmentation of service delivery has important effects beyond undermining local autonomy in a particular sector. Present arrangements weaken one of the most fundamental advantages of local governments—the ability and incentive to think holistically and territorially across sectors compared to central agencies. Prevailing sectoral silos may—in the narrow pursuit of specific service delivery goals—undermine the type of broader-based thinking about local development and prioritization processes (within budget constraints) that capture the essence of true local government and can create real accountability links to local constituents. Examining these issues in more detail could provide a basis for undertaking useful and workable reforms such as criteria-based processes for increasing unconditional transfers in conjunction with multiactor processes designed to create viable local development visions. Fourth, local revenue generation generally has deteriorated, even predating the high-profile GPT suspension in 2006. The central government blames local governments, saying they are incompetent and corrupt. The local governments blame central control and interference, which they claim makes it difficult for them to raise local revenue. Citizens do not pay local taxes and charges because they see limited value for money. It would be beneficial to understand better

Paul Smoke, William Muhumuza, and Emmanuel Ssewankambo

247

the nature and importance of the various constraints on local revenue generation, and look for ways to overcome them in conjunction with developing the political credibility needed for local businesses and households to willingly pay local taxes. Fifth, proliferation of district governments has diluted their capacity and increased overall administrative costs. District creation is political, and stopping or reversing it may be difficult. At the same time, there may be politically acceptable ways to capture scale economies and improve service efficiency. Options include, for example, use of existing provisions for regional governments or developing mechanisms for districts to work together to deliver services. Mechanisms for interdistrict cooperation are likely to be less politically contentious and involve less dramatic changes than new regional governments. But a strong legal framework would be needed and a number of practical issues, such as how to govern partnerships and where to locate specialized staff serving multiple districts, would require resolution. Given the problematic consequences of district proliferation, the value of exploring offsetting options seems clear. Finally, development partner interventions have remained fragmented and inconsistent, often reinforcing antidecentralization tendencies of central actors. Some key donors began to coordinate decentralization support over time, but there has been little linkage between these efforts and donor programs for other public sector reforms such as PFM and SWAps. Better harmonization of reform efforts could lead to more effective public sector institutional development and improved outcomes, both generally and decentralization specific. Inherent tensions underlie the present situation, but there certainly are ways to promote greater consistency than has occurred to date. The issues outlined here are obviously not mutually exclusive, and some of the proposed reform areas are highly interdependent. There is no obvious “best approach” to reverse the deterioration of integrated systems and processes in the face of both considerable constraints and entrenched positions that result from complex political and bureaucratic dynamics. Given the broad concerns about decentralization, however, it is important to look for common ground that could serve as the basis for taking steps to improve on the status quo. The question going forward is whether the more centralized system that is now in place will persist, or if well-considered efforts might strategically and appropriately reempower local governments in ways that increase the probability of better and more sustainable governance and development performance.

Notes 1. See Smoke (2000); UNCHS (2002); Onyach-Olaa (2003); UNCDF (2003); UNDP (2003); World Bank (2003); Saito (2004); USAID (2004); Okidi and Guloba (2006); Asiimwe and Musisi (2007); and Kauzya (2007).

248

Uganda

2. The legal framework for decentralization in Uganda is robust (see below), and the critical role of decentralization is highlighted in documents such as the Poverty Reduction Strategy Paper (Government of Uganda 2000); Fiscal Decentralization Strategy (Government of Uganda 2002); and National Development Plan (Government of Uganda 2010). 3. Uganda’s political history is covered in more detail in Kanyeihamba (1975); Byrne (1990); Apter (1997); and Sejjaaka (2004). 4. Francis and James (2003) provide a useful discussion of the National Resistance Movement. 5. For detailed treatments of decentralization in Uganda, see Byrne (1990); Villadsen and Lubanga (1996); Azfar, Livingston, and Meagher (2007); Government of Uganda (2001); Francis and James (2003); Onyach-Olaa (2003); Saito (2004); Steffensen, Tidemand, et al. (2004); Wunsch and Ottemoeller (2004); Okidi and Guloba (2006); Asiimwe and Musisi (2007); and Muhumuza (2008a). 6. Some elements in the NRM do not agree with the creation of many small districts, arguing that they are not feasible and increase administrative costs. These sentiments are occasionally captured in news articles, individual statements, and public discussions. An example is a news article from the government newspaper The New Vision, which quoted then-minister of local government Hon. Kahinda Otafiire criticizing the creation of more local governments (“District Number Worries Otafiire,” September 22, 2008, by Raymond Baguma and Joel Ogwang, p. 3.) 7. There is currently no city under the local government setup, as Kampala is now governed by a separate Kampala Capital City Authority Act passed in 2010 and that entered into effect in 2011. 8. For example, Buganda is requesting the reintroduction of the federal arrangement that operated between 1962 and 1966. 9. See, for example, Muhumuza (2003, 2006a, 2006b). 10. A general framework on political economy of decentralization is presented in Eaton, Kaiser, and Smoke (2011). 11. Please refer to Steffensen, Tidemand, et al. (2004) for a detailed review of decentralization experiences up to May 2004. 12. See Ablo and Reinikka (1998); Jeppson (2001); and Reinikka and Svensson (2004). 13. See Jeppson (2002); Kasumba and Land (2003); and Wunsch and Ottemoeller (2004). 14. Details of the budgeting process are outlined in the Fiscal Decentralization Strategy (Government of Uganda 2002). 15. For more details, see Muhumuza (2008a). 16. Francis and James (2003) discuss the dual-mode system of technocratic- and patronage-driven governance; Smoke (2008b) discusses the politics of local revenue. 17. Ahmad, Brosio, and Gonzalez (2006) elaborate on this point. 18. The government created a hotel tax and a local service levy, but these do not make up for the loss of GPT and are important primarily in urban and tourist areas. 19. There is a detailed discussion of the history of district creation in Green (2008). 20. This was widely reported in the press; the version of events came from a National Public Radio story, “Kingdom, Government Clash in Uganda” February 14, 2010, www.npr.org/templates/story/story.php?storyId=123710230. 21. See Craig and Porter (2003) for a discussion of this issue.

12 African Decentralization in Comparative Perspective J. Tyler Dickovick and Rachel Beatty Riedl

Since 1990, many African central governments have initiated or deepened processes to transfer authority, power, responsibilities, and resources to subnational levels. Why have they done so, and what have the consequences been for the quality of democracy and governance in Africa? In this concluding chapter, we outline several key findings from the ten cases examined in detail in this book. With respect to the framework used throughout the book, we find mixed results of decentralization processes: meaningful achievements in instituting decentralized authority on a lasting basis de jure, but limited advancement with respect to enhancing subnational autonomy, downward accountability, and governing capacity. Notwithstanding the impressive frameworks for decentralization that exist on paper, several top-down central institutions—especially dominant governing parties and centralized states—play leading roles in politics after decentralization. In other words, decentralization appears to have changed the structure of government more than the quality of governance. We highlight one key function that decentralization, for better and worse, appears to have played in the African politics of the past two decades and more: it has helped to maintain and consolidate regimes and party states across many cases. Decentralization is theorized to disperse power, distributing it more broadly, making governance more democratic, and making governance more responsive to the citizenry. Yet in one case after another, the authors in this volume have found that governing regimes have either made use of decentralization to extend their authority to local levels, or have found ways to delimit decentralization if it did not serve the purposes of the center. As a result of this contingent commitment to decentralization on the part of African central governments, decentralization’s impact on deepening local democracy and local capacity has been modest at 249

250

African Decentralization in Comparative Perspective

best. In fact, the paradox is that its impacts on the quality of local governance have been less significant than its impact on sustaining regime types and strengthening state power at the national level. An apparent outlier is Mali, where decentralization was followed by a coup, but even here it seems that any causal link is more about the weakness of decentralization leading to a return to instability rather than an excess of it. After decentralization in Africa, the center holds. And it often does so with increased—not decreased—power. A set of processes initially trumpeted as national reform with local impact has inverted in several countries: it is local politics with national consequences.

Comparative Findings The chapters in this book have shown that African decentralization consists of a set of meaningful changes to national politics mixed with a number of limitations and enduring impediments to change. We summarize several of these changes and continuities before turning to the consequences of decentralization in subsequent sections of this chapter. With regard to changes, decentralization has advanced considerably in the devolution and deconcentration of legal authority to subnational governments and subnational administrative units. In establishing decentralization framework laws, these processes have advanced in the political, fiscal, and administrative dimensions of decentralization. Politically, all countries in this study now hold subnational elections (see Tables 12.1 and 12.2). On the fiscal side, decentralization gave SNGs important expenditure responsibilities in public services, and guaranteed access to some of the revenue resources needed to meet these responsibilities. A major achievement here is the creation of regularized, formula-based revenue transfers from the center, though these are relatively small in several cases (see Table 12.3). On the administrative side, central governments have undertaken extensive processes to deconcentrate authority from central government offices to administrative units controlled by the center. These changes were meaningful in their initiation, and as yet there has been relatively little in the way of explicit recentralization of power in subSaharan Africa. This suggests that decentralization is stickier than simply being subject to reversal at the whims of sovereign central governments (see Dickovick 2011). We find no cases of reversing political decentralization by returning to the appointment of previously elected SNG officials, nor do we find significant reversals of fiscal decentralization to devolved governments once it has been established. In some instances, deconcentrated authority has

251 Table 12.1 Political Decentralization in Ten African Countries Year of Political Decentralization Law

Elected Levels of SNG

Botswana

1965

1

Elections held for councils in all areas, unelected land boards, tribal authorities exist in some areas.

Burkina Faso

1998

2

Elections held in 33 urban communes in 1995, and nationwide in all urban and rural communes in 2006.

Ethiopia

1993

3

Elections in kebeles, woredas. and states in all areas, also at the zone level in one regional state.

Ghana

1988

2

Districts: 70% of seats elected, 30% appointed by president; local councils not fully elected.

Mali

1995

3

Elections held directly in communes, indirectly for cercles, regions, and High Council.

Mozambique

1997

1 (or 2)

Elections held in autarquías (in some areas) since 1998 and in provinces for the first time in 2009.

1976, 1996

2

Local government reforms in period from 1976 through partial democratization of the 1990s.

South Africa

1994

2

Municipal council elections are hybrid ward and proportional representation; representatives to National Council of Provinces elected by regional assemblies.

Tanzania

1996

1 (or 2)

Elections held at district level in all areas, with wards as constituencies, also at village level in rural areas.

Uganda

1993, 1995, 1997

3 (or 5)

Elections in village, subcounty, districts (local councils 1,3,5); parish, county levels in some areas (local councils 2, 4).

Country

Nigeria

Notes

252

African Decentralization in Comparative Perspective

Table 12.2 Subnational Government in Ten African Countries Country Botswana Burkina Faso Ethiopia Ghana Mali Mozambique Nigeria South Africa Tanzania Uganda

Country Botswana Burkina Faso Ethiopia Ghana Mali Mozambique Nigeria South Africa Tanzania Uganda

Village Level

Municipal Level

District Levela Councils

Subkebele Unit committees

Commune Kebele Councils Commune Autarquia (some areas) Local government Municipality

Woreda District Cercle

Village (rural) Village (local council 1)

Subcounty (local council 3)

District District (local council 5)

Regional Level

Second Chamber

Deconcentrated Units

Region Regional state Region Province State Province

House of Federation High Councila

District administrations Provinces; regions Zones Regions Regions; cercles Provinces; districts

Senate Natl. Council of Provinces Regions County (local council 4); parish (local council 2)

Notes: The High Council for Collectivities in Mali is a consultative body without legislative power. a. In many anglophone countries, the names of districts sometimes differ in urban or semiurban settings, including Botswana (urban councils and rural councils); Ghana (metropolitan assemblies, municipal assemblies, and district assemblies); and Tanzania and Uganda (city, municipality, town).

been pulled back up the hierarchy, and in some (e.g., Botswana and Uganda) recentralization has been explicitly attempted. Yet the achievements of would-be recentralizers are remarkably modest. Apart from reassertion of central control over the civil service in Botswana, Uganda, and several other countries, and some furthering of central controls in South Africa, decentralization has largely stuck.

J. Tyler Dickovick and Rachel Beatty Riedl

253

Table 12.3 Fiscal Decentralization in Ten African Countries SNGs Percentage of National Expenditure (approx.)

SNGs Revenue from Own-Source Tax (approx. %)

Botswana

< 10

Burkina Faso Ethiopia

Major SNG Tax Bases and Sources

Major Intergovernmental Transfer Sources

10–15

Property tax

Revenue Support Grant