Electoral System Design and International Trends in Electoral Reform

A country’s choice of electoral system is one of the most important institutional decisions for any democracy as it establishes the link between the political preferences of citizens and the policy choices made by government. The rules of the electoral system determine how we choose our representatives, who gets elected to govern on our behalf, who makes our country’s laws and who sets policy and holds government accountable for its performance. While not a silver bullet for all of society’s challenges, appropriate electoral institutions are foundational to the quality of democracy and can nudge the political system in the right direction through appropriate institutional design.

This chapter provides an overview of the Electoral Reform Consultation Panel’s (ERCP) research, including an introduction to electoral system design, the effects of different design choices to achieving specific values and international trends in electoral reform. The first section discusses the key elements of electoral system design, including the broad categories of electoral systems, the size of electoral districts (constituencies), compensatory seats, ballot structure and electoral thresholds. This is followed by a section on the evidence for the effects of different designs on achieving the guiding principles adopted by the ERCP. These guiding principles were adopted by the ERCP to assess electoral system options in terms of the country’s Constitutional values including inclusivity, fairness, accountability, simplicity, electoral manageability, transparency and stable and efficient government. The final section provides a summary of electoral reform trends around the world.

Electoral System Design Elements

In simple terms, an electoral system is the method and rules that determine how the votes cast in an election are translated into the seats won in legislatures by political parties and candidates. Key design elements include the electoral system type (majoritarian, mixed, or proportional systems), the size and boundaries of electoral districts, formulas for seat allocation, the structure of ballots, and a range of other smaller design elements.

Issues of election administration are separate considerations from electoral system design and are beyond the mandate of the ERCP. This includes questions around the use of electronic voting, the number and location of voting stations, voter registration and candidate nomination procedures. The choice of electoral system may, however, have implications for electoral functionality insofar as it affects ballot structure and ballot numbers per voter, counting procedures and the declaration of results.

Electoral System Types

Electoral systems are broadly categorised into three main families: plurality/majority systems, mixed systems and proportional representation systems.⁶⁶ Plurality/majority systems include first-past-the-post (FPTP) electoral systems such as that used in the United States, the United Kingdom, and various other former British colonies such as India, Botswana and South Africa pre-1994. The French electoral system, based on a two-round system of voting, presents another plurality/majority variant. In such a system a second round of voting takes place if no candidate wins the first round of voting outright – that is, 50 percent plus one. In these systems, voters in a defined electoral region choose one candidate, and the candidate who receives the most votes becomes the sole representative for that area.

Systems of proportional representation (PR) allocate legislative seats in proportion to the number of votes received by each party and most commonly use party lists from which candidates are elected.⁶⁷ In systems using party lists, voters choose between party lists prepared by political parties rather than voting for individual candidates. The winning candidates are taken from party lists according to the number of seats won in proportion to their share of vote share. South African national and provincial elections make use of a party list PR system, which also allows for independent candidates to contest elections. PR systems are common among European and South American countries, as well as a range of African countries, including our neighbours Namibia and Mozambique.

Mixed electoral systems combine elements of plurality/majority and PR systems. In such a design these systems run alongside each other. In mixed systems, some candidates are typically elected in single-member constituencies⁶⁸ while others are elected in multi-member constituencies.⁶⁹ South African local government elections use a mixed system whereby candidates are elected in wards and from party lists.

Mixed systems differ according to whether competition in single-member constituencies and PR lists are linked or not. In parallel systems the results of the two elections are distinct, and seats are allocated independently with no interaction between the results in the two systems. This is practised in countries such as Japan, Italy and Ukraine. Other mixed systems have a compensatory component in which the results of these two types of elections are linked. Some see this as a variant of PR as it will still result in overall proportionality. This type of system is most commonly referred to as mixed-member proportional representation (in New Zealand) or personalised proportional representation (in Germany). In these systems the allocation of party list seats is linked to the results of contests in single-member constituency seats, and allocations of party list seats are adjusted accordingly to ensure overall proportionality. Variations can be found in countries such as New Zealand, Germany and Lesotho. South African local government elections use a mixed variant where proportional representation seats are determined by the aggregation of both ward and party list ballots in a two-tier compensatory system.

Electoral Districts (Constituencies) and District Magnitude

These broad “families” provide a shorthand for identifying some of the main features of different systems, but there is often significant variation between systems sharing the same broad category. Beyond these broad families electoral systems also vary in the size of electoral districts in terms of the number of seats allocated for election in a particular area. An electoral district, or constituency, refers to the geographic boundaries within which a group of voters vote for a set number of candidates. An electoral district can be for the election of one or more representatives to an elected body. In some countries, this is also referred to as a constituency, which is used as a synonym for electoral districts.

Plurality/majority systems typically make use of single-member constituencies,⁷⁰ while all PR systems make use of multi-member constituencies. In mixed systems representatives are usually elected from a combination of overlapping single-member and multi-member constituencies. The size of multi-member constituencies is a key design feature differentiating systems of PR in terms of the number of candidates elected from a specific constituency. This is referred to as the district magnitude of a constituency. Party lists are prepared according to the district magnitude of multi-member constituencies, and the number of seats in a constituency is generally proportional to the number of voters in that area.

The Constitution of South Africa sets a maximum of 400 seats for the National Assembly, while provincial legislatures can have a minimum of 30 and a maximum of 80 members,⁷¹ depending on the population size of each province. In line with the Electoral Act, the National Assembly has 400 seats. Of these, 200 seats are divided among nine multi-member constituencies aligned with the boundaries of each of the nine provinces (the regional ballot). The remaining 200 seats are compensatory (including both results from the national and regional ballots).

Within these multi-member constituencies, voters were asked in May 2024 to elect representatives from regional party lists, ranging from five representatives in the Northern Cape to 47 in Gauteng – the largest constituency. Provincial legislatures have a single-tier multi-member constituency; Gauteng and KwaZulu-Natal have the maximum 80 seats, while the Northern Cape and Free State have the minimum of 30 seats. Voters receive a single ballot to vote for the provincial legislature.

Countries using PR systems vary in their approach to deciding the boundaries of multi-member constituencies and, in most cases, follow existing administrative boundaries. Multi-member constituencies can either be at the national level as seen in small countries including Israel and The Netherlands, at the provincial level such as in Brazil, Spain and Switzerland or at the local government level as seen in Norway and Portugal. Some countries such as Sweden and Turkey follow administrative boundaries but subdivide major cities into smaller multi-member constituencies to avoid large discrepancies in the size of constituencies to bring representatives closer to voters.

The size of constituencies is a key design choice within the family of PR systems. The degree of proportionality and the closeness between voters and their representatives depend on the size of constituencies. Electoral districts with a higher district magnitude (the number of seats in a multi-member district) increase the proportionality of an electoral system. However, larger constituencies weaken the linkage between voters and their representatives.⁷² This linkage includes defined responsibilities to an area, the ability of voters to identify and engage with their representatives and the degree to which their vote affects the re-election chances of representatives (i.e., their ability to vote representatives out of office for non-performance).

In understanding the effects of district magnitude on the proportionality of electoral results, it is important to recognise that improvements in proportionality diminish as the size of multi-member districts increases. The biggest improvements in proportionality are achieved as the number of seats in a constituency increases from one to five, after which improvements in proportionality flatten out. Any increases in the size of constituencies beyond eight seats show limited improvement in overall proportionality.⁷³

An overview of multi-member constituency variations in proportional representation systems can be found in Appendix 1. On average South Africa has the highest district magnitude (largest constintuencies) and the second largest average population per constituency being served among all PR systems.

Compensatory and Two-tier Systems

Where an electoral system uses compensatory seats, these seats will be decisive in determining overall electoral outcomes and proportionality. South Africa’s electoral system for the National Assembly makes use of a two-tier system whereby voters vote for candidates drawn from national and regional party lists, or independent candidates. The higher tier national party list is compensatory and seats are allocated based on the overall results of votes on both the regional and national ballots. The allocation takes into consideration the number of seats won in constituencies to determine the final allocation of compensatory seats. Parties that receive too few votes on regional ballots to qualify for one of the regional constituencies can still win national seats based on their overall vote share on both ballots. In the 2024 elections, for example, several smaller parties received too few votes in the regional constituencies to secure a seat but received enough votes overall to win seats in the National Assembly (based on their allocation of compensatory seats).

Compensatory seats, also known as adjustment, levelling or equalisation seats, can be used in mixed-member proportional and in PR systems such as ours. Compensatory seats aim to increase the overall proportionality of electoral systems and are usually introduced to address any disproportionality caused by the use of small constituencies (especially single-member constituencies in mixed-member proportional systems). Appendix 2 provides an overview of countries that make use of two-tier compensatory systems.

The allocation of compensatory seats considers nationwide votes to determine the overall seat share, or entitlement, for parties based on their total vote share. The overall seat entitlement represents the total number of seats due to a party based on its proportional share of the vote. This entitlement is subtracted by the number of seats won by a party in constituencies to determine the number of compensatory seats due to a party. For example, if the overall seat entitlement of a party is 40 seats and it wins 18 constituency seats, the party will be compensated with 22 compensatory seats to make up the difference and return overall proportionality. As far as party seat shares are concerned, the outcome is identical to a straight nationwide allocation. The size of constituencies then becomes irrelevant to the final results and overall proportionality of the system as long as enough compensatory seats are allocated at a second stage to correct for any disproportionality at the constituency-level.⁷⁴

A key question for two-tier systems is how many compensatory seats are needed to overcome the disproportionality arising from constituency-level results. This will depend on the degree of disproportionality generated in constituency results. The more disproportional the lower tier, the greater the share of compensatory seats needed if full compensation is desired.⁷⁵ Factors such as the average district magnitude and the number of seats in the legislature will affect the degree of disproportionality in the lower tier.⁷⁶ Consequently, very large compensatory tiers are “unnecessary to produce a substantial reduction in disproportionality,” where constituency-level results are already proportional such as where multi-member constituencies are used.⁷⁷

Ballot Structure (Open/Closed Lists)

Another key design feature differentiating PR systems is the ability for voters to indicate candidate preferences on party lists. South Africa uses closed party lists, meaning that the order of candidates on a party list is determined by political parties. In such a case, voters are unable to express preferences for individual candidates. Voters choose between pre-determined party lists rather than preferred individual candidates. In open list systems, voters can indicate preferences for individual candidates on party lists. Currently, 40 of the 82 PR systems use open lists – a near even split between open and closed list systems. Most countries using open list systems are substantially smaller than South Africa, and only three – Brazil, Indonesia, and the Democratic Republic of Congo – have populations larger than South Africa.

There are many variations of open list systems which are distinct in how votes are cast and the extent to which preferences determine the final allocation of seats to individual candidates. In terms of voting methods, the most common form is for voters to vote for a list and for one or more candidates on that list. These candidate preferences are then taken into consideration to determine the order of candidate positions on lists to a greater or lesser extent, depending on the rules of the system. In fully open list systems, the number of votes received by candidates determines their order on the party list. “Flexible” lists introduce a candidate vote threshold that a candidate must pass to move up a party list, usually expressed as a percentage of that party’s votes.

Open lists can give voters a greater choice over the election of individual candidates, which can improve representation and accountability. This can lead to the election of better candidates although this is not guaranteed. The design of these systems, including questions of ballot structure, should be carefully considered to determine whether this can be effectively implemented in a particular system. A key issue is how party affiliation, candidates’ names and/or photos will fit on a ballot paper, which becomes more challenging as the size of constituencies and the number of competing candidates increases.

The adoption of such an open list system in South Africa may be impractical under the current electoral framework. This is mainly due to the large size of constituencies in the most populous provinces, and the large number of political parties and independent candidates contesting elections. In the case of Gauteng, political parties can put forward 47 candidates on their party lists for the regional ballot, and 52 political parties as well as independent candidates contested the 2024 election. This would have practical implications for the preparation and ease of use of ballot papers. For these reasons, open lists are more common in smaller countries, those that use smaller constituencies and those that include a legal threshold.⁷⁸

Thresholds

Every electoral system has a minimum number of votes that political parties or independent candidates must reach to win a seat. In some countries, there is a legally imposed threshold, usually expressed as a percentage of the vote, to discourage excessive fragmentation of party systems. Some countries allow a degree of leeway in which winning a single seat in a multi-tier PR or mixed system is sufficient to secure representation without meeting the percentage-based threshold, or do not apply thresholds to independent candidates. Some two-tier systems have no threshold for the lower tier but include a threshold for the allocation of compensatory seats from the national list.

The question of a legal threshold has become relevant for South Africa in the context of the recently prepared Municipal Structures Amendment Bill, 2024. The bill includes a provision proposing to introduce a legal threshold for local government elections. Such thresholds are relatively common, with 52% of PR systems and 53% of mixed systems applying them. While proportionality is generally regarded as good, most electoral systems have some way of limiting it, with electoral thresholds being the most explicit barrier to entry. This is usually justified in terms of preventing excessive fragmentation of party systems, which can undermine the formation and stability of coalitions as well as the development of institutionalised political parties.

Another manifestation of the threshold concept is to use the seat quota as a legal threshold such as used in the Netherlands. In this form, the quota becomes the threshold where only parties that achieved the quota in the first instance are considered for further allocations, meaning that parties cannot win seats only based on remainders. This concept is loosely referred to as a quota-threshold.⁷⁹ It promotes electoral merit in that contestants need to meet the quota of votes per seats for any consideration of seat allocation.

Countries without a legally imposed threshold instead have effective thresholds based on how easy/hard it is to win a single seat. This effective threshold is determined by the size of elected bodies (more seats that can be won) and the permissiveness of electoral rules that make it easier to win seats. This is generally a by-product of electoral system designs with very high proportionality. South Africa’s electoral system for the National Assembly is highly permissive and has previously had an effective threshold as low as 0.17% – among the lowest in the world.⁸⁰ This should not be confused with quotas, such as Droop or Hare, which are used in calculating the number of votes needed for a whole seat and formulas for allocating remainders.

Considerations for South Africa

These various design choices have implications for the degree of proportionality in electoral systems, the linkage between voters and their representatives, the ease with which small parties can be elected and the precision with which voters are able to choose their representatives. In some cases, this may require trade-offs between priorities. However, the relationship between principles is often non-linear, and the effects of a design choice in one area can be counter-balanced with choices in another area of electoral system design to find an appropriate balance between priorities for the South African context.

Principles and Evidence for Assessing Options

When considering electoral system options for South Africa it is important that local circumstances be considered as electoral systems that work in other countries may not be appropriate for the South African context. To navigate the myriad of electoral systems available, the ERCP adopted seven guiding principles for evaluating proposed electoral systems for South Africa. These principles draw on the country’s founding values set out in section 1 of the Constitution to establish a multi-party system of democratic government based on accountability, responsiveness, and openness to uphold values of human dignity, non-racialism, non-sexism and the achievement of equality. The seven guiding principles are: inclusivity, fairness, accountability, simplicity, electoral manageability, transparency, and stable and efficient government. These principles are not exclusionary and instead serve as a starting point for assessing electoral systems and other principles that may also prove applicable to the South African context.

Inclusivity

Given South Africa’s diversity and history, inclusivity is indispensable. The electoral system choice should encourage participation across the broad political and socio-economic tapestry of South Africa to foster diversity of opinion and encourage representation of all South Africans across demographic and geographic boundaries. This includes gender representation in line with values of non-sexism and equality.

Majoritarian/plurality systems more often tend to result in competition between two main parties, while systems with an element of proportional representation tend to lead to a greater range of parties winning seats and usually result in the formation of coalition governments.⁸¹ In this way, electoral systems with an element of proportional representation, including PR and mixed systems, are more likely to promote political diversity and representation of minority views, while more closely reflecting the public at large.⁸²

Diversity, representation and inclusion can take several different forms:⁸³

• Geographical representation implies that each region, be it a town or city, province or electoral district, has members of the legislature representing their area who are ultimately accountable to their community.

• Ideological representation includes the diversity of political perspectives held in society, which contributes to policy deliberations.

• Representation of the political situation that exists in the country in the sense that the legislature should be representative of the relative levels of support for different political views – i.e. outcomes should be broadly reflective of the will of voters to represent views held in society as a whole.

• Descriptive representation whereby the legislature should to some degree “mirror the nation,” including men and women, the young and the old, the wealthy and the poor and reflect the different religious affiliations, linguistic communities, and ethnic groups within a society.

The representation of different communities is a key concern in diverse societies. Systems with high levels of proportionality, allowing for the participation and representation of small parties representing particular communities. South Africa’s use of compensatory seats is a powerful tool for ensuring the overall proportionality of electoral results, allowing for the inclusion of small parties. In the current system, most small parties representing minority groups primarily receive seats from the compensatory tier, which considers votes received on both the regional and national ballot to determine a party’s seat entitlement across the country as a whole. This means that demographic groups that are not geographically concentrated can still gain the same level of representation despite being dispersed across multiple constituencies. In terms of electoral system design, South Africa’s use of compensatory seats is the main design element that allows for the political representation of diverse communities in the legislature.

Gender parity forms another important dimension of representation and inclusion. South Africa currently does well on gender representation based on international comparisons. There is 43% female representation in the National Assembly, 38% across provincial legislatures (ranging from 30-47%), and 37% female representation at the local government level (with similarly wide variations across regions). Fewer women are, however, elected in ward seats (27%) than on party lists (46%) at the local government level.⁸⁴ Many factors influence gender representation in legislative assemblies, including societal norms and values, candidate selection processes and the rules surrounding electoral system design.

Both political parties’ candidate selection choices and voters’ own biases/preferences in casting their ballot affect gender representation. International experience generally regards elite bias, rather than voter bias, as the main factor responsible for women’s underrepresentation, with candidate selection processes being a key driver of underrepresentation.⁸⁵ Most studies suggest that voters are as willing to vote for women as for men, and in the South African context, voters are positive to female leaders, with the latest Afrobarometer results showing that 56% of respondents strongly agree and 25% agree that women should have the same chances as men to be political leaders.⁸⁶ Only 15% thought that men make better leaders than women.

Electoral systems with a party list component outperform plurality/majoritarian systems in terms of gender representation. PR (32%) and mixed-member proportional (36%) systems have the highest female representation (in part because these systems are more likely to have gender quotas, either voluntary or formal, in place)⁸⁷ – compared to 22% among plurality/majoritarian systems. Parties’ candidate selection processes and the incentives created by electoral rules affect rates of gender representation, with evidence showing that single-member districts disadvantage women relative to multi-member districts.⁸⁸

Single-member districts present a zero-sum game where parties are often more likely to put forward male candidates with the assumption that they have a better chance of winning (despite evidence showing that this is a poor electoral strategy). In turn, parties are more likely to put forward a diverse candidate list in multi-member districts. The second reason why single-member districts can favour male candidates is related to incumbent advantage and in countries with a low turnover in their legislative assemblies, fewer opportunities are presented for women to enter the legislature. However, in the case of electoral reform where single-member constituencies are introduced for the first time, there would be no incumbency advantage in specific constituencies and this effect would not be seen.

The effects of other aspects of electoral system design are nuanced. While single- versus multi-member districts are seen to have a gendered effect, the size of multi-member districts (district magnitude)⁸⁹ does not appear to affect gender representation. Statistical tests show no correlation between district size and rates of gender representation.⁹⁰ In the South African context, there is little difference between rates of female representation in the regional constituencies (45%) and the national compensatory list (43%). Similarly, differences in the size of regional constituencies show no clear pattern, with the Northern Cape as the smallest constituency showing the highest rate of female representation (80%), while KwaZulu Natal as the second largest constituency, has the lowest rate of representation (37%). These provincial patterns are similarly reflected in provincial and local government results, suggesting that candidate nomination procedures account for the differences in the levels of representation rather than electoral rules. Appendix 3 elaborates on rates of gender representation in South Africa.

Scholars find that “party magnitude” (the number of seats won by a party) is a better concept for understanding rates of gender representation as parties make calculations about whether to include women and where to put them on the ballot based on the number of seats they expect to win.⁹¹ Parties that expect to win more seats put forward more female candidates while the party lists of small parties tend to be headed by male candidates.

In terms of open or closed party lists, evidence shows that closed list electoral systems would be better suited for female candidates, but only where voters show a bias against female candidates.⁹² In the South African context this bias is limited and it should be recognised that 55% of voters in the 2024 election were women.⁹³ In countries where voter attitudes are favourable to female candidates, more candidate-centred systems (including open list systems) can facilitate the election of women without the need for quotas.⁹⁴

Equitable opportunities for independent candidates to compete present another dimension of inclusivity. This aspect was raised by various submissions to the ERCP. This is in line with the Constitutional Court’s New Nation Movement judgement, which determined that individuals should be allowed to contest elections as independent candidates. Consequently, the choice of electoral system should enable independent candidates to compete on equitable terms to allow for the exercise of their political rights.

Finally, we can also consider inclusivity from the perspective of voters’ ability to participate in elections. The simplicity guiding principle aims to ensure that the choice of system is accessible to voters. Another dimension frequently raised is that of voter turnout in terms of promoting participation in the democratic process, particularly in the context of rising voter apathy and declining turnout in South Africa. A common impulse is that electoral reform can address challenges with voter turnout. However, international experience with electoral reform shows that in cases where turnout improved after reforms, this tended to be short-lived. Voter turnout rates are instead context-specific and driven by a range of factors beyond the electoral system choices. These include demographic factors, voter identification with political parties and parties’ ability to mobilise voters, the state of the economy and the ease of voting. In the 1990s there was a view that PR performed better in terms of turnout, but further experience with the introduction of democracy in new regions has revised earlier conclusions, and the evidence remains inconclusive. Both PR and majoritarian/plurality systems provide distinct incentives for promoting voter turnout.⁹⁵

Fairness

The principle of fairness gives effect to universal suffrage and values of equality and human dignity. Every eligible voter should have the reasonable opportunity to participate in elections and as much as possible, votes should be of equal value. Election results should closely represent the will of the people and be fair in the allocation of national and provincial seats without undue bias or discrimination against candidates or parties.

To apply the words of Judge Albie Sachs in August v Electoral Commission:

The universality of the franchise is important not only for nationhood and democracy. The vote of each and every citizen is a badge of dignity and of personhood. Quite literally, it says that everybody counts. In a country of great disparities of wealth and power it declares that whoever we are, whether rich or poor, exalted or disgraced, we all belong to the same democratic South African nation; that our destinies are intertwined in a single interactive polity.⁹⁶

The question of fairness is closely associated with the proportionality of an electoral system, which in practice means limited deviation between the votes cast for a party or candidate and the final allocation of seats. Proportionality may also have the benefit of voters being more likely to participate in the electoral system that leads to proportional outcomes, thus limiting the feeling of their votes being inconsequential in not having impact on the winning of a seat. Key variables affecting the proportionality of an electoral system include:

• the number of seats in a legislative assembly (more seats increase proportionality),

• the average district magnitude (higher district magnitude increases proportionality),

• the allocation formula (Hare and Droop formulae are regarded as more proportional than D-Hondt), and

• the use of a second compensatory tier, including the size of the second tier.

One of the key trade-offs in electoral system design is the closeness of the relationship between voters and their representatives, and the overall proportionality of an electoral system. Increasing the size of electoral districts (both in terms of the number of seats and geographic size) increases the distance between voters and their representatives, weakening the linkages between them. But this also increases the overall proportionality of an electoral system. The addition of compensatory seats, however, removes the trade-off between these two principles and is commonly used to address disproportionality caused by small districts at the lower level.

In cases where a compensatory tier is included, the lower tier of seats loses importance in determining overall proportionality, as the compensatory tier is decisive in determining the final allocation of seats. Taagepera and Shugart find that “the magnitude of the basic district becomes irrelevant to the final votes-to-seats conversion, if sufficient numbers of remainder seats or compensatory seats are allocated at a second stage, so that they compensate for district-level deviation.”⁹⁷ In such cases, any disproportionality caused by small or single-member districts can be effectively removed through the use of compensatory seats. If sufficient seats are allocated at the compensatory level, the system effectively functions as a nationwide allocation, leading to a high degree of proportionality reflective of the overall support for each party.

The proportionality of electoral systems is commonly measured using the Gallagher Index of Disproportionality to measure the difference between the percentage of votes and seats each party gets following an election.⁹⁸ By modelling the effects of different electoral rules using the 2024 election results, it is possible to determine how proportional the electoral system options identified by the ERCP would be. This modelling shows that all options put forward would be equally proportional and would result in the same political outcomes with the same parties represented to the same degree due to the nature of compensatory seat allocations (Figure 2). As mentioned, the compensatory nature of these electoral system options means that they effectively function as a straight nationwide allocation of seats.

Accountability

Accountability is a key democratic value based on the principle that elected representatives are appointed to govern on behalf of voters. This relationship places an obligation on elected representatives to enact the will of the people and accountability is the idea that representatives are answerable in some form to citizens to explain their performance and justify their decisions as part of the social contract. Accountability can take both individual and collective forms and is closely linked to the question of responsiveness. While the choice of electoral systems can shape relationships of accountability, it is recognised that electoral systems cannot guarantee that elected representatives will act with accountability. A closer association between voters and representatives may be desirable.

Accountability is a fundamental democratic principle that ensures government officials and institutions are held responsible for their actions and decisions. It is crucial for maintaining trust and legitimacy in democratic systems. Parliament’s Oversight and Accountability Model⁹⁹ understands accountability as a social relationship where an actor has an obligation to explain and justify their conduct to another – whether a person, persons or institution – and ideally, the second actor should have the ability to hold the first accountable in some way for their performance. This relationship includes both giving account of one’s performance/actions and being held accountable in terms of punishment or rewards for performance. Accountability in democracy includes both vertical and horizontal dimensions. Horizontal accountability involves the checks and balances between state institutions, including the legislature’s oversight of the executive and organs of state. Vertical accountability refers to the mechanisms through which citizens hold their government accountable, with electoral processes forming a key aspect of the vertical accountability relationship.¹⁰⁰

Both dimensions of accountability are critical and complement each other. Horizontal accountability includes the use of parliamentary instruments to ensure that all organs of state are accountable to the National Assembly in terms of Section 55 of the Constitution. This requires the National Assembly to maintain oversight of the exercise of national executive authority, and any organ of state – including provincial and local government entities. This further includes the possibility of criticising the performance of the incumbents by opposition parties and voting against the policies and bills within the parliamentary processes.

Vertical accountability is best understood within a context of the relationship between voters and representatives. Voters delegate responsibility to elected representatives to govern on their behalf, while representatives remain accountable (answerable) to voters. A common way of understanding accountability relationships is through the principal-agent model, where one actor (the principal) delegates responsibility to another (the agent) who acts on the principal’s behalf.¹⁰¹ The act of delegation creates a risk that agents (elected representatives) may not act in the best interest of their principals (voters) who appointed them. Efforts to strengthen accountability are ultimately about adopting appropriate institutional mechanisms and incentives to reduce the gap between the actions of agents (elected representatives) and the interests of their principals (voters).

In terms of electoral system design, key design questions that will affect the accountability relationship include:

• The process for selecting individual representatives (the initial act of delegation).

• The effects of electoral system design on incentives for representatives to respond to voters (their responsiveness to voters’ interests).

• The institutional incentives for representatives to account to voters and for voters to monitor and assess the performance of representatives to determine whether representatives have acted in the best interest of voters.

• Voters’ ability to sanction or reward representatives based on their performance.

This relationship and its dimensions are illustrated in Figure 3, including the selection/delegation process, the incentives for representatives to respond to their principals, the act of accounting and assessing performance and finally, rewarding or sanctioning representatives based on assessments of their performance.

Starting with the act of delegating responsibility, an important consideration is the specificity with which voters can choose individual representatives that they believe would be best able to represent their interests. This is to ensure that the gap between voters’ interests and the actions of elected officials is minimised in the first instance. Factors such as the use of preferential voting (open lists) and the ease with which voters can pick or reject individual candidates will affect the specificity with which voters can choose representatives that best represent their interests. The size of electoral districts is a key consideration here. In the case of single- or small multi-member constituencies, voters are better able to distinguish between candidates and have greater influence in choosing the candidates they regard as best qualified to represent their interests.¹⁰² Greater emphasis on individual candidates likewise increases incentives for political parties to put forward the best candidates.

South Africa’s own experience at local government, however, warns against the assumption that smaller constituencies guarantee good governance and accountable government. There are too many factors that mitigate against this being the norm, irrespective of the electoral system and the factors leading to weak governance at the local government extend beyond the electoral system.¹⁰³ Despite apparent governance challenges at the local government level, members of the public often cited the local government system in public consultations with some indicating support for such a system despites these challenges. It is clear from the public consultation processes that what voters seek is not only broader accountability that is exercised at the end of the five-year term, but also responsiveness.

Responsiveness entails voters’ ability to assess the performance of individual representatives, which shapes incentives for representatives to be sensitive to the interests of their constituents. Electoral system rules that incentivise representatives to cultivate a personal reputation (open lists and smaller constituencies) promote greater responsiveness by representatives to the concerns of voters.¹⁰⁴ As the size of constituencies increases, the importance of individual candidates diminishes as longer party lists and the use of closed lists inhibit voters from distinguishing between candidates within parties. This disincentivises representatives’ responsiveness to voters. Evidence shows that representatives spend less time attending to requests from constituents in such these cases.¹⁰⁵

Where electoral rules make it more difficult for voters to identify individual representatives and assess their performance, voters have less information to determine whether representatives have acted according to their interests and appropriately reward or sanction them (i.e. hold them accountable). Candidate recognition is lower in closed list PR systems compared to other systems, particularly in cases where large constituencies are used.¹⁰⁶ Afrobarometer data among 20 African countries shows 58% of respondents were correctly able to identify their legislative representative among majoritarian systems, compared to 40% in mixed systems and only 21% in PR systems.¹⁰⁷ In South Africa, only 3% could identify their legislative representative, the lowest rate among the 20 countries.

Where representatives have a weak link with their constituents, there are limited incentives for representatives to respond to the interests of voters. The extent to which voters can monitor the performance of their representatives and ultimately reward or sanction them for their actions is key in shaping incentives for the behaviour of individual representatives.¹⁰⁸ Representatives have stronger incentives to act in the interests of voters where they are more reliant on their constituents for re-election, such as in open list systems or where smaller constituencies are used. Accordingly, such representatives spend more time serving their constituents.¹⁰⁹

In closed list PR systems, by comparison, representatives are more reliant on party leadership for re-election and can be protected from electoral punishment by voters. The use of large, closed lists reduces the electoral vulnerability of individuals. Placement in top positions in party lists mitigates against electoral punishment by voters so that poorly performing representatives close to party leaders can still be elected even if the party as a whole loses electoral ground.¹¹⁰

It is worth stressing that voters also want representatives to be responsive throughout the term. That voters hold parties accountable on election days has been evident in South Africa’s electoral outcomes, especially in the recent 2024 national election. However, as evident in the views expressed throughout public consultations, voters seek ways of ensuring that accountability is enforced on an on-going basis. To a large extent, participants in public hearings seemed to associate accountability with the proximity of the elected representative to the voting public. There appears to be two aspects to this concept of accountability. Firstly, closer proximity to the elected representative is likely to result in the voting public knowing who their representative is. Secondly, the representative is likely to work harder to serve the citizens if they are known to the voting public. Both factors are likely to improve the quality of representation.

This is where electoral rules can encourage, or discourage, individual accountability by shaping the range of principals to whom elected representatives respond. This will change based on the extent to which candidates are exposed to direct electoral pressures from voters.¹¹¹ Political parties play an important role in candidate selection and holding representatives accountable in any electoral system. But electoral rules that give voters greater leverage over the re-election of individual representatives will strengthen the linkage and accountability relationship between representatives and voters. This is referred to as shortening the route to accountability,¹¹² as illustrated in Figure 4. In more party-centred systems (closed lists and larger constituencies), the accountability relationship is mediated through political parties and voters have limited mechanisms for holding individual representatives accountable. By comparison, systems which expose candidates to direct electoral pressures from voters (open lists and/or smaller constituencies) expand the range of principals to which representatives respond. This provides stronger mechanisms for voters to hold individual representatives accountable.

Simplicity

A degree of simplicity is necessary to ensure that the electoral system is accessible to citizens to enable the exercise of their democratic rights. This requires adopting voting procedures that practically every voter can reasonably understand in order to make their choice(s) reduce spoilt ballots and allow for their effective participation in the democratic system. It should, however, be acknowledged that South African voters should not be underestimated and that voters are familiar with multi-ballot systems and distinguishing between individual candidates and political parties such as that used in the local government context and in the 2024 elections. Similarly, the number of spoilt ballots has been low over the last thirty years with only 1.3% of ballots spoilt in the 2024 general election – despite the introduction of a third ballot.

Electoral systems with more complex rules, such as the use of preferential voting, can be more challenging for voters. It is crucial to consider the accessibility of the system in voting and understanding how seats are allocated. While more complicated electoral rules may offer certain advantages, they must be balanced against the potential difficulty they may pose to voters. Overall, the goal should not be to maximise simplicity but rather to ensure that the rules are sufficiently accessible to allow voters to effectively exercise their democratic rights.

South African voters have, however, demonstrated their ability to engage effectively with both proportional representation and the mixed system used at local government level. Similarly, the changes introduced in the 2024 election did not pose significant challenges to voters. The 2024 Election Satisfaction Survey showed high levels of satisfaction and trust in the conduct of the election.¹¹³ The number of spoilt ballots in South African elections remains low by international standards. Based on the latest data, the international average for invalid votes is around 3.6%¹¹⁴ and South Africa performs much better in both general- and local government elections. Literacy rates in South Africa have also improved significantly over the last three decades with a national literacy rate of 95%, well above the global average of 87%.¹¹⁵ With this in mind, South African voters should not be underestimated when choosing an electoral system appropriate for the country context.

While South Africans have engaged effectively with different types of electoral systems and multiple ballots, the possible introduction of preferential or open list voting should be carefully considered. The feasibility of introducing open list voting and the ease with which voters will be able to engage with the system will be influenced by the number of seats contested in a constituency and the number of political parties and independent candidates contesting it. Choices around the size of constituencies and the possible introduction of a threshold or changes to election participation requirements to discourage frivolous contestations, would affect the practicality of introducing open list voting.

Election Manageability

While the principle of simplicity supports voters in the exercise of their democratic rights, electoral manageability is aimed at ensuring the effective and efficient conduct of the electoral process. The choice of electoral system should enable effective election management to ensure the credibility of the electoral process and limit frivolous contestations. This principle should be carefully balanced against the achievement of other principles and should not be equated with administrative convenience. Its aim to ensure the integrity of the election process to foster trust in electoral outcomes.

Much like simplicity, the objective of election manageability is not to choose the simplest system. Instead, the intention is to consider what the implications would be for effective election administration and whether the choice of system would be practically implementable based on the financial and technical capabilities of the Electoral Commission. The adoption of an open list system or two-round electoral systems, for example, would have practical implications for election administration that would need to be considered. The participation of a very large number of parties and candidates in an open list system would have practical implications regarding the complexity of ballot papers and the counting of votes, particularly in large constituencies. For this reason, open list voting is more commonly used in countries with smaller constituencies and in most cases, where an electoral threshold is in place.¹¹⁶

The Administration and Cost of Elections Project offers nuanced guidance on evaluating electoral manageability in selecting electoral systems, emphasising that simplicity in the short term does not necessarily equate to long-term cost-effectiveness. A cheap and easy to administer electoral system may not meet the needs of a country and could lead to disastrous outcomes if it failed to address the country’s needs. Conversely, a system that may be more costly and complex to administer may ultimately contribute to the country’s stability and support democratic consolidation over time.¹¹⁷ In this way, questions of cost and ease of administration need to be carefully balanced against what would be most appropriate and responsive to the South African context.

Transparency

Transparency is critical for the credibility of the electoral process and voters’ trust in the system. Processes for determining the allocation of seats for political parties and candidates based on election results should be clear to voters. In cases where voting boundaries are delimited, the process and rules must be transparent, easily understandable and ideally consistent with existing governance boundaries, to foster trust in the system and ensure that citizens regard the electoral system as fair and legitimate.

Transparency is particularly important for the delimitation of electoral boundaries to ensure that both voters and candidates perceive the system to be legitimate. The use of compensatory seats and the overall proportionality achieved by the ERCP’s electoral system options will eliminate the prospect of gerrymandering boundaries to distort electoral results in favour of individual parties or candidates.

Delimitation also has implications for ensuring that electoral boundaries are drawn in such a way that they form cohesive communities and are readily accepted by voters and candidates. To achieve this, it is crucial that the processes and rules surrounding the delimitation of electoral boundaries are transparent and easily understood by communities. Schedule 1 of the Municipal Structures Act, which establishes the criteria for municipal ward delimitation, provides guidance that could be considered in determining processes for the delimitation of electoral boundaries for national and provincial elections. In consultation with the Electoral Commission, the Demarcation Board is tasked with delimiting municipalities into wards, ensuring each ward has approximately the same number of voters while avoiding the fragmentation of communities. This process is based on community consultations to secure buy-in from communities in determining electoral boundaries. Similar processes and criteria could be considered in cases where electoral boundaries for national and provincial elections need to be drawn.

Beyond transparency in the delimitation processes, it is important that the processes for the allocation of seats are transparent. This issue is closely related to the principles of simplicity and electoral manageability. It is based on the idea that processes for determining the allocation of seats for political parties and candidates should be clear to voters. In terms of different electoral systems rules, there is little evidence that some electoral system designs are more transparent than others. Instead, the focus should be on how to incorporate transparency in associated processes, including the delimitation of electoral boundaries and transparency in results reporting.

Stable and Efficient Government

An electoral system on its own will not determine the emergence of a stable and efficient government, but appropriate institutional design can facilitate stability. The avoidance of political instability is a common concern in electoral system design and includes issues such as the ease of government formation, whether government can efficiently enact legislation and govern, and the overall legitimacy of the political system based on the perceived fairness of political outcomes.

Systems based on proportionality (including mixed-member systems) tend towards a larger number of parties gaining representation in the legislature, often requiring the formation of coalition governments in cases where no party wins a majority. The process of coalition formation relies on parties finding common ground on policy issues and agreeing on the allocation of portfolios across the coalition partners. Where legislatures are more fragmented, with seats distributed among many different parties, the process of coalition formation can become more difficult in finding common ground among parties. The extent of fragmentation will inform how many coalition partners are needed to form a majority, increasing the complexity of coalition formation in finding common ground and agreeing on policy issues across multiple actors with distinct interests.

The benefits of coalition government include broader representation in governance, checks and balances within government and more inclusive policymaking. On the other hand, coalition governments tend to be more unstable, particularly where coalitions struggle to balance multiple competing interests. This can undermine effective governance and, in cases of prolonged political instability, undermine democratic legitimacy. A key issue affecting the stability of coalition governments is the number of parties needed to form a government, and the inclusion of more parties means more potential conflict channels that can undo a cabinet. Numerous studies have found a strong relationship between the number of parties in a coalition¹¹⁸ and how long the government lasts.¹¹⁹

The question of how many parties are needed to form a working majority in the legislature is strongly linked to the nature of the electoral system which has a large influence in determining the extent of party system fragmentation. This creates a direct link between electoral system rules and government stability.¹²⁰ The degree of fragmentation in a party system will also depend on the structure of interparty competition and the nature of society itself, which may encourage the formation of many parties, but electoral rules act as a brake or accelerator for fragmentation, depending on how permissive or restrictive electoral rules are.¹²¹

The Seat Product Model presents a methodology for measuring the permissiveness of electoral rules and the extent to which the electoral system will allow for the fragmentation of the party system.¹²² This will be affected by the size of electoral districts, the total number of seats in the legislature and whether the system uses compensatory seats. Figure 5 shows the Seat Product Model for PR and mixed-member proportional systems as a measure of the permissiveness of electoral system rules, with a higher score indicating that the structure of the electoral rules is more permissive to fragmentation.

The permissiveness of the electoral system shapes incentives for the entry and exit of new parties, which will encourage fragmentation if too many, especially small parties are elected. For this reason, most electoral systems have ways of limiting the fragmentation of party systems.¹²³ The most direct approach is to introduce a legal threshold that parties need to pass to receive seats in the legislature, and this is most commonly introduced where the effective threshold is deemed too low. Countries with a legal threshold are marked in red in Figure 5. The figure shows that countries with more permissive electoral rules (larger electoral districts, more seats in the legislature and the inclusion of compensatory seats) more commonly introduce legal thresholds to contain fragmentation. South Africa has the most permissive electoral system in the world that does not include a threshold.

Interaction between Principles

It is important to recognise that these guiding principles have the potential to be interpreted differently, that options may at times require trade-offs between principles, and that principles should be considered in relation to each other. A system that seeks to maximise inclusivity can increase the complexity and difficulty of running elections if it leads to an unmanageable number of candidates and political parties on ballot papers. In turn, if taken too far, simplicity can undermine values of inclusivity, fairness and accountability.

Consequently, the intention should not be to maximise the attainment of any one principle. Instead, the goal is to find an appropriate balance between these guiding principles to give effect to the country’s founding values in a way that can be considered broadly legitimate to South Africans and allows for the establishment of functional governments. Some of these values may be considered more important to uphold when considering the relative balance, such as values of inclusivity or fairness. In contrast, others may only require a minimum level of sufficiency – i.e., the system should be simple enough that voters can easily exercise their democratic rights.

It should also be recognised that the relationship between principles and the attainment of specific goals is sometimes non-linear (i.e. changes in one variable do not lead to an equivalent change in another). This is particularly the case for two-tier compensatory systems where many of the conventional relationships between outcomes (for example, the size of constituencies and proportionality) are superseded by the compensatory nature of the system. In such a system, proportionality will remain the same regardless of the size of constituencies, which means that the closeness of the relationship between voters and representatives can be changed without any loss in proportionality. This has important implications for understanding how different aspects of electoral system design interact and affect the achievement of the guiding principles.

International Trends in Electoral Reform

Many countries have introduced electoral reforms in their efforts to deepen democracy, improve the outcomes produced by their electoral systems and respond to new challenges as their societies evolve and face new political pressures. This section examines international trends in electoral reform to understand how and why countries around the world adopt electoral reforms and the effects these reforms have on improving the quality of democracy in different countries to consider lessons applicable to South Africa.

Electoral reform can take many forms, including changing from one “family” of electoral systems to another (such as moving from a majoritarian system to PR) or changing rules within the same “family” of systems (such as changing the size of constituencies, the ballot structure or introducing electoral thresholds). In the 1990s, electoral reform was considered a rare occurrence with only 10 countries transitioning to a new electoral system family in the period between World War Two and the Third Wave of Democratisation (starting in the 1990s with the end of the Cold War and a wave of political reforms globally). Since the Third Wave of Democratisation, however, at least 44 countries have transitioned to a new electoral system family. The most common changes were from PR to mixed (13), mixed to PR (12), and plurality to mixed systems (11) as seen in Table 1. In many cases these changes have been driven by reforms aimed at deepening democracy, a better understanding of electoral system effects and in response to context-specific challenges.

Changes within the same electoral system type have also been common. This includes PR systems, with 74 cases of countries changing the number and size of their constituencies within the PR family. Cases where countries change the number of constituencies by one or two will likely have minor effects, but at least 36 of these reforms within PR would have had substantial effects on the functioning of electoral systems, including questions of proportionality and the closeness between voters and representatives. Of these 36 cases, 22 have decreased the size of constituencies to bring representatives closer to voters (increasing the number of constituencies) as seen in Table 2.

Another reform seen within the same family of electoral systems is where multi-tier systems rebalance the allocation of seats between the two tiers. This includes mixed and PR systems with a compensatory tier. Among mixed systems there are 31 instances of systems rebalancing the relative allocation of seats between the two tiers – either increasing or decreasing the number of compensatory seats. Of the 13 PR systems with a compensatory tier, around half have made changes to the relative allocation between tiers, and these changes have usually accompanied a change in the size of constituencies.

These reform case studies are worth considering for understanding the relationship between the two tiers and how they interact (Table 3). As compensatory seats are generally used to correct for any disproportionality created by small constituencies, countries have either increased or decreased the number of compensatory seats based on changes in the size of constituencies in the lower tier. For example, Poland removed its compensatory tier when the country increased the size of its constituencies (and therefore increased proportionality in the lower tier), while Nicaragua introduced compensatory seats when the country reduced the size of its constituencies (and therefore introduced more disproportionality in the lower tier).

Regional trends show similar patterns in response to context specific challenges. Here it is worth considering the three regions with electoral systems most similar and relevant for South Africa where proportional systems are most commonly found including Africa, Latin America and Europe.

In Africa, many countries in the post-independence period retained electoral systems first introduced by colonial authorities, including the British FPTP system or the French two-round system but there are a range of examples of reform across the continent.¹²⁴ This includes several cases where countries adopted electoral reforms as part of a political settlement, such as in South Africa, Namibia and the Democratic Republic of Congo. In other cases, countries introduced electoral reforms to address political instability and concerns over government legitimacy. These include Lesotho, where the country introduced a mixed-member proportional system to address distortions in election results caused by the FPTP system and Benin which introduced various changes to the system aimed at curbing extreme fragmentation. There are also various examples where countries initiated inquiries into electoral reforms but lacked the political will to follow through such as Zambia, Mauritius and Malawi.

Latin America similarly saw a wave of electoral reforms in the 1990s driven by concerns over economic stagnation and the perceived inefficiency of state institutions. Countries such as Bolivia, Panama and Venezuela moved from PR to mixed-member proportional systems, while Brazil, Colombia, Costa Rica, Ecuador and Peru introduced open lists. Chile, Nicaragua and Guyana reduced the size of multi-member constituencies. These reforms were generally aimed at strengthening the constituent-legislator bond to enhance responsiveness to local constituencies and increase individual accountability by exposing legislators to accountability pressures from outside their parties. These reforms occurred in response to declining trust in political parties, citizen disillusionment with a lack of political responsiveness and accountability of parties, and finally, a desire to reduce the growing alienation between parties and society by bringing representatives closer to voters.¹²⁵

Reforms in Europe since the late 1980s have similarly seen a trend towards the personalisation of electoral systems – that is, giving voters more of a say over the election of individual candidates.¹²⁶ Whereas several Latin American countries introduced mixed systems to incentivise personal accountability and strengthen voter-representative bonds, European electoral reforms have usually remained within the confines of PR and have instead focused on strengthening voters’ ability to choose individual candidates off party lists and reducing the size of constituencies. This is understood in the context of a widely recognised trend of the “personalisation of politics” among advanced democracies where party identification has declined. Instead, voters increasingly consider factors beyond parties to determine their choice, with an increasing emphasis on the quality of candidates available to them.¹²⁷ Electoral reforms in Europe have responded to this trend based on public demand for a greater say in choosing their representatives.

Internationally, the tendency of electoral reforms has been to converge on the middle ground with party-centred systems becoming somewhat more candidate-centred, while candidate-centred systems have become more party-centred.¹²⁸ This includes both PR and majoritarian systems shifting to mixed electoral systems as well as greater personalisation among PR systems. This shows a trend of bringing representatives closer to voters and strengthening the linkage between the two. The direction of reform has generally been context-specific, with reforms responding to challenges including concerns over accountability, democratic legitimacy, voter apathy and political instability. Although electoral reform was once considered a rare occurrence, since the 1990s it has become much more common as countries work to deepen democracy and strengthen institutions. It should be recognised that such reforms usually take many years, and where changes are introduced, it often takes multiple electoral cycles for changes to take effect.