The Democratic Republic of the Congo (DRC) is an ethnically/tribally divided society that has recently emerged from years of war (although sporadic fighting continues). Organizing and conducting elections in the DRC that will be viewed by all of the major stakeholders as free and fair presents a major technical and logistical challenge to the United Nations (UN)[1]. One of the many problems facing the UN in the DRC is the decision as to whether to delimit districts for the election of Members of Parliament (MPs), and if so, how.
Background
Since 1997, the DRC has been divided by ethnic strife and war. The hostilities, precipitated by a massive flow of refugees from the fighting in Rwanda and Burundi, led to the toppling of former dictator Mobutu Sese Seko by rebel leader Laurent Kabila in May 1997. Laurent Kabila’s regime was subsequently challenged by a Rwanda and Uganda-backed rebellion in August 1998. Troops from Zimbabwe, Angola, Namibia, Chad and Sudan intervened to support the Kabila regime. A cease-fire agreement was signed in July 1999 by the DRC, Zimbabwe, Angola, Uganda, Namibia, Rwanda, and Congolese insurgent groups, but sporadic fighting continued.
Joseph Kabila, who succeeded his father when Laurent Kabila was assassinated in January 2001, persuaded occupying Rwanda forces to withdraw from eastern Congo in October 2002. Two months later, an agreement (Global and Inclusive Agreement, 17 December 2002) was signed by all remaining warring parties to end the fighting and set up a transitional government. Ugandan troops officially withdrew from the DRC in May 2003. Localised violence (particularly in the Great Lakes region) continues, however.
Divided Society Battling ethnic groups (Tutsi, Hutu, Lendu, Hema and other ethnic groups) in the eastern portion of the country, supported by military forces from neighbouring countries, initiated much of the current conflict. Although the divisions within the DRC are not based solely on ethnicity, the large number of ethnic groups [2] – and the competition among them for limited resources – has certainly served to fuel clashes within society.
Lack of Resources Despite the vast potential of natural resources and mineral wealth, the DRC is one of the poorest countries in the world, with a per capita income of about US$90 in 2002. This is the result of years of mismanagement, corruption and war. In addition, the country has a high illiteracy rate (according to 2003 estimates, 41.7% of the population has had no schooling at all and an additional 42.2% has had only primary schooling) and limited technical expertise to draw on.
Little Intact Infrastructure The DRC is an enormous country (2,345,410 square kilometres; 905,063 square miles) with an almost complete lack of infrastructure. The existing roads (relatively few in number) have been badly damaged and many have yet to be tested for landmines. Although the UN Mission in the DRC, MONUC (Mission des Nations Unies en République Démocratique du Congo), has all existing airstrips operational, planes can reach a very limited number of areas in the DRC. In addition, the communication system is inadequate – although some radio and television broadcast stations are operating in the DRC. Conducting elections under these conditions will be quite challenging.
Lack of Legal Framework There is currently no electoral law in place. Nor has the transition parliament adopted laws on such issues as decentralisation and nationality – issues that are clearly controversial but must be decided before an election can be held. The transitional parliament, appointed by the signatories to the 17 December 2002 agreement, must reach agreement on these issues and promulgate laws and a new constitution before elections can proceed beyond the formative planning stage.
Time Constraints The transitional constitution, adopted on 2 April 2003, is of limited duration. It expires 24 months (with 2 six month extensions possible) after the inauguration of the transitional government, which occurred 30 June 2003. Elections must therefore be held by July 2005 (or July 2006 at the latest, if the two six month extension options are exercised). This is a very brief time period in which to organise a host of elections (including a referendum, general elections and local elections), even in optimal conditions – which certainly do not exist in the DRC.
Deciding on an Electoral System for the DRC
In newly emerging and transitional democracies, especially those where society is divided along ethnic, regional, or other factional lines, political institutions – particularly the electoral system – are especially significant. Political institutions in such societies are the most prominent channel of communication between antagonistic groups; if these institutions exclude significant groups from the “table” then differences between these groups cannot be resolved through negotiation and mutual accommodation. This could lead to the resumption of civil war and to the breakdown of the fledgling system.
Comparative experience suggests that the most important electoral requirement for democratic transition, particularly in a divided post-conflict state like the DRC, is an electoral system that maximises inclusiveness and is clearly fair to all parties. This goal is best achieved by a proportional representation (PR) electoral system, usually in conjunction with some form of overall power-sharing agreement within the government.
Given that some form of proportional representation is clearly in the best interest of the DRC, two obvious election system options exist. There are advantages and disadvantages associated with both of these systems which need to be considered when deciding which system to adopt in the DRC.
List Proportional Representation (PR) There is historical precedence for using List PR in the DRC: legislative elections held immediately following independence utilised this type of system (this was the system bequeathed to the DRC by the Belgium colonialists). Furthermore, List PR is the most common choice for newly emerging and transitional democracies so there is a wealth of experience to draw on for planning and conducting the elections.
Another advantage offered by List PR is that there would be no need to draw new electoral boundaries, although a decision would be required as to which set of existing administrative boundaries to use – province, district or territory – for regional List PR (assuming a national List PR system is not adopted).
A major drawback to using List PR is the current multiplicity of political parties – if this remains a feature of the DRC landscape, ballots could be far too long (especially if an open, rather than closed, party list is adopted). Political parties (most of which in the DRC have an ethno-regional basis and no ideological foundation) will have to unite, form coalitions or disappear for a List PR ballot to be manageable [3].
Another important disadvantage associated with List PR is that the geographical areas from which representatives are elected are usually quite large; hence the link between voters and their representatives is not as strong as it would be under a system offering single-member or small multimember constituencies. Many Congolese officials interviewed simply assumed that representatives would be selected from the level of the territory [4] (this has been the practice for the past 25 years at least) and felt that having representatives associated with specific territories was beneficial as it facilitated communication between voters and the government.
Mixed Member Proportional (MMP) The major advantage offered by an MMP system is the clear geographic link that is established between a representative and his or her voters. This specific concern was raised numerous times by Congolese interviewed. Not only would a distinct connection between a representative and his or her constituency facilitate the exchange of information, it would also empower Congolese voters. With time, Congolese voters would come to recognise their ability to hold their representative accountable (returning their MPs to office if they performed well on their behalf and voting them out of the legislature if their MPs failed them).
One important disadvantage with this system is the need to delimit a large number of constituencies. This could be done by either adopting administrative units smaller than provinces or districts [5], or by drawing unique electoral constituency boundaries. If such existing administrative units as territories were to be used, the population data associated with these units are even more problematic than the data at the provincial and district level. (The issue of data reliability is discussed at greater length below in “Construction of a Database”.)
Another disadvantage sometimes associated with an MMP system is the complexity of the ballot (MMP systems often require that two votes be cast, one for a constituency representative and one for a political party). In fact, however, the ballot can be designed so that voters cast only one vote; this single vote is used both to elect a constituent representative and to designate a party preference [6]. (Exercising the one-vote option would also limit the number of political parties to a manageable level, at least in the long run.)
Boundary Delimitation Options
Regardless of the electoral system adopted, some decisions will have to be reached concerning the delimitation of electoral boundaries. Although the size of the geographic area encompassed by the electoral constituency will vary depending on the type of electoral system (for example, List PR could adopt provincial lines as the electoral constituencies, while an MMP system would require smaller constituencies), some delimitation will almost certainly be necessary [7]. This delimitation could be as simple as adopting existing administrative boundaries (provincial, district or territorial boundaries, for example) as electoral constituencies and then allocating parliamentary seats to these constituencies on the basis of population; or it could be as complex as drawing new electoral constituency lines specifically for election purposes.
The importance of the delimitation process (and the rules that bind it), varies depending on the type of electoral system. Because plurality and majority systems can produce disproportional election results, the structure and rules established for the process are quite important. Although somewhat less important in the context of proportional representation systems, it is still essential that the law specify the process by which electoral constituency delimitation should occur.
Three alternatives exist for delimiting electoral boundaries in the DRC:
- Use existing administrative boundaries (for example, provincial, district or territory boundaries) for electoral purposes.
- New administrative boundaries could conceivably be drawn (and may in fact be proposed as part of a decentralisation package being debated by the transitional government in the DRC) and these could be utilised for electoral purposes.
- Electoral constituencies could be drawn that are unique (separate from the administrative structure).
Current Administrative Units Used as Electoral Constituencies in the DRC
The DRC is subdivided into several layers of administrative units, many of these in existence since before independence in 1960. The largest of these units are provinces, next in size are districts, then territories; the smallest administrative unit (for which data is collected) are collectivities. According to the Administrative Divisions Directorate of the Ministry of Interior, the present number of administrative units for each layer is as follows [8]:
Provinces (including Kinshasa) 11 Districts 27 Cities (namely ±98 communes, ±1,249 districts and Kinshasa) 29 Territories 145 Communities (spread out over 476 sectors and 261 chiefdoms) 737 Groups (spread out over ± 60,000 villages) 5409
In addition, six urban groups having more than 100,000 inhabitants which do not have a “city status” have been identified. They are spread out in the provinces of Katanga (Kalemi, Kamina and Kipush), Nord-Kivu (Kayna and Kanya-Bayounga in a single group), Orientale (Bunia) and Sud-Kivu (Uvira).
These administrative units are interlinked and used not only for government administration purposes but also to conduct scientific and administrative censuses and to determine demographic projections. They also are employed to carry out various operations at the national, provincial, regional and local levels such as medical immunization campaigns.
Maps identifying the boundaries of these administrative units are readily available, but are out-of-date because centres of population have moved. (The administrative boundaries themselves have not changed, at least not recently, but because of massive population movements as a result of the war, administrative boundaries may cut through the middle of new population centres).
Population projections for these administrative units exist, although these projections are not particularly reliable. (This issue is discussed at greater length in this portion of the study, in a section entitled “Construction of a Database.”)
Co-opting a layer of administrative units (for example, provinces, districts, or territories) for use as electoral constituencies has several advantages:
- Using already existing boundaries would negate the need to draw an entirely new set of electoral boundaries (which would be an extremely expensive and time consuming task).
- There is population projection data (albeit less than reliable) associated with these existing administrative units, making the exercise of allocating seats to constituencies easier, and perhaps more accurate, than would otherwise be the case.
- These administrative units have traditionally been used in the DRC for representational purposes (and many Congolese officials interviewed assumed that these units would be used again for the upcoming elections).
On the other hand, there are several disadvantages associated with the use of existing administrative boundaries:
- Administrative units were not designed to encompass communities of interest, and in fact often cut across tribal lines and divide homogeneous groups that should be united in a single electoral constituency.
- Some changes made to administrative boundaries (particularly at the territorial level) over time have been prompted by the desire to divide certain ethnic groups [9].
- The population projections that do exist for these administrative units are clearly not reliable, especially given the war and other unanticipated events (i.e., the AIDS epidemic).
If the decision is reached to use current administrative units for electoral purposes, then the question remains as to which set of units to employ for electoral purposes: provinces, districts, territories, or some smaller unit. Of course, a large part of this decision is dependent on the type of electoral system adopted. For example, if a List PR system is selected, then the choice of administrative units is limited to provinces, districts or possibly territories since the electoral constituencies must be large enough to permit the allocation of several seats to each constituency. On the other hand, if an MMP system is adopted, then electoral constituencies would need to be much smaller in size – territorial units would probably be the largest possible unit that could be employed.
A series of simulations were performed for illustrative purposes only (using the 1984 census data and employing no projections or adjustments to this data) to determine what the seat allocation would be to each electoral constituency under three scenarios: using provinces as constituencies, using districts as constituencies, and using territories as constituencies. The Appendix contains the detailed results of these simulations.
If provinces were to be used for electoral constituencies (this would only be possible under List PR or for the PR portion of an MMP system), for example, the range in the number of seats allocated to the provinces would be from 11 (Maniema province) to 58 (Orientale province) using the 1984 census data and hypothesising a 400 seat parliament.
If electoral constituency boundaries were to coincide with administrative district boundaries, and no districts were combined, the range in seats allocated would be as few as 0 or 1 (district of Bandudu in the province of Bandudu) to as many as 26 or 27 (district of Kwilu in the province of Bandudu).
If territories were used as electoral constituencies, a great many territories would have to be combined in order to be accorded representation. This is true whether approximately half of the representatives were to be assigned to electoral constituencies (as would be the case with an MMP system) or all of the representatives were constituency-based (as would be the case with any plurality-majority system). On the other hand, some territories would form rather large multimember constituencies, with as many as 8 MPs assigned to a single constituency.
The simulations in the Appendix illustrate a number of important points:
- The choice of which set of administrative units to utilize as electoral constituencies will have significant ramifications for the electoral system (how proportional the election results are likely to be), for the representation of communities (which communities would be assigned more and which less representatives, and which communities will have to be combined with others for representational purposes; and how many communities would be divided by electoral constituency lines), and for the electorate (how complex the ballot will be given the number of candidates competing and the number of seats to be filled).
- Some boundary “delimitation” would likely have to be engaged in, at least to the extent of deciding which administrative units to combine for electoral constituency purposes – at least if territories (and possibly districts) are chosen as electoral constituencies. If a level below the territory is selected, then most certainly “delimitation” will have to occur.
- The choice of a formula for allocating seats to electoral constituencies matters – an electoral constituency can receive more or less seats depending on what formula is used (i.e., compare our simple example of assigning seats when more than .51% of 74122 voters are assigned a seat, or when a seat is assigned for each 74122 voters)[10].
- The population data upon with the seat allocation will be based is important. Obviously the more reliable and less controversial the data, the better: the process will be more accurate, and will be perceived as more legitimate, if the population data on which the allocation is based is not in question.
New Administrative Units Used as Electoral Constituencies
In the event of a decentralization of the government into administrative entities other than the existing ones, new administrative units would have to be created. If these new administrative units are also to be utilized as electoral constituencies, then the criteria for the effective representation of electors should prevail during the administrative delimitation process.
Of course, even if these electoral criteria are not considered when the new administrative units are drawn, it is still possible to use the new administrative units as electoral constituencies. However, the same disadvantages would hold for the new administrative units as for the current administrative units (see list above), with the additional problems of having even less reliable population data for these new administrative units, and less time to prepare for the elections.
If, on the other hand, the new administrative units are designed with such criteria as population equality and communities of interest in mind, then the adoption of these units as electoral constituencies would make a great deal of sense. This approach would require the collection of new demographic and sociological data, however – an expensive and time consuming operation. (This issue is discussed at greater length in the section below entitled “Conduct a New Census.”)
Delimitation of a Set of Unique Electoral Constituencies in the DRC
The delimitation of constituencies in the DRC specifically for the purposes of the election of representatives to parliament would be an enormous undertaking (both in terms of the time needed and the resources required). The process would involve a number of steps, including (1) the construction of a database minimally composed of maps and population data; (2) the formation of constituencies by allocating parliamentary seats to sub-regions of the country and creating unique electoral constituencies within these sub-regions; (3) the evaluation of the proposed redistricting plan and the adoption of a final redistricting plan.
Construction of a Database Delimitation requires the collection of several different types of information. The two essential pieces of information are population data and maps. The population data, which may be in the form of census enumeration data or voter registration data, provide the only means of creating constituencies that are relatively equal in population. Maps are needed to ensure that only contiguous geographic population units are assigned to constituencies and that constituency boundaries do not divide communities of interest unnecessarily.
Possible sources of delimitation data in the DRC include:
- Using the currently existing census projections (based on 1984 census) and cartographic information
- Updating and consolidating the census projections and cartographic information using local expertise and technical assistance
- Conducting a new census (or, alternatively, a “light” census)
- Using information obtained from voter registration process to update currently existing data
Using Currently Existing Population and Cartographic Data Much of the data needed for delimitation purposes in the DRC is out-of-date and, because of the drastic changes the DRC has undergone in the past ten years, unreliable.
The last scientific census of the population in the DRC was held in 1984, whereas the most recent administrative census dates back to 1996. Since then, the National Statistics Institute (INS) has produced demographic projections that estimate the number of electors aged 16 and over at ±25,600,000. These data on electors are broken down, by province, as follows:
Bandundu 2,915,000 Bas-Congo 1,555,000 Équateur 2,807,000 Katanga 3,429,000 Kasai Occidental 1,876,000 Kasai Oriental 2,160,000 Maniema 762,000 Nord-Kivu 1,982,000 Orientale 3,447,000 Sud-Kivu 1,784,000 City of Kinshasa 2,902,000 TOTAL 25,619,000
Although these population projections are estimated to be marginally reliable at the national and provincial level, they are deemed to far less reliable for the lower administrative levels (districts, territories, collectivities, etc.). Population projections in the DRC are problematic in large part because of the prolonged civil conflict and the displacement of the population and higher than presumed mortality rates that the war entailed [11].
The cartographic data currently available for the DRC suffers from the same defect: much of it is out-of-date because of large population shifts, particularly in the last five years. Although the geographical coverage index of the territory of the DRC, available to the specialists of the Geographical Information Centre of MONUC, is diversified, some of this information has not been updated for twenty years.
Updating and Consolidating Census Projections and Cartographic Data MONUC is
equipped with a spatial reference geographical information system which, based
on the information compiled and entered, can locate demographical and
geographical elements, establish their exact number and distribute them in
space. This system specifies the administrative boundaries to the smallest
territorial entity; it also makes it possible to pinpoint hydrographical, rail,
road networks, etc. This structured set of data makes it possible, at least in
theory, to use a geographic information system to delimit electoral boundaries.
However, the demographic and sociological data necessary to use this system for
redistricting purposes is missing.
The main difficulty confronting MONUC specialists when maximizing the operation of the spatial reference geographical information system lies in the demographic and sociological data that are missing or incomplete or whose reliability must be validated. The collaboration of experts from the DRC would allow database specialists to obtain, verify and enter the required information.
Conducting a New Census The third option available for obtaining data for delimitation is to conduct a new scientific population census. However, this census operation would be quite costly and would require a minimum time period of two to three years to complete. Of course, the information obtained would be useful for endeavours far beyond the delimitation of electoral boundaries.
An enormous organizational and operational structure would have to be put in place to accomplish this task – the National Statistical Institute is ill-equipped to undertake a substantial statistical exercise at this point in time. For example, during the census held in 1984, use was made of 28,185 census-taking areas and 2,924 control areas spread out over the entire territory. In the event of a new census, these census-taking and control areas will first have to be checked and adjusted with respect to both the geographic territory encompassed and the number of anticipated respondents within them. The massive population movements in recent years, whether towards cities or elsewhere, require such a prior intervention.
Using Voter Registration Data The fourth option available is to use the information obtained during the voter registration process (assuming a voter registration procedure is conducted) to delimit electoral constituencies. A voter registration exercise designed to reach every household in the country could be used not only for a head count but could provide a geographical location for every potential voter in the country – invaluable information in a delimitation exercise.
The major drawback to using voter registration data is that much of the demographic and sociological data collected in a census would be missing from a voter registration database. Another problem is that the collection of this data would be completed rather late in terms of the election calendar, making delimitation on the basis of this data a challenge. Even if delimitation were to occur prior to the completion of the registration process, however, the voter registration counts could still be used to modify seat allocations if the need were to arise.
Formation of Electoral Constituencies Once a database has been prepared, the next step in the delimitation process is the formation of electoral constituencies. This is usually composed of one or two phases: the allocation, or apportionment, of parliamentary seats to regional entities such as provinces (this process is also referred to as “redistribution” in many countries); and the delimitation of electoral constituency boundaries within these regions.
The apportionment phase of the delimitation process is usually relatively mechanical [12], with the number of seats assigned to each sub-region usually dependent on the relative population of that sub-region. In countries that do not delimit single-member or smaller multimember constituencies, apportionment may be the only step taken to equalize population across electoral constituencies.
In countries that do delimit smaller constituencies, the second phase of the process is the creation of new electoral constituencies within the sub-regions themselves. (In countries that do not allocate seats regionally, this is the only phase in the delimitation process.) This is the step where the line drawers create a redistricting plan by assigning geographic units such as cities, towns and villages (or city blocks) to constituencies. A redistricting plan is complete when all geographic units within the given territory are assigned to a constituency and all constituencies in the plan meet the predetermined redistricting criteria.
Evaluation of Redistricting Plan Once the boundary authority has successfully completed a redistricting plan by assigning all geographic units in the territory to an electoral constituency, summary information for the plan should be produced. This information is used to evaluate the plan. A summary description of a redistricting plan should include information such as a description of the plan listing the geographic components of each constituency, map(s) of the plan showing the constituency boundaries, and a report summarising the most relevant statistical information for each constituency in the plan.
The summary information should allow the boundary authority, political parties, legislators and governmental officials, citizens, and other interested stakeholders to evaluate a redistricting plan according to the established criteria. Public hearing may be held to solicit the comments of these stakeholders. If, for example, the redistricting criteria adopted specify that constituencies be as equal in population as possible, information should be available regarding the population of each constituency the degree to which the population deviates from the electoral quota. The production of maps would allow interested parties to determine if communities of interest have been taken into account in promulgating the constituency boundaries.
After evaluating a proposed redistricting plan, including the solicitation of comments on this plan, the authority in charge of delimitation should endeavour to take these comments into account, and modify the redistricting plan accordingly. The final stage of the process is the adoption of the new redistricting plan; provisions for how this is accomplished should be described quite explicitly in the electoral law. In fact, the entire process (who should draw the constituencies, what criteria should be followed, etc.) should be mapped out as clearly as possible beforehand in the Election Act in order to guide authorities in charge of the process.
Problems with Delimiting A Unique Set of Constituencies The delimitation of constituencies in the DRC specifically for the purposes of the election of representatives to parliament would be an enormous undertaking both in terms of the time needed and the resources required. In fact, it is not likely to be technically feasible, given the lack of data and the current time constraints, to delimit a unique set of constituencies for the 2005 parliamentary elections. Furthermore, the delimitation of unique electoral districts – especially single-member constituencies – could well prove a political nightmare and is therefore not recommended for the DRC.
Conclusion
Only proportional representation (PR) electoral systems were discussed as viable options for the upcoming transitional elections in the DRC. This is because an essential condition for democratic consolidation in deeply divided, post-conflict countries such as the DRC is the inclusion of as many significant groups as possible, as proportionally as possible, in the parliament. A winner-take-all system that over-represents one of the political parties to the disadvantage of the others could only lead to more discord.
Regardless of what form of PR is adopted – regional List PR system or an MMP system – some delimitation of constituencies is likely to be required. Given the current status of population data in the DRC, it would be wise to consolidate and update census projections and maps for the currently existing administrative units (at least down to the territorial level) and use one of these sets of administrative units as constituencies for the 2005 parliamentary elections.
