1Development agencies criticize African political systems for being states with dis torted structures. Corruption, nepotism, embezzlement or an inflexible, inefficient bureaucracy are seen as typical features of African states. These are often called cleptocratic or clientelist. The proposed remedies are numerous and contain different ingredients, like democracy, good governance, the rule of law and decentralization.
2Decentralization, in particular, is a development project with a long history. However, successful cases of decentralization are hard to find. The reason for this constant failure can only in part be found in the decentralization programs themselves. Processes of decentralization are, like any other institutional and political innovation, shaped by the political and institutional context within which they take place.
3Corruption, nepotism, embezzlement or an inflexible, inefficient bureaucracy can be found across the board in Africa. However, stressing these similarities at the level of symptoms does not take account of the differences which exist among African states. We focus on the issue of decentralization in Mozambique to try to develop an approach for a more systematic analysis of political and institutional innovation in Africa.
4Our paper tries to accomplish two tasks. First, based on a short summary of the decentralization debate we develop a simple framework for the analysis of the political systems in which decentralization takes place. Second, we apply this framework to Mozambique discussing decentralization efforts against the background of a changing political system.
5Our analytical framework seeks to simplify the complexity and diversity of political systems. This is partly due to the fact that this is work in progress, i.e. unfinished business that needs further refinement and differentiation. We do think, however, that simplification can highlight central issues and trigger off critical discussion.
6Since the 1960s and 1970s the focal point of decentralization has remained the same. There is the belief that implementation of development policies can/should be supported by a redistribution of tasks from central governments to the regional and local level.
7The hopes were, and still are, that people will be reached more directly and development activities can react better and quicker to specific needs at the local level. Furthermore, it is believed that this can make the implementation of policies more efficient and effective.
8Currently, the concept of decentralization goes together with the general demand for «good governance», the promotion of the private sector in services (e.g. water) and with «democratization». It is expected that the shift in decision-making to the local level will make room for people’s participation at the same time as effecting a change to a more needs-oriented policy.
- 1 See: Barkan, 1989; Cheema, 1983; Illy, 1986; Illy, 1988; Lachenmann, 1991; Lachenmann, 2001; Leonar (...)
9The experience of former decentralization programs shows typical problems1.
10Analyses usually highlight among other aspects the lack of financial resources at the local level. With only limited income, so the argument goes, additional tasks can hardly be managed. Another common problem is the bad shape in which equipment and administrative infrastructure are, shortage of qualified staff as well as problems in creating reliable, transparent and accountable local institutions.
11A closer look shows that decentralization processes have contradictory outcomes. Decentralization does not guarantee more local autonomy, but may strengthen the national government’s position at the local level (devolution). Even when decentralization takes effect, decision making is delegated to the local level and local communities gain some local autonomy, this does not automatically lead to more participation, more efficiency or even more orientation towards the needs of the population. We may also find new forms of intermediary rule of local leaders who only in part accept the nation state. Power may still be based on military strength, control of local resources, traditional authority and/or on the ability to act as brokers for external resources from the national state or non-governmental organisations (Rösel, 1999). Decentralization is often associated with the image of a local community living in harmony. This is far from reality. Often, it is even unclear who is part of a local community and who is not, who has the right to speak or to decide. Therefore, local gains in autonomy may accentuate local conflicts and open up new political arenas for political disputes among powerful local actors. In such cases, we find a dynamization of local-level politics. The chances for the population to influence local politics vary substantially according to the specific situation.
12Up to now, the official discourse of development organisations has seemed to have underrated the political dimensions of decentralization. The concept of decentralization as used by development organizations implies that the restructuring of administrative organization could be an instrument for a well targeted change of patterns of administrative action and policy implementation. But the intended change can only be attained if the administration accepts and follows the new rules. In Africa, at least, this cannot be assumed. More often than not African administrations have strong personalized elements and do not follow established rules strictly. In fact, the impact of decentralization programs depends not only on the decentralization program itself, but also on the political system at work. At the same time, however, a decentralization program may change the political system. Therefore, we analyze processes of decentralization in a wider framework that includes the political system and the interactions between decentralization and the political system.
13Our theoretical perspective focuses on an instrumental concept of politics. To this end we shall highlight participation and the implementation of political decisions as the central analytical features of a polity.
- 2 Political participation is a necessary but not sufficient condition for democracy.
14Thus, we describe political participation as the active commitment of people in practical-political processes with a given chance of influencing political decisions2. Political participation may take place in formal and informal institutions and according to formal procedures or informal processes. People may participate directly in small communities in a common decision-making process or in larger units via a plebiscite. They can participate indirectly by delegating decisions to representatives. These may be notables, elders, leaders with personal authority or formally elected delegates. This indirect participation may be fueled by public discussions that influence the setting of an agenda, promotes specific decisions and criticizes other political decisions. The role of public opinion highlights an important second element of indirect participation, i.e. accountability. Accountability makes political representatives responsible for their decisions and the impact they have and ties representatives to their constituency or supporters. Except for the rare cases where everyone is directly involved in political decisions participation breaks down into two elements, namely (a) influence on the decision making process and (b) accountability of decision makers. In our view one implies the other, i.e. there can be no accountability without peoples’ influence and vice-versa.
15The opposite of participation is authoritarian rule (non-participation), where decisions are taken by a leader or a small group in power. In such a set-up the people at large are excluded from the decision-making process and decision makers are not held accountable.
16The second analytical dimension of a polity is how political decisions are implemented. We assume two ideal-types: (a) a Weberian-type bureaucracy under the rule of law and (b) an arbitrary mode of implementation of political decisions.
17(a) A Weberian-type bureaucracy under the rule of law is based on a strong administration which implements decisions taken by local political decision-makers. The implementation process is regulated by rules and laws embedded in a general constitution. Briefly, the whole implementation process is under the rule of law. At least ideally, this Weberian-type bureaucracy follows universalistic principles and acts irrespective of persons. We call this the Weberian-type bureaucratic mode of implementation. In the framework of law and established rules the administration acts according to a routine and most activities can be carried out without the requirement that a political decision be taken for every new case. There is a strict normative hierarchy with the constitution on the top followed by laws, by-laws and personal decisions at the bottom (Elwert 2001, 422). This legally bounded autonomy can be seen as a condition for an administration to work efficiently. There is, however, a caveat, namely that administrative efficiency does not guarantee «good» or «successful» policy.
18(b) The opposite of this is the arbitrary mode of implementation, where administration does not follow strict rules. All decisions are single-case matters as a result of the arbitrariness of the authority in power. Therefore, all activities of the administration are highly personalized in a double sense. First, decisions depend on the authority of persons in power and not on rules or laws. Second, the administration always takes into account who will be affected by a decision or activity. In the context of an arbitrary mode of implementation decisions and activities depend basically on power, personal influence, and personal relationships, and are not generally predictable. Compared to the Weberian-type bureaucracy there is a reversal of the normative hierarchy (Elwert 2001, 422): highest are personal decisions of those with authority, followed by by-laws; laws and the constitution are at the bottom. There is no general administrative routine. Every decision risks revision by new political decisions.
19On the basis of a combination of the two analytical dimensions, mode of implementation and political participation, we have constructed four highly simplified ideal types of political systems: (1) liberal democracy under the rule of law, (2) authoritarian state under the rule of law, (3) command state and (4) competitive political clientelism (see table 1).
TABLE 1: TYPES OF POLITICAL SYSTEMS
20These are ideal-types which provide analytical instruments with which to describe reality with reference to specific theoretical constructions. The Weberiantype bureaucratic mode of implementation and the arbitrary mode of implementation constitute a continuum of more or less bureaucratic and arbitrary elements. This holds true for the dimension of participation, too. The question is not participation or no participation, but rather how much participation and accountability, and what type of participation. Our analytical instrument is a two dimensional scale (mode of implementation, participation). The ideal types represent the combination of extremes (see figure 1).
21A Weberian-type bureaucratic mode of implementation combined with political participation describes the ideal type of liberal democratic constitutional states under the rule of law. Decisions are the outcome of a process that includes political participation. They are formulated and implemented as general laws and rules with general validity. Western democracies claim to represent this type of political system.
22A Weberian-type bureaucratic mode of implementation without political participation describes the ideal type of a non-democratic, centralized, authoritarian, constitutional state under the rule of law. Examples that present the main elements of this type are Prussia in 19th century or late colonial Hong Kong.
- 3 Elwert (2001) gives a detailed analysis of typical strategies and features of arbitrariness in Afri (...)
23An arbitrary mode of implementation without political participation describes a strictly centralized, authoritarian state with an arbitrary use of power. Even decisions of minor importance are taken at the top level. Civil servants follow only direct orders (commands) from their supervisors («par ordre du moufti») to whom they are accountable. The top level is overloaded, and minor decisions at the lower levels are taken without clear commands. The top level faces a potential control deficit which is then compensated for by secret service institutions and the promotion of denunciation. However, lower levels gain some freedom of decision on their own (Elwert 2001, 430f). The result of this is the creation of chances for corruption and embezzlement on every level. Arbitrariness goes together with an uncontrolled power of the state apparatus which often results in cruelty. We call this type a «command state» using term Elwert’s (2001; see also Bierschenk et alii 1993) which was applied to Benin during its «socialist» phase, but which can also be an accurate description of other African states (e.g. Rwanda under Habyarimana)3.
24An arbitrary mode of implementation combined with political participation describes a kind of «soft state» (Myrdal 1979) with a low degree of centralization and elements of political competition (see also: Chabal/Daloz 1999). Political decisions may be taken on all administrative and political levels according to their importance and the personal authority and power of office-holders. The striking difference from the command state is a double accountability, both upwards and downwards. Political decisions and administrative activities are directed to the political constituency of politicians and/or to people in the political process who can offer power, influence and resources. Due to participation politicians must serve their constituency, their power base, or else risk losing their position. The power of politicians and administrators is limited by this erratic accountability to the people. We see groups of people fighting for national or local resources in a «competitive political clientelism».
- 4 The voluntary element may be limited because clients have only a limited or no choice between diffe (...)
25This type is hardly found in its pure form. At the same time we find in African political systems elements of this participative-arbitrary mode, albeit combined with elements of authoritarian control. A good example of this mix was the competitive political clientelism of Kenyatta’s Kenya. In the 1990s, when the wave of democratization hit Africa, elements of this last type soared in importance. Political clientelism is based on voluntary personal relations between a patron and a client, two partners of unequal status. They exchange goods and services according to a logic based on the control of the exchange of resources (Roniger, 1994 ;Weber Pazmino, 1991)4.
26The patron offers protection, land rights, security, brokerage or political representation. The client offers labor, services, loyalty and political support. Democratic elections, in particular, offer people the chance to choose their patron. The election of a local member of parliament can include the installation of a patron-client relationship: People offer themselves as clients to competing patrons (politicians running for parliamentary seats). From the client’s point of view there are two main questions to asses the patron (Neubert, 1999a; Neubert, 1999b): what chances does a patron have of acquiring power/influence? will a patron be willing to share the resources he may gain?
27In addition to this two-dimensional framework the results of a decentralization process are also influenced by the relationship between the national state and the local level. Two aspects must be considered in this respect: the presence of the state in local everyday life; the control over (financial) resources.
28The presence of the state in local everyday life can be analyzed following John Dewey’s description of the state as a public space which arises as a result of attempts to render social life predictable and ordered (Dewey 1991 (1927)). In this sense, the presence of the state in local settings can be indicated mainly by the following aspects:
29– the assertion of the monopoly of violence,
30– the regulation of the local legal order,
31They can only decide whether they accept a patron-client relationship or not. However, without a minimum of voluntary decision the relationship may change to slavery, serfdom or suppression. Often, this voluntary element of patron-client relationships is not effected in the discussion.
32– the ability to make local people subject to national taxes or the influence of state services like technical infrastructure (roads, water, electricity) and social services on everyday life (schools, health services), or extension services.
33We may distinguish between areas and states where the national state is present at the local level (capitals, states like Rwanda, South Africa) and those with an absent state (Northern Kenya, rural areas of the Democratic Republic of Congo, Central African Republic).
34The control over (financial) resources highlights a simple fact: as long as the local level relies on government subsidies, it will remain under the scrutiny of central government. Gaining more local autonomy is linked to the ability to create own resources over and above such subsidies. This may be in the form of local taxes, control over natural resources or direct access to development finance.
35This analytical framework should help us describe the varying impact of decentralization programs in a systematic way in relation to the political system. A few simple examples where the basic features of the political system remain unchanged can show this relationship. Decentralization may have different effects in authoritarian and participative systems. Under authoritarian rule decentralization may be used by the central government to penetrate into the local level, especially in areas where the state may previously have been only weakly represented (devolution). Decentralization may also strengthen the state’s control over its citizens and limit local autonomy. Decentralization in a participative system opens up or strengthens local level politics, especially when the local level government has, or can find, access to own resources. In cases where the national state controls key resources this counterbalances possible gains in local autonomy.
- 5 Even in countries that claim a Weberian-type bureaucratic administration like Western European stat (...)
36Decentralization does not automatically influence the mode of implementation (Weberian-type bureaucracy or arbitrary), quite the opposite. The arbitrary mode of decision-making and implementation may be intensified by decentralization, for local level politics generally includes personalized decision-making and implementation5.
37The combination of more administrative efficiency and democratization that should lead to good governance will work best in a participative-Weberian-type bureaucratic system. Whereas the discussion on decentralization has tended to focus on financial, technical or organizational problems this shows that decentralization is also an issue that has to do with the basic features of a political system.
38Key issues in this regard are:
39How does a given political system and the activities of a decentralization project interact?
40And is decentralization a useful means to support good governance, efficient administration, democratic participation and needs-oriented local level policy?
41Our case study draws from the experience of decentralization in Mozambique. We wish, in particular, to highlight the importance that the nature of politics may have for the failure or success of a process of democratization. Mozambique is interesting for several reasons. First, because it has been undergoing a process of decentralization under conditions typical of many African countries. These include a structural adjustment program and the increasing importance of development and humanitarian aid. Secondly, though, after independence Mozambique chose a revolutionary path to economic and social emancipation typical of very few countries in Africa (Mali under Modibo Keita, Angola, Mengistu’s Ethiopia, Zimbabwe, etc.) the failure of which may have posited decentralization not so much as devolution, but rather as building the state anew.
42In our discussion of the revolutionary state we start with a description of the nature of the post-colonial political order. In the process we will seek to show that the ensuing revolutionary state had two distinct stages, both of which were the result of shifts in the mode of implementation of political decisions. Second, we want to argue that the immediate context for decentralization in Mozambique was the shift in the mode of implementation and that, therefore, to properly grasp the significance of the process of decentralization it is necessary to wrestle with the political system.
43When Mozambique became independent in 1975 Frelimo inherited a political system which in terms of the analytical framework suggested above can best be described as one based on a Weberian-type bureaucratic mode of implementation without political participation. Portuguese colonial rule was based on a rational bureaucratic apparatus modeled on the Portuguese civil service. From the time when Portuguese sailors landed on the coast of Mozambique at the end of the fifteenth century and declared the territory part and parcel of the Portuguese seaborne empire their rule had been based on formal laws, edicts and decrees passed by the Crown. Moreover, those sent to the colony for service, whether military or civil, were part of an established bureaucratic hierarchy based on merit as legally defined in Portugal.
44Portuguese rule in Mozambique drew its strength from the monopoly of the means of legitimate violence which the colonial state had been able to achieve by 1920 after the defeat of the last focus of resistance in the north of the country (Newitt 1995). This did not translate into an accountable political system, though. Decision making institutions such as the civil service, the governor-general and district officers were staffed by people appointed in Lisbon. Following hard on the steps of fascist practices in Portugal itself, settlers and Africans had no political rights, except for the possibility of electing council officials as far as the former were concerned. Mahmood Mamdani’s description of the colonial system as one based on the distinction between citizens and subjects (Mamdani 1996) captures the essence of political participation in colonial Mozambique. Indeed, while the European population could exercise some kind of political rights through formal participation at the local level and thereby enjoy some form of «citizenship», political participation for the majority of the African population was constrained by colonial regulations which contemplated traditional political authorities as the proper arena for Africans’ political participation.
45The Portuguese colonial system had elements of the British system of «indirect rule». The Portuguese revived, strengthened and stabilized traditional political authorities and made them the instrument of their colonial rule. In this sense, the so-called «regulado», the traditional authorities, turned African political participation into an artefact of bureaucracy. In other words, the «headmen», i.e. those who were supposed to be the political representatives of the population, were actually the representatives of the colonial administration before their own communities.
46Frelimo’s revolutionary government sought to dismantle the edifice of colonial rule by abolishing the «regulado» and by defining the new state as one based on people’s power (Machel 1975). General free elections in a liberal democratic sense were not seen as necessary. Frelimo justified the abolition of the «regulado» on the grounds that «headmen» had been «collaborators» who had taken part in the subjugation and oppression of their own people. A combination of nationalist rhetoric and the modernization discourse on which Frelimo’s Marxist ideology was premised seems to have been behind this hostility towards «headmen». Traditional authorities were seen as an aspect of the traditional society which had to be overcome if socialism was to be built. Nationalist rhetoric, for its part, drew from the need to form a national consciousness out of the peoples and cultures which colonial rule had brought together. The segmentary structure of the «regulado» was seen as an obstacle to the formation of this national consciousness.
47«People’s power» consisted of the idea that a political system that aimed at doing away with the «exploitation of man by man» should strive for forms of political participation which devolved power to the «oppressed masses» (Machel 1976). In order to achieve this the liberation movement was transformed into a party in 1977 at its third congress (Munslow 1983). On the same occasion it declared itself «the vanguard» of the people which, at least in theory, meant that it represented its aspirations. Moreover, the state was defined as the instrument through which the people, through the party, would exercise its power. In practice, however, the «vanguard party», supported by constitutional provisions which outlawed political parties, exercised its power over the people through the state. The principle of «democratic centralism» was introduced which lent legitimacy to the leadership claims of the party and made room for local decision-making units directly accountable to the party. Much like in the colonial period, people’s political participation was limited to forms of authority which were not accountable to the people itself, but to a centralizing bureaucratic apparatus. Indeed, under the revolutionary government Africans remained, to use Mamdani’s apt words, «subjects» and hardly became «citizens».
48To the extent that Frelimo’s rule was painstakingly codified within the general framework of building a socialist society it rested on highly formalized bureaucratic procedures. These procedures defined the work of the state, which consisted in the main in the implementation of party decisions. The state had an overbearing presence in society not only through rules and regulations but also through the way in which it concentrated economic and social activities in its own brief. The nationaliza tion of important social services such as health provision and education as well as a planned economy along Soviet lines defined the framework within which the state operated. Indeed, in the early years of independence this welfare role which the Mozambican state took upon itself became an instrument for lending legitimacy to «people’s power».
- 6 I (EM) remember a rally in the late seventies where the present head of state, Joaquim Chissano, pr (...)
49Consistent with the declared socialist goals the state took it upon itself to fashion a nation of «peasants and workers» (Saul 1985) out of a society fragmented along lin eage, racial and regional lines. To this end social relations were codified along Marxist lines down to the smallest detail rendering the state and the party omnipresent in everyday life. Curiously, however, the cavalier manner in which the state and the party treated the population at the level of implementation of political decisions – building a socialist society of workers and peasants – did not initially translate into less legitimacy for the political system. In fact, if popular enthusiasm for the new political structures, language and goals is anything to go by, people seemed to submit to the new political authority quite voluntarily. Within the rigid political arrange ments of the revolutionary era people could hold politicians accountable. The early Frelimo of the immediate post-independence period was a «puritan» party which observed a strict moral and political code of conduct (Hanlon 1984). This often found expression in mass rallies in which the people were encouraged to pinpoint the moral failings of the leaders, often leading to the demotion or outright sacking of the latter6. These events were a kind of symbolic and strictly limited form of participation that excluded critique of the party or party policies as such.
50In terms of the ideal-types we use in this paper the early revolutionary stage of the post-independence period could be classified as one based on an authoritarian state form under the rule of law. Building socialism in Mozambique in those days was a highly bureaucratic project firmly entrenched in law – euphemistically called «socialist legality» which failed to provide room for the articulation of dissent and alternative visions of society. Political participation was limited to the rigid channels of the prevailing political order, a fact which severely constrained participation. The revolutionary period had a later stage, which was marked by a shift in the mode of implementation of political decisions from authoritarianism under the rule of law to command-like state based on personalized decision-making and contempt for the rule of law. This shift in the mode of implementation of political decisions was directly related to the evolution of the political, economic and social situation, which from the late seventies until the mid-eighties took a turn to the worse. In a sense, this marked the failure of the revolutionary model. The form that this failure took provided the immediate background against which decentralization took shape and acquired meaning.
51In the literature on Mozambique there are several explanations for this failure. They range from a critique of the political orientation itself all the way up to a critical appraisal of central policies. Michel Cahen, for instance, questions the Marxist credentials of Frelimo and argues that the ideology was simply used as a cover for a small «creolised» elite’s will to power (Cahen 1987). Joseph Hanlon sees the reasons for the failure in the increasing bureaucratic orientation of the state which severed the party from what he believes to have been traditional links with the population (Hanlon 1984; see also Adam 1992). This argument is taken up by John Saul, who deplores serious mistakes in the process of building socialism (Saul 1985; Saul 1993). Several other authors take issue with Frelimo policies such as what they see as «forced» villagization (Bowen 2000; Chingono 1996) and ignoring the cultural traditions of the country (Geffray 1990) as well as too rigid and orthodox an interpretation of Marxism (Kößler 1992).
52To be sure, all these factors played a role in the failure of the revolutionary project in Mozambique. Indeed, it was the explosive combination of all these factors that brought the country to its knees. The war of aggression waged by the Apartheid regime, for instance, combined with armed insurgency, popular discontent in the face of increasing economic hardship, rural insecurity, natural hazards and political stag nation to deal a heavy blow on Frelimo’s hopes of a socialist society. While a defini tive account of the reasons for this has still to be written, there are reasons to assume that no single factor stands out, although the civil war ranks among the most important ones.
53The post-revolutionary period starts actually in 1979/80 with two significant events. First, the refusal of COMECOM countries to accept Mozambique into their fold and, second, the start of the civil war in earnest. These events would lead eventually to Mozambique’s turn towards the West in the form of the introduction of a structural adjustment program in the mid-eighties and a growing dependence on development aid. Reflecting the euphoric mood of the revolutionary period Frelimo had declared the eighties the decade in which underdevelopment was to be eradicated. This was based on a modernization blueprint that would have been the pride of any development theorist (Egerö 1987). This blueprint assumed continued support from Eastern bloc countries to compensate for the lack of access to capitalist dominated markets and a performing agricultural sector that would fuel industrial development (Schoeller 1992; Wuyts 1989). COMECOM’s refusal to accept Mozambique severely undermined the premises of the blueprint. Furthermore, with majority rule in Southern Rhodesia the Mozambican armed gangs which had been created by the Rhodesian secret services were taken over by the South African military and turned into an efficient terrorist and destructive machine that made any economic activity in the rural areas unfeasible. Added to this the country was assailed by a devastating drought that by 1985 had claimed 100,000 lives.
54Our analytical model allows us to capture the dynamic nature of politics in Mozambique. As indicated above the post-independence period should be divided into two stages. While in the early stage the revolutionary state was near to our type of an authoritarian state under the rule of law in the later stage it increasingly resembled a command state. The authoritarian state had placed emphasis on the so-called «socialist legality» and on ensuring that its decisions were taken and legitimated within a legal-bureaucratic framework that left little or no room to personal and arbitrary authority. This almost pedantic reference to rules and regulations gave the revolutionary state a level of legal-bureaucratic coherence which placed it nearer to a Weberian-type bureaucratic mode of implementation of political decisions. While political decisions were constrained to observe ideological purity, and thereby sometimes conflict with more rational courses of action, from a pragmatic rational point of view the state discharged its functions within a clearly defined and predictable legal framework. This framework left no room for real participation.
55At the same time, the early revolutionary state enjoyed a high degree of legitimacy. This legitimacy, however, was directly related to the ability of the state to dis charge its functions. With growing economic and social hardships, however, it became increasingly difficult to justify party political claims over the state. In a sense, the legitimacy of the revolutionary project rested mainly on the ability of the state to deliver. The gradual errosion in state capabilities placed the party under strain and its response to the crisis was increasing reliance on arbitrary forms of political decision-making. Indeed, there was a shift away from a Weberian-type bureaucratic mode of implementation to a an arbitrary one based on the personal authority of the leader of the ruling «vanguard» party. The later revolutionary state became in this way more like a command state owing still very little accountability to the people. As the situation continued to deteriorate well into the eighties the ruling party sought assistance from Western countries, all of which imposed a structural adjustment pro gram with the IMF and the World Bank as a precondition for any kind of economic aid. At the same time as these demands were made from outside there were also domestic calls for more accountability and popular influence on the decision-making process. Enter decentralization. Decentralization as one of the mainstays of development came to Mozambique via several roads. These roads had to do with the gener al critique of Mozambique’s modernization strategy, lending legitimacy to develop ment aid and the pre-conditions for economic growth.
56As we have seen, at independence Mozambique embarked upon a modernization project strongly influenced by Marxist ideology. This modernization project was based on two strong assumptions, namely, first, that the goal of modernization was not only the eradication of underdevelopment, but also the creation of a socialist society based on a workers-peasants alliance and, second, that the project was aimed at creating a «new man», i.e. one emancipated from the oppressive weight of tradition. The first assumption found political and institutional underpinnings in the policy of «democratic centralism», i.e. the concentration of political decision-making power in the «vanguard party». The second assumption was predicated on the creation of new organizational structures at the local level which saw no role for traditional political institutions.
57As the revolutionary state encountered mounting economic, political and social problems those who set about searching for the causes leveled responsibility on the assumptions behind the modernization strategy itself. This critique was particularly articulated by those who felt that Frelimo’s radical politics was making tabula rasa of traditional African culture (Geffray 1990). This was seen as a recipe for disaster. The critique concentrated especially on Frelimo’s policy of forced villagization and the dismissal of traditional political authorities (the so-called «regulado») from power. Both policies were seen not only as having contributed towards the alienation of the majority of the people from the party and state, but also as a setback to the declared goal of emancipating the people from the shackles of tradition. Decentralization pro grams in the country ritually include references to the need to rehabilitate traditional cultural institutions as the safest and politically sound way to consolidate democracy.
58Added to this is the ongoing debate in Mozambique concerning what people from the north and centre of the country perceive as the concentration of power by southern elites. Indeed, this debate has become a central issue in national politics. The perception is that Frelimo is a southern party that used Marxism and nationalism to advance its sectarian goals. The main opposition party, Renamo, has been quite successful in portraying itself as the defender of the north and centre against southern hegemony. In spite of its chilling record of terror against civilians during the civil war (Finnegan 1992; Gersony 1988; Hall and Young 1997), Renamo captured a sizable portion of the vote in the two general elections held in Mozambique since the signing of the general peace agreement in 1992. Election analysts in Mozambique (Serra 1998) believe that this share of the vote correlates with Renamo’s regional discourse. Consequently, calls for decentralization on the part of the opposition are also attempts at reducing the power of southern elites over the country.
59Up until 1980 the revolutionary state had been able to remain in overall control of the development process (Hanlon 1991). This was partly due to the fact that up until then the self-confidence of the authorities, bolstered by good economic and social indicators as well as the general euphoria over the building of a new society, was such that foreign agencies had to look for niches for themselves within an established institutional framework. With increasing economic difficulties made worse by a devastating drought and the ever mounting insecurity in the wake of the civil war, Mozambique became more and more dependent on foreign relief agencies for providing succor to its own population. With mounting confidence relief agencies began to see the Government of Mozambique and its bureaucratic procedures as an obstacle in their efforts to provide relief and initiate development at the local level (see particularly Hanlon 1991).
60First relief organisations and then development organisations became increasingly vocal in their criticism of Mozambique’s state organization. They resented having to submit their relief operations and small development projects to the scrutiny of central state institutions and pressed for direct access to the population in need. Their unease with the role of the state and the pressure they exerted for more freedom of movement gained the quality of justified arguments in the international neo-liberal context of the 1980s and early 1990s. Ironically, the rise of the New Right in Britain and the US with its neo-liberal rhetoric helped aid and relief organisations to argue that only they had the expertise and adequate institutional arrangements to create the conditions for market forces to respond. The problem with Mozambique, they argued, was that the state wanted to regulate everything, even those areas where aid and relief agencies had proven more competent elsewhere. Decentralization, as far as these organisations were concerned, was not only a way of reducing the constraining impact of state regulation on the provision of relief and development impulses, but also a strategy for limiting the influence of the state at the local level in order to render the latter more autonomous and a legitimate partner of development agencies.
61The mark left by neo-liberal rhetoric in international development has come to be known as the «Washington consensus». The IMF and the World Bank arrived in Mozambique in the mid 1980s when the country was in the throes of civil war and general economic chaos. True to their standard practice the policy recommendations which both institutions made and enforced assumed that Mozambique’s way out of the crisis would have to be an increase in exports and, more particularly, a scaling down of the state. The latter assumption was captured in the policy of «deregulation». In Mozambique deregulation took the form of dismantling the remnants of the revolutionary project through privatization, cuts in the social budget and encouraging private enterprise. Accordingly, the state disengaged from the economy by selling state companies, withdrew from its social commitments by cutting down on its health and education expenditures and, more generally, narrowed its economic policy focus to macro economic issues relinquishing, thereby, its ordering role in society to a heterogeneous combination of private voluntary organisations, private enterprises and a whole host of foreign official development agencies (for a more extensive discussion of these issues see Abrahamsson and Nilsson 1995; Hanlon 1996; Knauder 2000; Kyle 1994; Macamo 1999; Pitcher 1993).
62In this neo-liberal context decentralization appeared as a necessary precondition for economic development. Structural adjustment was not only about rebuilding the country’s economy, but also about doing so according to the terms laid down by the «Washington consensus», namely deregulation and privatization.
- 7 Sabine Fandrych’sstudy (2001) provides a well-researched analysisof theprocess ofdecentralisation i (...)
63The institutional program that gave coherence to all of these three roads was a decentralization project funded partly by the German Technical Cooperation agency (GTZ). This was based in the Ministry for State Administration (MSA) and was budgeted at over DM 8 million over an initial five year time-frame7. The main policy decisions taken by the ministry in the context of this decentralization project were consistent with the first two roads, namely critique of modernization strategy and lending legitimacy to development aid. The decisions were of a double nature. Firstly, they reopened the debate on the role of traditional authorities in the process of democratization. To this end researchers at the MSA argued that the traditional system of political authority had been perverted by the Portuguese colonial system (Lundin, Machava, and Mozambique. Nucleo de Desenvolvimento Administrativo. 1995) and that by abolishing it at independence the revolutionary state had acted upon a caricature. The subsequent development of the country, so the argument went, had shown that these political structures maintained their relevance to the people, as the massive support of the rural population for the armed rebellion against the central state seemed to suggest. One practical result of this policy was the approval of a law introducing elected council bodies in 33 cities and towns in 1997. Secondly, the policy decisions were geared towards rehabilitating the state itself before society by introducing the principle of subsidiarity in the administrative apparatus. «Civil society» in the form of voluntary private organisations were seen as key intermediary actors in this undertaking whose small-scale presence at the grassroots level would help empowering local communities and devolving responsibility for policy failures. This approach was particularly obvious in legislative initiatives such as the reform of land law, which gave local communities primacy over the central state in decisions over land entitlements.
64While it seems too early to assess the impact of decentralization rhetoric as well as decentralization projects such as the one carried out by the GTZ there are several remarks that can be made concerning the context and future scenarios. As indicated further above decentralization was not merely yet another piece of developmental mumbo-jumbo in Mozambique. There was an internal social, political and economic context which called for it. The recent political history of the country provided the immediate background. Subsequently, though, the different interests which were articulated within this context seemed to have acted as constraints on the process itself. One particularly important set of interests was that articulated by the main opposition party which manoeuvred itself into a dilemma somehow representative of the limits of decentralization in developing countries.
65The cease-fire agreement under the auspices of the UN (Alden 2001; Cabrita 2000; Chan and Venâncio 1998; D’Agnino 1999; Durch 1996) which ended Mozambique’s bloody war provided for democratic elections. Renamo had to transform itself into a political party, a process which included the formulation of a political program. Given that throughout the civil war it had argued that it was fighting for the rehabilitation of traditional cultural values its re-invention as a political party was premised on the articulation of the grievances of all those who bore a grudge against the central state. Accordingly, Renamo sought to convey the public image of a party of local communities and their traditional political structures pitched in battle against an intolerant state. This identification of the party with such interests proved an effective motivational gimmick, the long-term success of which was premised on Renamo’s conquest of central state power. Herein lies a fundamental contradiction. Renamo’s power base was by definition away from the central state, but in order to translate its authority into real power Renamo had to seize state power. Accordingly, Renamo consistently mobilized all its resources towards this objective. This has included, among other things, the refusal to take part in council elections (which would have provided it with an institutionally legitimated local power-base) as well as the rejection of election results as fraudulent even though they were deemed free and fair by international observers, some of whom were traditional Renamo backers. At the same, though, Renamo has been pressing for the right to nominate provincial governors and district administrators in those regions where it was the most voted political party.
66The strategy of the main opposition political party follows in part an internal logic which has to do with Renamo’s difficulties in re-inventing itself as a political party. At the same time though, it is a response to the rhetoric and practice of decentraliza tion in the country. As such it is a commentary on the shape of things to come, for it expresses a dilemma faced by those pressing for decentralization. On the one hand there is an acknowledged need for devolving political decision-making and investing local structures not only with nominal power, but also with resources to muscle up that power. On the other hand, however, devolution under these terms is only possible within the framework of a central state strong enough to pursue the goals of the program in a consistent manner and against claims to authority which undermine its integrity.
67The demands of the opposition are not for more autonomy for local communities and lower level administrative structures. They are rather arguments for more power for itself in those levels. Similarly, when development agencies call for more decision making competency in lower levels the outcome is not more local level democracy, but rather more freedom for these agencies in the implementation of their development and humanitarian activities. The same goes for international finance institutions such as the World Bank and the IMF. Decentralization does not increase the ability of local authorities and communities to formulate their own investment and growth priorities, but rather it allows private businesses to circumvent the central state in their bids.
68The rhetoric and practice of decentralization in Mozambique provides thus an institutional framework for the articulation of particularist interests. These interests do not necessarily reflect those of the local communities on behalf of whose autonomy decentralization is being implemented. In the present context of decentralization such particularist interests serve as an intermediary between the central state and local communities. The resulting arrangement is a potentially clientelist political order. Indeed, in their search for financial resources local communities may seek sponsorship from NGOs which, for their part, are also in search of projects to be funded. As a result of these overlapping interests there may be a supply-led process of development (see Neubert 2000) within which local communities do not necessarily enjoy direct political participation. Rather, they may exercise their right to be heard through the agency of patron institutions, i.e. NGOs, with a certain leverage on the central decision-making organs. Via this indirect chance participation an element of accountability towards the people is introduced into the political system. This development is not entirely consistent with the aims of decentralization, and yet it reflects what is possible within the historical framework of the Mozambican political system. In this sense, it is fair to assume that the decentralization of the revolutionary state goes hand in hand with the transformation of the command state into some sort of competitive political clientelism. In other words, decentralization is not leading directly to devolution, but to the creation of institutional and informal mechanisms for local communities to have their voice heard in central organs. This is mainly done by particularist interests canvassing for space away from the state.
69We started from the theoretical premise that decentralization is a process that is deeply related to the particular political system in itself. While this may appear obvi ous it is important to stress its relevance by drawing attention to the fact that the fortunes of decentralization are to a large extent linked to the political context. Decentralization in a polity characterized by a Weberian-type bureaucratic mode of implementation is less problematic than decentralization in a polity based on an arbitrary mode of implemmentation. In the former decentralization is indeed about devolution and rendering administration more efficient, whereas in the latter it may pri marily be about escaping state influence or reconstructing state forms.
70Our case study seems to confirm our theoretical premise. Indeed, decentralization in Mozambique seems to have been linked to the overall changes in the political system. While in the immediate post-colonial period Mozambique’s political system showed features of a Weberian-type bureaucratic mode of administration with low participation the current political arrangements seems to have more participation. There are reasons to believe that decentralization does not match the expectations of the development agencies. Especially the simple hope that decentralization plus political participation will enforce each other and lead to a liberal democracy did not materealise. There were three reasons for this.
71First, the evolution of the state after independence was marked by a shift from an authoritarian state to a command-state. In other words, the political system evolved from a mode of implementation strongly anchored in law and rules to an arbitrary one. We suggested that this shift might be accounted for by the economic and social problems faced by the revolutionary state.
72Second, decentralization, which came in many guises, actually strengthened the arbitrary mode of implementation by encouraging and giving legitimacy to forms of authority not directly sanctioned by the state. Interestingly, though, this increase in arbitrariness was compensated by a corresponding increase in participation. The presence of development agencies and their support for decentralization made room for the articulation of local concerns within political grass-root contexts. However, the support accorded to local level politics by development agencies may have undermined the power of the state over the distribution and application of external funds in local settings.
73Third, and finally, access to development funds at the local level can be counterproductive as far as participation is concerned. Donors’ criteria for the allocation of funds may gain supremacy over local priorities and, moreover, slow down activities geared towards tapping local resources. This may have the effect of promoting «extraversion» to use Bayart’s term, i.e. the general orientation of local actors towards maximizing their external dependence.
74The failure of decentralization to match the expectations of development agencies is not a comment on the idea itself, but rather on the failure to take the political system into account, against the background of which such policies are implemented. In Mozambique it is precisely this failure that leads us to assume that decentralization heralds a new political system based on competitive clientelism.