1The panel heard and discussed seven papers, not all of which were equally closely related to theme of the panel. It is, in fact, reasonable to ask whether we are talking about the same thing when we address the problems of municipal government in an industrial city of four million inhabitants like Johannesburg (Etzo) or the national politics of remembrance in a capital city like Windhoek (Gewald) and the issues concerning outlying rural communities of several thousand people in Senegal and Mali (Kaag, Nijenhuis).

The papers

2In his paper entitled "From the Old Location to Bishops Hill : the politics of urban planning and landscape history in Windhoek, Namibia", Jan‑Bart Gewald (Cologne) analyzed national politics of remembrance as expressed in city development in Windhoek, Namibia. He showed that the establishment of a new suburb (Hochland Park) with a completely new street pattern wiped out the memory of an important event in local and national history, i.e. the forced and violent removal of African town‑dwellers in the 1950s, while both the colonial conquest and the anti‑colonial struggle are represented in urban space by public monuments. Thus, official memory appears to be highly selective. In the ensuing discussion, the question was asked ‑ but not conclusively answered ‑ as to whether there was a political agenda behind this selectivity, or whether the latter should be seen as the effect of a multiplicity of contingent causes.

3Felicitas Becker's (Cambridge) paper on "Corruption and the decline of local politics in Southeast Tanzania" started with a brief summary of state‑locality links in Tanzania. Against this historical background, she then argued that with the increasingly marked local presence of the state, the former semi‑autonomy of local political arenas and local roots of political legitimacy are being increasingly eroded and direct access to the state becomes a major source of local political power. Despite the fact that the central state in a way overreached itself with the villagization campaign, it penetrates deeper into the lives of villagers than the colonial state ever could. The scope of local intermediaries acting between the state and villagers has diminished as a result.

4These findings were partly contradicted by EIke Grawert (Bremen) in her paper on "The decline of service provision in rural Tanzania and attempts to reverse it". Her main subject was the contradictory nature of decentralization, which involves, on the one hand, the disengagement of the central state from service provision, while on the other hand the state maintains control over local politics through powerful commissioners who are not only centrally appointed, but also remain loyal to the ruling party (giving rise to a situation of de‑facto one‑party rule in a formally multi‑party system). While the official motive for decentralization was to increase government legitimacy, the primary driving force behind it was the foreign donors. Strong vested interests in the government and party apparatuses actually tried to block it. The main problem surrounding decentralization currently remains the financial viability and autonomy of local elected bodies. Funds come mainly from foreign donors which makes these bodies more accountable to the donors than to their political constituents. The main local beneficiaries of the financial extraversion of local elected bodies are groups and individuals who manage to organize themselves and thus become the privileged owners of aid. Citizens can not rely on service provision as it depends on complex negotiating procedures between local councillors, representatives of the central state, foreign donors and, in some cases, wealthy local patrons.

5In his paper on "Multiparty municipal administration and corruption in Cameroon today", E.S.D. Fomin (Buea) addressed the potential link between multipartyism and corruption at local level. Based on three case studies, he showed that the extent of local corruption is not linked to the democratic (elected municipal council) or undemocratic (local government via central government delegates) character of the local political system, and is, in particular, independent of which party (the nationally dominant CPDM or an opposition party) governs locally. Elected local politicians in general and the mayors in particular have a certain room for manoeuvre, whatever the type of party regime. While there is, in fact, generally a high level of corruption at local level, not all municipal councils are equally corrupt ; they can play the political game differently, irrespective of whether the y are members of a democratically elected opposition party or CPDM delegates. In other words, similar political configurations can produce different degrees of corruption. Thus, multipartyism at local level is not per se a barrier against corruption.

6In her study on "Local government and social movements in South Africa", Sebastiana Etzo (Naples) examined a social movement fighting for service provision, in this case electricity supply in Soweto, Johannesburg, as part of a general campaign directed against the privatization of public services. The citizens' organization in question (the Soweto Electricity Crisis Committee) was the only case of a textbook "Civil Society Organisation (CSO) ", so dear to the World Bank and other donors, presented at our panel. Paradoxically the CSO in question turned its back on World‑Bank‑sponsored privatization policy. In fact, when read against the background of the other papers presented at this panel, Etzo's study would appear to confirm the thesis of South Africa's uniqueness within the continent in other respects also : e.g. the sheer size of the municipality (four million inhabitants), the high level of service provision (70% of the population have access to electricity) and the relatively high legitimacy of a democratic state strengthened by the "liberation ethos".

7The panel concluded with two papers which were particularly closely related thematically, both to each other and to Eric Hahonou's paper on "La chefferie coutumière face au projet de decentralisation dans une localité de l'Ouest africain" which was presented at the plenary session. All three papers deal with small rural communities in the West African Sahel and concern francophone countries (Niger, Mali, Senegal). In her paper on "Exploring the context of service provision in Senegal : Social dynamics and decentralisation in the Senegalese countryside" Mayke Kaag (Amsterdam) argued that decentralization is not the direct road to democracy but feeds into ongoing processes of inclusion and exclusion : while it will open up new arenas for political contest, it will not automatically prevent elite continuity. Decentralization will not automatically lead to greater transparency and efficiency as is claimed by the "democracy and decentralization narrative" which dominates the development discourse ; it may also contribute to a proliferation of norms and ideas, thus leading to greater administrative and political complexity and opaqueness instead of the increased accountability of local politicians and increased control of local people over local political arenas.

8This theme was also taken up by Karin Nijenhuis (Leiden) in her paper on "Decentralisation creating a new manoeuvring space and new power positions for local traditional authorities : An example from Southern Mali".

9Based on a detailed ethnographic account, she argued that decentralization most probably increases the fragmentation of local power structures and that it may even create a local power vacuum. Old institutions do not automatically disappear when new institutions are introduced. Instead, a configuration of multi‑layered power institutions is created which interact and compete in flexible and complex ways.

Conceptual discussions

10As is habitual in this kind of exercise, the presentations prompted conceptual discussions of key terms. In fact, all of the key terms used in the title of our workshop, from "local" to "politics" and "decentralization", were seen as problematic ‑ indeed, sometimes extremely problematic ‑ by one or the other of the participants.

11In particular, it was proposed that the notion of "locality" should be "deterritorialized" and that the notion of "politics" should take into account the differentiation between "government" and "administration" proposed by Mamadou Diouf in his keynote speech. However, as is equally common in these contexts, the participants were more active in pointing out the weaknesses in proposed definitions than in providing alternative and better definitions themselves. For example, while the differentiation between government and administration appears suggestive at an initial glance, on second thought, I would be inclined to believe that in the context of empirical studies or actual political practices as opposed to theoretical discourse, it would probably be difficult to identify administrative practices that are completely devoid of politics .

12On the other hand, it is certainly true that the notion of decentralization makes more sense in a context informed by French Jacobinism than one inspired by British concepts of local government. In this respect, the participants correctly remarked that the workshop failed to sufficiently discuss a perspective to which it was particularly well suited on the basis of its composition : the systematic differences, if any, between decentralization politics in French‑speaking and English‑speaking parts of Africa. With regard to the term "decentralization" itself, it should be kept in mind that in official development discourse, it is usually used in connotations originating in management and not political theory.

Policy considerations

13On a more policy‑oriented level, the panel stressed that any true decentralization process necessitates the decentralization of funds. As long as local elective bodies are underfinanced, they cannot be expected to fulfil the roles assigned to them. The question was also raised as to whether it was constructive to reduce the notion of democracy exclusively to multipartyism (as is currently the practice), or whether other forms of democratic representation might not be better adapted to the realities in many African countries.

General conclusions

14On a more general level, the workshop was quite successful in ‑ to adopt the current social science jargon ‑ "deconstructing normative decentralization narratives", which are fashionable in the development world at present. These narratives include, in particular :

• the democracy and grassroots narrative, 

• the accountability narrative, 

• the governance narrative.

15To start with the last, i.e. the governance narrative, decentralization does not constitute a barrier against corruption (as shown by Fomin). Under certain circumstances, it can even increase corruption, especially under the conditions of multiparty democracy. With regard to the democracy and accountability narratives, it would appear that decentralization often leads to a hybrid, composite form of local government and produces further fragmentation of local politic al arenas and informalization of political practices (a point made most forcefully by Kaag, Nijenhuis and Hahonoll ‑ the latter in the plenary session). This not only leads to a high degree of dilution of power at local level, with different veto powers bloc king each other, it also creates a need for constant negotiation between politic al actors, thus reducing the predictability of political processes, and the accountability of local political institutions. The norm appears to involve a composite or hybrid form of local governance with blurred boundaries between the state and private organizations (including Northern NGOs), the national and the local. Consequently, local political arenas after decentralization may be characterized by even lower capacities for regulation and mobilization than previously existed.

16The question was also raised whether the study of decentralization informs us about long‑term changes in African societies, in particular with respect to processes of state formation. No firm conclusion could be reached on this point as the ambiguity of the situations discovered and analyzed makes it difficult to read the exact direction the trends are taking. This difficulty was, of course, partly prompted by our preference for the analysis of micropolitics, e.g. the local political dynamics around decentralization, and our failure to pay sufficient attention to their significance with respect to changes in vertical relations between the local and the central state. We can conclude, however, that decentralization undoubtedly complicates political games at local level. While decentralization signifies the opening up of new political opportunities and the creation of space for political expression for previously excluded actors, ultimately, it may simply increase local veto powers and the number of people profiting from the state. While decentralization, which in theory signifies increased procedural homogenization, represents more intensive state presence at local level, in practice, it may result in the greater fragmentation of political arenas and greater procedural heterogeneity. Its outcome might well be new forms of local informal despotisms which, however, would remain rather impotent in terms of their capacities for regulation and mobilization.