The politics of municipal fragmentation in Ghana

The scholarly debate over the rival merits of local government consolidation and fragmentation is an old but enduring one. However, in this debate very little attention has been focused on the political dimension of council amalgamation and fragmentation – yet political considerations play a central role in both the formulation and outcomes of de-concentration policy. The purpose of this article is to fill a gap in the literature by examining local government fragmentation in Ghana from 1988 to 2014. The article does this by identifying the key players and analysing their interests and gains, as well as the tensions arising from the fragmentation exercise. The implications from the Ghanaian case for more general theories of fragmentation are drawn out.


Introduction
Local government fragmentation, which consists in the political and geographical division of the local authority into smaller jurisdictions, has been used as an instrument for decentralising government power in many countries.This instrument has been chosen because its proponents believe it is a promising way of realising the goals of decentralisation -although decentralisation defies a precise definition since different scholars look at different dimensions of the concept.The discourse and hugely diverse literature on decentralisation have generally focused on: its nature and scope; its relationship to development; its organisation as a system of governance; monitoring of its programmes; and case studies of countries, regions, sectors and programmes (Smoke 2003;Egbeya 2010;Olowu and Wunsch 2004;Dickovick 2011).Rondinelli (1981) refers to decentralisation as the transfer of authority to plan, make decisions and manage public functions from a higher level of government to any individual organisation or agency at a lower level.Similarly, to Smith decentralisation denotes 'reversing the concentration of administration at a single centre and conferring powers on local government' (Smith 1985: 1).In political studies, decentralisation means the territorial distribution of power.Its focus is on the extent to which power and authority are dispersed through the geographical hierarchy of the state, and the institutions and processes through which such dispersal occurs (Ayee 2008a).Decentralisation involves the sub-division of a state's territory into smaller territories and the creation of political institutions in those areas (Smith 1985).
These definitions highlight the transfer of power, authority and resources from central government and its agencies to lower-level government structures to efficiently and effectively carry out responsibilities of local relevance in a participatory manner at that level.Among other benefits, it is expected that decentralisation will enhance participation by citizens in local development decisions, ensure greater accountability of government to the population, and improve political stability through the legitimisation of differences in local needs and perspectives (Ayee 2008;Mawhood 1995;Schonberg 2010).Against this, possible limitations of decentralisation that have been highlighted in the literature include, among others, that it encourages separatist or fissiparous tendencies, it accentuates existing inequities in a country and results in power concentration at the local level, and an increase in corruption (Ayee 1999;Smith 1985;Griffin 1981).
However, despite its weaknesses decentralisation continues to be a preferred policy option aimed at achieving efficiency and effectiveness in the conduct of government business in many countries.Thus in 1988 Ghana embarked upon an unprecedented decentralisation project -one that seems to be unparalleled in Africa -when the Provisional National Defence Council (PNDC) military regime launched a decentralisation programme which aimed to 'give power to the people' and bring 'democracy to the doorstep', reflecting the political thinking of government at the time (Ayee 2008a(Ayee , 2008b)).Later, from 2001 to 2007, the New Patriotic Party (NPP) government rolled out an invigorated decentralisation programme to quicken the realisation of these policy goals.The current government under the National Democratic Congress (NDC) has promised and acted to reinforce and increase decentralisation to accelerate the pace of development (Ayee and Dickovick 2010).
However, even though there has been a swathe of literature examining Ghana's decentralisation programme since 1988, none has explicitly focused on fragmentation -the political and geographical division of the state into smaller jurisdictions -which is arguably one of the key processes of decentralisation (Ayee 2008a(Ayee , 2008b;;Crawford 2008;Awortwi 2010;Egbenya 2010).The only author who has examined fragmentation, including its politics, is Ayee (2013) but this work is contextually limited.However, mapping the contexts of problems and issues is important for understanding the multiple dimensions of knowledge, beliefs, and power, as well as the meaning and values that frame policy-making and implementation.
Accordingly, this paper addresses a gap in the literature: the scant attention paid to local government fragmentation and the politics of fragmentation, despite their very real consequences for policy outcomes.These politics play out in a number of ways, including: contestation over who has power to create districts; conflict over the distribution of benefits and costs; discontent over inadequate public consultation on the need for and nature of reforms; tensions and revival of traditional rivalries between towns; and the perceived electoral advantages and disadvantages accruing to political parties from fragmentation.The implications of the politics of fragmentation for general theories of fragmentation and political participation are examined below.
The remainder of the paper is organised as follows.Section two examines the theoretical basis on which municipal fragmentation is predicated.Section three deals with the rationale for local government fragmentation, while section four focuses on actors in the fragmentation exercise, their interests and controversies, and the winners and losers from the process.The final section covers the discussion and conclusion.

Theoretical basis for local government fragmentation
For over a century, proponents of amalgamation and fragmentation have offered dramatically opposed frameworks for organising local government, with each camp stressing its own model's advantages.

Amalgamation arguments
Proponents of local government amalgamation (the merger of smaller districts into a larger local government jurisdiction) argue that large local government areas bring tremendous benefits: more effective local problem-solving; greater opportunities to raise revenue without disadvantaging any group; enhanced functioning in a variety of tasks; and increased capacity for effective exercise of political responsibility due to a greater capacity to balance local needs and resources (Schoburg 2010).
Representational costs are lower in larger councils since there are fewer elected councillors for a given population -an issue that has featured prominently in amalgamation debates in New South Wales (a state in Australia) (Dollery et al 2006;Dollery and Johnson 2005).In Australia and New Zealand amalgamation has also been hailed as strengthening the capacity of local authorities to plan and co-ordinate responses to environmental issues such as watershed zones, or conurbation problems such as transport management (Holzer and Fry 2011;Aulich et al 2011).
However, amalgamation also has its limitations, and is criticised for making it difficult to build citizen identity and participation in new, larger units of local government, and for the inability to deliver projected cost savings in the new larger council areas (Byrnes and Dollery 2002;Dollery and Johnson 2005).Critics also argue that large local jurisdictions can become 'super governments' that may be insensitive to the needs of communities (Schoburg 2010).Moreover, compulsory amalgamation is almost always accompanied by high levels of acrimony and division in affected communities (Leland and Thurmaier 2010).

Fragmentation arguments
Those who favour local government fragmentation claim that it improves local public service delivery, creates more effective local governments and brings local political leaders closer to their constituents (Dollery et al 2006;Dollery and Johnson 2005).It ensures equitable distribution of national resources and accelerated development; allows for local difference, diversity and democratic choice; and serves as a 'laboratory of democracy'.Fragmentation also provides a context where political skills are nurtured; permits service innovation; affords opportunity for diversity in service delivery leading to government innovation; and decentralises government power to the local level (Ayee 2008a;Warner 2012).Moreover, 'small councils typically produce services at lower per capita costs' (Dollery and Johnson 2005: 83).Jimenez and Hendrick (2010) and Bish and Ostrom (1973) agree, insisting that the decoupling of production and provision through contracting-out enables local government to save on cost via scope and scale economies, while retaining control over important policy decisions such as expenditure for a particular service.Boyne (1998: 252) also concludes that 'the broad pattern of evidence suggests that lower spending is a feature of fragmented and deconcentrated local government systems', whereas 'consolidated and concentrated systems tend to be associated with higher spending'.
Evidence however also exists that some fragmented councils have struggled to survive (Tordoff 1980), due to competitive disadvantage and lack of revenue generation sources, requisite professional staff, or community interest in local government activities (Ayee 2000).There is also general agreement in the literature that fragmentation of general-purpose governments leads to lower spending, but fragmentation of special-purpose governments produces the opposite effect (see Berry 2008, Craw 2008and Stansel 2006 among others).There is likely to be greater competition among multi-purpose governments that offer different bundles of services, compared to special-purpose governments that offer single types of services (Boyne 1998;Stansel 2006).Also, since many singlepurpose governments are established to provide capital-intensive services to a large population or area (for example, water provision or reclamation), their fragmentation creates diseconomies of scale (Dowding and Mergoupis 2003) -although this argument has limited application to Ghana where almost all local governments are multi-purpose sub-nationals.But even in Ghana, after more than two decades of continuous council fragmentation, competition among local governments has not materialised due to the small catchment area of many districts.For example, fragmented authorities are unable to respond in a co-ordinated manner to urban sprawl (Carruthers 2003).

Political aspects of council fragmentation
Academic debates over the rival merits of council amalgamation or fragmentation have concentrated on technical aspects rather than exploring their inherently political dimensions.Thus, arguments focus on financial efficiencies and greater community participation whereas in reality the process is deeply political from the central government down to the local level, and in practice this should not be overlooked.This is particularly true of fragmentation policy, where the academic debate has focused on economic efficiencies of scope and scale but has failed to take into account the political dimension.
Yet the incessant tensions surrounding the creation of new local governments in Ghana can be described as a 'politics of fragmentation'.This political dimension has led to sub-optimal decisions in relation to the viability of the newly created districts.It has also engendered conflict because it has altered power relations, redistributed resources, gone against the grain of tradition and custom, and infringed on the interests of certain groups whose support is critical to the success of the policy.
This study suggests at least four important implications for general theories of fragmentation.First, even though decentralisation is often used as a strategy to reduce the monopoly of the central level over decisions of local relevance, the unfettered exercise of the president's power to create local governments may in fact be having the opposite effect.Second, the appointment rather than election of mayors and one-third of local government councillors has undermined local capacity-building and local democracy.Those who have the acumen to perform efficiently and effectively are side-lined in favour of party functionaries with fewer abilities.This is despite the fact that one of the specific objectives of fragmentation reform is to strengthen local public administration and local service delivery.Third, whichever way decentralisation reform is designed and implemented there will be winners and losers, as in any political struggle -emphasising the fact that redistributive outcomes can be contentious.Finally, fragmentation policies have exacerbated central public spending inefficiencies and created disincentives for local public financial accountability.This feeds into some of the key criticisms of fiscal decentralisation: namely that it leads to macroeconomic instability, lack of fiscal discipline, inefficient public spending due to weak local administrative capacity, local corruption and capture by local elites.

Methodology
The empirical basis for this study is content analysis of published documents from 200 interviews in four local government districts (40 interviews with non-elite members and ten interviews with elite members in each district -see below).The sample for the non-elite interviewees in each district was derived using a random sampling method.A list of households generated by the Ghana Statistical Service was used as the sampling frame for the selection of communities and households in each district.In addition, ten elite interviews were conducted with functionaries of local governments in each district.The respondents were selected because of their responsibilities connected with the design and implementation of fragmentation policy.These included the district co-ordinating director (chief bureaucrat of the local government), district finance officer, district planning officer and two councillors.Two district assemblies from the northern part of the country and two from the southern part were chosen.The choice of the four councils was dictated by the desire to ensure regional balance and by the fact that access and participation of respondents had to be negotiated and these varied from district to district.The cases are Sissala East District Assembly (SEDA) and Sissala West District Assembly (SWDA) -both created from the former Sissala District Assembly (in the northern part of Ghana) -and Ga East District Assembly (GEDA) and Ga West District Assembly (GWDA)both born out of the former Ga District Assembly (in southern Ghana).This paper, therefore, discusses the outcome of local government fragmentation based on evidence from these four districts.

Rationale for municipal government fragmentation
Despite the limitations of fragmentation as noted in the preceding section, Ghana opted for it because the perceived advantages outweighed these limitations.Successive governments since independence have therefore created new local government areas with a consequent increase in the number of councils from 65 in 1988 to 215 by 2012 (see Table 1).One of the principal reasons for the fragmentation of local governments is fragmentation's perceived ability to drive local governance and development.For example, the increase in the number of districts from 65 to 110 in 1988 was to 'facilitate economic programmes and to ensure equitable development throughout the country' (Yeboah 1987(Yeboah -8: 1020) ) and thus should be seen as the practical translation of the ideals of former President Jerry John Rawlings' military PNDC government, which was to ensure that 'not only a large majority of the people actually have a say in the assemblies [local governments] but also contribute positively towards nation-building' (Rawlings 1988a: 9422).The justification of former President JA Kufuor for embarking on local government fragmentation was to: Improve administration; deepen democracy and governance as well as the equitable distribution of the national cake as well as accelerated and overall national development (Kufuor 2008: 1).
The creation in 2012 of 45 additional local governments, it was envisaged, would :

Legal framework for local government fragmentation
The legal basis of council fragmentation is the 1992 Constitution and the Local Government Act 1993 (Act 462).Article 241 (2) of the Constitution states that Parliament may by law make provision for the redrawing of the boundaries of districts or for reconstituting the districts.
However, Act 462 gives the president the leading role in the creation of districts.Section 1(2) vests in the president the power by executive instrument to (a) declare any area within Ghana to be a district; and (b) assign a name to the district.The criteria for creation of a new district among others are economic viability and population size as follows: (a) for a district, there should be a minimum population of 75,000 people; (b) for a municipality, the minimum population is 95,000 people as well as a geographical area which consists of a single compact settlement; and (c) for a metropolis, the minimum population is 250,000 people.Within these criteria the law has given the president unfettered power to create districts, which has led to concerns about the sustainability and viability of unabated council fragmentation.

Local government system
Ghana is a unitary state whose government is structurally organised into four layers: national government, including line ministries and de-concentrated entities; ten regions with regional coordinating councils (RCCs) responsible to the centre; 215 metropolitan, municipal and district assemblies (MMDAs); and sub-district structures -urban, zonal and town councils; and unit committees (see Figure 1).The chief executive chairs an executive committee comprising one-third of MMDA members.The MMDAs are nominally non-partisan, as candidates are not sponsored by political parties, and the stated aim is to ensure consensus-building and promote development.However, evidence suggests that presidents routinely appoint people from their parties (Ayee and Dickovick 2010).Below the MMDAs come several sub-district structures, which operate at three levels.First, the metropolitan assemblies are divided into sub-metropolitan district councils.Second, there are urban councils where populations are greater than 15,000 and town councils where populations are between 5,000 and 15,000.These councils are not elected bodies.They are composed of representatives from the district assemblies and unit committees (UCs) as well as representatives of the metropolitan/municipal/district chief executives (MMDCEs).They promote citizens' participation in planning, implementation and monitoring of local services.UCs covers both urban and rural areas.Their members are partly elected (ten members) and partly appointed (five members appointed from among the MMDCEs).

Elements of municipal fragmentation
The council fragmentation exercise in Ghana comprises four components: (i) the establishment of new local government areas; (ii) the siting of local government jurisdictional capitals; (iii) the naming of local governments; and (iv) the elevation of districts to either municipal or metropolitan status (Ayee 2008b).The first two elements have been very contentious, raising a number of questions, including: • Why does the creation of new local government areas occur in election years or a year before elections?• Why is it the president who determines the creation of new local government areas?
• Why are one-third of councillors appointed by the president?
• What are the financial impacts of the creation of new local governments?
• Why is there no consensus on the criteria and timing of the creation of new local governments?• Why is there disagreement over the selection of district capitals?
• Who benefits and who loses from the creation of new local government areas?This paper focuses, exclusively in the case study districts, on the first two elements and the questions they raise as well as the consequences attendant on these issues.

Creation of new local government areas in election years
The president has been accused by opposition parties and civil society organisations (CSOs) of gerrymandering, by creating new local government areas in election years or a year before elections.
Gerrymandering can be defined as 'the practice of establishing a political advantage for a particular party or group by manipulating geographic boundaries to create partisan or incumbent political districts' (Ayee 2013: 12).For Ingraham (2015) gerrymandering refers to the 'drawing of political boundaries to give your party a numeric advantage over an opposing party'.This was precisely what happened in Pennsylvania in the US.In 2012, Democrats took 51 percent of the popular House ballot yet they won only 5 out of 18 House seats -less than one-third.This absurd outcome was blamed on Pennsylvania Republican government who redrew the state's congressional districts and created highly irregular districts (Ingraham, 2015).The validity of this perception was assessed in the fragmented case study districts.(Electoral Commission 1996) in those districts.In order to unseat the NDC in these districts, the NPP in the 2004 election year split Ga District into GEDA and GWDA, and Sissala District into SEDA and SWDA.This paid off because the NPP captured GWDA (winning 45.95% of the presidential votes to beat the NDC at 42.95%, and also winning 45.95% of the parliamentary votes (Electoral Commission, 2004)) and SEDA (where it polled 40.13% of the votes for the presidential election, but lost to the PNC who received 74.20% of the votes) in the parliamentary election, with the NDC placing second with 25.78%.The NPP won a negligible proportion of the votes (Electoral Commission, 2004).-).The NDC recaptured GWDA in both presidential and parliamentary elections in 2008 and 2012 (polling presidential vote shares of 51.9% and 60.47% respectively, and parliamentary vote shares of 42.95% and 51.82% respectively (Electoral Commission, 2008, 2012)).
The NPP held on to SEDA in the 2008 presidential elections (winning 44.70% of the votes) but lost it to the NDC in the parliamentary elections (the NDC winning a vote share of 38.91%) (Electoral Commission, 2008).The national picture is similar to the outcomes for the case study districts.To illustrate, the NPP government established 28 and 32 new local government areas in 2004 and 2007 (election years) respectively, that is, a total of 60 local authorities in its two terms in office (a duration of eight years).
Likewise, the sitting NDC government created 45 districts in the 2012 election year (see Table 1).The creation of new local governments gave rise to new electoral constituencies in both 2004 and 2012 (see Table 2).
In 2004, 30 new constituencies were created, which benefited three political parties, namely the NPP, the NDC and the PNC, which won 16, 13 and one seat(s) respectively in the general election that year (see Table 3).

Determination of new local governments by president
CSOs registered their displeasure over the creation of the new local authorities, claiming they were unconstitutional and describing their creation as gerrymandering.Some held seminars and workshops or issued press releases decrying the exercise, while others took legal action alleging that the creation of the new local governments flouted the provisions of the Local Government Act 1993 (interviews).
I was a member of a pressure group called Concerned Citizens of Ghana (CCG) which took the incumbent NDC government to court over its creation of 45 additional districts in 2012.' (GEDA interviewee)

Financial consequences of fragmentation
The financial repercussions of creating new local governments and constituencies exacerbated the controversy.In 2012 an establishment grant of GH₡ 1m was transferred to each of the newly created 45 districts in addition to GH₡ 4m for use in developmental projects (Ghana News Agency 2012c).
Nationwide, under the Fourth Republic, ex-gratia payments made to MPs between 2005 and 2008 alone amounted to GH₡ 14,871,954 with each of the 230 MPs taking at least GH₡ 64,661.
Other costs included salaries, accommodation for the new MPs, car maintenance and fuel allowances, furniture, office space and car loans, all of which totalled GH₡ 14,400 per MP per annum.Even though the MPs are personally liable for how they the money, the government provides collateral security against default.For a country that depends on donor funding to finance about 40% of its budget, additional MPs as a result of the creation of new constituencies can add additional strain to already stretch public finances (Public Agenda 2012).
The creation of more constituencies also means the admission of more representatives to parliament.
This means the then parent Ga District Assembly and parent Sissala District Assembly, which used to have a member of parliament (MP) each, have doubled their representation following fragmentation, since each has been split into two local governments with one MP each.

Lack of consensus on criteria for and timing of constituency creation
Under article 47(4) of the constitution the factors to be taken into account in the creation of constituencies are population quota (defined as the number obtained by dividing the number of inhabitants in Ghana by the number of constituencies into which Ghana is divided), communication, geographical features, density of population, areas and boundaries of the regions and areas and boundaries of other administrative or traditional areas.These criteria were further augmented by the Electoral Commission to include population growth and movement, urban expansion, and demand for the creation of additional electoral districts (Electoral Commission 2005).The controversy under discussion here arises because an increase in the number of local governments leads to a corresponding increase in the number of constituencies, which critics say amounts to gerrymandering.
Around two thirds of the respondents in the four case study districts queried that the criterion, "demand from communities for district creation", which the Electoral Commission smuggled in as an additional eligibility requirement for municipal fragmentation, is not only inappropriate but also unfortunate as that criterion is not a provision under Article 47 (4) of the 1992 Constitution.

Disagreement over the location of district capitals
Choosing local governments' capitals has been contentious ever since the launch of the fragmentation policy in 1988, for two reasons.First, unlike for the creation of local governments, there is no legal basis or provision for choosing jurisdictional capitals.Nevertheless, the MLGRD has over the years applied the criteria of population, centrality, accessibility, social infrastructure, geographical location, and resource base.However, it is almost impossible to find reliable statistics on all these criteria, which has created room for manipulation and arbitrariness on the part of central government.Second, the choices made have revived traditional rivalries between towns which see the selection of a capital as a win-or-lose affair.
This study investigated whether the issues relating to the location of the new district capital outlined above had arisen in the case study districts.In GEDA controversy arose between the people of Madina and Abokobi over the location of the district capital.Madina is an urban area and is the twelfth most populous settlement in Ghana, with a population of 137,162.Despite this Abokobi, a loose rural agglomeration of less than 1,850 people was chosen by the then NDC government as the district capital.According to many interviewees this infuriated the people of Madina, until then a stronghold of the NDC, and they threatened to withdraw their support from the party.According to one Madina informant, the government 'took the threat seriously and that was why it carved out Madina from GEDA to form a municipality of its own in 2012'.However, no comparable controversy was reported during the fragmentation of Sissala District Assembly.
Similar controversies erupted in other parts of the country.For example, in 1989 the creation of Tano District revived traditional rivalries between the towns of Bechem and Duayaw Nkwanta.Although the capital was sited at Bechem, there was an odd compromise that 'some of the administrative machinery be centred at Duayaw-Nkwanta' (Ayee 1994: 169).Likewise rivalries were also spurred between Kade and Akwatia (Kwaebibirem District), Akosombo and Atimpoku (Asuogyaman District), and Tolon and Kumbungu (Tolon-Kumbungu District) (Ayee 2013).

Gainers from municipal fragmentation
Some commentators have argued that those who gain most from local government fragmentation in Ghana are politicians, traditional leaders and youth groups.This was investigated in our four case studies.

Parliamentary seats won
The creation of new local governments led to a corresponding establishment of new electoral constituencies in 2004 and 2012 (Ayee 2013).The greatest gainers of the new constituencies in the case study districts were the NDC winning four of the eight seats, across the two general elections of 2004, 2008 -with PNC and NPP political parties, winning just 3 and 1 respectively.The CPP was unable to take any of the seats.However, in the 2012 elections the NDC took all four seats in the four fragmented districts in our study, meaning the NPP, PNC and CPP did not win any seats that year.
This finding supports the claim that the creation of new local governments by the NDC president amounted to gerrymandering.

Employment for party functionaries
The creation of new districts and constituencies has also created employment for some functionaries.The district chief executives (DCEs) in GEDA, GWDA, SEDA and SWDA were all party activists who were appointed by the president under the provisions of the Local Government Act 1993.For instance, the current DCE of Ga East District was a known NDC activist who was redeployed from his bureaucratic career position as district director of the National Commission on Civic Education (GEDA councillor interview #3, 2014).Other party functionaries were chosen to contest national elections after winning their parties' primaries.For example, the former Minister of Roads and Highways, who hails from Sissala West District, is a party activist who, having been selected in the NDC primaries, eventually contested and won that constituency seat (GWDA councillor interview # 2, 2014).The inference to be drawn from this is that the creation of constituencies may be seen as a patronage-building exercise for the president.

Youth employment
Some young people also secured jobs in the administrative offices of the new districts, working in semi-skilled and unskilled positions as cleaners, messengers and drivers as well as conservancy labourers and security guards.According to some interviewees, those who could not secure jobs criticised the recruitment process as unfair and based on party affiliations.It is certainly true that appointments to unskilled positions are decentralised, meaning that the part-affiliated DCE has control over who is selected to fill a vacant position.

Discussion
This study has analysed the subject of local government fragmentation and the controversies, tensions and conflicts that have attended its implementation in Ghana.The questionable design of the fragmentation policy, its unabated application with little public consultation, and its inequitable distribution of costs and benefits are the primary causes of these controversies, tensions and conflicts.
A number of design issues merit discussion.Firstly, it is not clear who has power to create local governments.While the 1992 Constitution vests that power in the legislature, the Local Government Act 1993 gives it to the executive.Moreover, whereas the constitution calls for a law to empower the legislature to create local authorities, the Act settles for an instrument.This inconsistency has implications for the 'separation of powers' doctrine -arguably a key building-block of democratic governance.Conflict would have arisen between the legislature and the executive if the former had asserted its right to create local governments, and this would have been injurious to Ghana's young democracy.
Fortunately, majority rule in Ghana has forestalled such an event, but at the same time has given unlimited power to the president to create local authorities.The withdrawal of legal instruments from parliament by successive governments suggests that council creation has been done hurriedly and without adequate consultation with key stakeholders, and impartially the affected public.This goes against one of the key tenets of the decentralisation programme, which calls for mass participation in all decision-making and development at local level.By contrast, successive governments in Ghana seem to prefer an elite approach to decision-making, which claims that involving the public can be disruptive, costly, time-consuming, and inefficient, because ordinary people are not capable of participating meaningfully (Steelman 2001).This is paradoxical: on the one hand, by launching the decentralisation programme the state is encouraging citizens to get involved, but on the other, by not adequately consulting them it is leaving them dependent on government for access, information and action.This makes a mockery of participation as a key tenet of democracy and decentralisation (Woodford and Preston 2011).
Another design problem is the appointment rather than election of mayors and one-third of local government councillors.Critics see these posts as 'jobs for the boys' exemplifying a client-patron relationship designed to keep the government in power, rather than a way to get the best personnel to take local decisions and solve local problems, as envisaged by the decentralisation programme.The mayor, together with the appointed councillors, is instead supposed to garner local support for the government party and implement the policies of the latter at local level.In return, the patron is supposed to assure appointees of security of tenure and the delivery of privileges.Because they are appointed, mayors and appointed councillors are unlikely to enjoy adequate political/popular support to provide the requisite level of legitimacy for changes to local policy design and implementation.The tensions, controversies and conflicts between different stakeholders (towns, youth groups, traditional authorities, political parties) generated by the fragmentation policy remind us that 'any hard and fast distinction between politics and administration can only be false' (Smith 1985: 201).It may be tempting to view decentralisation as a technical or an administrative tool; and even to assess it, and examine changes, as if these are issues which can be resolved by technical arguments about optimum areas, administrative efficiency and managerial performance.However, it should be noted that outcomes expressed in forms such as regional and local government decisions are the result of political conflicts (Ayee 2008a).The interests of individual towns are in direct conflict with each other; and so too are those of different political parties (which may be advantaged or disadvantaged by the fragmentation exercise) and youth groups or individuals (who may get or not get employment in the created district).The outcomes of any conflict are determined by the strategies, resources and power positions of each stakeholder.
One important insight that can be derived from these observations is that the type of policy being carried out is a function of the kind of political activity witnessed at the implementation stage Agyeapong and Adjei, 2008;Touti et al, 2007).Fragmentation programmes like that of Ghanafeaturing inequitable and divisible gains such as the distribution of power and resources, demarcation of local authority boundaries and location of jurisdictional capitals -actually aggravate conflict, tensions and competition among those affected and are therefore quite a challenge to implement as planned.Grindle noted that 'policies and programmes will be more difficult to manage and predict and even more subject to alteration in the Third World than elsewhere' (Grindle 1980: 19).This is because in developing countries there is a skewed level of individual and collective demand-making in favour of the powerful, resourceful and voice ferous as well as well-organized interests to the neglect of the marginalised segments of the population.This marginalisation creates and spurs conflicts as well as high levels of individual and collective demand-making, representation of interests, and emergence and resolution of conflicts.The implication, therefore, is that the launching of fragmentation reform is unavoidably complex due to the many different factors and interests involved and in particular because fragmentation is a highly political process.
During the 1980s and early 1990s, it was the public finance objective of reducing (centralised) public sector spending which drove many fiscal decentralisation (or 'anti-centralisation') reforms in developing economies, supported by the World Bank and other international financial institutions.
These early reforms did focus on the establishment of effective sub-national institutions.However since the mid-90s, while most public financial management reform programmes and public expenditure/fiscal accountability reviews continue to consider sub-national finances, more often than not sub-national expenditures are relegated a secondary role in such exercises (Boex 2009: 4).
To the extent that the use of intergovernmental transfer resources and local revenues are considered in public finance reviews, the emphasis tends to be on the relative inefficiency of local public spending rather than on the inefficiency and ineffectiveness of central spending programmes (Boex 2009: 5).
This certainly seems to be the case with the fragmentation exercise in Ghana.The financial implications of creating new local governments and constituencies have been noted to be grave.The colossal sums that need to be paid towards newly created MPs in the form of salaries, and allowances makes fragmentation an expensive decision .So fragmentation, instead of reducing central public spending as envisaged by the decentralisation programme, has rather increased it.
Moreover, the huge sums of money allocated to newly create local governments, most of which are not economically viable at local level, militate against fiscal decentralisation policy.These local governments cannot generate sufficient funds to run their affairs without central transfers.Boex notes that the 'primary role of fiscal decentralisation in the broader decentralisation reform is to ensure that funds flow to the sub-national level' (2009: 5).He points out, however, that this 'requirement sometimes comes at the expense of the level of attention that is paid to the design of the intergovernmental funding flows and the efficiency aspect of public finances' (Boex 2009: 5).This certainly seems to be the case with the fragmentation exercise in Ghana, where the indiscriminate allocation of monies to newly-created local governments, without applying any specified formula for intergovernmental funding flows, or increasing the capacity and ability of local government to raise their own revenue has implications for the efficiency of public finances.The recipient local authorities have spent the money but have little to show for it in terms of outcomes (Ayee 2008a), due mainly to the grants' general nature and also to the perception of the funds as a windfall gain.This evidence is consistent with Boex's observation that, in many developing countries, enthusiasm for democratic reforms has led to the emergence of a decentralised system 'where governance was placed as close as possible to the people' -but which has resulted in 'highly fragmented and inefficiently small jurisdictions incapable of planning for, or delivering, the local public services assigned to them in an efficient manner' (Boex 2009: 5).This negative outcome prompted many developing countries in the 1990s to consolidate sub-national jurisdictions or reassign expenditure responsibilities to higher government levels after finding out that some jurisdictions were too small to effectively perform their duties (Boex 2009).Yet the fragmentation exercise continues unabated in Ghana and policy-makers seem unaware of or unperturbed by the policy reversal in other developing countries.

Conclusion
This article has analysed the outcomes of local government fragmentation in Ghana.The study has demonstrated that the fragmentation exercise, like any public policy, is not merely a technical exercise but also an inherently political activity.The technical criteria for the delineation of local governments are nuanced, but play second fiddle to the pressure of vested interests in final decisions.Undoubtedly, the incessant tensions surrounding the creation of new local governments in Ghana stem from the tendency to over-concentrate on political aspects of fragmentation, to the benefit of incumbent governments.In so doing, the viability and rationality of the policy have often been undermined.Yet, for most local residents the rationale for decentralisation, and within that fragmentation, is that local government will be able not only to effectively deliver existing services (and also expand them), but also maintain their level and quality, and introduce new ones.Regrettably, Ghana's new local governments have not been able to meet these hopes.
There are four implications from this study for more general theories of fragmentation.First, the ambiguity in the legal framework that assigns the power to create local governments both to parliament and the president has the potential to put the legislature and executive on a collision course, and this could be injurious to Ghana's young democracy.The ambiguity has also given the president unfettered power to create districts, which is exercised without much consultation with key stakeholders -and thus, even though decentralisation is often used as a strategy to reduce central government influence over local decision-making, the present use of the president's power is doing the opposite.There should therefore be a determined effort to tighten up the legal framework and make the creation of local authorities the responsibility of the legislature, in line with international best practice.This would prevent the president manipulating the power to his own party's benefit, as well as eliminate the associated perception of gerrymandering.
Second, the appointment rather than election of mayors and one-third of local government councillors has undermined local capacity-building and local democracy.Those who have the acumen to perform efficiently and effectively are side-lined in favour of party functionaries who lack such skills.Yet one of the objectives of fragmentation reform is specifically to strengthen local public administration and local service delivery.Thus the current fragmentation process, instead of promoting functional decentralisation, undermines it.
Third, the introduction of any kind of decentralisation reform is bound to generate conflict because of the many invested interests at stake, and because of decentralisation's inherently political nature.
There will always be winners and losers, as in any political struggle.It needs to be acknowledged, therefore, that redistributive outcomes can be contentious, and that mitigating measures need to be put in place.
Fourthly and finally, one of the main objectives of decentralisation is to reduce central public spending inefficiencies and free up funds for transfer to local governments so they can be prudently address a critical priority need of the people and … ensure even and balanced development throughout the country, especially for the deprived and underserved district areas . . .It has been shown over the years that when you create districts, you bring development into the area . . .attract support from development partners to channel needed resources for the development of the new districts in various aspects(Ayee 2013: 1).The political and economic motivations for the creation of districts are therefore in line with the fragmentation literature.Another important consideration, however, is the viability of the newly created local government areas.The indicator of economic viability (that is, the ability of an area to meet basic infrastructural and other developmental needs from the monetary and other resources generated in the area) has not been rigorously employed in the establishment of local governments(Ayee 2013).Consequently, in over 92% of metropolitan, municipal and district assemblies (MMDAs) there is over-dependence on the District Assemblies' Common Fund (DACF) (a fund which receives 10% of central government revenue for distribution to local governments), to the detriment of locally generated funds(Crawford 2008).The scant attention that has been paid by successive governments to viability issues when creating new local authorities has prompted a former Minister of Local Government and Rural Development, Mr Kwamina Ahoi to criticise the uncontrolled power of the president in this regard, and to call for a review to introduce changes that will give due weight to local authorities' economic viability as outlined in section 46 (5) of the Local Government Act 1993 (Act 462).

Figure 1 :
Figure 1: Local government structure and responsibilities Before 2004 both Ga East District Assembly (GEDA) and Ga West District Assembly (GWDA) were part of a parent local government area called Ga District Assembly, while Sissala East District Assembly (SEDA) and Sissala West District Assembly (SWDA) were subsumed under the then Sissala District Assembly.The National Democratic Congress (NDC) and the New Patriotic Party (NPP) are Ghana's dominant parties, and have alternated power since Ghana's return to constitutional rule in 1993.Other smaller parties are the People's National Party (PNC), Convention People's Party (CPP), and Progressive People's Party (PPP).The four fragmented case study districts had been the stronghold of the NDC, which won both the presidential and parliamentary elections in 1996 and 2000 Mayors and appointed councillors are further circumscribed by party discipline, which makes them sympathetic to central policy preferences at the local level rather than assessing them on their merits in the local context.This undermines the vibrancy of local democracy and by extension the goals of fragmentation and decentralisation, which require unconstrained generation of ideas and ingenuity to solve local problems.Moreover, since one of the motivations for fragmentation reform is specifically to strengthen local public administration and local service delivery (Boex 2009), it can be argued that, by appointing rather than electing mayors and councillors, governments in Ghana have undermined local capacity-building.
used to speed up community development.But Ghana's fragmentation exercise has done the opposite by exacerbating central public spending inefficiencies and disincentivising local public financial accountability.Intergovernmental fiscal transfer arrangements are loose and there is ample opportunity to misuse public funds.This fuels the arguments of critics of fiscal decentralisation, who point to the practical problems and potential costs of decentralised finance: macroeconomic instability, lack of fiscal discipline, inefficient public spending due to weak local administrative capacity, local corruption and capture by local elites.

Table 1 :
Creation of Local Governments from 1988-2012 The MMDAs have 86 specific responsibilities ranging from provision of basic services in education, health, water supply and sanitation, to public safety and revenue collection.In formal terms, MMDAs can be seen as: political and administrative authorities; planning authoritiesAwAwortwi 2010authorities; budgeting authorities; and rating authorities(Ayee 2013;Awortwi 2010).They constitute a hybrid form of decentralised authority, combining elected and appointed officials: 70% of MMDA members are elected by universal adult suffrage and 30% are nominated by the president.The chief executive, the political head of the district, is nominated by the president and subject to approval by two-thirds of the MMDA members present and voting(Ayee   2013;Awortwi 2010).

Table 2 :
Creation of Constituencies in Ghana