In the name of bringing services closer to the people? Explaining the creation of new local government districts in Uganda

Abstract

Many governments in Africa and elsewhere in Asia and Latin America have created new local government (LG) jurisdictions as part of their decentralization policies. However, most decentralization studies have focused on fiscal, political and administrative assignments between levels of government. Much less attention has been given to the number and size of LG jurisdictions. Often, these are considered to be an accident of history, but the reality is not so. This article pursues five propositions concerning the rationale for creating LG jurisdictions and examines their relevance in the Uganda context. The article concludes that creation of LG jurisdictions in Uganda neither conforms to the policy objective of bringing services closer to the people nor to promoting participatory democratic governance. Instead, the practice conforms to central government gerrymandering tactics of forging an electoral alliance with small jurisdictions and to extend neo-patrimonial networks to win votes in order to stay in power.

Points for practitioners

Donors and development practitioners have often expressed their frustrations on decentralization policies and outcomes in sub-Saharan Africa (SSA) because they have analysed the policy based on what governments say their intentions are. Until they understand the political economy of decentralization in Africa they will always be bewildered not only by the mismatch between policy objectives and outcomes but also by many unspoken rationales for implementing decentralization reforms. Decentralization can be a facade behind which quite different practices take place. In many countries in SSA, it is a narrowly political process that ruling governments pursue for their own benefit (or accept to pursue when pushed by donors where it hurts them least) and not for the nicely written policy statements of improving services.

Keywords

administration and democracy administrative organization and structures administrative theory

Introduction

Over the past several decades many countries have adopted decentralization as an overarching policy for state reform (Smith, 1985) and public management (Pollitt, 2005). In western developed countries, decentralization was promoted as a means to streamline government administration under ever-tightening fiscal conditions and public dissatisfaction with bureaucracy and service delivery (Burns et al., 1994). In Latin America, decentralization has been an essential part of the democratization process as autocratic central regimes were replaced by elected governments operating under new more democratic constitutions (Burki et al., 1999). In East Asia, decentralization appears to be motivated by the need to expand service delivery to large populations and by the recognition of the limitations of central administration (White and Smoke, 2005). In sub-Saharan Africa (SSA) the rationale for decentralization varies, but in the late 1990s the spread of multi-party political systems created demand for more local voice in decision-making (Crook and Manor, 1998; Smoke, 2003).

There are many reasons for a country to embark on a decentralization policy, but whatever its rationale may be the process invariably involves redefinition of structures, procedures, territorial subdivision of nation states and their functional responsibilities, as well as the relationship between central government (CG) and sub-national governments (SNGs). Decentralization may be defined as delineation of geopolitical divisions of the state into smaller sub-national jurisdictions and providing these jurisdictions and their populations with a government that is closer to them. Many governments in SSA and elsewhere in Asia, Latin America and Europe have created new sub-national jurisdictions as part of their decentralization policy objectives (Alesina et al., 2004; Baldersheim and Rose, 2010; Mawdsley, 2002; Suberu, 1991). For instance, Burki et al. (1999) report the number of municipios in Brazil increasing from 3000 to nearly 5000 in the 15 years following the return to democracy in the country. In Venezuela, they report the number of municipios increasing from 202 to 330 within a decade. However, most decentralization studies have focused on fiscal, political and administrative assignments between levels of government (vertical dimension). Much less attention has been given to the horizontal dimension; namely, the number and size of local government (LG) jurisdictions. Often, these are considered to be an accident of history.

Uganda is an interesting example with which to study this phenomenon. It is a country that (especially in the 1990s) was seen by the World Bank and major donor agencies as a ‘model case’ or ‘best practice’ of preparing, formulating and implementing a decentralization policy (Steffensen and Trollegaard, 2000). Although there is no official and explicit policy of creating small LG jurisdictions, in recent times there has been a veritable proliferation of new districts, all in the name of ‘creating effective administration and the need to bring services closer to the people’ (Constitution of Uganda, Art. 179(4)).

Decentralization and the creation of LG jurisdictions in Uganda

Uganda's LG system has evolved and transformed over the years from a centralized kingdom and non-kingdom system during colonization, to autonomous local administration during independence in 1962, to recentralization in 1967, and back to broader decentralization in 1986 (Awortwi, 2011; GoU, 2006). Against the background of diverse people, cultures and institutions, the central issue of these changes has been tensions between greater autonomy of LGs and tight CG. State-centric literature on decentralization in Uganda tends to describe President Museveni as having transformed the local governance landscape in the country (Nsibambi, 1998; Tukahebwa, 1998). In a 2002 stocktaking survey on decentralization in Africa, Uganda ranked second among 35 countries in Africa, with only South Africa performing better (Ndegwa, 2002). Nevertheless, other authors suggest the opposite outcomes (for example, Francis and James, 2003; Golooba-Mutebi, 2008). Wherever one stands in this polarized debate, President Museveni is often praised for decentralization reform in the country. However, one area that he has been criticized for is the frequent creation of new LG districts without regard to cost or clear policy objectives.

When President Museveni took power in 1986, there were 33 districts and a population of about 15.2 million people, making an average of about 460,000 people per district. The Mamdani Commission, which was established to advise the government on the type of decentralization and LGs that Uganda should have, recommended that only four new LG territories be created in addition to the existing ones, although it had received 11 requests (GoU, 1987). The Commission bore in mind the cost of administering districts, both in terms of creating an administrative infrastructure and payment of personnel. However, Museveni decided not to create any new districts, only relenting in 1990 to create Kalangala, which comprised the Ssese islands in Lake Victoria. Kalangala's isolation from the mainland justified it becoming an autonomous district.

The cautious approach to district creation ended in 1996, and from 1997 President Museveni started to create more new districts, to the extent that by mid-2010, 67 new districts had been created, bringing the total number to 112, more than three times the number he inherited in 1986 (Bekunda and Namutebi, 2010; Kato, 2009). Figure 1 shows the growth in the number of districts in Uganda from 1986 to 2010.

Figure 1.

District creation in Uganda (1986–2010)

The analysis shows that from 2005 to 2010 Uganda created no fewer than eight new districts per year, and from mid-2008 to early 2010, an average of more than one district every month, which is highly unusual in any country if the increase is just to accommodate population increase or maintain a particular service standard in planning. If the increase in the number of LG jurisdictions is not an object of policy with particular objectives and with a particular target in mind, it may be an endogenously driven process ‘from below’, or alternatively other political objectives may drive the process ‘from above’. This article seeks to answer two basic questions: (a) What drives the increase in the number of LG jurisdictions in Uganda? (b) How have strategies behind the ‘rationale’ for creating districts being implemented?

Rationale for creating smaller LG jurisdictions: a review of the literature

The literature provides a number of reasons for creating on average smaller LG jurisdictions within a state system. In the sections below we highlight the most relevant ones.

Promotion of citizen participation in democratic governance

In countries with a history of autocratic regimes, the creation of LG jurisdictions can be an important first step in opening a space for citizen participation. The basic uniting principle is subsidiarity: what can be done best at a particular level should be done at that level and only the residual should be passed to higher levels. Political scientists will argue that the smaller the size of an LG, the closer the distance between the LG and the local community, the greater the disposition of community members to participate in the political affairs of their jurisdiction. The larger the size of the LG, the lower will be the degree of participation (Borck, 2002). Turning this around, one could argue that if people wanted to participate more, they would want to do so in smaller jurisdictions. CGs could promote this by facilitating the creation of smaller jurisdictions. Creation of LGs is therefore based on the application of democracy to a geographic scale. In Latin America, decentralization policies have created more and small-sized local jurisdictions with the explicit aim of increasing political participation of hitherto marginalized communities (Burki et al., 1999). Contrary to the pursuit of smaller jurisdictions, the literature also provides situations where a number of small jurisdictions are amalgamated in order to enjoy economies of scale for the purpose of improving efficiency of service delivery and reducing unit cost (John, 2010). For example, in 2005, the number of districts in Rwanda was reduced from 106 to 30 and provinces from 11 to 4 (MINALOC, 2007). Efficiency is also achieved because of a reduction in the number of decision-making centres.

Ethnic-linguistic conflict management

In multi-ethnic societies where people easily identify themselves with their ethno-linguistic lines, the arguments for the creation of SNGs are often based on the management of ethno-linguistic tension and conflict (Treisman, 2007). For instance where ethnic tension has escalated to outright conflict, territorial decentralization often becomes the institution of first resort for domestic or international actors in their attempt to bring peace (Lake and Rothchild, 2005). By granting ethnic groups autonomous jurisdictions, peacemakers aim to mitigate fears of political exploitation and inter-group violence and at least in part to satisfy local demands for cultural and religious autonomy. There are several country cases in the literature (Kosovo, India, the Philippines, Belgium, Spain, Indonesia and Nigeria) where territorial decentralization along ethnic lines emerged as a key solution to contain contemporary civil and ethnic conflicts (Brancati, 2006; Mawdsley, 2002).

Patronage-induced territorial decentralization

In recent times ‘Africanist’ scholars (Bayart, 1993; Chabal and Daloz 1999; Hyden, 2008) have revisited patrimonial theory to explain the nature of the state–society relationship in Africa. They argue that African states operate less according to procedural rules and more through personal relationships of clientelism. One of the strategies that CG politicians use for obtaining electoral support is the creation of LG jurisdictions to favour interest groups and further extend CG resources to the groups even if it means flouting formal rules and procedures that govern such development. In his examination of district creation in Uganda (from 39 to 80) in less than two decades, Green (2010) concluded that patronage is the main reason for the creation of districts to enable President Museveni to stay in power after structural adjustment had whittled down economic opportunities available for patronage. Patronage creation of local jurisdictions appears to go hand in hand with the small size of local jurisdictions and the growth in the number of jurisdictions over time. According to Khemani, under patrimonial decentralization newly created LGs would not raise their own revenues because their expenditures would be financed by national grants. Studies in a range of countries as diverse as Albania, Brazil, India, Mexico, Peru, Kenya, Ghana and the Philippines provide evidence of significant targeting of CG grants on the basis of electoral characteristics (Khemani, 2010).

Gerrymandered decentralization

Gerrymandering is the practice of manipulating electoral district boundaries with political parties' interests in mind with the aim of neutralizing the opponents' potential vote to capture power (Gumbel, 2008). The strategy involves mixing the spatial distribution of party support and/or forming party strongholds. In this approach, the ruling government and national politicians divide the electorate into smaller groups so that the strength of swing voters is dispersed across jurisdictions and a minimum winning coalition of jurisdictions is forged (Khemani, 2008: 71). The incumbent party, in a position to control redistricting, weakens the opposition party by ‘packing’ as many opposition voters as possible into a minimum number of conceded districts, and/or ‘cracking’ opposition voters among numerous safe districts where they are in the minority. For more readings on gerrymandering redistricting, see Grace's (2007) analysis on Malaysia; and Burton's (2001) on Singapore.

Efficient administration and effective provision of services

The central policy considerations for the creation of LG units are efficient administration of government and effective delivery of services to citizens/constituents. The assumption behind the efficiency argument is that the smaller the district, the more easily the services can be provided to the majority of the population since the distance between the service centre and where the people live is reduced. The effectiveness in service provision argument is that the LGs may be able to access better local information and tailor services to the preferences of local citizens. Citizens receiving quality services attuned to their preferences are then more likely to pay for these services (Breton, 1996). In the era of the New Public Management, having multiple local jurisdictions would not only promote competition among LGs but also ensure that users/clients played important roles in exacting accountability from service providers.

The above review of the literature suggests that there are multiple rationales for increasing the number of LG jurisdictions. In the next section we examine five propositions which are derived from the above discussions and which are relevant for empirical analysis in Uganda.

In search of explanations for increasing number of districts in Uganda

The authors undertook a comprehensive data and evidence search using published and unpublished government and research reports, and conducted qualitative field interviews with CG and LG politicians and bureaucrats as well as with a number of informants from academia for their perspectives on the subject.

Proposition one: Broadening citizen political participation in democratic governance

If this hypothesis is to be confirmed there will have to be at least two kinds of evidence: (i) that particular disenfranchised groups are incentivized to participate in democratic politics if districts are created to represent them. This is the essence of affirmative action in much of the political participatory process. (ii) People will participate more in political processes in smaller jurisdictions. Below we analyse these two issues.

Do demands by interest groups for representation drive district creation?

Uganda's record with regard to political inclusion and representation of interest groups is high. The national Constitution and LG Act make provision for women, youth and persons with disabilities to be represented at both the national and district levels. Also at the national level there is representation from the army, workers and other identifiable groups. Along with the creation of new districts, the number of constituencies and members of national parliament has also increased. When a new district is created, a female district MP automatically emerges as part of the constitutional requirement. As a result, the number of women parliamentarians representing districts has grown in just the same way as the number of districts. Women MPs as a percentage of all MPs have gone up from 14 percent in 1996 to 25 percent in 2008 and 33.9 percent in 2011. However, in speaking to a cross-section of women parliamentarians and interest group representatives, we found no evidence to suggest that any of them demanded the creation of new districts in order to increase their democratic participation.

Increasing citizens' participation in small LG jurisdictions drives the creation of LG districts

If this hypothesis is to be confirmed there would be evidence that districts with a small population or area will have higher citizen political participation than districts with larger land and population sizes. It is also expected that newly created districts will experience higher participation than old districts. What is the evidence in Uganda? While presidential and parliamentary elections show high voter turnout (presidential: 60.7 percent in 1996; 55.7 percent in 2001; 67.8 percent in 2006; and 59.3 percent in 2011); (parliamentary, 72.9 percent in 1996; 69.7 percent in 2001; 69.2 percent in 2006; and 72 percent in 2011), voter participation at district level elections is declining. Voter participation went down from 59.7 percent in 2001 to 46.3 percent in 2006 (Golooba-Mutebi, 2008: 152) but rose marginally to 46.7% in the 2011 elections. From 2000 to 2010, the number of LG districts increased from 56 to 112 and the average district size declined from 4229 km² to 2114 km² while the average population size per district was reduced from 427,786 to 289,285. Using district-level voter participation in 2011 presidential elections, the results show that citizens' participation did not increase significantly with the reduction in the size of population per district (see Table 1).

Table 1.

Citizen participation in presidential election of 2011

Citizen political participation	Voter turnout 2011 election (%)	National average
Small districts with population <100,000 (n = 18)	61.04 (a)	*59.3% 61.63%
Medium sized districts with population 100,000–200,000 (n = 48)	62.68 (a)
Large districts with population >200,000 (n = 46)	60.79 (a)
Small size districts with land area < 1000 km² (n = 35)	65.35 (b)	2114.8km
Medium size districts with land area 1000–3999 km² (n = 43)	60.77 (b)
Large size districts with land area >4000 km² (n = 34)	59.32 (b)
Newly created districts (2005) n = 22	62.95 (c)	*59.3% 61.63%
Newly created districts (2010) n = 33	61.71 (c)
Old districts created before 2005 n = 57	60.32 (c)

Source: Election results from National Electoral Commission (NEC).

Notes: (a) and (c): ANOVA confirms null hypothesis. (b) ANOVA tests confirm null hypothesis except for small districts vis-a-vis medium and large sized districts.

The NEC indicates voter turnout at 59.3 percent but using their own district data we arrive at an overall average of 61.63 percent.

However, political participation rises with the reduction in district size as voter turnout increased by 4 percentage points when the district was small (less than 1000 km²). Also, the participation rate in newly created districts increased by 1.3 percentage points above the national average and 2.6 percentage points above the rate of the old districts, but is reduced by 2.3 percentage points below the national average if the size of the district is above 4000 km². Statistically, the difference in voter turnout between categories of districts is insignificant in all instances except for small size districts that differ from medium and large ones. So decreasing the land size of districts to increase voter participation can be used to justify the creation of more districts, but this argument is hardly made explicit by any CG politician or the academic analysis of Uganda's decentralization to explain the creation of districts in the country.

Proposition two: Creation of less heterogeneous districts in order to reduce ethnic conflict

If this is so, two factors will be evident: (i) minority ethnic groups in a multi-ethnic district will protest against perceived or real marginalization, leading to frequent tension in the district, and hence (ii) the CG will attempt to resolve ethnic conflicts by splitting a multi-ethnic district into homogenous small districts. What is the evidence in Uganda?

Ethnic¹ tension in heterogeneous districts leading to frequent conflicts and demand by minorities for autonomy

Ugandan ethnic minorities include Alur, Kakwa, Karamojong cluster, Lugbara, Ma'di, Nubian, Bagungu, Bakenyi, Bavuma, Ik (Teuso), Soo, Batwa and Bakonzo. According to Wairama (2001), these groups are marginalized from Uganda's economic, political and social-cultural life. But is there any evidence to suggest that there are ethnic tensions in heterogeneous districts leading to minorities agitating for autonomy so as to promote peace? During our field visits to Uganda, LG officials gave anecdotal evidence concerning ethnic tensions in some districts (Yumbe, Hoima and Luwero). Peaceful settlement of these ethnic conflicts resulted in the creation of Kibaale from Hoima district; Nakasongola district from Luwero district; and Yumbe district from Arua district. In contrast, the creation of Tororo, Kabarole and Pallisa districts led to ethnic conflicts. The literature on Uganda provides high profile publicity of some ethnic groups' using bizarre forms of protest to demand autonomous districts (Buwembo, 2005; Manyak and Katono, 2010; Mwenda and Mugisa, 1999). Where such demands coincide with President Museveni's interests, he acquiesces to ethnic demand. For example, at the launch of Nakaseke district in August 2005, President Museveni said, ‘Although a new district takes a lot of money, this is the democracy we fought for. People must ask for what they want and get it’ (Ocwich, 2005). Fifteen out of 32 districts that were created since 2006 could be classified as heterogeneous districts (that is, districts inhabited by more than one ethnic group large enough to form a new district) but there is little evidence to suggest that the carved out districts had ethnic tension or conflict before they were divided. So we concur with Green's conclusion that districts are not created for the purpose of reducing ethnic tension and conflict (Green, 2008, 2010).

Is district creation ethnic balkanization?

If this hypothesis is to be sustained, evidence must show an increasing trend of splitting heterogenous districts in such a way that newly created districts or the rump districts now depict ethnic homogeneity. According to our literature review above, such a strategy is to reduce ethnic-linguistic conflict. Green's data from 2006 show that of the 46 districts created before 2007, only 13 or 28.3 percent have led to a minority ethnic group becoming a majority in either the new or the rump district (Green, 2008: 10). Our current analysis of the overall pattern of district creation shows that the tendency to create new districts was almost the same in the two subgroups. Taking 1986 as the base year, out of the 15 ethnically homogenous districts, 39 new homogenous districts were created and the 18 ethnically diverse districts contributed 40 new districts. Another 16 new homogenous districts were formed from the 18 ethnically diverse parent districts, and two of the ethnically diverse parent districts became ethnically homogenous as a result of separation of other ethnic groups, bringing the total to 72 homogenous districts (see Table 2).

Table 2.

Ethnic diversity of districts (1990–2010)

Districts based on ethnic groups	No. of districts as at 1986	Parent districts	Newly created districts by 2006	All districts by 2010
Homogenous	15	(15 + 2)	(39 + 16)	72
Diverse	18	(18 − 2)	24	40
Total	33	33	79	112

Source: Estimates derived from comparing district maps with the geographical distribution of ethnic groups.

However, of the 40 new districts that were created from the 18 ethnically diverse parent districts, 24 (or 60 percent) remained ethnically diverse and 16 of the 18 parent districts of this subgroup also remained ethnically diverse. Thus, on balance there were proportionally more ethnically homogeneous districts in 2010 (64.2 percent) than in 1986 (45.4 percent) when new district creation began. Although it would seem hard to sustain the assertion that proliferation of districts was aimed to balkanize or to eliminate local ethnic diversity, the evidence does suggest that the creation of small districts has (un)intentionally resulted in Uganda districts becoming less heterogeneous than they were. To suggest ethnic balkanization, ethnic tension would have to support this evidence, but as the analysis shows, there was no major or widespread ethnic conflict and demand by small ethnic groups for autonomous districts. Unlike Green (2010) who rejects completely the ethnicity hypothesis, our sub-hypothesis on ethnic balkanization does provide some evidence to suggest a cautious approach to the ethnic factor. The fact that Uganda has not experienced ethnic conflicts on the same scale as those of its neighbours (Sudan, Rwanda, Democratic Republic of Congo and even Kenya) is worthy of greater attention to how the country has managed potential ethnic conflict.

Proposition three: Proliferation of districts is an expression of neo-patrimonialism

The key conclusion by Green (2008, 2010) was that the increase in the number of districts is a patronage strategy to enable President Museveni and his party to win votes. While his analysis was based on anecdotal evidence of job creation and political support, our analysis takes that further and provides hard data to support that conclusion and analyses four additional factors.

Manipulation of formal rules regarding creation of LG jurisdictions

Article 179 (section 1-4) of the 1995 Constitution empowers parliament to alter the boundaries of districts. The legal process is clearly laid out in the LG Act in the form of bottom-up approach. However, the manner in which districts have been created shows that the legal procedures are not always followed. Contrary to the bottom-up approach, many of the newly created districts have been top-down arrangements. That was how local residents described the creation of Dokolo district. Sometimes the President, in an effort to appease local people, pronounces an area a district and instructs the Minister for LG to ensure that the formal procedure is followed. That was how Koboko district came into being. It is widely referred to as the ‘condolence district’ because Museveni granted district status to the area when he went there to mourn the death of the then Attorney-General and Minister of Justice and Constitutional Affairs and MP for the area, Francis Ayume, who had died in a vehicle accident (Sooma, 2008).

Creation of LG units to help the rise of new local elites to support the regime

After revoking the ban on presidential term limits, one of Museveni's strategies has been to create conditions to sustain the goodwill of his old allies that he started the resistance movement with and to solicit new political support from new elites that were emerging from the LG system that he created. So, behind the facade of deepening decentralization is the unproclaimed use of district creation to appease new groups as well as perpetuate a network of political mobilization (Larok, 2008). This led to the ending of the cautious approach to creation of districts in 1996.

The creation of districts resulted in the generation of government jobs at the local level to be shared among the party faithful and to entice new supporters. When a new district is created, it is allowed to recruit 224 technical and clerical staff. By 2010 about 17,500 bureaucrats were employed at the LG level, an increase of 150 percent from 1997 when there were only 7020 bureaucrats in 45 districts. From 2006 to 2010, the number of bureaucrats increased by 35 percent at the LG level. These figures show only employees at the district offices and exclude teachers and health workers (field data, Ministry of LG, 2011). In addition, new districts breathe economic life into poor areas by bringing an immediate influx of financial resources. The CG provides each new district with between US$280,000 and US$560,000 to construct district offices and purchase vehicles, office equipment, and other amenities (Manyak and Katono, 2010: 6) although this clearly can be seen as an additional cost to the state.

The NRM has also found a way of minimizing competition from disgruntled members. Given that NRM primaries to select candidates for parliamentary seats are highly competitive, many NRM candidates who lose the primaries usually contest as independent candidates in the general election, thereby splitting NRM votes. To avoid this happening frequently, candidates are promised the position of LC5 chair (district mayor). Eighty-six (77 percent) of the 122 district chairpersons elected in 2011 are NRM candidates, while an additional 14 are independent. The creation of more districts therefore enables the NRM to distribute political positions among a number of its members who are unable to get positions at CG level. Besides the colossal amount of money that is transferred to districts, other political largesse that is bestowed on a new district includes 11 government scholarships for university education. In addition, UGShs10m (US$435,000) is given to each MP as an Annual Constituency Fund to be used for a project of his/her choice. These strategies ensure that some of the electorate who benefit from the largesse switch allegiance to the NRM.

Creation of unviable small local jurisdictions that depend on CG for grants in return for votes

The LG Act of 1997 provides revenue sources and discretionary powers to LGs to mobilize local revenue. Provision was also made for the CG to provide grants to LG. But given that many of the districts have limited sources of revenue, continuous division of their jurisdictions into smaller units makes them unviable without CG grants. In July 2005, the ‘Graduated tax’ that traditionally had contributed over 75 percent of LG revenue was abolished by the CG. As a result, LGs are dependent on central grants for 95 percent of their expenditure needs. In addition, about 90 percent of the grants are earmarked for specific services identified by the CG as a national priority. Meaning LGs have only 10 percent discretion in the use of their grants (Awortwi, 2011). The increasing proportion of conditional grants from 79 percent in 2000 to 88 percent in 2004 coincides in timing with pressure to return the country to multi-party democracy. Critics of the regime argue that the proportion of the conditional grant was increased to enable the President to have a stranglehold of LG finances so he can claim the credit for any infrastructure and services that the LGs provide.

Proposition four: District creation is a gerrymandering strategy by Museveni in order to stay in power

If this is so, three factors could be at play: (a) the creation of LG jurisdictions will provide electoral votes to CG politicians; (b) demarcation of districts will take place immediately before an election; and (c) the creation of districts will not be even across regions of the country

LG jurisdictions matter in serving the electoral interests of CG politicians

As we argued above, gerrymandering is a strategy to win electoral votes by creating new electoral boundaries where ruling governments are expected to win substantial votes from the citizens. As shown in Table 3, whenever Museveni creates a district he wins substantial votes over and above the national average.

Table 3.

Presidential election results for new districts

Election year	Type of district	Museveni (%)	Other candidates
1990–96	New district n = 6	89.2	10.8
	Uganda average	74.3	25.7
2001	New districts n = 16	72.5	27.5
	Uganda average	69.4	30.6
2006	New districts n = 13	73.6	26.4
	Uganda average	59.3	40.7
2011	New districts n = 32	75.0	25.0
	Uganda average	68.38	31.62

Sources: 1990–2006 (Green, 2010: 98); 2011 results from Uganda Electoral Commission.

So there is no doubt that the creation of districts matters in serving the electoral interests of President Museveni and his NRM party. Both Green's (2010) and our expanded analysis confirms that Museveni benefits politically in terms of votes from creating new districts, but to show that there is a deliberate gerrymandering strategy, we expanded our analysis to additional factors that the literature suggests must be present:

Redistricting is needed to ward off the threat against CG political interests

In 1996 Uganda held its first direct presidential election in 16 years, in which three major political forces vied for the presidency. President Museveni was re-elected, receiving 75.7 percent of the vote, while his closet rival Ssemogerere secured only 22.9 percent. Parliamentary elections were held three months after the presidential election, with all the candidates standing as individuals and not on party platforms. The opposition alliance boycotted the parliamentary elections. NRM-affiliated candidates obtained 156 seats as against the opposition's 120 seats – only 36 short of Museveni's NRM (African Election Database, 2010). Thus, in contrast to Museveni's presidential election landslide victory, the NRM was left with only 56 percent of the elected seats in the legislative assembly and Museveni needed to find strategies to neutralize the opposition MPs. One strategy that the NRM was alleged to have considered was the creation of new districts (and thus possible new parliamentary seats) whose population would be thankful to the NRM and therefore likely to vote for pro-NRM candidates. From 1997, that is, a year after the presidential and parliamentary elections, President Museveni began to create more districts. Six new districts were created that year. Three years later, that is, a year before the next general elections were to take place, 11 new districts were created.

Another concern for Museveni was the possible outcome of a public referendum, planned to take place in 2000, on the introduction of a multi-party system in the country. However, the outcome of the referendum was overwhelming support (90 percent) for the NRM system and that eased the pressure on Museveni. No more districts were created until 2005, when the ban on multi-party elections was lifted; immediately 14 new districts were brought into existence in one fell swoop. In 2006, seven new districts were created, 2008 saw another 28, and by 2010 the total number had risen to 112 (including Kampala) ready for the 2011 national elections. Forty-two out of 56 (75 percent) districts were created from 2005 to 2010 because of intense multi-party political activities that must have made Museveni rather insecure. During the election campaign, new districts were offered as a political enticement in return for votes. On 19 August 2008, The Daily Monitor reported that when the President was touring Eastern Uganda, the people of Ngora County in Kumi district (which was known to be an opposition stronghold) demanded district status. While addressing a political rally at Ngora county, Museveni told residents that he had no problem raising Ngora to a district status.

I will discuss the idea with the movement caucus and recommend to them that Ngora be considered for district status because I love the people of Ngora. Whether your MPs cross to NRM or not I will give you the district. (Malinga, 2008; reported in The Daily Monitor, 19 August 2008)

Kumi LC5 chairman Ismail Orot told the President that residents pledged to support his fourth-term bid in return for district status. ‘The people of Ngora support your bid for the fourth term, but they need a district in order to bring services closer to them.’ A district councillor, Omoding, also made the same promise: ‘We are not bad people, as the President might think. If he gives us a district, we will also know how to reciprocate come 2011.’ What happened afterwards? Ngora was given district status in 2010, and in the 2011 elections Museveni won the district votes (48.2 percent) as against 27.4 percent of the votes he got in 2006 when Ngora was a subcounty. His closest rival got 47.3 percent of the district vote. That was the first time that the people of that area had given Museveni more votes than his rivals.

In addition, out of the 29 districts created from 2009 to 2010, that is two years immediately before the election, Museveni won all except two (93.1 percent). In Kitgum district, another stronghold of the opposition party, in 2001 Museveni won only 21.3 percent of the votes cast. In 2009 and 2010, he created two districts (Agago and Lamwo, respectively) from Kitgum and Pader districts and succeeded in winning both districts, receiving 50.2 percent of the votes in Lamwo and 59.57 percent in Agago, while in the parent districts (Kitgum and Pader) he lost, receiving only 38 percent and 46 percent of the votes. In Lira district, Museveni got a paltry 7.9 percent of the votes in 2006. In 2010, he created Alebtong district from Lira and got a majority vote of 48.57 percent in 2011 elections as against 37.67 percent for his closet rival. Similar instances are found in almost all the districts that he created in 2010 for the 2011 election. Two years after the 2011 election, no new district has been created.

Uneven regional distribution of districts

As part of an overall gerrymandering strategy, the pattern of new district creation will not be uniform in the country but skewed towards areas where the government is likely to win. Applied to Uganda, this raises the question why Eastern and Northern regions are getting significant increases in the number of districts (see Table 4).

Table 4.

Creation of districts per region

Regions	% share of pop	No. of districts 1997	No. of districts 2008	No. of districts2010	% increase of districts (1997–08)	% increase of districts (1997–2010)
Central	26	11	16	24	45	118
Western	25	12	19	26	58.	116
Eastern	26	12	24	32	100	166
Northern	23	10	21	30	110	200
Total	100	45	80	112	78	148

Sources: Bekunda and Namutebi (2010), Kato (2009), Larok (2008), Ministry of Local Government (2010).

Eastern and Northern regions are opposition strongholds so more districts are created there in exchange for votes. In 2011, for the first time, President Museveni's party won 42 out of 75 electable seats there.

Proposition five: Smaller districts bring services closer to citizens

The belief that large and monopolistic public organizations are inherently inefficient was a critical force behind the evolution of New Public Management (NPM) that sought to introduce customer-oriented practices into LG services. In Africa, poor physical infrastructure makes transportation a key factor in the determination of accessibility to public services. Elsewhere, the government policy of bringing public services closer to the people or improving efficiency and effectiveness of public service delivery result in the creation of small or large LG jurisdictions (John, 2010). If LGs in Uganda are created with the same intentions, then two issues may be eveident.

Deliberate state policy to bring services closer to people

Article 179 of the 1995 Constitution of Uganda makes provision for the creation of new districts based on ‘the necessity for effective administration and the need to bring services closer to the people’. The Constitution suggests that such an act may take into account the means of communication, geographical features, density of population, economic viability and the wishes of the people concerned. Because of this constitutional provision, ‘bringing services closer to the people’ has become the mantra of LG and CG politicians and bureaucrats to justify the creation of districts. To them more districts mean that more basic services (healthcare, schools, roads, district headquarters, and so on) are brought closer to the grassroots since LGs are responsible for the provision of these services.

The following quotations typify their understanding:

LGs were constituted during colonial rule for the purposes of administrative and service delivery efficiency. When we started reforming our LG system in 1987, we needed more grassroots participation and administrative efficiency, hence below districts we also established lower LGs. As population grows, what we're seeing is that the administrative efficiency is no longer there. Secondly, when you look at the delivery of basic services, there are sections of the population which will never be visited by government officials precisely because they neither have the time nor the energy anymore to do that because it's so wide. Now, by compressing the land area and by reducing the population requirement, we fulfil the basic demand of the constitution in terms of creating districts, which is to deliver basic services and to make it more administratively efficient. (Interview with a senior administrator at the Ministry of LG in December 2009)

Everything is poor in Buvuma. We have poor roads, poor health and education systems. This is why we need to split from Mukono district. We shall get our own facilities when we become a district. (Buvuma MP, 2 December 2009)

If creating districts means extending services closer to the people, let's create 10,000 districts. (Disaster Preparedness Minister and Sheema North MP Tarsis Kabwegere, 22 April 2010)

From the above, it is clear that the creation of districts is a policy that is linked to the understanding of improving service delivery, but where is the evidence of improved services?

Efficiency and effectiveness of services provision

Analysis of concrete evidence that the creation of more districts has improved administrative efficiency can be seen in terms of the Ministry of LG annual assessment of district performance. Districts are assessed based on 18 sets of performance indicators that include quality of the development plan and linkages with budget; local revenue; functionality of education, health, water and sanitation departments, etc. LGs that perform well are rewarded with a 20 percent increase in funding, while those that perform poorly are penalized with a 20 percent reduction. But in all its evaluations, the Ministry of LG has shown that older districts outperform newly created ones by a wide margin in meeting minimum conditions of efficient administration. In 2004, compared to a Uganda average of about 60 percent, only 50 percent (8/16) of districts created since 1997 passed these minimum standards and an even lower score of 45.5 percent (5/11) for those created since 2000 (GoU, 2004). In the 2007 assessment report, 81.5 percent (38/44) of older district governments were able to meet minimum conditions while only 66.7 percent (22/33) of the districts created since 2000 were able to do the same (GoU, 2007). Local and CG officials admit that the creation of new districts usually brings with it many administrative and logistical problems from the beginning but as time goes on they manage to solve them. But recent evidence suggests that this is not so. In fact, the CG is recentralizing many of the key administrative powers that it had earlier decentralized to LG on the basis of LG capacity problems (Awortwi, 2011).

Further evidence that the creation of LG jurisdictions has not improved service provision both in the new and old districts comes from analyses of reports of the provision of some of the key basic services (health, education and water) that have been decentralized to LGs (GoU, 2008). Green (2010) uses the UN's Human Poverty Index (HPI) and Human Development Index (HDI) across Uganda to examine the nexus between district creation and public service delivery. Despite some weaknesses of that approach (indirect route of assessing service delivery plus problems of aggregation of metrics that do not take into consideration wide inequalities between Uganda districts), Green concluded that there was no evidence that new districts have on the whole been created in deprived areas or that their creation has benefited citizens (Green, 2010: 89). Our analysis uses community report cards that provide direct assessment of the quality of services between old and new districts by citizens. In the era of NPM, citizens' perception of the quality of services in terms of efficiency, effectiveness and responsiveness is as important as technical assessment by LGs (Andrews and van de Walle, 2012). In a survey conducted by the Ministry of LG to establish citizens' opinions and satisfaction with LG services and administration in seven districts, the results show that citizens are generally dissatisfied with the quality of services they receive from their LGs irrespective of whether the district is new or old (Ministry of Local Government, 2010). On a scale of 1–5 (1 being bad and 5 being very good), only local administration was scored ‘good’ in all the districts. In terms of service provision the general assessment was ‘poor’ (see Table 5).

Table 5.

Community report card of LG services

LG Services		New districts (established after 2000)			Old districts (established before 2000)
LG Services		Oyam	Isingiro	Nakapiripirit	Nebbi	Masaka	Pallisa	Kasese
Private edu. services	Availability of teachers	Bad	Fair	Bad	Bad	Good	Fair	Fair
	Adequacy of classrooms	Fair	Fair	Fair	Fair	Bad	Fair	Fair
	Quality of teachers	Fair	Good	Fair	Fair	Good	Good	–
PHC	Availability of drugs	Bad	Fair	Bad	Bad	Fair	Bad	Fair
	Availability of staff	–	Fair	Bad	Bad	Fair	Fair	V. bad
	Access to ambulance	V. bad	Bad	Bad	V. bad	Bad	Bad	V. bad
	Opening hours and punctuality of staff	Bad	Bad	Bad	Bad	Fair	Fair	Fair
Water	Access to water (bore hole)	Bad	Bad	Bad	Bad	Fair	Bad	Fair
	Funding for maintenance	V. bad	Bad	Bad	V. bad	V. bad	Bad	Fair
	Operation and maintenance	Bad	Bad	Bad	Bad	Bad	Fair	Good

Source: Ministry of Local Government (2010).

In Dokolo district, which we visited, education and health facilities have not increased since its creation. Over 42 percent of the population of Dokolo still travel over 5 km to reach health services in Lira district (from which it was split off) because no new health centre was built for Dokolo (Dokolo District Local Government, 2008/9). In Nakasongola district, the health facilities that existed when it was a county in Luwero have not increased. Only 6.25 percent of the population have access to health facilities within a radius of 5 km (Oyuku-Ocen, 2008). Surely if LGs are established to bring services closer to the citizens, then one would expect better results than this. So in general the proposition that districts are created to improve efficient administration and bring services closer to the people is not supported by the available evidence.

Conclusions

In this article we have reviewed the literature on the rationale for creating more LG jurisdictions in a country. Using structural adjustment and democratization reforms as entry points with the concept of patronage as an overarching framework, Green explored four propositions: (a) the improvement of service delivery; (b) ethno-linguistic conflict; (c) the inability of the CG to resist local demand for new districts and (d) patronage, job creation and electoral politics, and he concluded that the recent creation of new districts in Uganda can primarily be explained by the need for President Museveni's government to create political patronage. Green argued that decentralization gave Museveni new avenues for patronage, as structural adjustment closed down other avenues, especially those via state enterprises and government interventions in the market. While admitting that other explanations may have some merit, the evidence Green used attributed district creation to patronage. Our study differs from Green's analysis in a number of respects; while not discounting his conclusion, it broadens the explanation using decentralization theory as the entry point, it extends the analysis to the year 2011 to cover 112 districts, instead of Green's focus up to 2006 with only 56 districts, and it uses new sources of evidence. Furthermore, we expand the analysis to include two new major propositions and a number of sub-propositions in addition to three advanced by Green.

Our results reject two of the five propositions: (i) the creation of districts brings services closer to the people as the GoU wants its citizens to believe. This outcome resonates with Green's conclusion; and (ii) the creation of districts aims at increasing political participation of citizens. Political participation as measured in voter turnout is lower at district level and is declining over time. Moreover, there are no significant differences between smaller and more populated districts, nor is political participation in more recently created districts significantly different from old established districts.

Evidence for the third hypothesis – the creation of districts aims at promoting less heterogeneous districts in order to reduce ethnic conflict – is inconclusive because only one out of three sub-hypotheses postulated is supported by conclusive evidence, and is insufficient to accept or reject the whole hypothesis. Our study confirms the propositions that proliferation of districts is an expression of neo-patrimonialism and of gerrymandered strategies by Museveni to stay in power.

The creation of new districts helps President Museveni to maintain political support and the affection of the people as it generates political positions and largesse to be shared among the party faithful and sympathizers in exchange for votes. Our findings on the period when redistricting occurs, the uneven regional distribution of district creation targeting opposition areas, and the subsequent winning of parliamentary seats there all bear the hallmarks of a gerrymandering strategy which started after 2006 elections. This was also the time when the regime felt heavily threatened by growing opposition. President Museveni's determination to survive multi-party competition at all costs has led him to pursue these two strategies. And no doubt the benefits to him and his regime have been substantial.

President Museveni is not alone in this league. Several African leaders have used a combination of patronage and gerrymandering strategies to perpetuate their stay in power. In all the countries, decentralization has been the overarching policy and bringing services closer to the people has been the veil; beneath it lies the real objective – to maintain a regime or entrench the dominance of the ruling political party.

Donors and development practitioners have often expressed their frustrations on decentralization policies and outcomes in SSA because they have analysed the policy based on what governments say their intentions are. Until they understand the political economy of decentralization in Africa they will always be bewildered not only by the mismatch between policy objectives and outcomes but also by many unspoken rationales for implementing decentralization reforms. Decentralization can be a facade behind which quite different practices take place. In many countries in SSA, decentralization is a narrowly political process that ruling governments pursue for their own benefit (or accept to pursue when pushed by donors where it hurts them least) and not for the nicely written policy statements of improving services.

Footnotes

Note

Nicholas Awortwi (PhD) is currently the Director of Research, Partnership for African Social and Governance Research (PASGR) in Nairobi, Kenya. Prior to that he was a Senior Lecturer at the International Institute of Social Studies of Erasmus University, The Netherlands and a visiting senior lecturer at F.H.R Lim A Po School of Public Administration & Governance, Suriname, and Nsamizi Training Institute of Social Development, Uganda. His research interests are decentralization and local governance, public sector reforms and basic services provision, in which he has over 15 publications.

A.H.J. (Bert) Helmsing is Professor of Local and Regional Development at the International Institute of Social Studies of Erasmus University Rotterdam, The Netherlands. His current research interests focus on institutionalist approaches to governance and local economic development. He recently contributed to the Journal of International Development (2012) on solidarity economy, co-edited a book on value chains, social inclusion and economic development, published by Routledge (2011) and contributed to the Journal of Modern African Studies (2010) on the Ethiopian flower industry.

References

African Election Database (2010) Elections in Uganda. Available at: http://africanelections.tripod.com/ug.html (accessed 24 August 2010).

Alesina

Baqir

Hoxby

(2004) Political jurisdictions in heterogeneous communities. Journal of Political Economy 112(2): 348–396.

Andrews R and van de Walle S (2012) NPM and citizens' perceptions of local service efficiency, responsiveness, equity and effectiveness. COCOPS Working Paper 7. http://www.cocops.eu/wp-content/uploads/2012/08/COCOPS_workingpaper_No7-.pdf.

Awortwi

(2011) An unbreakable path? A comparative study of decentralisation and local government development trajectories in Ghana and Uganda. International Review of Administrative Sciences 77(2): 347–377.

Baldersheim

Rose

(2010) Territorial Choice: The Politics of Boundaries and Borders, London: Palgrave Macmillan.

Bayart

J-F

(1993) The State in Africa: The Politics of the Belly, London and New York: Longman.

Bekunda

Namutebi

Newly-created districts begin operations. New Vision 2010. 1 July.

Borck

(2002) Jurisdiction size, political participation and the allocation of resources. Public Choice 113: 251–263.

Brancati

(2006) Decentralization: Fueling the fire or dampening the flames of ethnic conflict and secessionism? International Organization 60(3): 651–685.

10.

Breton

(1996) Competitive Governments: An Economic Theory of Politics and Public Finance, Cambridge: Cambridge University Press.

11.

Burki

Guillermo

Dillinger

(1999) Beyond the Centre: Decentralizing the State, Washington, DC: World Bank.

12.

Burns

Hambleton

Hogget

(1994) The Politics of Decentralisation: Revitalising Local Democracy, Basingstoke: Macmillan.

13.

Burton

(2001) The Economist. 1 November.

14.

Buwembo

(2005) What's the fuss about new districts? The Monitor. 20 June.

15.

Chabal

Daloz

J-P

(1999) Africa Works: Disorder as Political Instrument, 1st edn. Bloomington, IN: James Curry and Indiana University Press.

16.

Crook

Manor

(1998) Democracy and Decentralisation in South Asia and West Africa, Cambridge: Cambridge University Press.

17.

Dokolo District Local Government (2008/09) Three Year Development Plan, Dokolo, Uganda.

18.

Francis

James

(2003) The contradictions in Uganda's decentralisation program. World Development 31(2): 325–337.

19.

Golooba-Mutebi

(2008) Politics and local government in Uganda. In: Saito

(eds) Foundations for Local Governance: Decentralization in Comparative Perspective, Heidelberg: Physica-Verlag, pp. 137–164.

20.

GoU (1987) Report of the Commission of Inquiry into the Local Government System, Kampala: Government of Uganda.

21.

GoU (2004) Annual assessment of minimum conditions and performance measures for local governments, 2003, Kampala: Government of Uganda.

22.

GoU (2006) Decentralization policy strategic framework, Kampala: Ministry of Local Government Kampala: Government of Uganda.

23.

GoU (2007) Annual assessment of minimum conditions and performance measures for local governments, 2006, Kampala: Government of Uganda.

24.

GoU (2008) Uganda service provision assessment survey, 2007, Kampala: Ministry of Local Government.

25.

Grace J (2007) Malaysia: Malapportioned districts and over representation of rural communities. Available at: http://ssivaprabu.blogspot.com/2007/12/malaysia-malapportioned-districts-and.html (accessed on 14 September 2011).

26.

Green

(2008) District creation and decentralization in Uganda, London: Crisis States Research Centre Working Paper 24.

27.

Green

(2010) Patronage, district creation and reform in Uganda. Studies in Comparative International Development 45: 85–103.

28.

Gumbel

(2008) Election fraud and the myths of American democracy. Social Research: An International Quarterly 75(4): 1109–1134.

29.

Hyden

(2008) Institutions, power and policy outcomes in Africa. Africa Power and Policy Program (APPP), Discussion Paper No. 2 (June). London: ODI.

30.

John

(2010) Larger and larger? The endless search for efficiency in the UK. In: Baldersheim

Rose

(eds) Territorial Choice: The Politics of Boundaries and Borders, London: Palgrave Macmillan, pp. 101–117.

31.

Kato

(2009) Uganda: Why more districts for Buganda? The New Vision. 2 December.

32.

Khemani S. (2008) Gerrymandering decentralization: Political selection of grants-financed local jurisdictions. Paper prepared for IEB's 5th Symposium on Fiscal Federalism, titled Regional Fiscal Flows, ‘Balance-Sheet’ Federalism, and the Stability of Federations, Barcelona, 19–20 June.

33.

Khemani

(2010) Political capture of decentralization: Vote-buying through grants-financed local jurisdictions, World Bank Policy Research Working Paper no. WPS5350. Washington, DC: World Bank.

34.

Lake

Rothchild

(2005) Territorial decentralization and civil war settlements. In: Roeder

Rothchild

(eds) Sustainable Peace: Power and Democracy after Civil Wars, New York: Cornell University Press, pp. 109–132.

35.

Larok

(2008) Deepening decentralization or regime consolidation? What explains the unprecedented creation of new districts in Uganda?, NGO Issues Paper no. 1. Kampala: National NGO Forum.

36.

Malinga

(2008) Museveni promises district status to Ngora residents. Daily Monitor. 19 August.

37.

Manyak

Katono

(2010) Decentralization and conflict in Uganda: Governance adrift. African Studies Quarterly 11(4): 1–24.

38.

Mawdsley

(2002) Redrawing the body politic: Federalism, regionalism and the creation of new states in India. Commonwealth and Comparative Politics 40(3): 34–54.

39.

MINALOC (2007) Rwanda Decentralisation Strategic Framework: Towards a Sector-wide Approach for Decentralisation Implementation, Kigali: Ministry of LG.

40.

Ministry of Local Government (2007, 2009) Ministerial Policy Statement to Parliament, Kampala.

41.

Ministry of Local Government (2010) Survey to establish citizens' opinions and satisfaction with local government services and administration, Kampala: Ministry of LG.

42.

Mwenda

Mugisa

(1999) ‘War’ looms over break-up of Toro. The Monitor. 23 June.

43.

Ndegwa

(2002) Decentralization in Africa: A stocktaking survey, African Region Working Paper Series no. 40. Washington, DC: World Bank.

44.

Nsibambi

(1998) Decentralisation and Civil Society in Uganda: The Quest for Good Governance, Kampala: Fountain Publishers.

45.

Ocwich

(2005) Can Uganda's economy support more districts? New Vision. 8 August.

46.

Oyuku-Ocen

(2008) Decentralization and the dance of creating districts: Theory and practice in Uganda, Unpublished MA thesis. The Hague: Institute of Social Studies.

47.

Pollitt

(2005) Decentralization: A central concept in contemporary public management. In: Ferlie

Jr Pollitt

(eds) The Oxford Handbook of Public Management, New York: Oxford University Press, pp. 371–397.

48.

Smith

(1985) Decentralisation: The Territorial Dimensions of the State, London: Allen and Unwin.

49.

Smoke

(2003) Decentralisation in Africa: Goals, dimensions, myths and challenges. Public Administration and Development 23: 7–16.

50.

Sooma

(2008) Museveni grants Luuka, Buikwe district status. Daily Monitor. 15 July.

51.

Steffensen

Trollegaard

(2000) Fiscal Decentralisation and Sub-national Government Finance in Relation to Infrastructure and Service Provision: Synthesis Report on 6 Sub-Saharan African Country Studies, Washington, DC: World Bank.

52.

Suberu

(1991) The struggle for new states in Nigeria. African Affairs 90(361): 499–522.

53.

The Daily Monitor (2008) Uganda: All constituencies should be districts. 7 May.

54.

Treisman

(2007) The Architecture of Government: Rethinking Political Decentralization, Cambridge: Cambridge University Press.

55.

Tukahebwa

(1998) The Role of District Councils in Decentralisation. In: Nsibambi

(eds) Decentralisation and Civil Society in Uganda: The Quest for Good Governance, Kampala: Fountain Publishers.

56.

57.

White