ANALYZING DECENTRALIZATION: A FRAMEWORK WITH SOUTH ASIAN AND WEST AFRICAN ENVIRONMENTAL CASES by Arun Agrawal and Jesse C. Ribot Decentralization is a word that has been used by different people to mean a good many different things. But what do we see in practice? Experiments with local government that end in chaos and bankruptcy; ‘decentralized’ structures of administration that only act as a more effective tool for centralizing power; regional and district committees in which government officials make decisions while the local representatives sit silent; village councils where local people participate but have no resources to allocate. Philip Mawhood, Local Government, 1983 INTRODUCTION Since the early 1980s, decentralization has reemerged as a valued political and economic goal in most developing countries.1 According to a recent World Bank study, “out of 75 developing and transitional countries with populations greater than 5 million, all but 12 claim to be embarked on some form of transfer of political power to local units of government.”2 Advocates of decentralization justify it on grounds of increased efficiency, more thoroughgoing equity, and/or greater participation and responsiveness of government to citizens.3 Despite these claims, most decentralization efforts end up not significantly increasing the powers of local authorities or peoples.4 Decentralization of natural resource management, the focus of this paper, is especially intricate because it is not only about providing services efficiently. It also requires the devolution of real powers over the disposition of productive resources. In addition, it requires the resolution of divergent interests among a host of actors so that externalities associated with natural resource management are not disproportionately borne by any subgroup. 1 This global pursuit of decentralization, even if it appears novel, is not the first time governments have attempted the decentralization of fiscal responsibilities or institutional powers. In francophone West Africa, for example, since 1917 at least four distinct waves can be discerned. In each case, decentralization accompanied and perhaps was sparked by fiscal crises of the state. (See Ribot 1999:51.) In South Asia, at least three waves are visible since the mid-1800s. The first of these occurred as early as the 1860s (Fisher 1991). The second wave of decentralization can be seen to have taken place after independence through efforts at community development. A third wave is currently under way (Ribot 1999). 2 Dillinger 1998:1. Wondering about how to think of the 1990s in the context of development, Uphoff and Esman (1974:607) see them as possibly the “decade of deregulation, democratization, and decentralization.” Several other authors also speak of decentralization in the same breath as democratization. (See Totemeyer 1994:49-61; Blair 1998.) But clearly the two are quite different. (See Fox and Aranda 1996 and Wang 1997:1431-42). 3 For a discussion examining the justifications advanced for recent attempts at decentralization around the world see Agrawal, Britt and Kanel 1999. For a more historical discussion of francophone West Africa see Ribot 1999:23-65. 4 For references dating to the mid-1800s, see Crook and Manor 1998; Crook and Manor 1994; Parker 1995; Leonard 1982; and Cohen, et al 1981. 1 New institutional economics and public-choice literature indicate that decentralization has the potential to achieve greater efficiency and equity in public decisionmaking by internalizing externalities, deploying all available information,5 and better matching service provision to needs.6 In this paper, we suggest that representation and accountability are critical if devolved powers are to serve local needs efficiently and equitably.7 We conclude, analyzing four case studies, that the presumed benefits of decentralization become available to local populations only when empowered local actors are downwardly accountable. Actors, powers, and accountability emerge as essential elements of a framework that can help evaluate the effectiveness of decentralization. However, in many instances around the world, decentralization reforms do not attend to these elements. In Senegal, responsibilities in forest management were devolved to local elected councils without devolving access to the related commercial profits.8 In Burkina Faso, powers to cut, sell, and manage forests have been devolved to private project-based committees, rather than to representative bodies.9 In Zimbabwe’s CAMPFIRE program, powers were transferred to District Development Committees who were largely under the control of central government.10 In Nepal, one can point to projects that view decentralization as being accomplished simply by directing a stream of monetary benefits toward a group of resource users rather than attempting to create institutions that allow durable decision-making powers to local authorities.11 Perhaps all who conduct research on decentralization are familiar with such examples. 5 Decentralization strives to allocate decision-making powers to those who have the greatest information about a particular resource. But the availability of greater information does not necessarily lead to the use of that information in favor of the people on behalf of whom decisions are made. Information can also be used to favor the decision-makers themselves or their friends and cronies. 6 Ostrom, Schroeder and Wynne 1993. 7 In theory, decentralization can increase efficiency by helping internalize costs and reducing transaction costs. See the 1988 World Development Report which asserts that “decentralizing both spending and revenue authority can improve the allocation of resources in the public sector by linking the costs and benefits of local public services more closely” (World Bank 1988:1). The devolution of decision-making to local actors can reduce administrative and management transaction costs via the proximity of local participants to decision-makers and the access to local skills and information (see Crook and Manor 1998) and (Huther and Shah 1998:1). See also Bienen et al (1990: 61-75), who argue that in Nepal, decentralization worsened information flows, thereby increasing transactions costs. Decentralization is usually believed to increase effectiveness of coordination and flexibility among administrative agencies and in development/conservation planning and implementation. It is also regarded as a way to increase equity at the local level by allowing greater local retention of revenues and more equal distribution of benefits from local activities. Slater (1989:501-31) emphasizes the role of decentralization in reducing inequalities. Responses can be seen in Rondinelli (1990: 494-500) and Prud’ homme (1994). 8 Ribot 1995:1587-9. 9 Delnooz 1999. 10 Thomas 1992; Murombedzi 1997. 11 There are, of course, success stories. Decentralized healthcare provision in Ciera, Brazil is argued by some to be a great success (Tendler 1997). Local governments in South America have experienced successful fiscal decentralizations. (See Fiszbein 1997.) Uganda instituted elected representatives at the village level in the mid-1990s. (See Karlstom 1996.) These successes, however, are usually partial: Uganda’s new representatives lack powers; in Brazil the efficiency of health services provision has increased but central control has scarcely declined. 2 This paper provides a framework to examine whether the policy choices being made even constitute decentralization. Governments often perform acts of decentralization as theater pieces to impress or appease international donors and NGOs or domestic constituencies. Our framework can be seen as an analytical lens for assessing reforms made in the name of decentralization. It can be used to identify shortfalls in decentralizations—design flaws or political obfuscation. Identification of such flaws can allow advocates of decentralization to push reforms beyond proclamations and closer to action on the ground. The framework can be applied to single sectors, pointing up sectoral shortcomings in an otherwise well-crafted institutional initiative, or to a core set of decentralization laws affecting all sectors. However, the framework does not analyze issues of training, physical infrastructure, or education at the local level that may also be necessary if decentralization is to be successful in its stated aims. Instead of identifying decentralization simply as an institutional reform in the political, fiscal, or administrative realm as is commonly done,12 our framework shows how a particular reform can be analyzed by referring to changes in actors, powers, or accountability. Using four case studies from South Asia and West Africa, we compare instances of decentralization of resource management and assess the utility of our framework by applying it to understanding the extent to which decentralization actually occurred in each case.13 The analysis of the cases in light of our framework leads us to focus on downward accountability as a key aspect of decentralization.14 12 Manor 1999; Binswanger 1999. In their recent book Cohen and Peterson (1999:69-70) propose a framework in which accountability figures centrally. However, they characterize it as an outcome rather than as an element of administrative design. 13 The comparative strategy we use is best classified as the “comparative case-oriented method” described by Ragin (1987). See also Skocpol 1979; Skocpol and Somers 1980:174-97; Lichbach and Zuckerman 1997:3-16. 14 In their collection of essays on accountability, Fox and Brown simply state that accountability is the process of holding actors responsible for their actions (1998:12). Paul (1998:2) puts forth a discussion of problems of accountability in public agencies (1992). Crook and Manor (1998:2) see accountability as a “crucial link between enhanced participation and enhanced institutional performance.” 3 DEFINITIONS AND JUSTIFICATIONS OF DECENTRALIZATION Decentralization has been defined as any act in which a central government formally cedes powers to actors and institutions at lower levels in a political- administrative and territorial hierarchy.15 Devolving powers to lower levels involves the creation of a realm of decision making in which a variety of lower-level actors can exercise some autonomy.16 Deconcentration (or administrative decentralization)17 is said to occur when powers are devolved to appointees of the central government.18 Political decentralization19 is different from deconcentration since powers in this case are devolved to actors or institutions that are accountable to the population in their jurisdiction. Typically, elections are seen as the mechanism that ensures accountability in political decentralization.20 We propose definitions of political decentralization and deconcentration that foreground accountability more centrally. When powers are transferred to lower-level actors who are accountable to their superiors in a hierarchy, the reform can be termed deconcentration. This is true whether lower-level actors are appointed or elected because elections can still be structured in ways that make elected officials upwardly accountable. When powers are transferred to lower-level actors who are downwardly accountable, even if they are appointed, the reform is may be categorized as political decentralization. Critical to understanding the process, then, is empirical examination of the structures of accountability in which actors are located. 15 Mawhood 1983; Smith 1985. We should note that the formal transfer of power to lower levels of government may sometimes be a centralizing act if the powers being devolved were earlier exercised informally by non-state actors. 16 See Booth (1995:89-105) for a discussion of decentralization in France and how it led to greater autonomy for local governments, but at the same time also prompted a struggle for the redefinition of the roles of different levels of government. See Smoke (1993:901-02) for a short statement on the degree of local autonomy in several developing countries, and the relationship of such autonomy to the colonial experience. 17 “Bureaucratic decentralization” is another name given to deconcentration (Rolla 1998:27-39). 18Adamolekun (1991:285-86) points out that deconcentration often takes place in the name of decentralization and that the two are confused. 19 Political decentralization is also called democratic decentralization by some authors (Manor 1999). Blair (1998:1) writes of democratic decentralization as “democratic local governance.” 20 When powers are ceded from the state to non-state bodies such as private individuals or corporations, the process can be termed privatization, which we do not consider to be decentralization. When, under governmental supervision, powers and specific responsibilities are allocated to public corporations or any other special authorities outside of the regular political-administrative structure, it is called delegation. (See Ostrom, Schroeder, and Wynne 1993: 166.) Devolution describes “the increased empowerment of local organizations with no direct government affiliation” such as NGOs, private bodies, corporations, community groups, etc. (Maniates 1990: 1). 4 Most justifications of decentralization are built around the assumption that greater participation in public decision making is a positive good in itself or that it can improve efficiency, equity, development and resource management.21 By bringing government decision making closer to citizens, decentralization is widely believed to increase public– sector accountability and therefore effectiveness.22 At its most basic, decentralization aims to achieve one of the central goals of just political governance—democratization, or the desire that humans should have a say in their own affairs.23 In this sense, decentralization is a strategy of governance to facilitate the transfer of power closer to those who are most affected by the exercise of power. In the rest of the paper, we use ‘decentralization’ as a shorthand for its political/ democratic form. 21 For arguments defending decentralization on the basis of greater participation of citizens in democratic governance see De Tocqueville (1945[1835]), J.S. Mill (1993:3) and more recently Dahl (1981:47-49). Webster (1992:129) is only one of the later figures to argue that decentralization is “seen as a means by which the state can be more responsive, more adaptable, to regional and local needs than is the case with a concentration of administrative powers” (Bish and Ostrom 1973; Weimer 1996:49-50.) Riker, in a counterview that does not see in decentralization any necessary benefits of responsiveness or freedom, says, “to one who believes in the majoritarian notion of freedom, it is impossible to interpret federalism as other than a device of minority tyranny” (1964:142). For an argument favoring efficiency through decentralization see Boeve (1996:94-117). For efficiency arguments in the context of public choice see Tiebout (1972) and Oates (1972:11-2). 22 Fox and Aranda 1996:1; World Bank 1997. 23 The relationship between decentralization and democracy is discussed by Nzouankeu 1994:214-15 and Souza 1996:529-55. 5 A FRAMEWORK FOR ANALYZING DECENTRALIZATION Many analyses of decentralization consider the transfer of powers in three sectors to be necessary for success. Manor, for example, argues, “If it is to have significant promise, decentralization must entail a mixture of all three types: democratic, fiscal, and administrative.”24 Binswanger, in almost exactly the same terms, asserts, “The three main elements of decentralization—political, fiscal, and administrative—should be implemented together.”25 While political, fiscal, and administrative arenas are critical for statecraft, new tools are necessary to understand the principles that make decentralization effective. We suggest that three distinct dimensions underlie all acts of decentralization: actors, powers, and accountability. Without an understanding of the powers of various actors, the domains in which they exercise their powers, and to whom and how they are accountable, it is impossible to learn the extent to which meaningful decentralization has taken place. In our conceptualization, the political and administrative domains of decentralization are characterized by the mix of these three underlying analytical dimensions. Fiscal powers, through this optic, constitute just one of the types of power that may be devolved in administrative or political decentralization.26 It is not an analytically distinct type of decentralization. Actors in Decentralization Actors in the local arena who exercise powers over public resources may include appointed or elected officials, NGOs, chiefs, powerful individuals, or corporate bodies such as communities, cooperatives, and committees. Each of these actors is typically located in particular relations of accountability and has certain types of powers. These relations depend on the historical, social, and political constitution of the powers of each actor, which may be based on ideology, wealth, heredity, election, appointment, or other factors. Actors may also be differentiated from each other by their beliefs and objectives, or if collective rather than individual, by the internal structure of their organization, their membership, funding sources and the laws to which they are subject.27 24 Manor 1999:7. Manor also suggests, against World Development Report 1997, that such tripartite mixtures are reasonably common. 25 Binswanger 1999:2. 26 Although accountability can be seen as dependent on the relationship between actors and the types of powers they exercise, we treat it as a separate dimension because of its critical role in the politics of decentralization. It is through different mechanisms of accountability that those exercising powers are held to account for their actions. 27 Many analysts do not consider devolution of powers to non-governmental organizations to be decentralization (World Bank 1997; Crook and Manor 1998; Manor 1999). We include NGOs in our list of possible actors since they are often recipients of public powers in the name of decentralization. They therefore deserve analytical attention. Whether a transfer of public powers to an NGO or any other non- state body constitutes decentralization is an empirical question. Such actors are often neither representative nor downwardly accountable (Fox and Butler 1994; Guyer 1994:215-229). 6 Actors are positioned at different levels of social action. Indeed, since decentralization is about changes in how actors at different levels of political authority exercise their power, by definition the actors involved must be located at different levels of action. In actual cases, any one actor or combination of actors may be seen as the appropriate legal bodies toward whom decentralization should occur. Because the dealings of particular actors are impelled by their interests,28 it is likely that the same types of powers devolved to different actors will lead to variable outcomes. Consequently, the nature of decentralization depends to a significant degree upon who gets to exercise power, and the accountability relations to which they are subject. Types of Power We distinguish four broad powers of decision-making as being crucial to understanding decentralization. These powers are a) the power to create rules or modify old ones, b) the power to make decisions about how a particular resource or opportunity is to be used, c) the power to implement and ensure compliance to the new or altered rules, and d) the power to adjudicate disputes that arise in the effort to create rules and ensure compliance. Enlarged powers of decision-making at lower levels of the political- administrative hierarchy in relation to any of the above four categories constitute some form of decentralization. These four types of powers correspond to three more familiar categories: legislative (creation of rules), executive (making, implementing, and enforcing of decisions), and judicial (adjudication of disputes). Further, the classical issues of separation of powers and checks and balances that apply to central governments also have their corollaries in the decentralized arena.29 The power to create new rules is usually held in some domain of decision-making over which governments seek to decentralize control, and in relation to some group of actors. Those who exercise the power to design new rules or modify old ones do so for some kind of resources, and for some groups of people. This set of powers allows decentralized actors to legislate principles that structure decisions and actions concerning who is to benefit from given resources or opportunities, as well as how, and to what extent, they are able to do so.30 Typically, greater powers to make decisions in some domain of action that influences others increases the autonomy of the actor who gains these powers and can be considered a form of decentralization. Such powers enhance the discretionary authority of local bodies, and directly affect the use of resources. Decisions of this type need not affect the behavior of others by prescribing what they must, must not, or may do. Thus if a local body comes to have a larger budget or greater powers of revenue raising, and or greater autonomy to expend the budget as it sees fit, a degree of decentralization has been 28 Interests of actors are always in the process of formation. The particular social, political, and economic contexts in which actors operate are instrumental in the formation and perception of interests, and in constraining interest-based actions. 29 See Oloka-Onyango (1994:463-518). Also see Mahmood Mamdani (1996:145-6) who attributes the failure of the rule of law in colonial Africa to the fact that “…judicial and administrative activity was fused.” 30 For a discussion of these categories in terms of property rights, see Schlager and Ostrom (1992). 7 achieved even if it does not exercise greater powers of rule making. Many decentralization programs result in some autonomy to local governments in raising revenues, or in their spending discretion.31 Such an increase in powers of revenue and expenditure can be seen as contributing to the decentralization of fiscal powers.32 Implementation and ensuring compliance to decisions and rules implies the power to execute, and to meter and monitor whether actors are carrying out the roles they are supposed to perform. It also includes the power to impose sanctions on those who do not subscribe to the tasks they are supposed to perform, and to enforce those sanctions. If a particular group of users are supposed to harvest certain levels of benefits distributed equally among themselves, the power to ensure compliance can include the determination of when they have actually harvested that level of benefits, and whether the distribution has been equal. Rule-makers may have decided upon particular types of sanctions to be imposed on those whose actions violate particular levels of harvest. The power to enforce compliance would also, then, include the ability to ensure that rule-breaking individuals conform to the sanctions imposed because of rule violations. In passing, we should note that devolution of powers to make decisions and rules without the devolution of powers to enforce them can be meaningless; these sets of powers are complementary. Further, all these executive powers require fiscal and administrative resources. It is certainly possible that in a particular decentralization effort, powers to enforce are transferred to administrative branches of the state rather than to representative local governments at the same level. Whether such a transfer of power leads to effective decentralization depends on the nature of accountability relations (see below), the mix of powers that a given actor holds, and horizontal relations among actors at the same level. We suggest that the division between powers of rule-making and enforcement can lead to effective decentralization if the actors who possess the powers to enforce are either easily accessible to those with the power to make decisions and rules, or under the control of those who have the power to make decisions and rules. The power of adjudication is significant whenever new rules are created, or there is a change in the type of decisions made by particular actors. Such changes also signify a modification in the powers of these actors. It is more than likely that when changes in powers take place, contests and negotiations will spill over into the arena of adjudication. Two aspects of adjudication, we suggest are important: accessibility and independence. Local populations who are influenced by devolved powers should have the possibility of appealing to accessible channels of adjudication. Further, these channels of adjudication should be institutionalized so that they do not have structural links with sectoral interests: constituents should be able to challenge the rules, decision, implementation, and enforcement by those who hold decentralized powers, and the outcomes of such challenges should not be biased in favor of power holders. What is critical about powers 31 Therkildsen and Semboja 1992:1101-13. 32 Martin Doornbos of the Institute for Social Studies at the Hague (personal communication, November 30, 1999) points out that decentralization may result in an increase of the power of the central state if, for instance, taxation powers are devolved in exchange for central transfers of funds and the local tax base is limited. 8 of adjudication is not that they be devolved to some representative bodies at the local level. It is more important they be exercised accessibly and without systematic bias. Accountability in Decentralization Rulers claim to be responsible to their people; people try to hold them to account. Accountability is thus the measure of responsibility. John Lonsdale, “Political Accountability in African History”, 1986:127 The allocation of different sets of powers of decision-making and rule-making to lower-level actors creates decentralization. The effectiveness of decentralization hinges on a third dimension: accountability. We suggest that if powers are decentralized to actors who are not accountable to their constituents, or who are accountable only to themselves or superior authorities within the structure of the government, then decentralization is not likely to accomplish its stated aims. It is only when constituents come to exercise accountability as a countervailing power that decentralization is likely to be effective. All modes of accountability are relational. To understand its nature, therefore, it is necessary to attend to the actors between whom relations of accountability exist. Accountability is also about the mechanisms through which countervailing powers are exercised by those subject to actors holding decentralized power. Accountability in this sense, to paraphrase Foucault, is not in a position of exteriority to power, but depends on the exercise of a counter power to balance arbitrary action. Since this paper focuses on the public actors to whom powers are devolved on behalf of a constituency, we are primarily concerned with the accountability relations of such actors downward to their constituencies. It is downward accountability that broadens participation. Actors can be held downwardly accountable to local constituencies in numerous ways. The most commonly cited are electoral processes.33 While elections may be important, they are not sufficient. Many elected officials are not accountable to their constituents—even when the electoral system is well crafted. Other mechanisms for increasing local or downward accountability—of elected or any other local actors— include: procedures for recall; referenda; legal recourse through courts; third-party monitoring by media, NGOs or independently elected controllers; auditing and evaluation; political pressures and lobbying by associations and associative movements; the provision of information on roles and obligations of government by the media and NGOs; public reporting requirements for governments; education; embeddedness of leaders in their community; belief systems of leaders and their communities; civic dedication and pride of leaders; performance awards; widespread participation; social movements; threats of social unrest and resistance; central state oversight of local government; and taxation. Although long, this is not an exhaustive list. All these mechanisms can contribute to local accountability. There is always also some degree of 33 For a study of competitive local elections in West Bengal, India that helped make policy more responsive to the poor, see Echeverri-Gent 1992:1401-22. For a similar argument from Colombia about the importance of competitive elections at the local level, see Fiszbein 1997a:1029-43. 9 upward accountability of appointed and representative actors. Upward accountability can also be structured through many of these same mechanisms.34 Vertical and horizontal ties among branches of government can also shape the relation of accountability between local government actors and their constituencies. Similarly, the relations between customary authorities and their administrative superiors can shape their downward accountability.35 However, downward accountability of those who receive powers from the central state on behalf of a constituency is the primary dimension of decentralization since it can broaden the participation of local populations and enhances the responsiveness of empowered actors. It is through such greater participation and responsiveness that the many lauded benefits of decentralization are realized. 34 For a more developed discussion of this list of accountability means, see Ribot 1999a and O’Donnell 1998:112-126. 35 Personal communication, Doug Porter, United Nations Capital Development Fund, Kampala, Uganda, October 1999. (Cf. Porter and Onyach-Olaa 1999.) 10