Chapter Two: Demarcation of Constituency
Boundaries

Do Electoral Constituency Boundaries Matter?

Disparities in the number of voters in small and large constituencies instil heated political controversies because EMBs, boundary commissions, incumbent governments, and national parliaments charged with boundary delimitation are not always willing or able to revise long-standing constituency boundaries. A common denominator of drawing electoral divisions in sub-Saharan African countries is the rendering of parliamentary elections by unequal suffrage, as single-seat and multi-seat constituencies are malapportioned and the principle of one man, one vote is not respected.5

In this chapter, I present an inductive analysis of a wide range of sources, including documentation on the delimitation of electoral boundaries and lessons learned by scholars and practitioners, to establish patterns of institutional malpractice in the demarcation of constituency boundaries. The demarcation of electoral constituencies is a slow-moving variable in the prevention of electoral conflict. Nonetheless, this study examines it as a trigger requiring policymakers’ careful attention, because controversies over malapportionment have actually triggered election disputes in countries such as Zimbabwe, the Democratic Republic of Congo, and Ghana and have worked to undermine election stakeholders’ confidence in election management bodies and other competent authorities involved in the drawing of electoral districts.6

Challenges and Key Factors in Redistricting

Delimiting electoral constituencies is a challenging endeavour. Numerous technical and financial issues must be considered to accomplish a process that is usually lengthy and costly. The exercise involves collecting and systematising accurate information, including maps and population data, prior to drawing the new constituency boundaries. Redistricting requires changes in voter rolls and, likely, the relocation of polling stations. African states intending to update constituency boundaries face, in addition to political considerations, financial constraints and outdated population data.

Two main factors must be considered for the redrawing of electoral constituencies: first, the district magnitude, and second, the alignment of electoral district boundaries with existing administrative boundaries. District magnitude refers to the number of legislative seats elected in a district, either single member or multi-member. Single-member districts must be redrawn periodically to ensure that they comprise a relatively equal number of eligible voters. Updated single-member constituencies tend not to be aligned with existing administrative boundaries. On the other hand, multi-member districts coincide more often with administrative divisions because the number of representatives per district can be adjusted based on population changes and without altering district boundaries, to ensure that a relatively equal number of votes is required for every seat in each district (ACE n.d.).

Boundary Delimitation in Sub-Saharan Africa:
Principles and Standards

International organisations and nongovernmental organisations (NGOs) have developed standards, based on numerous widely accepted principles, guiding the delimitation of constituency boundaries. The following table summarises the current proposed standards based on previous pivotal work by the International Institute for Democracy and Electoral Assistance (IDEA) and Lisa Handley: 7

Table 2.1. Standards of Boundary Delimitation

However, these standards do not provide specific guidelines to border authorities on setting an agenda with clear priorities and benchmarks for constituency demarcation. In fact, the political priorities of communities of interest, to some extent, sideline the principle of one man, one vote. All international standards drawn by international organisations refer in some way to the equality principle enshrined by article 25 of the International Covenant on Civil and Political Rights (ICCPR), United Nations Human Rights Commission, General Comment No. 25, paragraph 21: “The drawing of electoral boundaries and method of allocating votes should not distort the distribution of voters” (UN Committee on Human Rights 1996). While the Venice Commission stipulates that the permissible departure from the norm of evenly distributing seats among the constituencies should not be more than 10 percent and certainly not exceed 15 percent, the fourth edition of the Office for Democratic Institutions and Human Rights (ODIHR) Election Observation Handbook also puts the threshold at 10 percent.8 The Commonwealth Secretariat and the Electoral Institute of Southern Africa (EISA), while stressing that each vote should have equal weight, do not clearly indicate the extent to which deviation based on other general principles should be permitted. Other organisations such as the African Union have not actually formulated any concrete proposals so far (Dundas 2011). Nonetheless, the evidence outlined in this chapter shows the difficulty of addressing the uneven allocation of seats in many sub-Saharan African countries, due to the lack of a sound regulatory framework and clear regional standards on deviation limits.

At the national level, data compiled by the Administration and Cost of Elections (ACE) Electoral Knowledge Network show that in two-thirds of sub-Saharan African states, constituencies are specifically delimited for election purposes, whereas provisions in the national elections acts and constitutions entrust the drawing of constituency boundaries to either boundary commissions (fifteen), legislatures (thirteen), EMBs (nine), or other governmental and independent entities (eight).9 The remaining states, mainly francophone states with the exception of Mozambique and South Africa, use existing regional, provincial, or other boundaries. Overall, in over half of cases documented, electoral boundaries are drawn based on the equality of voting strength (53 percent); other widely used demarcation criteria are based on local jurisdictions (44 percent), the geographic size of districts (40 percent), and natural barriers (33 percent).

Beyond legislation on constituency boundaries, the electoral representativeness of communal groups defined by ethnicity, geography, religion, language, and other identity bonds underpins the political debate and informs political choices regarding the demarcation of constituency boundaries. Electoral constituencies based on disputed administrative divisions often involve underlying grievances among communal groups regarding central governments’ allocation of decentralised resources.10 In particular, perceptions that authorities instrumentalise boundary delimitations to favour their own communal group and to secure electoral outcomes exacerbate disputes among supporters of warring parties and armed groups, and usually emerge during electoral campaigns.11 In this sense, the demarcation of constituency boundaries can play a critical role in disputes among minority groups over access to resources and rights. Nonetheless, as Posner’s study on the Chewa and Tumbuka cleavage in Malawi and Zambia illustrates, the political salience of ethnic cleavages depends also on the group’s size and its electoral significance in the competition for state resources (2004, 529–45). A case in point is the protracted and deep-rooted dispute in Plateau State (Nigeria) that originated in 1991, when General Babangida subdivided the Jos Local Government Area into two local government areas (LGAs). The new Jos North LGA, which included the city of Jos, then became demographically dominated by the Hausa community, and the so-called “indigene” communities, which are predominantly Christian, perceived this as the Northern Muslim General’s deliberate attempt to favour the Hausa settlers (Ostien 2009, 1–42). Controversies over political representation also play a major role in the protracted conflict between militias in the east of the Democratic Republic of Congo. RCD-Goma threatened to suspend its participation in the transitional institutions when the Independent Electoral Commission (CEI) rejected its demand to make the locality of Minembwe a constituency in North Kivu so that the Banyamulenge ethnic group could obtain a National Assembly representative in the 2006 general elections (EISA 2007a).

Shortages of Transparency and
Gerrymandering Allegations

Although most African countries have endorsed widely accepted delimitation principles, the bodies mandated to draw constituency boundaries often fail to revise constituencies regularly due to political pressures, or revisions occur without duly informing all stakeholders of the criteria used. Opposition parties tend to mistrust state authorities that are unable to operate transparently in the delimitation of electoral boundaries or to use reliable, accurate data to resolve allegations of gerrymandering.12

Zimbabwe provides an illustrative case. Lack of transparency has been a lingering electoral challenge over the last decade. Although the president was to proclaim the parliamentary constituencies’ names and boundaries upon reception of a final report and maps produced by the Delimitation Commission according to article 60 (8) of the 1980 Constitution, observers noted that before the 2000 elections the Delimitation Commission briefed only the Minister of Justice, Legal, and Parliamentary Affairs on the progress of the delimitation process. This minister happened to be the current Secretary of Finance of the ruling party, and opposition parties protested that the Zimbabwe African National Union-Patriotic Front (Zanu-PF) had obtained advance knowledge of the electoral boundaries––valuable information for establishing local campaigns in each constituency. They accused the Delimitation Commission of gerrymandering in favour of the ruling party by incorporating areas of rural constituencies into urban ones, to weaken the chances of opposition candidates in their urban strongholds (Commonwealth Observer Group 2005c; NDI 2000).

The Delimitation Commission’s report on the 2005 Zimbabwean Parliamentary elections was also released late, and as a result, its dissemination through voter-education campaigns was limited; voters did not have enough time to check the voter rolls and identify their constituencies. Moreover, copies of the report, including constituency maps, were not made available to voters. As a consequence, many voters were turned away from polling stations on election day because they were in the wrong constituency.13 Opposition and domestic observers pointed out that political considerations were taken into account in the delimitation of constituencies. Observers noted that the Harare South constituency, the only one that Zanu-PF won in Harare, did not follow the “natural” boundaries, and the town of Norton was transferred from the Mhondoro constituency into the newly created Manyame constituency. As a result, the Movement for Democratic Change (MDC) no longer controlled Mhondoro, but Zanu-PF and Mugabe’s party won in Manyame as well. There were concerns about the accuracy of the number of registered voters in the MDC strongholds, Harare and Bulawo, where the number of seats was reduced. Moreover, to conform to the maximum deviation from the average size, which the constitution establishes as plus or minus 20 percent, the boundaries of constituencies were shifted, and observers noted that political choices may have influenced the decision to move the wards (e.g., Mutare South and Chimanimani) (Vollan 2005).

Although the Constitution of Zimbabwe Amendment (No. 18) Act of 2007 abolished the controversial Delimitation Commission and its functions were transferred to the Zimbabwe Electoral Commission (ZEC), the same issues emerged in the harmonised general elections of 2008: the final delimitation report was published several days after the date set by the nomination court, parties had only two weeks to determine their candidates, and there was not enough time to make voters aware of the new boundaries (Zimbabwe Human Rights NGO Forum 2008, 13; EISA 2008b, 30). Once more, stakeholders raised concerns that more constituencies were created in rural than urban areas, and claimed that some wards of urban constituencies, such as Harare North, were fictitiously populated given that the communal address system allows community members to register using the same address (PAP 2008b).

Institutional Hindrance and Passive Malapportionment

Lisa Handley points out that malapportionment is not always active but serves to avoid the periodic redrawing of boundaries. This research project finds that in several African countries, even when boundaries are revised prior to elections, the equality principle is not respected. Malapportionment has become a source of controversy over alleged gerrymandering, yet it remains a largely overlooked phenomenon. Thus, election stakeholders, media, and civil society have not always shown concern for substantial gaps in the number of voters per constituency, due to lack of information on the extent to which constituency boundaries influence electoral outcomes. Well-known studies emphasise that incumbents choose electoral formulas to maximise political representation; consistent with this finding, consecutive victories in electoral contests diminish ruling parties’ incentives to update constituency boundaries unless they perceive electoral gains to be at risk due to bottom-up pressure or population mobility. In this context, constitutional provisions to revise constituencies at regular intervals are often bypassed (Norris 2012, 15).

In the Gambia, we find a paradigmatic case of an incumbent government’s resistance to revise constituency boundaries. After a group of young officers of the Gambian National Army, led by the former president, Yahya Jammeh, staged a military coup in 1994, the so-called Armed Forces Provisional Ruling Council (AFPRC) overthrew the long-standing government of Dawda Jawara. The AFPRC pushed for election reforms, and according to Elections Decree 78 of 1996, electoral constituencies for the 1996 presidential and 1997 National Assembly elections were demarcated based on chieftaincy districts. Such demarcation produced large disparities in the number of voters in each constituency (Foroyaa 2010b). Section 50 of the Constitution of the Second Republic came into force after the 1997 elections and stipulates that “the Independent Electoral Commission shall demarcate the constituencies for the purpose of elections to the National Assembly,” based on standard delimitation principles such as equality and representativeness (Republic of Gambia 1997). However, when the chairman of the electoral commission, Bishop Johnson, commissioned a report to revise the constituency boundaries according to the criteria laid out by the constitution, its recommendations were not implemented; instead President Jammeh removed Bishop Johnson from office.14

In addition, parliament passed a constitutional amendment in 2001 stating that an act of the National Assembly was to establish a boundaries commission responsible for demarcating constituencies and the criteria for demarcation. However, the boundaries commission foundational act has not been passed, and constituencies have not been revised since 2001 (Commonwealth Observer Group 2006a). As a result, there were disparities over 56,000 voters between constituencies in the National Assembly elections held in 2007 and 2012 (see Commonwealth Observer Group 2006a; The Gambia Echo 2006; Foroyaa 2010a).

As the Ethiopian case study illustrates, institutional hindrance to the revision of constituencies can reflect ruling parties’ eagerness to cling to power by any means. The Ethiopian People’s Revolutionary Democratic Front (EPRDF) government has thus far not shown political willingness to revise constituency boundaries. The current 547 constituencies have not been reviewed since they were established in 1995; the number of voters per constituency for the House of Representatives was highly unequal, as constituencies for the 2010 elections varied from just over 40,000 in the Beneshangul-Gumuz Region to nearly 130,000 in the Afar Region (EUEOM 2005a, 2010b).

Southern Africa reveals other examples of institutional hindrance. In Swaziland, the only monarchic autocracy remaining in sub-Saharan Africa, although the 2005 constitution stated that the Elections and Boundaries Commission (EBC) would review the number and boundaries of tinkhundla (constituencies), the constitutional boundaries for the 2008 national elections remained as demarcated in 1993 (Commonwealth Expert Team 2008). On the other hand, the National Assembly of Mauritius has consistently refused to revise constituency boundaries. Mauritius is commended internationally for its democratic credentials15 and is the only African country rated as an unflawed democracy in the Economist Intelligence Unit’s 2011 Democracy Index (twenty-fourth worldwide).16 The Elections and Boundaries Commission (EBC) is mandated to report the revision of boundaries at least once every ten years to the National Assembly, which so far has rejected the commission’s recommendations in 1976, 1986, 1999, and 2009, leading to a very unequal representation of voters. The number of eligible voters in each constituency differs widely (EISA 2006a). For example, both the smallest constituency, Port Louis Maritime/Port Louis East with 22,488 voters, and the largest constituency, Savanne/Rivière Noire with 58,341 voters, on the mainland have the right to three representatives. Yet, Rodrigues Island, with 26,930 voters, is entitled to elect only two representatives (EISA 2010a). An electoral reform proposal known as the Carcassonne report, commissioned by a group of academic experts, emphasised the need for redistricting along with a proposal for Mauritius to endorse a party-list system of proportional representation to ensure fairer representation of communal groups (Bogdanor, Carcassonne, and Vilanova, n.d.).

Lack of Reliable Voter Data

The quality of the voter register is a matter of concern even in African countries with more-consolidated electoral institutions and procedures. An outdated or inaccurate voter register leads ineluctably to malapportionment since the number of voters in each constituency cannot be ascertained. The parliamentary elections held in Madagascar in 2007 provide an example. Reliable data could not determine the number of voters per constituency since the previous census had been conducted in 1993. Actually, the delimitation of constituencies and allocation of seats for the sitting National Assembly established by a presidential decree in July 2007 were not based on any known delimitation criteria (Toulou 2008, 5). A blatant disproportion in the allotment of seats for the constituencies resulted. Hence, Antananarivo II and III had two seats each for a population of 136,001 and 191,506, respectively, while constituencies regarded as opposition strongholds, such as Vohibato in the Haute Matsiatra region and Mananjary in the Vatovavy-Fitovinany region, were allotted only one seat each despite having estimated populations of 527,144 and 702,256, respectively (EISA 2008a). The problem was not addressed prior to the first round of presidential and parliamentary elections held on October 25, 2013, since the long-overdue population census was not undertaken by mid 2014 due to funding constraints.17

Lack of reliable population data also seriously challenges the equal and fair representation of citizens in sub-Saharan Africa and the demarcation of constituency boundaries in post-conflict environments. Sierra Leone provisionally adopted the district-block system, rather than the single-seat system envisaged in the constitution, to directly elect representatives in the 2002 and 2007 parliamentary elections. An equal number of seats was allocated to every district (eight) despite great differences in the number of registered voters in every district.18 Likewise, in Liberia, although parties won legislative seats in the 1997 elections based on the percentage of the presidential vote obtained, in 2005 a simple majority system was adopted, and each of Liberia’s fifteen counties was allocated two seats in the Senate and two seats in the House of Representatives; the remaining thirty-four house seats were distributed to the counties based on the number of registered voters, despite a large disparity in the ratio of constituents to representatives in various counties (NDI and Carter Center 2007). In 2010, electoral preparations were delayed because the parliament and the executive could not agree on the number of additional seats to be included in the threshold bill based on the 2008 national census. The House of Representatives was reluctant to reduce the number of seats in the least populous constituencies, whereas President Sirleaf did not want to commit additional state resources to substantially increase the number of parliamentarians. Eventually, a compromise was reached: nine seats were added to those districts with the highest population growth, while remaining districts kept their existing number of seats (Carter Center 2011a, 23). However, this political agreement did not fully address inconsistencies in the ratio of voters required to elect seats across the country; the Liberty Party presidential candidate, Charles Brumskine, an attorney and former senator, accused Unity Party loyalists in the National Election Commission of having gerrymandered Liberia’s constituencies in Sirleaf’s favour (Foreign Policy 2011).

In contrast, other post-conflict states have chosen to update population data prior to demarcating new constituency boundaries for elections. In the Democratic Republic of Congo (DRC), although the voter-register quality was not optimal, it was used to design a seat-apportionment formula in compliance with international standards (DRC 2006). In contrast, in Sudan the delimitation of constituency boundaries was heavily contested because it was built upon a disputed census published in May 2009.19 Sudan’s census bureau acknowledged that the census was incomplete or inaccurate in several states, including conflict-torn areas in South Sudan and the bordering disputed areas of Abyei and the Nuba Mountains (Carter Center 2010b). The law envisaged that constituencies were to be created within thirty days, and the National Electoral Commission (NEC) entrusted this cumbersome task to the High Electoral Commissions (HECs) at the state level. As a result, constituency boundaries were “vague, unmapped, and difficult for observers and election officials to comprehend” (Carter Center 2010b). The boundary reports compiling the results of each state’s delimitation did not include detailed constituency maps or comprehensive village lists, leaving room for manipulation (see also Gustafson 2010, 28). In fact, many constituencies did not respect the principle of equality of the vote, in breach of article 38 (b) of the National Elections Act, which stipulates that the total population in each district should not deviate from the national dividend by plus or minus 15 percent. Of 270 constituencies, the deviation surpassed the 15 percent benchmark in thirty-one states.20 Moreover, lack of transparency tainted the appeals process. Although 885 objections were submitted to the NEC and 363 were accepted, the process was not always perceived as impartial because the HECs had authority to address objections to their own decisions within their states, and stakeholders believed that successful objections were not addressed (Carter Center 2010b).

A General Trend: Passive Malapportionment
in the Context of Rapid Urban Growth

Demographic trends have mainly reflected an increased number of Africans living in urban areas. According to UN projections, Africa is likely to reach a 50 percent urbanisation rate by 2035, from 14.4 percent in 1950, which means that one in every five of the world’s urban inhabitants will live in Africa (see UN Social Affairs 2011). Rapid population growth in several African states has rendered obsolete the population quota allocated to unrevised electoral constituencies. Outdated boundaries undermine the principle of one man, one vote, particularly in rapidly growing urban areas. In fact, further research on this issue would be desirable; nonetheless, available data indicate that maintaining or increasing the number of seats in rural areas, where voting patterns reflect traditional bounds and ethnicity, tends to favour incumbents, whereas voters’ choices are less predictable and opposition parties tend to perform better in urban, multiethnic areas.

In this context, East African countries experienced rapid urbanisation alongside economic growth. In Zambia, rapid urbanisation has resulted in discrepancies in the number of voters per constituency because the last review of boundaries was conducted before the 1991 general election; in the 2011 general election, there were fifteen times more voters in some urban constituencies, mainly in Lusaka, in relation to less densely populated rural constituencies (EUEOM 2011d). In Tanzania, the National Electoral Commission (NEC) conducts the delimitation of constituencies and reviews ward and constituency boundaries at least once every ten years.21 A revision was conducted before the 2005 general elections (Masterson 2009), and although the population in Dar es Salaam had annual average growth rates of 4 percent between 2000 and 2005 (approximately 100,000 new citizens every year22) (UN Habitat 2010, 8), there was only one more constituency (232) and sixteen more wards (2552) compared to the last delimitation done in 1995, as some districts and town councils were realigned (TEMCO 2006). During 2005–2010, Dar es Salaam became the fifth fastest-growing large city in Africa, in absolute growth terms (UN Habitat 2010). In 2010, the NEC added seven new constituencies, but these changes were insufficient as the gap widened; in 26 percent of constituencies a candidate needed fewer than 10,000 votes to be elected (even fewer than 1,000 in one case), whereas in other constituencies more than 100,000 votes were necessary (EUEOM 2010c). According to the results of the 2012 Tanzanian census, Dar es Salaam added almost two million citizens to its population between 2002 and 2012 (United Republic of Tanzania 2013), and the United Nations Department of Economic and Social Affairs projects that the city’s population will increase by 1,754,000 citizens from 2010–2020 (UN Department of Economic and Social Affairs 2010). In this sense, given the substantial number of voters who support opposition parties and independent candidates in CCM-dominated Dar es Salaam, constituency boundaries are likely to become a source of tension among political parties unless authorities create new constituencies to address disparities in the number of voters in each constituency. In addition, the CHADEMA party won in Arusha District in the 2010 elections, including in Arusha town, which has one of the country’s fastest population-growth rates (over 5 percent according to Arusha City Council data).23 CHADEMA also won other large urban areas with increasing population, such as Mbeya Urban (Mbeya Mjini constituency) and Mwanza Urban (Nyamagana and Ilemela Districts) (United Republic of Tanzania 2006).

The huge disparity in size and population of electoral constituencies in Kenya, particularly between urban and rural constituencies, fueled allegations of gerrymandering against the administrations of Daniel arap Moi and Mwai Kibaki. The Electoral Commission of Kenya did not revise the constituencies ahead of the 2002 general elections, although a High Court in Nairobi stipulated that the constituency boundaries did not satisfy constitutional principles specifying that all constituencies should contain nearly equal numbers of inhabitants.24 The least populous constituency, Wajir North, had 8,977 voters, while the most populous, Embakasi in Nairobi, had 152,906 voters. Concerns were raised that the disparity in constituencies’ size and population favoured the ruling party (Carter Center 2003a). In 2007, malapportionment was even greater. As requested by the High Court ruling of 2002, the Electoral Commission of Kenya (ECK) put forward proposals to increase the number of constituencies for the 2007 elections, but parliament rejected those; the 2007 elections, aside from minor administrative adjustments, maintained the same constituency boundaries as those used for the 2002 elections. The least populous constituency, Lamu East, had 12,866 registered voters, while the most populous, Embakasi, had 249,903 registered voters (EUEOM 2008).

Urban growth rates have also been significant in African countries severely afflicted by the HIV/AIDS epidemic. According to census data, the population of Lesotho’s capital, Maseru, increased from 137,837 inhabitants in 1996 to 227,880 in 2006; thus Maseru’s growth rate was 65 percent between 1996 and 2006, and the country’s total urban population rose from 16.9 in 1996 to 22.6 percent in 2006. In contrast, in the same period Lesotho’s growth rate between 1996 (1,841,967) and 2006 (1,876,633) was 1.8 percent, with an annual rate of only 0.08 percent, and the HIV/AIDS epidemic was a major factor in the declining growth (Joint UN Programme on HIV/AIDS 2012).25 In the context of rapid urbanisation, it is self-evident that the principle of equality can be upheld only by conducting regular and timely revisions: the review of 1998, before the general election funded by the United Nations Development Programme (UNDP), was based on the two-year-old 1996 population census, and even the snap elections of 2007 were based on the 1996 census (IEC 2011). In 2010, the Independent Electoral Commission conducted the long-overdue revision of constituency boundaries according to constitutional principles, and these changes took effect in the 2011 local government elections and in the 2012 general election. In this respect, international observers claimed that criteria used to transfer voters to different constituencies were unclear (Lesotho Times 2010; Commonwealth Observer Group 2012).

Malawi has also suffered excess mortality and a lower population-growth rate due to HIV/AIDS. Nonetheless, population increases in urban areas have been much greater; the population of the country’s capital, Lilongwe, tripled from 1987 to 2008 (Republic of Malawi n.d.b). Section 76 (2) of the Constitution of Malawi requires the Malawi Electoral Commission (MEC) to demarcate constituencies so that they contain approximately equal numbers of eligible voters, subject only to consideration of population density, ease of communication and geographical features, and existing administrative areas. The stipulation empowers the commission to review existing constituency boundaries at intervals of not more than five years and to alter them in accordance with the principles in subsection 2 (a) (Republic of Malawi n.d., article 76 2[a]). Demarcation was last conducted before the 1999 elections, was sponsored by foreign donors, and became controversial because some opposition leaders argued that the ruling United Democratic Front (UDF) influenced the MEC to demarcate more constituencies in the southern region, which was the ruling party’s stronghold.26 The process was not consistent with the principles in the constitution, given that a small constituency in Nkhata Bay East contained approximately 4,000 eligible voters, compared to a large constituency of 40,000 voters in Machinga North. A constituency-demarcation exercise did not precede the 2004 and 2009 elections, in the former because funds were not available and in the latter because political disagreements prevented parliament from passing a new delimitation submitted in 2008 (Patel 2006; EUEOM 2009). Consequently, the rapid urban-growth rate has led to the under-representation of highly populated areas in parliament, compared to less populated constituencies. Of 193 constituencies, ninety-eight had less than 30,000 registered voters, while fifteen had over 50,000 voters. The largest disparity was between the Chitipa Wenya constituency (9,468 voters) and the Lilongwe City Centre constituency (93,608 voters) (Patel 2006; EUEOM 2009).

Malapportionment in Urban/Rural Constituencies
Drawn According to Administrative Divisions

The ratio of voters in urban versus rural constituencies is imbalanced in several African countries in which conformity with local jurisdiction boundaries remains at least one criterion applied to the drawing of electoral boundaries. Administrative barriers inherited in part from colonial times were drawn based on historical communities of interest, geographic features, and means of communications, but the principle of equal population does not inform these divisions. Moreover, there are concerns that decentralisation initiatives are actually a tool for ruling parties to maintain political support in the constituencies by allocating state resources to feed clientelistic networks.

In Uganda, the decentralisation process gave birth to an extensive network of local government units divided in district or city councils, county or municipal councils, and smaller units down to the village level.27 According to the constitution, each county approved by parliament must have at least one member of parliament. Hence, there are 194 counties and 238 directly elected members of parliament, plus 112 elected women to represent each of the 112 districts of Uganda and twenty-five representing special groups such as the national army, workers, youth, and people with disabilities. Over-reliance on administrative divisions has resulted in wide disparities in the number of voters in each constituency, particularly between urban and rural areas.28 The ratio between North East Uganda and Kampala in the 2011 general elections was thirty to one. Inequality of votes is even greater in the election of seats reserved for women; the number of voters per district ranged from 21,039 in Amudat to 1.18 million in Kampala, a ratio of fifty-six to one (EUEOM 2011c; Saxena et al. 2010; Uganda Law Reform Commission 2011).

Similarly, in Ghana the Electoral Commission (EC) grants at least one constituency to each electoral district. The endorsement of this apportionment principle is controversial, however, given that the Ghanaian constitution’s section on the demarcation of election constituency boundaries does not provide for it. Nonetheless, the constitution specifies that guidelines to draw boundaries shall respect the equality principle and regional boundaries. Either way, measures taken by the EC have not actually addressed malapportionment, and voters of the two most populated regions of Ghana, Ashanti and Greater Accra, are clearly underrepresented. For example, in the 2008–2009 parliamentary election, the number of voters per constituency varied significantly, from 13,679 in the Nadowli East constituency, Upper West Region, to 162,646 in the Weija constituency, Greater Accra Region (EUEOM 2009a; Smith 2011, 5).

As highlighted above, several francophone African countries rely heavily on administrative delimitations to draw constituency boundaries, but demographic discrepancies among administrative units can lead to electoral malapportionment unless adjustment mechanisms exist. In the Central African Republic, each sub-prefecture (sous-préfecture) is an electoral constituency, except the capital, Bangui, where each arrondissement is an electoral constituency. Taking into consideration demographic growth, the electoral code stipulates that new constituencies must be created for each 35,000 additional voters of sub-prefectures and 45,000 new voters of each arrondissement of Bangui (Central African Republic 2009). On the basis of a report addressing territorial administration and produced by the Ministry of Territorial Administration and Decentralisation, led by a prominent member of the ruling National Convergence party Kwa Na Kwa (KNK), the government issued Decree No. 10,049, which opposition parties criticised because it compromised the principle of one person, one vote by allocating fewer seats to their strongholds. Observers also criticised that the allocation was based not on the 2010 registration exercise but on the 2005 voter register. In this sense, the Central African Republic illustrates malapportioned seats across the board. Of six constituencies, the 531,763 voters of Bangui’s constituency, the capital city, were allocated the same number of seats as were three other constituencies, each with fewer than 326,00 voters, while the smallest constituencies were given three and four seats for populations of 37,595 and 98,881, respectively (EISA 2010c). Given the urban growth rates, if the allocation had been based on the 2010 voter register, the disparities would have been even more blatant.

The unequal allocation of seats in electoral constituencies drawn according to administrative divisions fuels regional tensions. In Togo, the number of seats allocated to northern and southern constituencies has become a highly controversial issue. The framework agreement (the so-called the Global Political Accord, l’Accord Politique Global [APG], of 20 August 2007) involving the ruling party, the Rassemblement du Peuple Togolais (RPT), and some opposition parties retained standing electoral constituency boundaries based on administrative divisions (so-called prefectures) for the 2007 parliamentary elections. The parties of the so-called National Dialogue also agreed to use party-list proportional representation, with a system of highest averages and the préfecture as electoral constituency; a minimum of two seats were allocated to each constituency, plus an additional seat for those constituencies with more than 100,000 inhabitants. However, additional seats were allocated so that a much higher ratio of voters was required to elect a seat in constituencies of the southern regions, particularly in the Togolese capital, Lomé, and its hinterland in the southern maritime region (Golfe Préfecture). In Golfe, 127,071 voters elected a seat, while 11,468 voters sufficed in Assoli Préfecture in the northern region of Kara (OIF 2007; see table in EUEOM 2007a, 28–29; PCAR and OPSRA 2010, 23).29

Concluding Remarks and Lessons Learned

The demarcation of electoral constituency boundaries in sub-Saharan Africa entails technical challenges and politically sensitive dilemmas. The political arena remains dominated by hegemonic parties that control the state’s bureaucratic machinery, limiting the capacity to revise constituency boundaries impartially. Numerous countries have replaced or reformed their delimitation bodies, and independent members appear to be appointed. Whether parliaments, executives, or independent commissions are in charge of demarcating constituency boundaries, evidence shows that most sub-Saharan African states still lag behind in the impartial delimitation of constituencies. They have not succeeded in applying transparent procedures to strengthen stakeholders’ confidence in the process and to dissipate gerrymandering allegations.

Even though over half of all sub-Saharan African states include constitutional and other legal provisions enshrining equality of voting strength as a core criterion for constituency demarcation, long gaps between revision exercises result in huge disparities in the number of voters between fast-growing constituencies and others. In short, the equality criterion is breached on a regular basis. As discussed above, while the equality principle should not exclude other criteria such as representativeness and nondiscrimination of minority groups, this research underscores the need to address malapportionment through the establishment of clear standards or guidelines specifying maximum deviation limits. Hence, policymakers, legislators, and boundary authorities willing to guarantee the integrity of the electoral process need to establish appropriate deviation benchmarks based on the advice of delimitation experts.

In the author’s view, deviations of more than 20 percent indicate a dysfunctional democratic system that does not truly reflect the will of the people. African ruling parties gerrymander to achieve electoral gains by marginalising ethnic groups, including minorities, who are perceived as followers of opposition parties; nonetheless, high deviations do not necessarily result from incumbents’ desire to gerrymander but, more generally, from a reluctance to open Pandora’s box, namely, to address demographic changes and challenge long-established boundaries in urban or rural areas. Thus, political discussion of the representativeness of electoral constituencies, beyond long-standing administrative boundaries, is often absent. In this regard, striking a balance among equality, representativeness, and nondiscrimination could prevent outbreaks of ethnic violence linked to contested constituency boundaries during electoral periods.

On the basis of evidence analysed in this chapter and other lessons discussed above, the African Union and subregional organisations should consider recommending standards to guide their member states on the use of unambiguous criteria based on the principles of impartiality, equality, representativeness, nondiscrimination, and transparency, for the demarcation of constituency boundaries. AU member states could explore ways to promote these standards, and if deemed necessary, the African Charter on Democracy, Elections, and Governance could be amended to this effect, as provided by its article 50 (AU 2007). Finally, the deficiencies identified in this research underscore the need for international actors to promote an international convention on constituency delimitation, under the auspices of the United Nations and an advisory body, to monitor member states’ compliance with their international and regional commitments regarding boundary delimitation.