How people elect parliaments
United Kingdom – pre-Reform Act (to 1832) – [not yet posted]
United Kingdom – mid-19th century (1833-1867) – [not yet posted]
Prussian three-class system (1870s to 1914) – [not yet posted]
Germany – Weimar Republic (1920-1932) – A bicameral Bundestag of which the lower house was the Reichstag, an assembly with a floating total of between 450 and 650 members. The Reichstag was elected by a party list seat allocation method applied within an integrated tier system, and with floating total numbers of members elected in the divisions, and thus in total. Members were not directly elected, but seats were allocated to parties in 35 region-based electoral divisions and also in 17 combined divisions, with each of the latter consisting of 2 (and in one case 3) neighbouring primary divisions. The method of allocating seats used a pair of arbitrary fixed quotas, rather than the more common highest averages method. In each of the primary electoral divisions, each party was allocated one seat for every 60,000 votes won. After such allocations were determined, within each of the 17 combined divisions, party totals of all votes which had not been used to secure seats in the primary divisions (including the whole vote for unsuccessful parties as well as the ‘remainder’ votes for parties which did win seats (after deducting multiple of 60,000 votes for seats already won)) were amalgamated, and one additional seat was allocated to each party which held 30,000 votes. Seats were then allocated to individual members from closed party lists. Terms were up to four years, but early elections were called frequently. In the 9 elections held between 1920 and 1932, the number of members elected ranged from 450 to 650.
France – post-war (1945-1994) – [not yet posted]
France – PR (1994-1996) – [not yet posted]
Hungary – (1991 to 2012) – A unicameral Az Orszag Haza or Országgyűlés (National Assembly or Diet) of which the sole house was a composite assembly of 386 members. There were three components of the assembly. 176 members of the assembly were directly elected in local divisions using a form of two-round runoff plurality voting system. In this system all candidates who won at least 15% of votes in the first round, and the top 3 candidates in any case, were eligible to enter the second round, in which winners are determined by the plurality vote method. (For second round contests, parties often strategically withdrew candidates who ran 2nd or 3rd in the first round, often as part of negotiated arrangements between two or more parties). 152 seats were allocated separately in 20 multi-member divisions by the closed party list system of seat allocation (using largest remainder rounding with a Hare quota) to parties which win 5% of the total national vote. Finally, a further 58 levelling seats were allocated to parties which won at least 5% of the national vote (using a national list) so as provide that final representation of each such party was as proportional as possible to its share of the national vote total. The counting method for the regional list category was such that in certain circumstances some of the regional list places fail to be allocated to any eligible party, in which case such seats (for example, 6 seats in the elections of 2010) were added to the number used in the allocation of seats to national lists.
Italy – post-war
Italy’s system for electing the Camera dei Deputati has evolved in several distinct phases since the Second World War, including:
The ‘Porcellum’ system (2005-2015):
For the Camera that were elected in 2009 and 2013 the nation was divided into 27 electoral divisions based on the 20 regions of Italy: Lombardy had three electoral divisions, Piedmont, Veneto, Latium, Campania, and Sicily each have two, and all other regions have one. Each of the electoral divisions was allocated a number of members in proportion to its share of the national population. One very small division – the Aosta Valley region – was allocated a single seat, which was filled by direct election using the plurality voting method.
The members for the other 26 divisions were not directly elected; seats were allocated within each electoral division by the closed party list system of seat allocation, using the D’Hondt divisor formula, to parties whose national vote exceeds one or more eligibility thresholds. For the purpose of the eligibility thresholds, two or more parties could choose to be associated as a coalition – a form of apparentment. Any individual party winning 4% of the national vote was eligible for seats, as was any party which won 2% of the vote if it was a member of a coalition. All the parties in a coalition which won a total of 10% of the national vote were eligible, and finally the party which won the highest vote in each coalition was eligible even if its vote was below 2%. Finally, any party which represented a regional linguistic minority was eligible if it won 20% of the vote in its electoral division.
The allocation of seats to parties was altered by a reinforced majority mechanism. The coalition which won a plurality of votes was guaranteed a minimum of 340 seats in the Camera. If the initial allocation of seats to parties according to their vote shares resulted in the leading coalition achieving that target, then that seat allocation was confirmed. If the lead coalition won less than that target, their seat allocation was arbitrarily increased to 340, and the remaining seats were re-apportioned among the other eligible parties in proportion to their vote shares. The distribution of such seats across the divisions was also proportional. Once final seat numbers for each party in each division were confirmed, seats were allocated to individual candidates in party list order.
A final 12 seats were allocated to Italian citizens living abroad, divided into four electoral divisions (zones): Europe (including Russia and Turkey), South America, North and Central America, and Africa/Asia/Oceania/Antarctica – each of which elected a share of the 12 seats proportional to the number of citizens enrolled in each zone. Terms were up to four years.
The ‘Italicum’ system (2015-17, never used):
Under the ‘Italicum’ system legislated in 2015, a total of 617 seats were to be filled not by direct election but (in common with the previous system) by a reinforced majority form of party list seat allocation. The nation was to be divided into 100 electoral divisions established within the boundaries of its 19 regions (not counting the Aosta region), with each division allocated between 3 and 9 deputies based on divisional populations.
There was no provision for individual, independent candidates outside of political parties. Registered political parties will publish a list of candidates for each division. The first candidate on each list was to be termed a ‘head-of-list’ candidate, and such candidates would be nominated on the lists for up to ten electoral divisions. All other candidates would only be nominated on the list of a single divisional.
There were two gender-equality rules relating to party lists. Firstly, on each party’s published divisional list each party’s candidates would be shown in a sequence alternating by gender. Secondly, across the divisions within each region, no more than 60% of head-of-list candidates for the same party could be of the same gender. At the principal round of voting the ballot paper would provide for voters to vote for just one party.
The party list approach would have be semi-‘open’ (or rather, it will be ‘open’ except that one lead candidate will be imposed per list). The head-of-list candidate will be displayed on the ballot, and voters will have the option to write in the names of up to two other candidates from the list of their chosen party as ‘preference votes’. If the two preference votes marked on a ballot are for candidates of the same gender, the second preference will be disregarded. Only parties which receive at least 3% of the national party vote total will be eligible to be allocated any of the 617 available seats.
The 2015 system was also based on identifying a ‘lead party’ (or lead coalition) arising out of each election result. If the party receiving the plurality of the national vote total also received at least 40% of that total, it would win the lead party status. If the plurality party fell short of that 40% threshold, a simplified second round of voting would have been held, at which voters would choose between the two parties that received the most votes in the primary round, and the winner of that vote would have become the lead party. However the lead party status is determined, that party would be allotted 340 of the 617 seats. (This allotment was strictly only a minimum number of seats; a very successful lead party might conceivable receive more seats, but a national vote share of around 55% or more would be needed, and such a result is highly unlikely in Italy’s present multi-party system.) Seats for all parties other than the lead party – which would have numbered 277 seats at most – would then have been allocated among the other parties proportionally based on national vote shares.
Proportional allocation of seats between all the non-lead parties (or conceivably between all parties including the lead party should the lead party receive a very high vote share) will use the largest remainder method. Once the national number of seats allocated to each party has been determined, these seat numbers are then projected onto the 100 electoral divisions so that each division has the predetermined number of representatives. Actual seats are then allocated to the candidates in each division with the head-of-list candidates guaranteed each party’s first seat (which guarantee may be waived as needed where such a candidate was nominated in multiple divisions), and with further seats being allocated to each party’s other candidates in order of the numbers of preference votes they received.
Outside of the main electoral system, voters in the Aosta Valley region would still have been allocated a single seat, to be filled by direct election using the plurality method.
As with the previous system, a final 12 seats would have been allocated to Italian citizens living abroad, divided into four electoral divisions (zones): Europe (including Russia and Turkey), South America, North and Central America, and Africa/Asia/Oceania/Antarctica – each of which electing a share of the 12 seats proportional to the number of citizens enrolled in each zone.
The Italicum system was legislated in 2015, but during 2017 a legal challenge to the law’s validity in the nation’s constitutional court was upheld, forcing the legislature to replace with system with a new model legislated in late 2017.
The debate over electoral law in Italy in 2015 was strongly influenced by the attempt by the former Renzi Government to reform the constitutional arrangements which make Italian governments simultaneously accountable both to the Camera and the nation’s upper house, the Senato. The then government argued that the system of dual accountability, with the Senato divided between the three current major Italian parties, renders national governance intolerably difficult. However a referendum to implement the proposed change was rejected by voters late in 2016, causing the government to resign.
Lebanon – (1960 to 2017) – Prior to long-deleyed electoral reforms adopted in 2017, the Lebanese Assemblée had an unusually complex combination of features including electoral divisions, the block voting method, a unique faith-based limitation on nominations – which is also used to alter the seat award rule used for the block vote method – and an unusual approach to ballot paper options which significantly impacts on campaign tactics and election results. In addition, Lebanon featured a highly diverse party configuration.
Seats were allocated in general proportion to population among 26 electoral divisions based on local administrative district boundaries (with some grouping and subdivision of districts), which elected numbers of members ranging from 2 to 10. Members were directly elected in these divisions by a variant of the block voting method. Voters were not required to exercise all of the votes they are entitled to, but may validly vote for a lesser number of candidates than the number of places available.
All nominations for election were be made under a system of ‘confessional distribution’ in which nominees were required to identify with one of ten religious communities or ‘confessions’ (six Christian and four Islamic). Within each electoral division every seat to be won is also associated with one of the ten confessions. Each electoral division had numbers of specific allocations of seats to combinations of the ten confessions ranging from 0 to 5 seats. The distribution of confession seats across the electoral divisions was such that there are a total of 64 Christian places and 64 Islamic places, a national compromise adopted by a political agreement in 1989. Candidates could only nominate for election if there was at least one seat available in their electoral division for their confession, and by implication citizens who do not identify with any confession provided for in the electoral division in which they resided could not be nominated for election to the Assemblée.
Voters, however, were not bound by the confessional identity system and may vote for any candidate nominated in their electoral division. To ensure that the seats were filled appropriately by the desired number of confessional candidates, the normal rules of block voting (whereby the places are awarded to the N candidates with the highest votes) were modified such that the highest-placed candidates associated with each available set of confession-based seats are elected. For example, if a division had 7 seats, including 3 for Maronite Christians, 2 for Sunni Muslims and 2 for Druze Muslims, the highest placed 3, 2 and 2 candidates in those confessional categories were awarded the available seats (which outcome may differ from that of the 7 highest-placed candidates overall).
In addition, voters were not provided with standard ballot papers listing all candidates, but could either insert names onto a blank paper, or lodge a pre-written list. The use of standardised pre-written lists provided by campaigning candidates was very common, and for reasons of tactical advantage under the block voting system, candidates with reasonable chances of success across diverse parties – even parties with strong rivalries – often collaborated to encourage voters to use complex multi-party pre-written lists. The impact of these tactics significantly complicated the usual role of parties in grouping votes, and also altered the significance of nominal party vote totals in each election.
In addition, and perhaps because of this system, Lebanon developed a highly diverse party configuration, with 19 separate parties and 11 independents represented in the Assemblée elected in 2009.
The last election under these electoral rules was in 2009, which was followed by a lengthy delay in elections until 2018.
New York City Council – (1937-1947) – A unicameral assembly of varying size. The members were directly elected in 5 electoral divisions (“boroughs”) electing 1 member for every 75,000 valid votes cast at each election. Seats were allocated within each borough by the single transferable vote (STV) method. Candidates with below 2,000 votes were eliminated at the beginning of the counting process. A fixed quota of 75,000 votes was used, typically resulting in the final available positions being filled by candidates with less than a quota of votes. Terms were two years. The city had the separated powers system of government, in which executive power exercised by the Mayor elected separately from the Council.
Japan – post-war (1950-1993) – [not yet posted]
Chile – Pinochet regime aftermath (1988-2015) – a bicameral Congreso Nacional of which the lower house was the Cámara de Diputados, an assembly of 120 members. Members were directly elected in 60 two-member divisions, using the unusual binomial system, an unusual form of party list system of seat allocation simplified for use with 2-member divisions. Electoral division boundaries showed a significant degree of malapportionment favouring rural areas.
Within each electoral division, each party nominated (up to) two candidates. Voters cast single non-transferable votes for one candidate among all those nominated. The party totals – the sum of the votes for the two candidates of each party – were used to allocate the two available seats, with the two parties having the largest totals each being allocated one seat, unless the largest party had twice the vote of the second-largest, in which case that party was allocated both seats. For a party allocated one seat, the successful individual candidate was the one who received the higher number of individual votes. This method tended to eliminate representation of parties (or coalition lists) other than the two dominant ones (except in regional areas where third parties have concentrated support), and also reduced the relative representation of the largest party and increases that of the second-largest, creating a strongly stable two-party system.
The former electoral system was introduced by the Pinochet regime in the 1980s. In addition to the significant pro-rural malapportionment it was also by its nature severely discriminatory against small parties, effectively forcing Chilean political parties to merge into joint coalition candidate lists at election time.
Key sources of information include on the assemblies given here include the following:
The Parline database: a web publication of the Inter-Parliamentary Union; highly authoritative, although sometimes a little out-of-date.
The Institute for Democracy and Electoral Assistance (IDEA): a UN-created international organisation which maintains authoritative databases on electoral system type.
The Office for the Promotion of Parliamentary Democracy (OPPD): an agency of the European Union which published information about electoral systems of EU members.
Mackie and Rose: an authoritative handbook of national electoral systems, but not current beyond the 1991.
Wikipedia: useful for a variety of background information, but the use of terminology is not always consistent and accuracy of complex nuances in specific entries can be unreliable.
The ElectionResources website: an independent psephological blog.