In political science, an electoral system is a set of rules governing an election. The study of electoral systems has tended to be restricted in a number of respects. First, most research focuses on elections to legislative assemblies and to political executives, even though, of course, many nongovernmental agencies also hold elections as the means of filling various offices. Second, many issues of central concern in the study of electoral systems – such as paradoxes emanating from the application of particular voting rules – are usually studied within the ‘theory of voting’ branch of rational choice analysis; in some respects, therefore the study of electoral system can be understood as more of an ‘applied’ social science, while the ‘theory of voting’ is ‘pure’ theory.
- Elements of Electoral Systems
- Ballot Structure
- Constituency Structure
- Electoral Formula
- Classification of Electoral Systems
- Proportional Electoral Systems
- Nonproportional Electoral Systems
- Proportional Electoral Systems
- Consequences of Electoral Systems
Elements of Electoral Systems
Following Rae (1971), many scholars distinguish between three main elements of an electoral system: ballot structure, constituency structure, and the electoral formula.
There are several different ways in which this can vary, including:
- The number of votes each voter is permitted. In some systems the voter casts as many votes as there are positions to be filled (e.g., plurality systems with multimember constituencies – as employed in many state legislatures in the USA before the 1960s); in others, voters may cast only one vote irrespective of how many positions are being filled (e.g., the Single Non- Transferable Vote system used in Japan before 1994).
- The kind of information the voter can reveal in voting. Some systems permit two different kinds of vote to be cast; for example, in the German Additional Member System (AMS) the voter votes for a party candidate in a constituency, as well as for a party for the party list component of the ballot. In many versions of party list systems the voter can reveal both which party they prefer, and which of that party’s candidates they prefer. Again, by combining two separate elections, a system may make it possible for voters to reveal extra information about their preferences. For example, in Israel since 1996, voters reveal not merely which party they prefer, but also which party they wish to form the government: the Israeli voter can vote both for a party in the election for the legislature, and for a particular candidate to be prime Minister following that election. Furthermore, in some systems, voters can reveal only their first preference (whether for a party or a candidate), while in others they can provide a rank ordering of the different parties (or candidates); the plurality system used in British parliamentary elections is an example of the former, while the Single Transferable Vote (STV) employed in the Republic of Ireland is an example of the latter. Finally, some electoral systems limit the expression of voter preference to those candidates or parties listed on the ballot, while others allow the ‘writing in’ of names not on the ballot, or allow the voter to register opposition to all those listed on the ballot by voting for ‘none of the listed candidates.’
- Whether the ballot is open or secret. In the twentieth century it has become the norm for national elections to be contested under a secret ballot, but until the late nineteenth century balloting was a public act. In the USA, parties in the nineteenth century printed their own ballot papers, and in several states the parties used different colors of paper for the ballot so they could tell who had voted for them. Whether the ballot should be secret or not was a major controversy in nineteenth-century political debates, and some of those who advocated an extensive franchise (e.g., John Stuart Mill) still favored an open ballot.
There are three main ways in which electoral systems vary in this regard.
- The election can be conducted on an ‘at large’ basis, or an election can be held in each of several constituencies. Generally, though with some exceptions, ‘at large’ elections are held only in smaller territories and when a proportional mechanism is used to allocate seats in an assembly (e.g., The Netherlands). Most national and subnational assembly elections are divided into constituencies.
- Whether the constituencies are organized on a territorial or some other basis. Although there have been debates about the desirability of a functional, rather than a territorial, basis for representation, most electoral systems use the latter. The odd exceptions include the ‘university seats’ of the British Parliament that were abolished before the 1950 election. Any territorial subdivision of a polity involves the drawing of boundaries for the constituencies, and how the boundaries are drawn may affect the overall election result. This issue becomes most significant in systems using nonproportional means for allocating seats, but it is still an issue in those using proportional mechanisms.
- Whether the constituencies are single-member or multimember. Systems that employ proportional means for allocating seats in an assembly have to use multimember systems, as does the SNTV; other systems do not. Generally, polities using the plurality system for allocating seats have moved towards the exclusive use of single-member constituencies during the twentieth century.
This is the way in which votes are connected to the office(s) being contested – in an assembly it is those rules governing how votes are translated into seats. There are three main kinds of formulae:
- Majoritarian. Under these systems, the winning candidate or party must obtain more than 50 percent of the vote. The two most commonly used means for achieving this are the Alternative Vote System (AV), and the double Ballot System (DB). The AV, which is used in Australia, involves voters showing their second, and subsequent, preferences on the ballot form; when no candidate obtains more than 50 percent of the vote, the lowest-placed candidate is eliminated and his/her voters have their second-choice votes counted. This process continues until one candidate attains more than 50 percent of the vote. The DB requires a second election to be held in all constituencies in which no candidate obtains more than 50 percent of the vote; this second contest includes only the two leading candidates from the first election.
- Plurality. Here the candidate(s) with the largest number of votes wins. In relation simply to its electoral formula, the SNTV system is merely a plurality system; however, its very different ballot structure – voters having only one vote each in multimember constituencies – mean that it is a radically different system in its operation.
- Proportionality. This formula attempts to provide a ‘fit’ between the share of the votes that the candidates or parties obtain, and the share of the seats the parties get in the assembly. There are two main devices for achieving this. One is to use a formula, of which there are a number of variants including d’Hondt and St. Lagüe, that distributes seats to parties on the basis of their share of the vote. The other method is to permit voters to rank-order all candidates, and to redistribute the ‘wasted votes’ of those candidates who have more votes than they need to get elected; that, broadly speaking, is the basis of STV. Disproportionalities in supposedly proportional systems can occur for three main reasons:
- the ways in which electoral boundaries are drawn, if the system is divided into constituencies, thereby giving an advantage to particular parties;
- relatively small district magnitude – that is, a small number of seats being contested in each constituency; the fewer the seats in each constituency, the more likely it is that, in aggregate the allocation of seats between parties will diverge from the distribution of total votes. (STV has to be employed in conjunction with relatively small district magnitude; it becomes unwieldy with large district magnitude because of the complexities involved in expressing preferences between a large number of candidates.); and
- in party-list systems, through the use of thresholds which require that parties failing to secure a minimum share of the vote will fail to win any seats at all. A threshold of between 2 and 3 percent of the total vote usually reduces the number of parties represented in a legislature quite dramatically, while a threshold of 5 percent (as used in Germany) restricts seriously the number of parties represented.
Classification of Electoral Systems
There are a very large number of possible electoral systems, including many hybrids of the most common types. (An example of such a hybrid is the system used for the French National Assembly, which is a variant of the DB; a second ballot is used, but it is not confined to the two leading candidates from the first ballot. Instead, any candidate with more than 12.5 percent of the vote is permitted, but not required, to take part in the second ballot.) Given this large variety, considerable attention has been given to classifying the main types of system. It is widely agreed that the main distinction is between ‘proportional’ systems and other types, but there is some disagreement on two matters.
The first is whether the principal division is a bifurcation between ‘proportional’ and ‘non-proportional’ systems, or whether a more complex main division is appropriate. The second disagreement concerns how to treat a variety of electoral systems that are sometimes grouped together under the heading ‘semiproportional’; these include the SNTV and the Cumulative Vote (CV). (The latter is a system used in Illinois for many years that, in a three-member constituency would give three votes to each voter but would allow the voter to give all three votes to one candidates, give two votes to one candidate and one to another, or to give a single vote to each of three candidates.) Sartori (1994), for example, argues in favour of treating the SNTV as a form of proportional representation (PR), but most scholars do not do so.
The most common classification is to distinguish between those systems that have mechanisms for generating proportional results that are internal to the systems themselves, and which do not depend for their proportionality on the behavior or strategies of voters or candidates, from those that do not.
Proportional Electoral Systems
Within the class of proportional systems there are two main variants – the party-list and the STV systems, and within the party-list subclass there are four main sub-subclasses:
- Pure party-list. In these systems the order of the party’s candidates on the list, and hence their likelihood of being elected, is determined exclusively by the party. The voter can vote only for a party and not for a candidate. (Such systems are relatively uncommon, but this system is used in Israel and was used in France in 1945–96, and in a slightly less pure form during the rest of the fourth Republic.)
- Voter-choice from party-list. These systems allow the voter to express a choice between candidates on the party-list, though great variations in the procedures are evident. In some systems it is actually difficult for voters to propel a popular candidate from a low position on the list to a higher one; in other systems it is much easier to do so. These types of system are the most common form of the party-list, having been used in many European countries.
- Mixed systems. The AMS is the principal variant of this type. A proportion of the seats (sometimes one half) are contested in single-member constituencies under plurality voting. The remaining seats are contested using a party-list, and the allocation of seats from the party-list is used so as to correct for any disproportionalities resulting from the elections in the constituencies. This type of system, which is used in Germany and New Zealand, produces election results that are rather similar to those that would occur under other party list systems.
- Panachage. Used in Switzerland, this system permits voter choice between candidates of more than one party. Some experts argue that it is not really a form of party-list system because of this rule, although its operation more closely resembles such systems than any other types.
Unlike party-list systems, STV achieves proportionality in the distribution of seats in a legislature through a mechanism that allocates seats to particular candidates. A party acquires seats through given candidates of theirs receiving sufficient votes to win seats, rather than the parties being allocated votes through a formula that distributes seats directly to the parties.
Nonproportional Electoral Systems
Among the nonproportional systems, there are three main subclasses:
- Plurality. The first-past-the-post system has been used in many Anglo-Saxon democracies, and it is still the system used in Canada, and most of the USA, as well as for elections to the UK Parliament.
- Majoritarian. The most common variants of systems requiring the winning candidate (or party) to win more than 50 percent of the vote are the AV and DB. Outside national elections, there are some majoritarian systems that require the winning candidate to receive more than 50 percent of the vote – one example being the Democratic National Convention which, until the 1930s, required that the nominee for the presidency receive two-thirds of the delegates’ votes. Since such systems can produce many rounds of balloting and still fail to produce a result, they are wholly unsuitable for elections at which electors cannot interact personally with each other in an effort to achieve a definite result.
- Semi-proportional. This is a subclass containing several rather different kinds of system, including the CV, SNTV and the Limited Vote (LV). The last is rather similar to the SNTV in that there are multimember constituencies, but, whereas in SNTV voters have only one vote, under LV they have more than one vote, but fewer votes than there are seats being contested. It is employed for some seats in the Spanish Senate, and was used in a few urban constituencies in Britain between 1867 and 1883. Semiproportional systems acquire their (rather misleading) name from the fact that they tend to generate greater representation for smaller parties than does the plurality system. However, the objection to classifying them as a form of proportional system is not that they tend to produce disproportional results – though in fact they do tend to do so – but rather that how proportional they are depends on the particular behaviour of voters and parties. In some situations they may not be proportional at all. By contrast, proportional systems are designed to generate a degree of proportionality.
Consequences of Electoral Systems
Public and academic interest in electoral systems tends to be much greater when the franchise is being extended (as in Britain in the 1860s), or new legislative assemblies are introduced (Britain in the 1990s) when democratic regimes have collapsed or are unstable (Europe from the 1920 to the 1950s), and when regimes are democratizing (Eastern Europe in the 1990s), or redemocratizing (Latin America in the 1980s and 1990s). One of the main focal points of attention is the likely consequences that one form of electoral system will have compared with others.
In the 1950s, and in light of the experience of collapsed democracies in the interwar years and with the potentially unstable regimes in postwar France and Italy, Duverger (1959) propounded an argument that has come to be known as Duverger’s Law: ‘the simple-majority single ballot system favours the two-party system’. He also argued that PR tended to be associated with multipartism. His argument was often linked to the widely held view in Britain and the USA at that time that PR was responsible for unstable governments with multiparty systems, of the kind found in the fourth Republic, whereas plurality systems were a recipe for stable two-party democratic government.
Since Duverger’s Law was first propounded it has had its critics, and at the time of writing many political scientists would point to its limitations. It does not apply when there are strong regional cleavages, for example; they also reject the popular view that PR was causally responsible for unstable government. That view does not fit well with the evidence: most European countries since 1945 have used some form of PR, and most have had multiparty systems and coalition governments, yet most have not had unstable regimes. Although there are some exceptions, notably Sartori (1994), many political scientists now argue that the electoral system has only a limited effect on a regime’s party system, and hence on its potential for instability. Increases in the number of parties in a system, for instance, tend not to follow the introduction of PR, but rather tend to precede it; moreover, the timing of the introduction of PR is significant in relation to whether party fragmentation will occur. An early switch to PR following democratization does tend to produce fragmentation – because the larger parties have had limited opportunities to develop loyal electorates. A switch much later does not lead to fragmentation, because the established parties now have loyal electorates who are not likely to defect to new or small parties.
Nevertheless, the precise form of electoral system used can have an impact on election results: changing the rules does change the political game. For example, the switch from plurality voting to AV in British Columbia in the early 1950s probably did prevent the socialist CCF party from gaining power in that decade. Similarly, the switch to PR from DB for the 1986 Assembly elections in France helped to minimize the Socialists’ losses and to establish the National Front as a credible party that could win seats in the Assembly. Electoral systems cannot transform the social bases of electoral politics, but they can make it more likely that some parties will win office, or share power in government, and make it less likely that others will do so. For that reason, politicians are tempted sometimes to change them. For the most part, though, they are not changed frequently – partly because those in office are the people who usually benefit from the working of the existing system, and partly because the full consequences will be unknown, and hence might actually run counter to the interests of those seeking change.
- Blais, A., Massicotte, L., 1996. Electoral systems. In: LeDuc, L., Niemi, G., Norris, P. (Eds.), Comparing Democracies. Sage, Thousand Oaks, CA, pp. 49–82.
- Duverger, M., 1959. Political Parties, second ed. Methuen, London.
- Lijphart, A., 1994. Electoral Systems and Party Systems. Oxford University Press, Oxford, UK.
- Rae, D.W., 1971. The Political Consequences of Electoral Laws, second ed. Yale University Press, New Haven, CT.
- Sartori, G., 1994. Comparative Constitutional Engineering. Macmillan, Basingstoke, UK.
- Ware, A., 2009. The Dynamics of Two Party Politics. Oxford University Press, Oxford.