Intergovernmental Relations in Ghana, 1 988- 1 995

co-ordination.

ful committee, the Executive Committee (EC), is the overall governmental district authority with responsibility for development planning and a very wide range of government services. A district treasury has been created to eventually control the budgets of all decon centrated departments and ministries. Below the districts are two fu rther tiers, the Urban/Zonal/Town Councils and Unit Committees, but these are essentially consultative bodies (partially nominated) with no budgets of their own, carrying out functions as de!egated.
The decentralization programme has both explicit and implicit objectives, which can be fo und in government documents and the speeches of political leaders, include empower ment, popular participation, accountability, effectiveness, efficiency, responsiveness, de congestion of the national capital, development and checking the rural-urban drift. I The implicit objective, on the other hand, is what one would call the politics of decentralization. In Ghana, the PNCD that seized power by force reckoned that aid donors on whom it depended for its Structural Adjustment Programme (SAP) would applaud the creation of elected local government councils as a move towards more open, democratic politics. But the PNCD appeared to see decentralized democratization as a substitute for an opening up of politics at the national level. Elected bodies near the local level would give people the sense that they were participating and exercising a little influence over government without threatening the regime's grip on the higher reaches of the state. This !ent a certain demo cratic legitimacy to the PNCD. In short, decentralization is used as an instrument of stabi lizing a political system in crisis and securing some of the political objectives of the government.
2 On the positive side, the politics of decentralization, insofar as it is aimed at political renewal or the opening up of a previously repressive regime, seemed to have created a huge number of opportunities for mostly young people who aspire to a career in politics.
The explicit objectives of the decentralization programme have not, however, been realized. Although democratization of the DAs through the elections of 1988/89 and March 1994 did achieve some success in political terms, this very success paradoxically produced deep fru strations at the institutional level, reflected in mainly marginal improvements in the developmental performance of the DAs. The lack of tangible development results under mined the other objective of the DAs, that is, the creation of a more legitimate and respon sive form of govemment at the local leve1. 3 The under-achievement of the explicit objectives of the decentralization programme is attributed to the uneven distribution of power relationships and distribution of tasks between the central govemment and the District/Metropolitan Assemblies. This paper, therefore, examines the question of how the decentralization policy affects relations between the central and district levels of govemment in Ghana. Specifically, it discusses this balance of political power by focusing on the nature and character of central-local relations fostered under the decentralization programme. in short, it deals with the central izing features of decentralization in Ghana. The paper is divided into three parts. The first part is devoted a discussion of the concept of intergovemmental relations. The second part identifies and evaluates the nature of intergovemmental relations in Ghana and the degree of autonomy that should be given to the decentralized units. The evaluation is based on following four indicators: (i) choice of levels; (ii) choice of decentralized authority; (iii) choice of task to decentralize; and (iv) choice of device used to decentralize power. 4 The third part draws on the lessons of central-local relationships in Ghana and their impli cations for relocating govemment administration.
The Concept of Intergovernmental Relations (IGR) W. Anderson 5 , one of the earIiest scholars on intergovemmental relations (IGR), has defined it as "an important body of activities or interactions occurring between govemment units of all types and levels within ... the federal system". In other words, IGR refers to the interactions that take place among the different levels of govemment with a state. Ander son's general definition has been fleshed out by Wright 6 , who identifies five distinct characteristics. First IGR recognizes the multi pli city of relationships between all types of govemment. Second, it emphasizes the interactions between individuals, especially public 4 5 officials. Third, these relationships are continuous, day-to-day and informal. Fourth, IGR insists on the important role played by all public officials, be they politicians or adminis trators. Fifth, it emphasizes the political nature of relationships and focuses on substantive policies, especially financial issues, such as who raises what amount and who spends it for whose benefit with what resultat. Summarizing these characteristics, Wright notes that "the term IGR alerts one to the multiple behavioural continuous and dynamic exchanges OCCUf ring between various officials in the political system,, 7 .
The concept of IGR is usually associated with states having a federal administration system where the relationships between the federal, central or national government and the major sub-national unit (state, region or province) are fo rmally spelt out in the constitution. In particular, the jurisdictional powers of each level of government are spelt out in the consti tution and any rearrangement must be through a constitutional amendment, involving both levels of government. Although the emphasis in the analysis of IGR is on federal-state relations, the fuH picture also inc1udes how both levels relate to the local government units established within each state. The result, then, is that a fuH analysis of IGR within a federal administration system must cover the following: federal-state, federal-state-Iocal, federal local, inter-state, state-local and inter-local relations. 8 With regard to a unitary state, like Ghana, IGR refers to the interactions between the central government and the sub-national government units. In some unitary states there is only one category of sub-national government commonly referred to as local government units. In such states, IGR refers to the interactions between the central (national) government and the local government units, and it is more commonly called central-local relations. In some other unitary states, like Ghana, sub-nation government units exist at two levels, the local and regional levels. In this second category of states the pattern of interactions inc1udes central-regional, central-regional-local, central-local, inter-regional, regional-local and inter-local relations. 9 Notwithstanding the fact that the concept of IGR exists in a unitary state, like Ghana, the concept is generally associated with federal states for two reasons. First, the formal consti tutional allocations of governmental fu nctions between the federal and state governments in a federal system is absent in a unitary system. modify these fu nctions without consulting the lower units. Second, and related to the fi rst reason, is the ability of the central government in a unitary state to determine unilaterally both the substance and style of intergovernmental interactions. What this means is that whilst some of IGR exists in a unitary state, it differs substantially from the IGR in a federal 1 0 state.
The concept of IGR covers all forms of decentralization, such as deconcentration, delega tion, devolution, federalism and regionalism. Decentralization refers to the distribution of power to lower levels in a territorial hierarchy, whether the hierarchy is one of governments within a state of offices within a large-scale organization. In other words, decentralization either refers to the real division of powers ll or an act of "reversing the concentration of administration of a single centre and conferring powers on local government" 12 . So defined, the term decentralization encompasses both political and bureaucratic/adminis trative decentralization or decentralization between levels of government and within each type of government. Decentralization, therefore, involves ad elegation of power as weil as changing the power relationships and distribution of tasks between levels of government. 13 Against this theoretical background, we now proceed to examine the nature of central-Iocal relations in Ghana on the basis of the four indicators earlier on stated.

Choice of Levels
The choice of levels has a strong influence on the degree of central infl uence and commu nity involvement and on the direction and extent of accountability. Basically, four admin istrative levels or tiers have been created in Ghana, even the three Metropolitan Assemblies have a five tier structure. They are (i) 10 Regional Coordinating Councils (RCCs), which occupy the apex or upper tier of the structure; (ii) 3 Metropolitan Assemblies (MAs), 4 Municipal Assemblies (MAs) and 103 District Assemblies (DAs); (iii) Urban/ZonallTown Councils, and (iv) Unit Committees (see Figure 1). Each of the 10 regions of Ghana has a Regional Coordinating Council (RCC). Unlike at the district level, the RCCs are not elected bodies. Membership consists of the fol lowing: (a) Regional Minister (RM) and his deputies -who are all government appointees; (b) the Presiding Member (Speaker) of each District Assembly (DA) and the Distriet Chief Executive (DCE) -a government appointee -of each district in the Region; (c) two chiefs from the Regional House of Chiefs elected by the chiefs at a meeting of the House; and (d) the regional heads of the decentralized ministries in the region as members without the right to vote.
The chairman of the RCC is the Regional Minister (RM), who represents the President and coordinates and directs public administration in the region. The secretary to the RCC is the Regional Coordinating Director (RCD), the head of the Regional Administration.
The main role of the RCCs is to serve as an intermediate coordinating and monitoring link between the central government and the districts. Specifically, the RCCs (a) monitor, coordinate and evaluate the performance of the DAs in the region; (b) monitor the use of all monies allocated to the DAs by any agency of the eentral government; and (e) review and eoordinate publie services generally in the region.
It may seem that the RCCs play a rather seeondary role in the system of loeal government in Ghana. Unlike the DAs, the RCCs do not eommand a loeal resource base in the form of loeally generated revenues. It has also few service delivery fu nctions. Despite its secondary role, the RCCs are very powerful in praetiee, especially when there has generally been no agreement on their "eoordinating and monitoring" fu nctions. They seem to direet and eontrol aetivities of the DAs in praetice to ensure that the DAs operate within limits set by legislation. This is a way of guaranteeing some measure of supervision from the centre. Sometimes the power to approve or reject the bye-Iaws of the DAs is delegated to the RCCs by the Minister of Loeal Government and Rural Development.

The District Assemblies (DAs)
The District Assemblies (DAs), which are eleetive bodies, are the basic units or the fu lcrum of the loeal government. They are found in the 110 districts. They are either Metropolitan (population over 250,000), Municipal (one-town Assemblies with population over 95,000) or District (population 75,000 and over). The eomposition of membership is made up of elected representatives, the District Chief Executive (DCE) -the chief representative of the central government in the district, Members of Parliament (MPs) whose constituencies fa ll within the area of authority of the DA ad 30 % of the total membership of the DA appointed by the President in eonsultation with traditional authorities and interest groups.
The DAs are the politieal and administrative authority in the districts and they exereise deliberative, legislative and executive fu nctions. They also provide guidance, give direction to, and supervise all other administrative authorities in the districts.
The general objectives of the DAs inc1ude all-encompassing fu nctions such as responsibil ity for the (a) overall development of the districts; (b) formulation of strategies for the effective mobilization of human, physical , financial and other resources; and (c) provision of basic infrastructure and municipal works and services. While success in achieving such goals is very difficult to measure, the legislative instru ments setting up each DA provide a very specific list of up to 86 particular duties. Such legal duties do form a useful bench-mark against which to measure the outputs of the DAs.

Sub-District Structures
They are essentially consultative bodies with no budgets of their own. They carry out fu nc tions as delegated by the DAs. They assist the DAs in the performance of their fu nctions. The sub-district structures, which are yet to be established under Legislative Instrument (LI) 1589 of 1994, are also to serve as instruments by which DA members can relate to their communities. The sub-district structures have two tiers: The Urban, Zonal and Town Councils They are not elective bodies. The Urban are created for settlements with populations above 15,000 and which are cosmopolitan in character, with urbanization and management prob lems, though not of the scale associated with the metropolises. 34 of such councils have been created. The Zonal and Town Councils, on the other hand, are established for settle ments with population between 5,000 and 15,000. The Urban, Zonal and Town Councils consist of representatives of the relevant DAs, Un it Committees (UCs) and government appointees, se1ected by the DCE on behalf of the President, after consultation with the Presiding Member and traditional authorities and organized productive economic group ings in either the urban area, municipality or the town. The Councils are supposed to be the rallying point of local enthusiasm in support of the development objectives of the DAs.

The Unit Committees (UCs)
The Unit Committees (UCs) are elected bodies, consisting of not more than 15 persons, made up of 10 persons e1ected by universal adult suffrage and 5 government appointees, again selected by the DCE on behalf of the President, after consultation with the Presiding Member and traditional authorities and organized productive economic groupings in the Unit. A Unit is normally a settlement or a group of settlements with a population of between 500-1 ,000 in the rural areas, and higher a population (l,500) for the urban areas. The UCs are supposed to perform roles, like education, organization of communal labour, revenue rai sing and registration of births and deaths. Elections to the UCs were postponed two times in 1995 as a result of inadequate education and preparation by the Electoral Commission.
While, from the national perspeetive, the distriet level appears appropriate for being sensi tive to loeal needs, from the grass roots it has appeared as a remote souree of policies and deeision. Moreover, there has been the tendeney of the DAs to eentralize within their boundaries, thus Iimiting the diseretion of eommunity-based organizations.. Thus, meeha nisms for eommunity involvement often remain as unresolved problems, espeeially when the sub-distriet struetures are not in existenee. To eurb them, open publie meetings and publicity drives are organized by the DAs in the rural areas to ensure so me relative partici pation of the people. These are still inadequate though.
The ehoice of levels reveals one signifieant thing worth noting. This is the faet that the eentral government has been given the power to appoint persons to every tier of the strue ture. Although the government in theory is supposed to appoint people on the basis of their expertise and eompetenee, in praetice most of the appointees are predietably loyal to it. This is to ensure eentral direetion and eontrol of the deeentralized units.

Choice of Decentralized Authority
The DAs, like loeal government units, take areal fo rm, with responsibilities for a wide range of loeal services. The DAs and the Unit Committees are a hybrid form of deeentral ized authority, eombining eleeted and appointed members. Neither are they fully eleeted and responsible solely to their eleetorate nor simply a sounding board for loeal publie or professional opinion. As indieated, the appointed members are to add expertise or make the deeentralized units more representative of their community. The RCCs and the Urban, Zonal and Town Couneils, on the other hand, are wholly appointed.
The question to ask is how does the hybrid form of decentralized authority affeet aeeount ability within the DAs. Or in more specifie terms, whose interest do the DAs represent? This question ean be answered by looking at the relationship between eleeted members and the electorate, eleeted members and the offieials and appointed members and their groups, Iike traditional authorities and other interest groups in the distriet.
The relationship between elected representatives and their electorate was far from being eordial. One of the major claims made by the governmcnt for the DA system was that it represented a significant contribution to the development of popular democracy at the loeal level. It is truc that there wcre novel fe atures in the system which justified the claim both in theory and in praetiee. Paradoxically, however, it was the very sueeess of demoeratization whieh led to pressures on the relationship between the eleeted representatives and their eleetorate pressures so severe that they threatened the long-term prospeets of the DAs. The DA members were under offieial injunetion to be responsible for the development of their areas, and were to be judged by their sueeess or fa ilure in eneouraging development.
They were also to be judged by how cIosely they maintained contact with their constituents, a task facilitated both by their local residence and the cIoseness of representation buiIt into the system -an average ratio of one DA member to 2,000 electors. All the pressure there fore was on the DA members to act as delegates for their communities, to fight for a share of the assembly's resources of their area at the excIusion of all else. 1 4 Popular enthusiasm for the DAs when they were first introduced was remarkably high. The policy seemed to herald aboost in both resources and government attention to the problems of the rural areas, albeit that it was linked by the government to the policy of encouraging self-help and community initiative. Expectations were high, and this was as true of the elected DA members as it was of the public. The expectations aroused could not be fu lfilled except in a tiny mi nority of communities, given the limited resources and the large number of "represented" communities in each district. Only a few of the DA members representing the large towns or those with special political influence or "connection" were able to succeed in "bringing development" to their constituents. For the influential DA members, there were two possible options. On one hand, they could emphasize the logic of general taxation, in the hope of getting a return through the "Iocal rebate" schemes tried (unsuccessfuIly) as tax-paying incentives. On the other hand, they could go in for organiz ing local self-help. Both options, of course, had their limitations.
DA members who put all their efforts into encouraging tax paying without succeeding in getting any projects for their areas ended up in a "no win" position. They were condemned by the government and the local administration fo r the low level of DA revenue generation. And they were condemned by their electorate both for faiIing to deli ver on their promises and for being involved in frequently heavy-handed "pay your tax" campaigns. That the legitimacy of local taxation had hardly improved since colonial times was shown by the levels of basic rate collection. An average of 13 per cent of adults had paid the tax in the 1 10 districts. 15 On the other hand, if DA members decided to emphasize the alternative dimension 01' government policy, namely, that communities had to develop the spirit of self-help, they were equally condemned. The electorate complained bitterly of the "double jeopardy" to which they were exposed. First, they were told they must pay their basic rate and other taxes or there could not be local development. But when they asked for a new school roof, a bridge to be repaired or their feeder road to be resurfaced, they were told to levy local contributions and organize local communal labour. Most communities quite rightly could not see the purpose of DA taxation when the only projects or services they obtained were through their own local efforts. The dilemma posed by the competing demands of community-based development contributions and DA taxes was a direct result 14 Crook; Ayee, 1994; Ayee, 1996 (note 1).
of the unresolved contradiction between participatory and direct democracy embodied in the DA concept. All that success in creating larger numbers of self-help projects achieved was to create fu rther demands for aid with building materials which the DA could not meet. It was not surprising therefore that many DA members feit demoralized. The pressure to become over-involved in tax collection had a profoundly destructive effect on their role. The more closely identified a DA member was with his electorate, the more keenly he feit the pressure of disappointed expectations. If the DA member had not been so closely identified with their communities and so exposed to an intense social pressure to deliver benefits to their electorate, they might not have fe it so embarrassed and fed up. Many of them had conscientiously kept contact through regular meetings as required by law. But these tended to be the ones who had either succeeded on the self-help front or who had more or less given up on the AD. Other members who had nothing to show had tven up atten d ing meetings with their constituents or lapsed into resentment at the system. ] The relations between elected DA members and officials, on the other hand, were also not very cordial. When a bureaucratically run local is either taken over or replaced by an elected authority, the first and most important change to be feit is in the formal structures of accountability. Officers now have to report to and be guided by the decisions of elected members; in the Ghanaian case they also had to work closely with elected members in subcommittees and report personally to the executive committee of the DA. DA members in Ghana struggled to achieve control over policy-making, policy implementation and the execution of the budget. The reason was mainly structural. The most important structural factors were the position of the District Chief Executive (DCE), which until the coming into force of the 1992 Constitution is known as the District Secretary (DS) and the failure of the government to establish a legal basis for a genuinely decentralized civil service.
With regard to the role of the DCE, the fu ndamental problem remained the apparent unwillingness of successive governments in Ghana to abandon the colonial district administration system. The DCE's political role is paramount. Any duty he performs in and out of the DA is defined primarily by the central position he occupies in the political life of the district. In the DA he is the representative of the central government; and his office is a conduit for transmitting central government concerns to the district at large. He therefore has dual allegiance which, because of his office, is heavily tilted towards the fo rmer. He is the most influential member of the DA, as ret1ected by his chairmanship of the Executive Committee, the nerve centre of the DA, which exercises its executive and co-ordinating fu nctions. Neither in law or practice does the Presiding Member (PM) ever present a real threat to the dominance and pre-eminence of the DCE. For this reason the PM is specifi cally excluded from membership of the executi ve committee to the full DA. Of course a PM who has secured his membership of the DA by election, as opposed to one appointed, ] 6 See note 13.
is inclined to see himself as wielding more power in the DA than the DCE, whose political life is tied to the fortunes of the government in power and is not dependent upon the existence of the DA.
The control exercised by most DCEs is manifest in their confrontational attitudes to DA members which, in the Adansi West District, contributed to low attendance at DA meetings. The Deputy District Co-ordinating Director (DCD) of the Afigya-Sekyere District, the second topmost administrator in the district, clearly shows the depth of the DCE's influence: "When a suggestion is made by the DCE it is difficult to kick against it. In his absence, the situation during the Assembly meetings is different; everybody is free to speak their mind. The presence of the DCE is detested by almost all the Assem bly members. Those who do not acknowledge his influence are the illiterates.,, 17 The influence of the DCE in decision making is not in doubt. For instance, in the Amansie East District, the DCE influenced decision reached; his counterpart in the Sekyere West District used his discretion on certain occasions to distribute development materials. This being so, it seems that the consensual style of decision making prevalent in most DAs is merely a matter of convenience. There is, then, a tendency for DA members to regard deci sions as being made in the interests of the DCE and the central government, rather than those of the pubIic. The current practice of the President appointing the DCE, with the prior approval of two-thirds of the members of the DA present and voting, is a simple way to get round the DCE's contentious role in Ghana's decentralization programme. But this leaves important questions unanswered: How can the DCEs reconcile their position as government representatives with their accountability to the DA whose members have voted to ratify their appointment? And how does the managerial and administrative authority suggested by their title affect their political authority? A DCE can only assuage power struggles and central government influence in the DAs but never eradicate them completely. Given the role of the DCE in previous decentralization rdorms, any official who replaces the DCE will not merely be a titular head but will eventually become a power broker. 1 8 The other structural impediment to the achievement of fu ll control by the elected DA members is the government's tardiness in implementing administrative and fiscal decen tralization. Even seven years after the DAs had been set up, a legal basis for the decentralization of departments and central mi nistries had not been enacted. As a result, staff in the "decentraIized" departments at the district level are still appointed by and responsible to their parent ministries through the civil service system. The District Co-  ordinating Director (DCD) to whom much of the DCEs work is delegated and who acts as secretary to the DA has to work very closely with DA members. But he remains in practice an established civil servant whose position and career are legally separate from that of the DA, despite the promulgation of the new Civi! Service Law (pNDC Law 327, 1993) and the Local Government Act (Act 462, 1992).
The relationship between appointed members and the groups they are supposed to represent is non-existent because there is no formal mechanism for consulting the groups or for being held accountable to them. Not surprisingly, appointed members find their position uncom fortable and fru strating. Though the government was keen to emphasize the skills and expertise appointees could offer, the real motive behind their inclusion in the DAs was political; it was to give them "political direction" (which in the language of the day meant nothing more than serving as government watch-dogs) and ensure that the presence of the government was feit in the DA fo rum. The government hoped that appointees, like the DCEs, would give some certainty that the DAs would operate within the confines of central government policy.
The government's hope of using the appointees to its own advantage has not fu lly material ized. Some have lived up to their image of providing expertise to their DA. Others have been the most active critics of the local administration and the government itself. Never theless, despite some appointees having successfully confounded the public view of them as "government agents or stooges" one cannot fai! to perceive the real threat to the autonomy of the DAs posed, as a class, by the appointed members. There is evidence in some of the DAs that the pro-government stance of some appointed members has not disposed them to a harmonious relationship with their elected counterparts.

Choice of Tasks to Decentralize
Irrespective of the type of decentralization, the tasks that are decentralized, and the relative balance of responsibilities for each between different levels, seem to strongly influence the degree of power that decentralized units wield, and the extent to which accountability lines run upwards or downwards. Those tasks are (i) revenue raising; (ii) policy-making and planning; (iii) resource allocation; and (iv) interagency and intersectoral co-ordination.

Revenue Raising
One aim of decentralization is to mobilize local-Ievel resources for development. It can occur in two ways -when eJected bodies impose taxes, and when citizens make voluntary contributions to development projects. Has this been the case in Ghana? The ans wer to this question requires fu rther probing. The decentralization of the government consisted of three main elements; namely, the transfer of power, the transfer of competence and the transfer of means, whereby means implied resources, such as finance without which power and competence would not be effective. Indeed, the decentralization policy had made the DAs take on additional expenditure responsibilities, such as the provision and maintenance of second-cycJe educational institutions, which hitherto were performed by the central government. As a way of maki ng the DAs financially buoyant therefore, the Local Government Law (PNDC Law 207 of 1988) set up a financial base that covered sic broad areas: local taxes, such as rates, fees, Jevies and licenses, which the DAs were empowered to impose and collect; bank borrowing within Ghana for investment; investment income; central government development grants; shares of revenues collected by the central government on specific natural endowments, for example, stool lands, and shares of seven central government revenues, such as casinos, betting, gambling, entertainment, daily transport, advertisement taxes and income tax (registration of trade and vocation), which were ceded to the DAs. The first five were used by previous governments. However, the sixth, the ceded sources were, until 1988, coJlected by both the central government and the district councils concurrently and shared between the two levels of government equaJly. The PNDC governments's arrangement differed from that of previous governments in that aJl the sources were to be shared among the 110 DAs and not between the two levels of 20 government.
Ceded revenue added f: 135,556 to the total revenues of the DAs in 1989, f: 151,555 in 1990, f: 294,000 in 1991, f: 1,050,000 in 1992 and f: 600,000 of the first half of 1993?1 In most cases the yield from the ceded revenue even exceeded the total revenues mobilized from the districts themselves by the DAs. Monies generated from the ceded revenue were shared among the DAs by a fo rmula developed by the Planning Division of the Ministry of Local Government. The factors of the revenue sharing fo rmula were: a flat equal sum for each DA, designated as "equality"; the population of each of the districts; and assessed development status or depri vation of each district, calculated in terms of availability of services, basic educational facilities, justice and security services and accessibility to administrative and fi nancial institutions. 22 20 Ayee, 1994 (note 1). The finaneial paekage under PNDC Law 207 did not improve the capacity of the DAs to perform the fu netions assigned to them. The DAs still looked to the eentral government to provide grants for the development projeets, but these were mostly not fortheoming. As a result most DAs resorted to illegal taxation, including levies on export erops, whieh, together with a general inerease on domestie rates, eost them eleetoral popularity. Although mueh emphasis was plaeed on improved revenue eolleetion from eeded revenue, market tolls, lieensing, fees, stool land revenues, property and basic rates and special levies, they did not produee adequate revenues. It is noteworthy that the PNDC, despite its so-ealled eommitment to deeentralization, refused to eoneede any of the eentral government's luerative revenue sourees, sueh as duties on exports and imports, to the DAs whieh were expeeted to assume responsibilities for loeal development. Thus, the insolvent financial position of many of the DAs rendered them ineapable of beeomi ng promoters of socio-. d eeonomle eve opment II1 t elr areas.
The DAs also had only limited sueeess in inereasing revenues from rates and other forms of loeal taxation for a number of reasons. First, most of the DAs are loeated in impoverished areas, whieh limited the aetual amount of mobilizable resourees. Seeondly, even if resourees were there to be mobilized, eleeted DA members naturally hesitated to impose fresh taxes on their eonstituents sinee they made the representatives unpopular. They were partieularly reluetant beeause most of the resourees whieh mi ght be tapped were in the hands of loeal elites, from whieh most DA members eome and on whom they often depend at eleetion time. Thus efforts to mobilize loeal resourees often amount to attempts to persuade loeal elites to tax themselves. Or they bring eleeted DA members into bitter eont1iet with their eleetorates. Thirdly, the long history of the fai lure of loeal government in Ghana and of the misuse of loeally mobilizable resourees has made taxpayers eynieal about wh at the DAs may do with loeal eontributions or tax revenues. They are therefore generally unwilling to pay rates. Fourthly, the seope of the fu netions alloeated to the DAs in relation to their size and resourees almost guaranteed that large numbers of the eleetorate would be disappointed. Even if the DAs had made a better job of revenue eolleetion it was impossible that every eommunity in the 110 distriets eould have its demands for large-seale infrastruetural and social development. 24 In the drafting of the 1992 Fourth Republiean Constitution, it beeame imperative that the financial base of the DAs needed to be strengthened if they were to perform their statutory general and speeifie fu netions. It needs to be emphasized that the decentralization policy of the govcrnment was meant to aehieve two things; first, to stimulate greater popular partiei pationlinterest in loeal government through enhaneed representation and influenee on loeal 23 Ayee, 1994; 1995 (note 1).
24 See note I. development projects; and secondly, to encourage greater readiness on the part of local populations to contribute to local development efforts (via, among other things, rates and other forms of local taxation). There was, however, a conflict between these aims on the one hand and, on the other, the continuous reluctance of the central government to allow very much real autonomy to the DAs and in addition pressure from the World Bank, through the Ministry of Finance, to keep down government expenditure as part of the structural adjustment programme which started in 1983. The result was a vicious cycIe whereby the DAs had little revenue for development projects, hence local populations lost interest and became reluctant to contribute more through local taxes to DAs which seemed to be doing little for them? 5 The establishment of the District Assemblies' Common Fund (DACF), under Section 252 of the 1992 Constitution, into which not less than five per cent of total revenues of the Ghana government are to be allocated to the DAs for development, was intended to provide a way out of this situation.
The DACF is administered by a District Assemblies' Common Fund Administrator (DACFA), who is appointed by the President with the prior approval of Parliament. The DACFA performs basically three functions: (i) proposes a formula annually for the distri bution of the Common Fund for approval by Parliament; (ii) administers and distributes monies paid into the DACF among the DAs in accordance with the formula approved by Parliament; (iii) reports in writing to the Minister of Local Government on how allocations made from the DACF to the DAs have been utilized.
The DACF is shared on the basis of a formula: (i) need -35 %; (ii) equalizing -30 %; (iii) responsiveness -20 %; (iv) service pressure -10 %; (v) contingency -5 %. Disburse ments from the Common Fund have yielded [600,000 for the second half of 1993, [ 17,000,000 for 1994 and f 23,500,000 for 1995. Since the introduction of the DACF in 1993, local revenue mobilization has on the average declined . Although proceeds from the Fund has also far exceeded locally generated revenue, regional ministers, DCEs and MPs have complained publicIy that the disbursements are not enough to take care of the numerous responsibilities given the DAs. They have, therefore, called for an increase of the 5 %. In 1994 proceeds from the DACF constituted 19.6 % of total development budget of 26 the government.
Under Section 9 of the DACF Act (Act 455) of 1993, the Ministry of Finance (MOF), in consultation with the Ministry of Local Government (MLG), was authorized to determine the category of expenditure of the approved budget of a DA that must be met out of the DACF. After the areas had been known, the DAs would se\ect projects to the value of their 25 Ayee, 1995;1996;Crook (note I). 26 Ayee, 1995 (note I). share and communicate this to the DACFA and the MLG. This stipulation was carried out in 1994 and 1995 by a directive from the MOF, in consultation from the MLG, to all the 110 DAs to concentrate their allocations from the DACF on specific projects in the areas of health, education and central government undertakings . This directive was seen as under mining the autonomy of the DAs. The proceeds from the DACF are specified for certain purposes and so restrict the freedom of the DAs to distribute resources according to their own sense of priorities. This reinforces the view that it is possible to equate financial dependence with centralized contro!. In other words, centrally-provided finance, such as the DACF, accompanied by other instruments of guidance and control, tends to erode the autonomy and independence of sub-national governments. As central fu nding grows, so does central contro! .
27 Without local autonomy to manage resources, there can be no genuine accountability. However, given the limitations on local mobilization of resources which afflict all the DAs in Ghana, a significant element of transferred central fu nding in f .. bl 28 act appears mevlta e.

Policy-making and Planning
The 86 fu nctions assigned the DAs include all-encompassing fu nctions such as responsi bility for the overall development of the district, the formulation of strategies for the effec tive mobilization of human, physical, financial and other resources and the provision of basic infrastructure and municipal works and services. They tend to be numerous, unclear and permissive as opposed to mandatory. Most of the fu nctions are regulatory, with little institution-building activity taking place. In addition, the fu nctions also show that the responsibility of the DAs is restricted to local issues. They can make recommendations to the central government but through official "clarification", directives and exhortations, the DAs have been made to understand that their domain of influence is the district. On malters that can even remotely influence national policy they have no power. In so far as the practi cal powers of the DAs are concerned, decentralization has been turned into centralization. In fact, with the exception of assisting local development through the provision of fi nance and materials, the DAs have not, in practice, been entrusted with any major responsibilities at all; they remain on the sidelines of central government activity. Their confinement to local issues has constrained their link role. They transmit the purposes of government to the people. But their ability to persuade the government to modify its policies to conform to popular opinion has been severely curtailed.
The DAs do not influence national policy. But government has the power to influence their decisions. The scope of their operations is defined by law promulgated by the government. But the law has not set out, in detail, what the DAs can and cannot do. And the interpreta tion of the powers of the DAs rests with the central government. Indeed, all bye-Iaws of the DAs are approved by the Ministry of Local Government before they can become operati ve. Any act of DA which the government considers to be irreconcilable with the law is declared illegal. This procedure has been used to set aside certain decisions of the DAs, even where they were in the interest of their communities. Moreover, the central government has the power to dissolve the DAs and replace them with interim management committees (a power that has been used four times since 1989) without consulting the local people who elected the majority of the DA members.
The DAs are designated District Planning Authorities (DPAs) while the District Planning . and Co-ordinating Units (DPCUs) integrate and co-ordinate the district sectoral plans into long, medium and short term plans. After two public hearings of the plan both at the sub district and district levels the plan is forwarded to the Regional Co-ordinating Council (RCC) for harmonization with plans of other districts within the region. The plans are then sent to the National Development Planning Commission (NDPC), which is given the responsibility for harmonising nationallsectoral policies and guidelines for the preparation of development plans at the local level. The discretionary power granted the NDPC to approve plans in accordance with national development goals may defeat the decentralized and partici patory aspect of the planning system. This is because there is the tendency on the part of the NDPC to impose plans on the DAs and thereby control them rat her than acting in a purely advisory capacity. In spite of the stipulation that the DAs conduct plan hearings at both district and sub-district levels. None has been conducted, while the DAs have been given up to April 12, 1996 to submit their plans to the NDPC through the RCC. The order by both the Ministry of Local Government and NDPC undermines the participatory and decentralized nature of the planning system and reinforces the top-down heaviness and centralized character of planning in Ghana since independence. One should not, however, lose sight of the fact that a planning system is frequently seen as a mechanism for retaining some central influence while permitting considerable local discretion. This is in theory rather the practice in Ghana.

Resource Allocatioll
Given the substantial dependence of the DAs on central fu nding, resource allocation proce dures are vital in affecting the degree of local autonomy and the extent to which the local level can adapt national policies to suit local preferences. Although by 1992 distriet treasuries had been set up, the idea of a district composite budget remained a fiction so long as the Ministry of Finance refused to relinquish control to the DAs. lndeed, central financial control over capital spending and official payments had become tighter as a result of structural adjustment. Perhaps even more importantly, the DAs' subordinate staff were not fu lly at their disposa!. The districts had been given strict instructions on manning levels by the Office of the Head of the Civil Service (OHCS), and the districts were not permitted to recruit without permission from the OHCS and the Ministry of Finance. Even the transfer of a grade I typist had to be done through the civil service machinery? 9 DA members often made efforts to participate in policy-making and in the monitoring of policy implementation and expenditure contro!. This was most evident in the proceedings of the executive committees and their subcommittees. The full DAs themselves often generated lively exchanges as weil, but the end results were frustration and an inability on the part of members to follow up committee decisions.
The major problem in the DAs was management of the budget, in both its revenue and expenditure aspects Revenue estimates were routinely so overoptimistic as to be incapable of fu lfilment, with shortfalls between 200 to 400 per cent . The lack of revenue meant that in practice expenditure decisions were often made on an informal basis. With regard to expenditures, there were virtually no day-to-day controls for ensuring that expenditures conformed to budget estimates. Before the introduction of the DACF in 1993, recurrent expenditure of all the DAs accounted for between 85 and 87 per cent of their spending. 30 Much of it went into administrative costs, including official travel and transport, entertain ment and the costs of decentralization reform itself: constructing or improving offices, providing housing accommodation for government staff and underwriting the expenses of the DAs and their communities. Some of these costs did not fall within the development budget at all but rather on "other administrative costs", including official entertainment. Richard Crook has pointed out that in the East Mamprusi DA in the Northern Region, the amount spent on official vehicles, travel and transport and entertainment in 1991 was four times greater than the whole development budget. The DAs generally fell far short of both their general and specific statutory fu nctions . This was particularly true in respect of their duties to provide and maintain water, roads, electricity, sanitation and schools . Moreovcr, as Crook 3 1 has emphasized "even services and projects that were provided did not respond closely to popular conceptions of needs, and were therefore not highly valued by the public". Although no controls exist for the allocation of recurrent fu nds, specific central approval is required for major capital expenditures.

Interagency and Intersectoral Co-ordination
The linkages between the different levels and institutions have not been weil established. The intended placement of 22 departments and organizations under the DAs, to promote functional specialization and technical expertise, has been fraught with difficulties. The personnel, who are central government employees, depend on their mother ministries and organizations for practically all resources needed for running of their departments in their districts. This has created problems of double allegiance. The original 22 departments are, however, replaced by 16, 13 and 11 departments under the Metropolitan, Municipal and District Assemblies respectively with the coming into force of a Legislative Instrument (LI) to be enacted by the Ministry of Local Government. Even then the more li mi ted deconcen tration would still not resolve the question of double allegiance which ultimately hinges on the question of who hires and fires the staff of the decentralized institutions.
Certain services have been deconcentrated to the districts. They are the Immigration, Police, Customs, Excise and Preventive, Internal Revenue, Prison, National Fire and Education Services. Although the Services do not come directly under the umbrella of the DAs, they help in the general governance of the districts. And yet there has been poor co ordination between the Services and the DAs, which often lead to conflict. This is the result of lack 01' proper definition of roles.
There is a grey area in fu nctional relationship between the RCCs and the DAs. The Local Government Act (Act 462,199 3 ) gives autonomy to the DAs in a manner that seeks to create the impression that the role of the RCC is to be de-emphasized. Yet when it comes to spelling out the fu nctions of the RCC the same Act makes the RCC responsible far monitoring, co-ordinating and evaluating of the performance of the DAs in the region . This raises the issue 01' competence of both the DAs and RCC. Autonomous local government is more likely to wark at a very local and correspondingly modest level, concentrating on small-scale projects and services . The larger infrastructural developments (secondary schools, water and roads) would be better provided by well-funded regional authorities which in the Ghanaian context at least had always been the most effective levels of close administration until the DA reform attempted to bypass them . De-emphasizing the region and giving greater autonomy to the district could be considered politically prudent but from the managerial point of view this may not be quite rational, since the district lacks the resources to be able to perform fu nctions assign it . Also, the vertical linkage between the DA and the sub-structures is very weak. The linkage is exercised more in the person of a DA member than in the form of an institutional Iinkage.
The role of the Ministry of Local Government (MLG) in the whole decentralization programme needs to be stressed . The MLG functions as a Local Government Secretariat under the Office of the President. Essentially, it is involved in political issues, Iike the President's appointment of 30 % of members of the DAs, appointment of DCEs, establishing DAs and monitoring and evaluating them. Seen in this way, the MLG has considerable control and leverage over the DAs. It is the nerve centre of the government's control mechanisms of the DAs. It issues guidelines to the DAs and approves their bye-laws before they are enforced.
Financial accountability is ensured through regular auditing of DA account. However, DA accounts have shown gross embezzlement and Misappropriation as indicated in the report of the Auditor General. This is mainly due to low calibre of accounting personnel in most districts and the MLG itself . Most accounting staff at the district level are book-keepers while Local Government Inspectors of the MLG, who are to check financial malpractices lack the expertise to do so.

Choice of Device Used to Decentralize Power
Chapter 20 of the 1992 Constitution, "Decentralization and Local Government", stipulates that there should be a decentralized system of administration in which functions, powers, responsibilities and resources should be transferred from the central government to local government units, called District Assemblies (DAs). Viewed in this way, there is a provi sion for a proper balance of power relationships between the central government and the structures below it. However, there appears to have been a gap between the theory and practice of decentralization. There is also a contradiction between the centralizing fe atures of decentralization and the government's portrayal of the decentralization programme as a manifestation of its commitment to participatory democracy and the grant of power to the people.
The legal framework for decentralization is provided by the following: The unusual incorporation of decentralization cJauses in the Constitution gives them exceptional prestige, signifying their recognition as fundamental rights. However, constitu tional provisions are by no means guarantees for a harmonious and equal intergovernmental relations.

Conclusion: Implications for Relocating Government Administration
The relationship between the central government and the DAs in Ghana is basicaHy one between a superior and subordinates rather than one of partnership and reciprocity. Central control over decentralized units is important in a number of ways. It guards against the disintegration of govemment machinery, conserves national unity, and maintains the supremacy of the central government. In a unitary state, like Ghana, districts remain in an integral part of the machinery of the central government, and so their powers and functions are subject to the overriding authority of the centre. Such direction and control are necessary to harmonize the activities of the decentralized units with the political purposes of the government. It often happens that the interest of the people is projected as the driving force behind central control over local government units. Thus the fear that the DAs might infringe their people's economic and civil liberties in raising revenue, passing bye-laws and pursuing general development activities partly underJines the need for central government to limit the exercise of their powers. This is particularly important in a country such as Ghana where democratic practice or ethos have not yet taken root. Central control therefore over local government units is not necessarily evil. It becomes evil, however, when it forsakes its proper sphere of constructive control and meddles with the details of local government processes.
In a unitary state local government units will never be independent from the central government. Decentralization will be a question of degree and type of autonomy. One distinction which is commonly made in this respect is the distinction between local control 32 over means and local control over goals. In the case of Ghana the local control over means is definitely low. The paper has shown that the central government has very tight control over finances, manpower and technical and managerial expertise. The simple fact that the whole administrative and technical apparatus on which the DAs are supposed to rely for implementing their policies is composed of deconcentrated element of the central government machinery serves to underscore this fact.
It may be more disputable to what extent central government also exerts strong control over the goal formulation of the DAs. Formally, the DAs have wide powers to decide on goals. This is, however, apower which is basically restricted to a freedom to formulate priorities and planning proposals to be submitted to higher level authorities within the framework of the national planning system. It is a right to participate in influencing the formulation of the national goals more than a power to decide on the actual goals for local development. Whatever input the DAs provide into the national system is supposed to be harmonized by central government authorities, like the National Development Planning Commission (NDPC) into a national plan. An equally important role of the DAs within this framework is to serve as the local implementors of this comprehensive national plan at the local level once it has been developed. More generally, the legislation is very dear, both in its form and in its intentions, on putting the DAs under obligation to serve as instruments for implementing national goals. Viewed in this perspective the over all assessment will have to be that also the real power to decide on goals is highly restricted.
The DAs by their fu nctions, responsibilities and relationship with the central government are not self-governing bodies. Rather the self-governing bodies are local branches of some of the non-governmental organizations (NGOs) and Sasakawa Global 2000. They operate to supplement the efforts of the DAs and the government. The NGOs have a rural focus and have assumed in increasingly important role in the process of service delivery and institu tionalization. They have provided water, dinics, latrines and schools to vulnerable communities in the districts.
Donor agencies, such as the World Bank, Canadian International Development Agency (CIDA), Overseas Development Administration (ODA), World Health Organization (WHO), European Union (EU) and United Nations Development Programme (UNDP) have assisted in areas of infrastructure programmes and capacity building. For example, the World Bank supports the most urbanized district infrastructure programmes. CIDA and the EU, on the other hand, through their funds for local initiative have provided assistance for basic needs to the DAs in the form of support for schools, dinics, water, latrine construc tion and income generating activities. The presence of donors in the districts have created legitimacy problems for the government because of the possibility of local people mistakenly crediting NGOs for every project.
In the conceptual discussion of decentralization, scholars have drawn a distinction between positive and negative autonomy. Autonomy in the negative sense means that the local community has an interest in "protecting" itself from central influence and "penetration". This is a defensive form of autonomy. It may be based on, for example, a feit need to protect local culture or local economy from outside pressure and influence . Autonomy taken in the positive sense, on the other hand, means that the problem confronting the local community is not so much its ability to cut itself off from the outside world and live its own life, as it is the question of eapaeity -or laek of eapacity -to get aeeess to the eentre's resourees and to influenee its environment. 33 The Ghanaian system of deeentralization is quite c1early based on the assumption that deeentralization is all about positive autonomy. The formal design and the politieal rhetorie of the system fo eus on enabling the loeal eommunities to partieipate in the national decision making system. It is also a system whieh is quite c1early designed so as to facilitate the diffusion of goods and values from the eentral to loeal level. It is, in short, a system whieh seems to be based on the idea that development depends on eentral distribution of goods and resourees. In this perspeetive the main problem eonfronting loeal eommunities is pereeived of as that of getting aeeess to the eentral level.
This perspeetive is fu rther underseored by meehanisms in the system c1early aiming at preventing loeal aetors from exercising negative autonomy, that is, to eut the loeal eommunity off from national governanee. Generally, the system of deeentralization as it is designed in Ghana is as mueh a system for penetrating traditional loeal structure and integrating them into a national proeess of modernization as it is a system for loeal influenee on national polieies. If deeentralization in Ghana is to be assessed on its own merits, and in view of its own objeetives, it has to be assessed as a system for facilitating positive autonomy, not as a framework for asserting loeal freedom in the sense of providing a framework fo r "Ioealism". The latter is something which the system quite clearly aims at preventing. The DAs eontinue to operate within a strueture of eentral politieal, administra tive and financial eontrols of a quite fundamental kind. It is not for want of trying that DA members are unable to ehallenge the power of the DCEs or the Ministry of Finanee and the ei vii serviee hierarehies.
The Ghanaian deeentralization programme has been deseribed as ambitious and ambivalent. As deeentralization of the "maehinery of government", as it is frequently referred to, it is ambitious. As deeentralization in the sense of really delegating power to loeally rooted institutions it is at best ambivalent, imposing so many eontrol meehanism and holding baek so many resourees that would be needed to aetually operate the system loeally, that it amounts to subverting the system altogether. Are these two fe atures of the deeentralization programme interrelated? It may be that the ambitions eoneerning the seope of aetivities and fu netions to be moved from eentral to loeal level is aetually so big that it beeomes unrealistie to let loeally based institutions take eontrol over them. Partieularly in a eountry with weakly developed common political norms and a weak democratic tradition the more fu ndamental need to preserve national unity (irrespective of the type of regime) might take it risky to create local bodies with the real powers wh ich are indicated in the more programmatic phrases of the decentralization policy. Perhaps decentralization in more real terms would have benefited from a less ambitious decentralization programme which, on the other hand, could have left more real autonomy to the local level conceming those tasks and responsibilities which were actually given to them. raises what amount and who spends it for whose benefit with wh at resultats. The paper argues that subnational governments, if they should have any political credibility and legitimacy, must have some measure of independence in the level of revenue they raise, the choice of public goods on which to spend it and the capacity to "hire and fire" staff.
An evaluation of the nature of intergovernmental relations in Ghana and the degree of autonomy that should be given to local management of resources will be based on four indicators: (i) choice of levels; (ii) choice of decentralized authority; (iii) choice of tasks to decentralize; and (iv) choice of device used to decentralize power.
Lean mallagemellt-Reforms in Developing Countries in Sub-Saharan Africa

By Hildegard Lillgllau
Public administrations often represent a major obstacle to development in developing countries in Sub-Saharan Africa. The analysis of public administration reforms in Benin and Uganda shows that public administration reforms are both necessary and fe asible. For reforms to be successful, they require the commitment of a reform-oriented government and a comprehensive approach to overcome structural dysfunctionality. Reform-and development-oriented governments of developing countries like Uganda give an impressive picture of how imaginative public administration reforms can be implemented effectively.
If, in contrast, it is not possible to break the dominance of self-enriching ruling classes and to restrict their power, perceptible improvements have yet to occur, as is the case in Benin.
The new public management debate and the lean management reform concept in particular offer important starting points and instruments for necessary reform projects and respective support measures in the framework of development cooperation.