DECENTRALIZATION, LOCAL GOVERNANCE AND THE DEMOCRATIC TRANSITION IN SOUTHERN AFRICA: A COMPARATIVE ANALYSIS.^©

James S. Wunsch

INTRODUCTION

What factors are required for viable, democratic, local governments in Africa? This is an important question for several reasons. First, in an era of continuing economic problems and structural adjustment, national governments have been forced to reduce the services they provide. While the private sector may pick-up some of these, collective and non-profit-earning social goods must be delivered or funded by sub-national governmental units if they are to be provided at all (World Bank, 1994). Second, research over the last two decades has suggested that highly centralized and top-down service delivery is expensive, cumbersome, inflexible, adapts slowly to new information (if at all), and is prone to political abuse (Esman, 1991). Third, government collapse and incapacity, with spontaneous patterns of local initiative in education, sanitation and marketing, suggest there is an untapped local capacity to make collective choices and take collective action (Green, 1995; Wunsch and Olowu, 1990; Fass and Desloovere, 1995). Finally, a growing body of research suggests that democracy must be rooted in functioning local, participatory self-governance institutions. This literature emphasizes the importance of the growth of civil society, development of public "ownership" of political institutions, mobilization of talents and resources into constructive patterns (positive sum rather than zero-sum or negative-sum political interaction), and countervailing power vis-à-vis national institutions (Harbeson, Rothchild and Chazan, 1994; Wunsch and Olowu, 1990). Despite these compelling reasons, most "experiments" in decentralization and local democratic governance suggest that African local democracy and governance has failed in virtually everyplace it has been tried (Olowu, 1990)!

Some important recent research suggests that these problems are rooted in specific policy choices and strategies pursued by African governments. These policies include deliberate withholding of resources, whether fiscal or juridical (Mutahaba, 1989), from local entities for political or ideological reasons; central bureaucratic hostility and weakness (Smoke, 1994); turbulent economic and policy environments which have undercut local institutions (Olowu and Wunsch, 1995; Wunsch and Olowu, 1996); absence of complimentary reforms needed in national administrative law and systems (Ayee, 1997); and underdeveloped local civil society that left local governments "rudderless" as they tried to develop policy and deliver services (Olowu and Wunsch, 1995; Wunsch and Olowu, 1996).

This paper reports the results of a 1995-1996 comparative analysis of local government in the context of current political reforms in South Africa, Botswana, and Swaziland in an effort to arrive at the preconditions for efficient and effective local institutions of collective choice and of service delivery. At a theoretical level, this paper reports the general patterns observed and the specific findings in each country. Since all countries are in periods of dynamic change, it will also review relevant contextual aspects.

DECONCENTRATION AND DEVOLUTION IN THE AFRICAN CONTEXT

Much of the debate, and perhaps the policy choices, regarding African local government has been structured by terminology first developed by James Fesler in the 1950s. Fesler argued it was important to distinguish between devolution and deconcentration in analyzing local government systems, and in describing patterns of "decentralization" reforms. Devolution refers to the distribution (or re-distribution) of authority to make decisions and to take action by local governments independently of central administrative oversight. Central governments might retain overall legal control (equal protection under the laws, voting eligibility, allocating authority to raise revenue, ensuring general law and order, and regulating fraud and corruption) and the authority to alter local government powers. Within those boundaries, devolution exists if local entities have substantial authority to hire, fire, tax, contract, expend, invest, plan, set priorities, and deliver the services they chose. Deconcentration, in contrast, occurs when local entities act largely as the local agents of central governments, manage personnel, and expend resources allocated to them by central government authorities. Deconcentration refers to essentially the redistribution of central resources to localities on the sufferance of those central authorities. These terms have, since Fesler, been revived, updated and applied by several scholars of Africa and elsewhere. Cohen and Peterson summarize much of this discussion in their recent monograph (Cohen and Peterson, 1996).

Debate and discussion about effective local governance in Africa since independence has often hinged on these as alternative strategies. For example, the operating policy of most African states has reflected a bias toward deconcentration. Kenya, Ghana, Tanzania, Zambia and others have several times trumpeted "decentralization" revolutions. Overtime, however, it became clear that these were really only modest policies of deconcentration. Scholars, in contrast, have often argued instead that real improvement in local government performance (e.g., efficiency, energy, effectiveness, and local participation) could only be expected when devolution was pursued (Wunsch and Olowu, 1990).

Recently, a third factor has also been found to be important for effective local governance: a viable local political process. Dele Olowu and Wunsch, for example, found that substantial decentralization efforts in Nigeria during the later 1980s and early 1990s were weakened by the absence of viable local political processes to convey information to the public about government decisions, to organize publics to be attentive to government actions, to mobilize public opinion regarding local government, and to hold local officials accountable for their performance. The absence of attentive local publics and of close linkages between officials and citizens appeared to explain many of the problems found in local governmental performance in Nigeria (Olowu and Wunsch, 1995; Wunsch and Olowu, 1996).

In fact, if one considers the arguments implied by all three strategies, they can be seen as complimentary aspects of a single model rather than as independent or mutually exclusive. For example, advocates of deconcentration emphasize the fiscal, skill, and personnel poverty of localities, and their need for central resources to function. They also emphasize the problem of providing effective direction for local administrators. Advocates of devolution emphasize that the residents of local governments will be unlikely to participate, invest attention or resources, or learn many of the lessons of governance unless their local governments can make real decisions with effective consequences. The same applies to local officials as well, who, furthermore, are unlikely to invest extra attention and energy in local affairs and programs unless they are accountable to localities rather than the center. In fact, most African local authorities are woefully short both of revenues and trained personnel, and have remained either largely passive or only spasmodically active (in limited areas), in the post-independence era.

The argument regarding a viable local political process combines both of these perspectives. Such a process is necessary to provide policy direction for local officials, information to allow learning and fine-tuning of policies and programs by local officials, and accountability regarding performance. Resources provide the "raw material" of local governance, authority provides the structure that allows local political entrepreneurs to combine resources to produce local services, and a viable political process energizes, directs, and redirects political entrepreneurs in their activities. Simultaneously, the political process provides an arena for political entrepreneurs to explain and market their activities, trying to build support and raise additional resources. When they fail, the political process is the mechanism that reviews and replaces them. Usually a viable local political process includes an active civil society, some general political organizations (e.g., factions, parties or their surrogates), a legislative arena, opinion leaders and their publics, and mechanisms to gather and spread information. With all this in mind, it may appear reasonable to hypothesize that effective African local governance requires deconcentration (of resources and personnel), devolution (of authority) and development of a viable local political process.

This working hypothesis was recently applied during nine-person weeks of exploratory field research on local government in Southern Africa: Mupmalanga Province of South Africa, Swaziland, and Botswana. Each research site was assessed for the resources available to selected localities (deconcentration), the authority devolved to localities (devolution), the health of a local political process, and the effective functioning of local governments. It was expected that the more fully developed were the units considered on each of these criteria, the more effective local government would be. The variables were measured via a multi-indicator, multi-method strategy. This included interviews with local officials, political personnel, leaders of civil society organizations, academic and donor personnel, site observation, including of equipment, maps, meetings, personnel, and business, and review of appropriate documents, including plans, budgets, personnel systems, audits, minutes of meetings, and the like.

SOUTH AFRICA: MPUMALANGA PROVINCE

Although the dramatic events in the Republic of South Africa may be familiar to most people, the laborious process by which a new structure for local government was hammered-out in negotiations that continued through 1993 is not as well-known. The Local Government Negotiating Forum (LGNF) was often overshadowed by the more dramatic national interim constitutional negotiations, but many believe its results will have as great (or greater) impact on the new South Africa (Pycroft, 1996). The major players in the LGNF were SANCO (South African National Civic Organization) and the National Party. SANCO was the national federal body that grew out of the informal "civics" that arose in the black urban areas of South Africa during the 1980s. The "civics" drove the apartheid governmental authorities out of the black urban areas and provided some level of local governance.

The Nationalist Party was, of course, the ruling party of South Africa from 1948 until the African National Congress (ANC) took power in the free election of 1994. Both were galvanized into discussions by the obvious conflict and stalemate that existed between Africans and the minority dominated local governments. Their search for a mutually acceptable solution led to the Local Government Transition Act of 1993, and relevant provisions of the 1993 Interim Constitution Act. The major consequences of these acts for local government was the establishment of varying institutional frameworks for local government (for e.g., metropolitan, urban, and rural), all of which provided for: non-discriminatory participation in local political affairs; a partial-proportional representation system for local elections in the former white, Asian, "Coloured," and black areas; a guaranteed and substantial role for local government; and a process for amalgamating former white, Asian, "Coloured," and black jurisdictions in an institutional framework that placed most operational responsibility on the formerly white executive and managerial structures, the ones which had been best resourced and were most institutionally developed (Cloete, 1994; Pycroft, 1996). The final constitution of 1996 essentially maintained this structure (Cameron, 1997).

Nelspruit, Mpumalanga Province (formerly the Eastern Transvaal) was the first research site. It is located in the eastern part of Mpumalanga Province, a generally fertile agricultural region of rolling hills. Only fifty kilometers from the world-famous Kruger Game Park, it has a diversified economy based on agriculture, services, tourism, and light industry. Formerly a largely white municipality of 24,000 (22,000 white, 1,000 Asian, 1,000 "Coloured"), it has grown to an estimated 250,000 with the addition of two former "homeland" townships, Kanyamazane and Matsulo, each approximately 65,000 people, and various peri-urban areas. Additionally, it has become more of a regional government, as it now stretches from Nelspruit to Matsulo, a distance of some 50 kilometers.

Along with substantial growth in population, racial diversity, and size, the new Nelspruit also faced several other significant changes. These included the recent elections of new legislative personnel and of the (largely ceremonial) mayor, vastly expanded service responsibilities for Nelspruit in the former black townships, and an enlarged personnel system which included personnel from both the former white and black local government units. All this occurred within continued ambiguity regarding, the role of the new provincial governments, the status of the former "homeland" areas still under rural council rule, and the size of central government revenue subventions to be allocated to upgrade and develop the former black townships. In this section, we review the four factors outlined in section II regarding local government in Nelspruit, Mpumalanga Province.

Resources: Over the long-run, the Nelspruit area has been a resource-rich area. The personnel staffing local government are highly educated, professionally qualified, and experienced. Their facilities are modern, spacious, comfortable, and equipped with computers, software, and personnel able to manage and use them.

The municipality supports generous services (public utilities, roads, economic development, planning, housing sites, etc.) to the Nelspruit population, which is virtually entirely white, from a local tax base that has generated typically 130 million Rand (around $45 million US per year). This budget has been adequate to provide services, to staff the local civil service personnel slots, and to service whatever capital borrowing the locality chose to make.

Future resource adequacy is less clear. The revenue base, which is composed of property taxes ("rates") and water and electricity service fees, has limited scope for expansion. Rates are set primarily on site value rather than on improvements, and payroll taxes, sales taxes, income taxes are practically difficult to implement and not under consideration at the regional and national levels which would need to authorize them for local use. Revenue allocations from the national government are uncertain and almost inevitably will be limited in size to a small portion of local needs. The much awaited grants from the Rural Development Fund (RDF) are likely to meet only a small portion of the needs (Cameron, 1996). But local needs will grow enormously with the vast increase in local population caused by municipal amalgamation. The possible addition of what were called the "traditional area," nearby areas formerly administered by the defunct African "homelands," will add to Nelspruit's responsibilities (Pycroft, 1996; Cloete, 1994). Revenue sources in the former black townships are inevitably limited given the poverty there and the culture of "non-payment," the refusal of Africans to pay water and electrical charges, and any rates/taxes due in urban areas to protest apartheid. Additionally, the municipality already experienced a substantial increase in personnel costs with amalgamation with the former black townships.

The most current budget estimates by Nelspruit's government suggested that capital investment needs for the former black townships added to Nelspruit is roads, water, sewers, and street lights were 220 million Rand (approximately 75 million dollars US). The current capital budget was 35 million Rand (approximately 12 million dollars US), leaving a 185 million Rand capital budget short-fall. Operating or current costs are estimated to be 150 million Rand (approximately 52 million dollars US) which leaves a 30 million Rand short-fall from the current operating budget of 120 million Rand. This was a bare-bones budget proposal, essentially only to continue those minimal services offered in the townships, not to upgrade the areas to Nelspruit standards. Overall, then, Nelspruit has been a well-resourced local government. However, it faces serious challenges in the immediate future given the changed ratios of resources to needs.

Authority: Historically, Nelspruit, like other white municipalities, was a powerful juridical body. Working within broad parameters set by central government, it hired personnel, levied and raised taxes, established budgets, set local service priorities and development strategies, borrowed funds (with central Ministry of Finance approval) for capital development, and the like (Cameron, no date). According to the interim constitution of 1995 (Chapter Ten and Section 245), provincial legislators must not "encroach on local government to such an extent that it compromises the fundamental status, purpose and character of local government" (Cloete, 1994, pp 44-45). Broad statutory, regulatory, and executive powers were also established in the interim constitution for local governments to enable them to provide for local welfare, and to raise money through various taxes, fees, levees and the like (Cloete, 1994). When one considers the new constitutional framework, it does not seriously change the authority of local government. However, the amalgamations that grew from the LGNF process fundamentally changed the political dynamics at the local level as they brought about black majority control.

The major potential change-agent on the horizon is the newly established Mpumalanga provincial government. While Nelspruit's civil service personnel are almost 100% white (except for the generally junior grade personnel absorbed from the black municipalities), the provincial personnel are virtually 100% black. How the provincial government will relate to Nelspruit (and other similar cities) over the long-run is unclear, as the agenda of the two groups might well be expected to diverge, and it might be expected over time to encroach on Nelspruit's accustomed independence.

In the short-run, however, the provincial government was in confusion stemming from its recent establishment. It was unclear about its own role and legal authority, preoccupied with hiring personnel and disagreements over priorities. So far, the provincial government had little impact on the municipal government, and personnel interviewed in the provincial government did not aspire to interfere much in the municipal government. Provincial personnel did not review municipal budgets, personnel decisions, municipal priorities, tax rates, etc. In 1996, the provincial personnel in Mpumalanga believed these issues were best left to the municipal political process to decide.

Political Process: The established local political process was largely rendered irrelevant by the "new dispensation" after 1994. The latter has probably quintupled the electorate, and introduced many "new" groups and interests to Nelspruit's government. However, the representative institutions (council) and linkage institutions (political parties and civil society) between the population and the municipality are still underdeveloped. Furthermore, institutional arrangements now in-place to facilitate these are flawed in their design, while the interest groups organized at the "grass-roots" level are still weak and fragmented.

The municipal council is the only representative body at the local level. The effective chief executive is a civil servant, the town clerk. The mayorship is an honorary office, filled by election from the council, and without any executive powers. While the council does have full authority over local budgetary and programatic decisions, in Nelspruit its effectiveness in encouraging and focusing a local political process remains unclear.

With forty members meeting only monthly, the council is a large, part-time body. It is not clear how well it will be able to connect public wants and concerns with municipal administration. It has no staff, and its members have no offices. Its only committee is a fifteen person "management committee" which meets three or four times monthly, but it also is without staff or office facilities. Furthermore, the personnel are largely new to public office: in Nelspruit 36 of the 40 members had never held public office before.

Existing municipal procedures reinforce the powers of the town clerk. The council only sees the budget as a finished document. The absence of functionally specialized committees reduces its effectiveness in analyzing the implications of the proposed budget for programs and services. Miscellaneous other problems also appear likely to hinder the council. Most of the members live long distances from the Nelspruit city facilities, and must use unreliable and slow public transport to journey from home to the offices. Furthermore, most must continue employment elsewhere, further pulling them away from time to learn and follow municipal government. At an intangible level, councilors face the challenge of constructing new roles both their own and for a formerly placid, service-oriented city government, and dealing with a limited budget and with constituents who have increased expectations of, but limited experience in the process of democratic governance.

The ANC, whose leadership heads Nespruit council, is the strongest single political party in the area. Nonetheless, ANC's local strength was compromised by a large Nationalist Party delegation in the Council and by alleged tension between it and SANCO. There was conflict in the former black townships between elected and defeated ANC council candidates. At the time of this research, the ANC did not appear to be developing a coherent agenda for local government, its policies, or activities.

A large number of non-governmental organizations and personnel from the "civics," the grass-roots organizations of the anti-apartheid era movement, have survived the transition and are more-or-less active. Though loosely organized, SANCO elected several persons to the council. Their cohesion and cooperation there, and any coordinating role for SANCO, appeared questionable at the time of this research. The political interest and capacity of the local voluntary organizations was also uncertain. Composed of women's groups, sports clubs, student organizations, anti-AIDS efforts, and the like, these organizations often depended on the efforts of a single person or a small leadership circle, were financially precarious and at times dependent on shrinking donor funding. They and the "civics" frequently looked to the national rather than the local government for their needs. In general, these groups were not yet emergent "local publics" which could attend to local government policy, nor had they been major players in the recent local elections. They appeared to be inadequately informed about local government policies, programs, and resources.

Probably the most coherent and organized component of the black African community was a number of local business-owners. They had been working for several years with Asian and some white business firms to increase their access to capital, and expand marketing opportunities via a local multi-racial chamber of commerce. How their agenda, and the pressing needs of the largely poor and unemployed or underemployed black population might cohere into a single political force was also not clear at the time of the research. Thus, a viable local political process after post-apartheid, 1995 local elections had not yet taken root in Nelspruit.

Local Government Effectiveness: The executive branch of the Nelspruit municipality is a highly organized and sophisticated governance institution. The planning, programming, budgeting, managerial, and auditing functions appeared to be well designed and implemented. Municipal officials were knowledgeable about the details of their budgets, each other's responsibilities, the general budgetary implications of the spatial reorganization, and the challenges of training and working with a town council that has many members entirely new to politics.

Budgeting, planning, and management support services (human resources, property management, purchasing) functioned well. Budgetary figures (estimates, expenditures, operating capital, and revenue budgets) were easily available, and estimates and actuals were very close, reflecting good planning and performance. "Structure Plans" (general land use plans) and "Town Planning Schemes" (zoning regulations) were available, up-to-date, and of professional quality. Revised plans were in progress. The Management Service Office had accurate and timely data on personnel budgets and was trying to solve the management problems of integrating civil service personnel from the old homelands (black townships) with the Nelspruit personnel. The town planner was occupied with the challenges of planning for the "new" areas of responsibility. Offices were well organized, well staffed, and had sufficient resources.

Similarly, the Town Clerk (city manager) was well versed in the conditions and priorities of the newly added areas. He could discuss the challenges he faced: balancing revenue versus expenditures; strengthening personnel management; reconciling traditionally European versus African communication patterns (i.e., formal/Gazette style versus grassroots consultations and discussion); educating the new town council; and developing working relationships with the new Mayor and the new Chair of the Council Management Committee (the key legislative-oversight role). In 1995, the relationships among all three were positive and cordial. All these patterns are indicators of a stable bureaucracy, with the capacity to identify problems, set priorities, develop programs and budgets, and implement them in a relatively transparent and efficient way.

However, beyond the purely managerial and organizational dimension, Nelspruit was not at all effective in delivering services to its new constituents. It is not even preparing to begin doing so. Specifically, while local officials were thinking a lot about this "new era," they did not appear to be engaging in serious reprogramming of resources to meet their new responsibilities. So far, some personnel had been redeployed, but only from the former black areas to the Nelspruit office, a 50 km. distance. There was no clear plan to reconfigure those offices to improve and expand services to the "new" areas, and adjust to becoming, in effect, a regional government. Similarly, officials appeared to be approaching the coming budgetary process in a largely incremental way. Existing programs and investments that focused on Nelspruit town would be marginally modified to reflect the reorganization, but there was no evidence of a fundamental reconsideration of resource or service flow, or of changed personnel responsibilities. Officials in Nelspruit repeatedly emphasized the need for revenues to expand in the new areas before services could be expanded. There were no plans to reconfigure service priorities to reflect the needs of the new areas.

This lack of plans could be considered in light of the fact that officials had already "swallowed" a substantial increase in fiscal responsibility in absorbing and achieving salary parity for the former homeland personnel under a one-time grant from the provincial government. With the new area-wide council that was to begin functioning in January of 1996, any specific planning might have been premature. Additionally, ambiguity regarding possible national development grants and such specific problems as the town's authority in traditional areas had also left the civil servants with uncertainty. Nespruit's executive branch personnel appeared to embrace the status quo.

SWAZILAND

The research in Swaziland targeted local governments in Mbabane and Manzani, the two largest cities in the country. Like the South African case, though perhaps less dramatically, these local governments are entering a new era owing to changes in the larger social and political environment.

Propelled partly by recent rapid urban growth and partly by donor encouragement, a major conference was held in Swaziland in 1990 on its urban government. This was funded by the United States Agency for International Development (USAID), the United Nations Development Program (UNDP), and the International Bank for Reconstruction and Development (the "World Bank" or IBRD). Among the conference participants were the Prime Minister of Swaziland, the senior executives of Mbabane and Manzani, the Acting Minister of Finance, most key urban decision makers, and senior representatives from each major, related ministry. It was the first major review of urbanization policy in Swaziland since the Urban Government Act of 1969, the legislation that then guided Swaziland's local government authorities. In 1969 only 14% of Swaziland's population lived in urban areas, and Mbabane and Manzani together held only 30,000 people. The 1990 conference participants projected that 45% of Swaziland's population would be urbanized by the year 2000, and Mbabane and Manzani would probably increase to at least 200,000 in population. There was widespread concern in 1990 about the ability of urban governments to manage existing growth, including sanitation, transportation, education, housing, and employment. Participants therefore recommended substantial revision in the statutory structure of urban government. This would include establishing a new ministry responsible for urban government, better coordination of housing, land, infrastructure and urban development by urban authorities, expansion of urban revenue sources, closer ties between urban councils and local populations, and election of local councils (USAID, 1990). Many of these changes would require amendment (or replacement of) the 1969 Local Government Act, a structure which established central government control over weak urban governments.

Swaziland has moved only tentatively toward democratic reform. Parliamentary elections were held in 1994 (the first since King Sobhuza suspended the constitution in 1973), and local government ("council") elections occurred in 1995. In neither elections were political parties allowed to contest. More importantly, the traditional elite's hold on power remains unchallenged. It controls most land, the monarchy still controls the national development fund, and it dominates local rural affairs via the traditional chiefly Tinkundla system (Sallinger-McBride and Picard, 1989). Together, the intangible influence of democracy in South Africa, burgeoning urban growth, deteriorating urban conditions, donor support for urban political reform, and resistance to change by a traditional elite, shaped the Government of Swaziland's program of urban governmental reform in the early 1990s. A new Ministry of Housing and Urban Development (HUD) was established by 1995, and several donors supported projects to upgrade urban managerial skills. Elections had replaced the formerly appointed urban councilors. Within the structure of the 1969 local government statutes, HUD expanded substantially the effective authority of urban governments. A revised "Urban Government Policy" which would substantially add to the authority and resources of urban areas, was presented to the Cabinet in 1995 (Ministry of Housing and Urban Development, 1995).

Resources: Mbabane and Manzini present a mixed resource base picture. Local revenue systems in Swaziland face several overall problems that restrict their capacity to meet local service and development needs. These include: (1) administrative practices which hamper collection of local revenues; (2) high expectations that the national government would "bail-out" local governments in serious deficits; (3) limited local tax base and authority, and resistance to raising rates on the existing base; and (4) poor utilization of other revenue sources.

Poor administration and collection of rates in the past led to a significant pattern of non-payment by 1995. No clear figures on actual collection percentages were available because the revenue rolls were disorganized and account records were not kept from year to year. Consensus among officials was that a large proportion of revenues were never collected. The locally powerful were a real problem. They used their political connections at the center to evade local collection efforts. Local governments' administrative weakness made them unable to sell property for taxes, and thus impeded recovery of unpaid rates. Poorer people, when approached directly by their local councilors, had a better record of paying rates.

Mbabane was particularly lax in its enforcement. Since its accounting system did not "age" ratepayer accounts, there was no record of accounts in arrears. The city also failed to follow the prescribed procedures for collection of unpaid taxes, and had taken no action over a six year period. Since 1995, Mbabane attempted to pursue delinquent ratepayers. The city posted announcements in the local newspapers to alert ratepayers who were two to three years in arrears that their property would be put up for auction. Three properties were sold over a two month period in 1995 with an additional thirty slated for auction in early 1996.

Virtually all local officials and leading citizens interviewed regarded the national government as the funder of last resort. Indeed, more than once local officials referred to the national government as likely to "bail-them-out" if they were in real financial difficulty. A pattern of large recurrent subventions (about 40% of the operating budget on average in the larger cities), capital grants, and central tolerance for accumulating deficits reinforced this attitude. Subventions are fairly stable; capital grants are more variable.

Revenue sources at the local level are not extensive. There is intensive resistance to further increases in property rates. Commercial and industrial ratepayers have organized in Manzini and Mbabane to contest recent rates increases (which were in the 50% range in 1995). Residential rates have not been substantially increased much yet, but would encounter resistance in view of the limited services that many residents receive. Currently, only the central government may collect revenue from business taxes, licenses, income, and petrol. To local officials, this is a situation when they feel they are obligated to provide services (street maintenance, refuse collection), but cannot collect taxes on those persons and businesses using these services. The amount of non-taxable central government property in Mbabane is also a matter affecting local revenue potentials.

Rents and fees are the only other sources of urban revenues. These could be greatly increased in some areas, but even when such increases have been budgeted, they have not been collected. The reasons for this are unclear, but lack of political will and administrative weakness are both likely explanations, and overall these are in severe deficit. Privatization may be the only answer to resolve these questions if the political will to collect fees that cover costs is absent. Business license fees, motor fuel tax, auto registration fees, and other new sources of revenue have potential to increase revenues, but these fees have not yet been authorized by the national government. Unless such new taxes are imposed, Swaziland's urban governments will have to cut costs in over-staffed areas, or eliminate subsidies in their services. As table I shows, revenue shortfalls in Mbabane have in the past been met by central Government subventions. This may not be possible in the future.

TABLE 1

Currency: Emalangeni (1 US$=5.0710 Lilangeni)
Source: City Council of Mbabane: Estimates For The Year, 1995/96.

The human resource base of Swazi local authorities is also weak. A few executives were clearly well trained. One town clerk, for instance, who held an MPA from Harvard, is familiar with his responsibilities and quite competent. But overall, personnel of such quality was rare. In Mbabane, an unqualified person held the position of acting treasurer. Most often, personnel other than the town clerk lacked appropriate skills, the situation being worst in smaller cities and towns.

Authority: Some progress toward devolution of authority has been made in Swaziland, largely because of leadership by HUD. Specifically, the larger urban areas (Mbabane and Manzini) may now hire and fire their own personnel, establish property tax rates, let contracts, prepare their annual budgets, and manage their own funds. However, there are important ways in which this authority is circumscribed. Although used sparingly, HUD retains authority to approve and disapprove all local budgets. Moreover, local governments cannot raise funds for capital investments. Authority to borrow is unclear, and there are currently no markets for municipal loans (e.g., insurance companies, pension funds, bond markets).

The three key institutions in the Swazi fiscal governance system are the Ministry of Finance, the Ministry of Economic Planning and Development, and the Ministry of Public Works. These three meet annually to review capital requests from all other ministries. Their dominant role in capital decisions has important implications for Swazi development as a whole and for local government. This lack of local autonomy in capital investment greatly reduces the effectiveness and political ability of urban government and makes forward planning difficult. Revenue flows from the center are highly variable and appear to be arbitrary as well. This also has implications for the political process, as the ability of local officials to build coalitions, to be taken seriously, and to concentrate local political activity in local institutions is seriously compromised by such arrangements.

Other national agencies are also problems at times. The Tender Board, which approves all contracts with the private sector, appears to be a bottleneck for the cities' public projects. In 1995, it had not met to review proposed projects for approximately six months. HUD's own National Housing Board, which is responsible for overseeing public housing projects, also has a mixed reputation for efficiently completing local projects. The Ministry of Home Affairs has disputed some urban authority policies like regulation of alcohol sales.

The recommendation to enact revenue-sharing mechanisms created serious inter-ministerial tensions. HUD supports a formula-driven allocation and a system of direct grants that would provide local governments with the opportunity to meet broad, national priorities, but to set their own priorities within them. In contrast, the Ministry of Finance is opposed to the concept of revenue sharing and supports instead tied transfer payments keyed to its revenue projections. The Ministry of Finance's proposal would continue the current problems for the local planning process and local government autonomy. In general, then, some progress has been made in devolving authority to Swaziland's larger urban areas. However, particularly in the fiscal area, important strings continue to be held by central ministries.

Political Process: Swaziland's local governments have only recently begun to develop a viable local political process. While elections were held for municipal councils in March, 1995, communication channels have not been institutionalized between the public and officials. There are a number of private, voluntary organizations active in Mbabane and Manzini: charitable organizations, sports clubs, student groups, trade unions, professional organizations, and women's groups. However, they have not institutionalized any relationship with local government that could make them assertive publics against local authorities. A businessman's association in each city comes closer to this, as do ratepayers groups, which largely duplicate the businessmen's associations.

Local councilors, in office since the elections of May, 1995, report having made initial direct contacts with their constituents, as well as attempts to influence municipal government to bring government services to them. The scale of needs (roads, electricity, water, sewage, housing sites) is so great while the resources are so limited, that it is not clear that the average urban dweller will see much improvement from this. Councilors were unclear regarding their roles vis-à-vis local government administration: they had no clear vision on the separation of powers between themselves, local executives, and HUD. Representatives of the public seemed equally uncertain on these questions. Nor was it evident anywhere that a process to develop this understanding was underway. The Tikundla system, the structure which organizes the authority of traditional chiefs, continued in urban areas, producing ambiguous authority and responsibility for municipal officials in areas nominally under the control of urban councils.

Local Government Effectiveness: Overall, these interviews and several reports recently completed on urban administration in Swaziland strongly suggested that urban policy-makers and administrators had not developed effective problem identification, planning, programming, budgeting, or implementation capacities (Dupuis, and Cooper and Lybrand, 1995; Ernst and Young, 1995; Ministry of Housing and Urban Development, 1995). The two larger cities seemed closely to follow previous year budgets instead of considering and consciously planning for future needs and changing priorities. Financial and personnel management was equally weak.

In the area of budgeting and financial management both Mbabane and Manzini faced serious problems. First, neither had any programmatic context included in their budgets. These were simple line item budgets which provided no programmatic direction to city activities, or goals and benchmarks against which to assess performance. Second, both Mbabane and Manzini tended to project seriously overly optimistic levels of rate collection and/or to engage in deficit spending. Neither city "aged" its rate collection accounts and thus made little if any effort to collect from historic non-payers. Mbabane had serious problems in tracking its expenditures, with accounts running several months late. In 1994-95 this led to a deficit of nearly Emalangeni 2.0 million (equal to around 800,000 dollars U.S.). This pattern recurred over several years, leading to a built up deficit of nearly E 8 million. Mbabane also shows consistent patterns of wide discrepancies in estimated versus actual tax expenditures, amounting to more than 50 percent in several areas (City of Mbabane, 1995; Ernst and Young, 1995). While all governments must adjust for unexpected contingencies, this pattern of wide discrepancies suggests that budgeting or management may be seriously flawed.

Fiscal management problems included running city commercial services (such as a slaughter-house) at severe losses. There was severe over-staffing in other service areas (building inspection), so that their costs exceeded fees by several times. Activities that could be contracted out (such as maintenance of vehicles) were continued by city departments at costs that seemed higher than market prices. Though several of these service areas are presumably commercial they were not financed from the fees they generated. These figures are illustrative of these problems:

TABLE 2

Currency: Emalangeni (1 US$=5.0710 Lilangeni)
Source: City Council of Mbabane: Estimates for The Year, 1994/95.

Finally, the fee collection system is also plagued by great discrepancies between revenue budgeted (i.e., fees anticipated to be collected) and actual collections. The 1994-95 estimates were generally 100-300 percent higher than the actuals collected. While 100 percent cost recovery is not likely in all areas (parks, sports and recreation), deficits in other areas could be reduced or eliminated either by raising fees or reducing personnel (Ernst and Young, 1995).

Personnel systems and practices were perhaps in even worse condition than financial systems in both the leading cities. In a recent report compiled by H.T. Dupuis and Associates and Coopers and Lybrand of Swaziland, personnel system deficiencies were discovered in virtually every area (Dupuis and Coopers and Lybrand, 1995). The current system lacked overall policy guidelines regarding the use of human resources, a medium-or long-term plan to meet human resource needs, a policy on in-service training or development, descriptions of positions or criteria for evaluating performance, industrial (labor) relations policies, and a performance appraisal and salary increase system. Also personnel records were absent or incomplete, discrepancies existed between many employees' grades and pay, and over-staffing was severe. There were no management information systems or internal managerial assessment and analysis (audit) functions. At least in Mbabane, the town clerk is aware of these problems but feels prevented from resolving them by insufficient personnel resources and a non-supportive council.

BOTSWANA

In October of 1994 Botswana marked its sixth open and competitive election since independence, a record unmatched throughout Africa (Danevad, 1995). With an equally impressive economic growth rate, the highest in the developing world from 1965-1989 (World Bank, 1989), Botswana offered an environment which was to prove very supportive for local democratic governance.

While observers debate the nature and extent of "democracy" in Botswana (Picard, 1987; Holm, 1987; Tordoff, 1988; Parson, 1991; Danevad, 1995; Good, 1996), none deny that Botswana has been exceptional in its political openness, competition and adherence to law (Africa Today, 1993). It is particularly notable and unusual in its acceptance of political competition at the local level, and has established a climate where a local political process has been unusually viable. Even given the elite's capture of a disproportionate share of Botswana's growth proceeds, it has offered an equally unusually high flow of resources to local governments.

As Picard (1987) and Tordoff (1988) note, local government has a long history in Botswana. While it suffered in the early post-independence era from many of the same problems local government faced elsewhere in Africa, strategic choices made by the political center avoided the centralization seen in most of Africa. In 1970, with central government confidence in local councils at a low point, and with a bureaucratic battle raging over the future of local government between the Ministry of Finance and Development Planning and the Ministry of Local Government and Lands, the Tordoff Commission recommended that Council staffing and finance be substantially improved. Its recommendations were adopted, and since then local government has been a major component of Botswana's political system (Picard, 1987). However, as this case study will indicate, local government there still faces challenges, and has room for improvement.

Local Resources: While overall resources are higher than found in most of Africa, local revenue sources are currently very limited. Rural districts raise funds primarily through fees for services, small business licenses, and rents for council-owned housing. Urban districts also have these sources, but their primary source of funding is property rates. Urban authorities cover their budgets better than rural districts do, mainly because of their use of property rates. Nonetheless, if Gaborone is typical of Botswana's urban authorities, even urban revenue collection is weak (Peters-Berries, 1995).

A serious problem is the failure to levy service fees comparable to service costs. One of the causes of this failure is the absence of an effective system to determine the actual costs of these services. Another is the limited authority that local government has to set fees, as the Ministry of Local Government Land and Housing (MLGLH) or other central ministries control fee levels for most local services. Another problem is a tendency of local personnel to assume that national government will cover their financial shortfalls (Peters-Berries, 1995). The Local Government Tax, an income tax, was previously in place but was rescinded at the height of the diamond mining boom. A new Local Government Tax and the extension of rate authority to the district councils appear to be the most viable options for new local revenue. For the moment, national revenue transfers are substantial and cover extensive local programs. Botswana's ability to sustain these in the future, through, may be a problem.

The quality of local government staff is quite impressive with regard to diversity of personnel slots authorized and filled at the local level (medical officer, planner, attorney, economist, chief of staff, etc.), their paper qualifications, and their demonstrated competence and professionalism. With the exception of "industrial class" employees (primarily laborers and other unskilled personnel), all local government personnel have been hired and managed by the Unified Local Government Service (ULGS) since 1973. ULGS is managed by an Establishment Secretary and Secretariat, housed in MLGLH. The salaries of these personnel (about 23-35 percent of local government personnel) are also paid by the national ULGS, which eases the local budget burden. Other personnel are the responsibility of local councils (SIDA, 1993; Dahlgren, 1993).

One concern regarding personnel and administrative systems is the breadth of coverage of qualified personnel. Other researchers have noted that less competent personnel are in-place in the remote areas, and that many intermediate posts are vacant throughout the ULGS. If these vacancies are widespread, they will weaken local governance (SIDA, 1993; Dahlgren, 1995).

Local Authority: This case requires more detailed attention than the other two, because the success and challenges of local government in Botswana are relevant to local government in Southern Africa in general. Botswana's unique pattern also offers much enlightenment on the general issue of decentralization and local democratic governance: it shows how apparent "authority" can be illusory, and its consequences.

The Ministry of Local Government, Land and Housing (MLGLH) is the "parent" ministry of all local government in Botswana. As Botswana is a unitary state, local government has no constitutional status and is a purely statutory creation. Within these statutes, MLGLH plays key roles in virtually every aspect of local government. These include controlling or supervising most key decisions regarding local personnel, budgeting, development planning, self-help projects; ensuring conformity with national policies and priorities; providing training; developing new revenue sources; and developing new managerial systems and procedures. Even when local governments seek greater autonomy, their primary spokesperson and advocate (and occasionally, foe) is MLGLH. Thus, this ministry is critical for all aspects of local governance in Botswana.

The ULGS hires, assigns, promotes, disciplines, discharges and transfers all non-industrial personnel. This system is responsible for ensuring that all areas have adequate personnel and discouraging any tendencies toward ethnic concentration of personnel. However, there are numerous criticisms of the system's appropriateness for local governance. Specific problems are the lack of control that local governments believe they have over selecting congenial and committed personnel, and the disruptive impact that the transfer system has on local affairs and on the lives of local personnel, particularly of women.

In interviews at three local government site visits located in different areas, the same complaints were made by legislative and administrative personnel: arbitrary and abrupt transfers of personnel were disruptive to local projects and programs, they hurt smooth administrative functioning, team-building, morale, and the employees' personal lives. Town secretaries and councilors in particular complained about being unable to ensure that they were assigned personnel committed to their goals and familiar with their particular needs and problems. These patterns were also noted by Picard (1987). Since ULGS' control extends to personnel training, development, and performance appraisals, local authorities rely on government training centers that are managed by MLGLH. Supervisors must receive permission from the ULGS to send employees to training workshops. The result is that a request for training must compete against similar requests made by all other local authorities. Limited resources led many requests for training to be denied or delayed.

Recent changes however are gradually expanding local control over junior appointments. By the end of the current national development plan, 85 percent of local government posts will be filled and administered by local authorities. However, this does not address the most pressing concern of local officials: the senior staff. Some type of national-local civil service system may be needed to give security and homogeneity of service, to help ensure professional competency, and to maintain links between sector ministries (i.e., health and education) and local officials. However, an end to arbitrary and rigid transfers, greater deference to hiring choices of localities and personnel, and improved status vis-à-vis the national civil service are also needed.

From one perspective, local governments already largely control their budgeting. Localities are entirely responsible for preparing their annual estimates. MLGLH reviews budgets primarily for arithmetic accuracy, to ensure that none are in deficit, and for conformity with national policies regarding personnel (that certain posts are filled, certain teacher-pupil ratios are followed, certain health facilities are staffed, etc.). Of course, pay levels must be met along with other benefits. Emergency conditions, like the recent drought, may require certain extra activities and funding by local governments. And, new and existing centrally directed capital investments (schools, bore-holes, sewage systems, clinics, roads) must be staffed and maintained. Still, within these guidelines, local governments prepare their own budgets.

But what does this local government control over their budgets really amount to? At every site visit, local officials (both professional/administrative and political) agreed that MLGLH's oversight actually left them with little autonomy. In fact, recurrent commitments are so high and so driven by policy decisions made in Gaborone, that local officials believed they had virtually no latitude at all to identify and respond to their own unique priorities and problems. Local needs that they deemed to be particularly pressing, like absence of storm drains to solve flooding problems in one area, health initiatives in another, low-cost housing in a third, were ignored because all revenues were already committed elsewhere. Salary and vehicle costs alone were so high that little was left over for any local initiatives (Peters-Berries, 1995). If a problem was not on the agenda of a national ministry with the ability to channel funds to meet the need, it was not addressed. If such ministries felt that something should be done, it was usually accomplished, regardless of local priorities (Briscoe, 1995; Channaux-Repond and Kanengoni, 1995). Local professional personnel felt discouraged by this pressure. Local political leaders felt that their work was futile and that they were unable to respond to their constituents.

In contrast to recurrent budgets, the status of capital budgets was unambiguous. These were clearly determined through a top-down process, with the Ministry of Finance and Development Planning (MFDP) telling the MLGLH how much was available in each sector area (health, education, roads, etc.). for local projects. MLGLH then apportioned it to the local governments. They, in turn, reported that most of these grants were consumed by investments required by national ministry norms. In any case, MLGLH decided which projects would be funded, so that local governments were uninvolved in policy-making, despite the fact that these decisions also had significant implications for local governments' recurrent budgets (Picard 1987).

Central regulation and oversight also impede the timely completion of capital projects. Before starting a project, local authorities must submit a project memorandum to the MLGLH. The ministry then forwards the memorandum to the MFDP for approval. There is no assurance that the project will be funded in a timely manner or in its entirety. For example, in Kweneng, 20 million Pula (approximately 8 million dollars, US) was requested for an approved capital project, but the district received half that amount. Lobatse often experiences a six to seven month delay in the receipt of capital funds after it submits project memoranda. Once a project memorandum has been approved, ceilings are provided for each capital project. The limits are often incompatible with the financial requirements of the proposed project. The central government also requires authorization and appropriation for each certificate of payment to release funds. Thus, projects with multiple contractors often force local authorities to make repeated requests and trips to the center for payment of invoices. The tedious and time-consuming process can be costly to local authorities, since contractors who have not been paid within 21 days of the execution of a contract may assess interest against the amount they are owed. Delays in the receipt of funds or the payment of invoices for contractors often cause problems for development projects. This, of course, follows the "repetitive budgeting" model noted by Caiden and Wildavsky (1974).

Thus, capital projects have remained a cumbersome, centralized process for local governments. A local process ostensibly designed to register, assess, and encourage local development (i.e., each locality has a senior economic development officer and engages in a grassroots development process) has in effect been prevented from fulfilling this promise, making local government appear rather irrelevant to any critical eyes.

In the past, district and urban planning were not at all coordinated with the national development plan. The district and national plans covered different periods and were prepared at different times. The consensus is that district plans rarely affected national plans, which ultimately controlled all investment decisions (Chanaux-Repond and Kanengoni, 1995). Indeed, local planning was mostly ignored (Briscoe, 1995). One highly placed respondent reported that local development projects essentially were approved when they fit into the national ministry plans. However, national and district plans (in 1995) were being prepared simultaneously for the National Development Plan - 8 , and it appeared that efforts were being made to incorporate some local priorities via what they are calling "matrix planning." It was too early to predict how well this might work at the time of this research, but strengthened input for localities in the capital investments of sectoral ministries is desirable if local governance and decentralization are to progress in Botswana.

As noted above, the Ministry of Finance and Development Planning (MFDP) plays a major role in local government. Because of local governments' overwhelming dependence on the central level for subventions to cover recurrent budgets, MFDP's role in determining the amount to be allocated to these grants (subventions) is extremely important. MFDP, of course, determines the amount of the capital development budget. While MLGLH allocates the shares to go to each local government, MFDP retains final authority to approve or deny any specific project (through the project memorandum requirement) and controls disbursement of revenues for each invoice as projects are implemented. In virtually every respect, local governments are subject to decisions made at the central level, through a complex and often cumbersome process. Other ministries with major impact on local government include Health and Education. Unlike most African countries, Botswana's local governments are responsible for elementary education and primary health care. Personnel involved in delivering these services are employees of the local governments. Day-to-day management is provided by the local authority which appears to discharge it rather competently.

There is, nonetheless, a strong ongoing role for the "parent" ministries in the areas of health and education. They set staffing levels and standards, determine minimum equipment levels, provide ongoing training and professional support, and set conditions of service. In the case of the most professionally qualified service-sector employees (medical doctors), their subordination to the local government "chain of command" is not entirely clear to the parties concerned. Much of the redundancy of this situation is desirable. When primary health care was essentially separated from the national Ministry of Health in Nigeria, standards and morale were severely reduced in the field, and local government support was often inadequate (Olowu and Wunsch, 1995).

Local Political Process: The local political process is far more developed in Botswana than in either Swaziland or South Africa. Botswana has had uninterrupted democracy and civilian rule since independence even though a single party has dominated until recently. Nonetheless, the local political process is unevenly developed.

Municipal (urban) councils are the most developed institutions involved in the local political process. Councilors interviewed at three research sites seemed well-informed and interested in local administration. Nonetheless, they face challenges (as part time officials) in keeping abreast of and contending effectively with full-time, professional administrators whom they do not hire nor fire, nor over whose programs have they really much control. The three councils visited utilize a similar committee structure to conduct business. Committees focusing on education, health care, social services, trades and licenses, and finance exist in Lobatse, Goborone, and Kweneng. Committee members select the chairperson, and in Lobatse and Gaborone the mayor sits as an ex-officio member of all committees. The frequency of committee meetings varies. In Gaborone council committees meet once per month, while in Kweneng the district council's committees meet in conjunction with the formal council sessions. The position of mayor currently has little authority in Botswana.

In Kweneng District councilors were articulate and clear as to their roles in local government. The district councilors identified the following functions as their primary responsibilities: meeting constituents and seeking feedback, informing the council of the needs of constituents, defending the council's decisions before constituents, advising constituents of council actions, meeting with village committees and village extension teams to determine needs, encouraging self-help among the constituents to avoid dependence on the council, increasing knowledge of government policy, and supporting each other as councilors.

Councilors employed several strategies to keep attuned to constituents. In Lobatse, they sponsored meetings in their wards to exchange information with constituents. Councilors also held membership in civic organizations and reported they used these affiliations to stay informed of the concerns of the community. The Lobatse town council is attempting to develop a newsletter on local government that will be widely circulated in the community. Councilors in Gaborone sponsored forums in their wards. The mayor of Gaborone also toured the city with members of Parliament and visited each ward. In total, the mayor convened 25 meetings and used the occasions as an opportunity to gauge public sentiment on a variety of issues. Kweneng District councilors met with constituents in conjunction with regular sessions of the council. There is little coordination of effort between district councilors and members of Parliament.

Councilors, however, are not as well prepared as administrators to discharge their duties. The separation of powers is unclear, and councilors do not have a command of pertinent public policy issues. The fact that citizens do not have a direct financial stake in local government may force councilors to develop innovative means by which to engage constituents. Citizens who do not see a direct link to local government may simply discount the importance of local authorities. Councilors must simultaneously seek to solidify their position in local government and increase confidence in their representation among constituents.

Presently a number of civic and voluntary organizations function in Botswana. Civic organizations such as the Lions Club, Rotary Club, Youth Council, and church-affiliated organizations encourage self-help and provide some limited local assistance. In Lobatse, the town council awarded grants up to P 18,000 to these small, community-based organizations. The Gaborone City Council recently received permission from the Ministry of Local Government, Land and Housing also to provide small grants to civic organizations. Some of the larger organizations have offices, but most do not. It is not clear that these organizations, in 1995, were serving as serious linkage mechanisms between citizen, elected official, and administrator. Some were dependent on donor grants to survive. However, a base is there on which such a process might develop.

Political parties were quite active at the local level at the time of this research. Indeed, in each of the two urban sites visited, the majority of the council was composed of the national opposition party. However, it is not clear from our interviews that this was energizing governmental-grass-roots interaction, or leading to any new policy directions. There appear to be few ideological and only nascent class differences between the two major parties (Danevad, 1995). Like civic and voluntary organizations, it appears as more a latent structure that might become vitalized, perhaps if local governments were to have more programatic and policy discretion.

Local Government Effectiveness: Local governance in Botswana presents a complex and mixed picture of local personnel and institutional capacity. On the positive side, a workable and working administrative structure is in place, and, at least in the more accessible areas, is generally filled with competent personnel. In the governance of city, town, and district council affairs, this structure has been able to prepare and execute budgets dealing with the personnel, supply, and maintenance functions of several large, complex departments and activities. For the most part, they appear to manage these functions well, although often requiring supplemental payments and allocations to complete each year. Several written studies validated these field findings (Peters-Berries, 1995; Picard, 1987; Tordoff, 1988; SIDA, 1993; Dahlgren, 1993).

Local authorities, furthermore, engage in local physical planning, prepare development plans with specific project proposals, and manage the implementation of a large number of capital projects on an annual basis. In doing this they must, among other things, deal with a slow, cumbersome bureaucratic process to obtain funds from MFDP, and balance the requirements of contractors, MLGLH, and MFDP. Achieving this balance in itself reflects substantial institutional and personnel capacity. In general, personnel interviewed were professional, knowledgeable about their responsibilities, and articulate about the problems of the local governance system. As a result, perhaps, local governments in Botswana have an excellent record of delivering such key services as water, education, and health (Tordoff, 1988; Holm, 1987).

Another area of local activity is implementing relatively small, labor-intensive projects connected with the drought relief effort. According to several interviews and one study (Meyer, 1995), local authorities exhibited some skill in doing this and in resolving personnel, fiscal, managerial, and technical problems in the process.

Consensus exists among written sources (Peters-Berries, 1995; Chanaux-Repond and Kanengoni, 1995; SIDA, 1993) and local respondents that financial administration is the weakest point in local governance. Budget projections are frequently inaccurate, audits often as much as three years behind, and funds sometimes misallocated (i.e., capital funds into recurrent expenditures). In 1989-90, only three of fourteen urban and district councils produced their final accounts on time. Only two received unqualified approval by auditors. Also, district councils have exerted few cost-saving measures, probably in part because of absence of incentives as well as shortfalls in capacity (Briscoe, 1995). Other problems include minimal capacity to collect local revenues and little evidence of ability to analyze expenditure efficiency. The extremely poor cost-recovery system for local services is an example of this. Also, estimates and actuals regarding locally collected revenues often differ by 100-200 percent (Briscoe, 1995).

Local governments regularly apply for supplementary grants. Since these appear to be regularly approved, this may be as much a financial strategy as a problem of competent budgeting and expenditure control. Also, the need for grants often results from unpredictable natural events such as drought (Briscoe, 1995).

The second area of weakness is personnel management. Respondents agreed that local governments were able to accomplish some routine tasks but believed them incapable of the range of personnel functions that complex personnel departments pursue. Personnel development, discipline to resolve difficult personnel problems, efficiency studies, reconfiguration of personnel assignments, and the like exceed present local capacity (SIDA, 1993; Dahlgren, 1993). If there were to be a phased increase in local personnel responsibilities, such capacities would need to be enhanced.

A final concern at the local level is weak management information systems and a lack of programmatic focus at the local level in the various sectors. Local personnel, one key study found, did not conceptualize their areas in a strategic, problem-focused, and systematic sense. This may stem from the habit of expecting direction from the center and the reality of central dominance over planning and capital investment decision. Regardless of the cause (central preemption or local passivity), there is little local strategic or programmatic initiative (Langlo and Molutsi, 1994; SIDA, 1993). These, however, are relatively sophisticated functions, and many local governments in the developed world could be similarly criticized.

CONCLUSIONS

This paper began by positing three variables to explain the performance of local governance. Drawing from the diverse literature which has accumulated on decentralization and local government in Africa, it argued that three factors were necessary for effective local government: resources, authority, and a working, grassroots-based political process. Greater levels of each of these were expected to be positively associated with improved local government performance, with the latter including internal operations and delivery of services appropriate for local needs.

Local governance in the Republic of South Africa is entering a dynamic era. In applying this paper's analytical framework, one finds it characterized by high levels of resources and authority, but with a political process undergoing rapid change. Historically, the political process was viable, but only for the enfranchised and empowered white minority. That political process is obsolete, though remnants of it are still visible. The current post-apartheid political process is weak at the local level in all respects: civil society, political parties, the legislative arena, information flows, and informed and active publics.

Local government performance in Nelspruit reflects this transitional pattern. It has, for the moment, retained the managerial and organizational capacity of the old resource-rich system, but it has yet to begin redirecting its services and activities to the populations for which it is now responsible. In these regards, our analysis of Nelspruit validates the model offered in this paper. Its performance, at least in delivering services, must be regarded as weak.

Our second case, Swaziland, is at the very beginning of the process of building viable local, democratic governance, and not surprisingly it is a work "in progress." While indicators of performance were, at the time of this research, quite discouraging, it must be remembered that the reforms that established this system were recent, with some (elections) less than a year old. As the model used in this paper would predict, Swaziland does not now have anything like effective local governance.

In our last case, Botswana, local government displays unusually high levels of performance, both in its internal management and its record of delivering services (e.g., schools, water, roads, relief, and health care) throughout most of the country. In each respect, it is unusual for Africa, and indeed for most of the developing world.

Contributing to Botswana's success are several key factors: a substantial and sustained flow of fiscal and personnel resources; a national climate which has remained open to local party politics; an open and critical media; and activity by diverse voluntary and civic organizations; a stable legal environment and a sustained commitment from the center to maintaining real local government; and significant responsibilities though, limited authority, for local governments. The outcome of these factors is that local governments are able to perform many activities and functions, but personnel (both political and professional) are frustrated by the encumbrances which prevent them from doing more. As a result of this last problem, it is not clear that the "time and place" information, which is the relative advantage of local governance units, is being used by the national ministries in their policy pronouncements. Nor is it clear that national ministries are inspiring localities to take initiative. It appears there is little or no flexibility for unique local problems, needs, and priorities, whether in diverging from national guidelines or in shifting resources from health or education into other areas. For example, health or education might be better served in an area by upgrading transportation rather than by building additional facilities or hiring more staff. Under the current system, recurrent budgets are largely prescribed by national policy, and capital budgets are entirely under national control. Nonetheless, while local autonomy is limited, and while there are still personnel and operational weaknesses in Botswana's local government, it clearly has made vast strides and appears the best of the three cases studies in providing quality local governance which responds to local needs.

With only three case studies, it is of course impossible to test all logically possible combinations of these variables. However, each of the three case studies offered a different pattern, and did allow a preliminary test of the general hypothesis: the higher the levels on these three variables, the better the performance of local government. They also offered a possibility to explore specific configurations of the three variables, how each uniquely affected local government effectiveness, and what their theoretical implications might be. Table 3 summarizes the paper's findings.

TABLE 3

Note: Evaluations range from: very weak, weak, moderate, strong, very strong.

The table clearly supports the paper's working hypothesis.

Beyond the general hypothesis, the three cases offer insight into several sub-patterns. In South Africa, for example, strong resources and authority, in a technocratic and legalistic political community, led to local governments which were highly efficient but de-coupled from the majority of their constituents. At the time of this research, 1995, the city surveyed in South Africa seemed almost robot-like in its continual attention to doing what it had historically done in the way it had historically done it! Professional city personnel seemed aware of the need to change direction, even anxious at times, but no real redirection had occurred. Perhaps the momentum of any large bureaucracy (e.g., habits, policies, standard operating procedures, existing clientele) explained this. It is unclear, however, what, if anything, will emerge to create the new political process for Nelspruit's local government. Without an external "push" from an active, sustained political process its technical and organizational excellence could be destroyed by public frustration and political rancor.

Decentralization in Swaziland is only a few years "old" and lacks resources, authority, and a local political process, and is performing poorly. In terms of the model presented in the introduction to this paper, Swaziland is deficient on all three key variables: authority, resources, and political process. As a result, its local governance performance is currently rather poor. Local management is weak and few services are being delivered. However, local governments are in place and functioning, and if nurtured may be expected over time to improve in their performance. For those who would abandon the policy, it would be well to consider that many similar criticisms were made of local governance in Botswana in the late 1960s and early 1970s (Picard, 1987), and it has traveled a long distance since then. What the Swaziland case should tell us, in conjunction with other cases, is not that decentralization and effective local governance are not possible, but that they are not easily achieved.

Much progress is needed in Swaziland: revenues must be expanded, personnel upgraded, fiscal and personnel management systems strengthened, town councils developed, ties to communities enhanced, services rationalized and upgraded, and the like. But a framework for local governance has been put in place by electing local officials, by outlining some limited local responsibilities, and by establishing at least a basic cadre of professional personnel. This is the minimum framework necessary to build on for future progress. Key investments need to be made now in developing local cadres and administrative systems in order to carry on decentralization. Without these investments, stagnation and regression may ensue.

The proposed national policy on urban government may resolve many questions of the legal status and clarify the role of local government in Swaziland. Key issues addressed by that policy include: (1) authorizing local governments to use revenue sources currently controlled by the central government; (2) releasing HUD and local governments from control over local capital improvements held by the key national ministries (Finance, Planning and Development, and Public Works); (3) clarifying the boundaries and role to be played by the Tikundla system in local governance. Improving the quality of local personnel (already a focus of several donor programs) paired with these expansions in authority and resources should enhance local governance. As local governments currently hire and discharge their own personnel and have control over their recurrent budgets, these are not issues. An eventual fourth issue might be the expansion of local authority to cover additional areas such as housing, primary and intermediate health services, and primary education. National control over these areas necessarily limits the impact that decentralization/localization can have on energizing grassroots involvement in capital development and enhancing service delivery.

Botswana, overall ranks the highest on our three independent variables, and has the strongest overall performance by local government. Nonetheless, its relative weakness in local authority seems closely related to serious discontent expressed by local officials (both professional and political) regarding their ability to fine-tune national programs to local needs. Weaknesses in planning and local initiatives may also be related to this. Also, the still low level of local resources mobilized in Botswana for local government may at least in part be explained by the still incompletely developed local political process. As for continued short-falls in management and operations, they can be explained by the relative youth of Botswana's local government systems, and overall continual weakness of the nation's personnel base. For Botswana's local government to reach the next plateau of performance, locally raised resources and local authority must be expanded. These might empower local administration at the same time that they energize local political process.

In summary, local governments in Southern Africa present a mixed but not discouraging picture. The variance among them can be explained by their respective environments and policy choices, and tend to confirm the model as hypothesized. At the theoretical level, South Africa must develop a viable political process. Swaziland must continue down the road it recently started, enhancing local resources, sustaining local democracy, and gradually expanding local authority proportionality. Botswana has the "easiest" challenge, gradually to enhance the authority and responsibility its localities already have. Of course, translating theory into policy is often difficult, as it involves such unknowns as political will, good luck, and astute leadership. These are rare in any political system.

WORKS CITED

Africa Today (1993). "Prospects for Institutionalized Democracy." Africa Today. 40: (1). 7.

Ayee, Joseph (1996). Intergovernmental Relations in Ghana, 1988-1995. Presented at the Third International Conference of the International Institution Administrative Sciences: Beijing, China (October, 1996).

Briscoe, Andrew (editor) (1995). Local Government Finance, Planning and Business Promotion. Friedrich Ebert Stiftung: Gaborone, Botswana.

Cameron, Robert. (no date) The History of Devolution of Powers to Local Authorities in South Africa: The Shifting Sands of State Control. Department of Political Studies, University of Cape Town (Mimeo): Cape Town, South Africa.

Cameron, Robert (1997). South African Final Constitutions: The Elevation of Local Government. Presented at the International Association of Schools and Institutes of Administration Conference, Quebec.

Cheneaux-Repond, Maia and Stan Kanengoni (1995). Some Strengths and Weaknesses of Botswana's Development at Local Authority Level. Friedrich Ebert Stitfung: Gaborone, Botswana.

Caiden, Naomi and Aaron Wildavsky (1973). Planning and Budgeting in Poor Countries. New York: Wiley.

Cameron, Robert (1996). "The Reconstruction and Development Programme." In Journal of Theoretical Politics, 8:2.

City Council of Mbabane (1995). Estimates for the Year, 1995-96. City of Mbabane: Mbabane, Swaziland.

City Council of Mbabane (1995). Management Report for the Year Ended March 31, 1995. City of Mbabane: Mbabane, Swaziland.

Cloete, Fanie (1994). "Local Government Restructuring." In Politikan, 21:1.

Cohen John and Stephen Peterson (1996). Methodological Issues in the Analysis of Decentralization Development. Discussion Paper No. 555. Harvard Institute of International Development, Cambridge, Mass.

Dahlgren, Stefan and Tyrrell Duncan, Allan Gustafsson and Patrick Molutisi (1993). SIDA Development Assistance to Botswana: 1966-1993. SIDA. Gaborone, Botswana.

Danevad, Andreas (1995). "Responsiveness in Botswana Politics: Do Elections Matter?" In Journal of Modern African Studies, 33:3.

H.T. Dupuis and Associates and Coopers and Lybrand Services (Swaziland) (1995). Report on Human Resource Management. Dupuis et. al.: Mbabane, Swaziland.

Ernst and Young (1995). City Council of Mbabane: Annual Financial Statements at March 31, 1995. Ernst and Young: Mbabane, Swaziland.

Esman, Milton (1991). Management Dimensions of Development: Perspectives and Strategies. Kumarian Press: West Hartford, Conn.

Fesler, James (1949). Area and Administration. University of Alabama Press: University, Alabama.

Good, Kenneth (1996). "Towards Popular Participation in Botswana." In Journal of Modern African Studies, 34:1.

Green, David (1995). Decentralized Public Service Provision in Sub-Saharan Africa: A False Start. Associates in Rural Development: Burlington, VT.

Harbeson, John, Donald Rothchild and Naomi Chazan (editors) (1994). Civil Society and the State in Africa. Lynne Rienner: Boulder, CO.

Holm, John D. (1987). "Botswana: A Paternalistic Democracy." In World Affairs, 150: 21-30.

Langlo, Marilyn and Patrick Molusti (1994). Decentralization and Health Systems Performance: The Botswana Case Study. Center for Partnership in Development (Norway) and University of Botswana: Gaborone, Botswana.

Meyer, Wayne (1995). Phase I Report: Public Works Programs Using Labour - Intensive Methods. A report to the Friedrich Ebert Foundation, Gaborone. Typescript .

Ministry of Housing and Urban Development (1995). Kingdom of Swaziland: Urban Government Policy; August 1995. Ministry of Housing and Urban Development: Mbabane, Swaziland.

Mutahaba, Gelase (1989). Reforming Public Administration for Development: Experiences from Eastern Africa. Kumarian Press: West Hartford, Conn.

Olowu, Dele and James Wunsch (1995). "Decentralization, Local Government and Primary Health Care in Nigeria: An Analytical Study." In Journal of African Policy Studies, 1:3.

Olowu, Dele (1990). "The Failure of Current Decentralization Programs in Africa." In James Wunsch and Olowu, The Failure of the Centralized State. Westview Press: Boulder, CO.

Parson, Jack (1991). "The Peasantariat, Politics and Democracy in Botswana." In Robin Cohen and Harry Goulbourne, editors, Democracy and Socialism in Africa. Westview: Boulder, CO.

Peters-Berries, Christian (1995). Aspects of Local Government Finance in Botswana. Friedrich Ebert Stiftung: Gaborone, Botswana.

Picard, Louis A. (1987). The Politics of Development in Botswana: A Model for Success. Lynne Reinner: Boulder, CO.

Pycroft, Christopher (1994). "Local Government in the New South Africa." In Public Administration and Development, 16: 233-245.

Salinger-McBride, Jan and Louis A. Picard (1989). "Rural Development Areas in Swaziland: The Politics of Intergraded Rural Development." In Comparative Politics 22:1.

Smoke, Paul (1994). Local Government Finance in Developing Countries. Oxford: Nairobi.

Swedish International Development Agency (1993). Decentralization in Botswana: Policy Paper and Action Plan. SIDA: Gaborone, Botswana.

Tordoff, William (1988). "Local Administration in Botswana." In Public Administration and Development 8, 183-202.

USAID, (1990). Strengthening Local Authority in Swaziland: Challenges of the '90s. USAID: Mbabane, Swaziland.

World Bank (1994). Adjustment in Africa: Reforms, Results and the Road Ahead. Oxford University Press: Oxford, U.K.

Wunsch, James and Dele Olowu (eds.) (1990). The Failure of the Centralized State: Institutions and Self-Governance in Africa. Westview Press: Boulder, CO.

Wunsch, James and Dele Olowu (1996). "Regime Transformation from Below: Decentralization, Local Governance and Democratic Reform in Nigeria." In Studies in Comparative International Development, 31:4.